Biosynthesis of Phycoerythrobilin and its Attachment to Phycobiliproteins

Transcription

1 Biosynthesis of Phycoerythrobilin and its Attachment to Phycobiliproteins Dissertation to obtain the degree Doctor Rerum Naturalium (Dr. rer. nat.) at the Faculty of Biology and Biotechnology Ruhr-University Bochum International Graduate School of Biosciences Ruhr-University Bochum Faculty of Biology and Biotechnology Department of Microbial Biology/Physiology of Microorganisms submitted by Andrea Wilma Ursula Busch from Greifswald, Germany Bochum April, 2011 First Supervisor: Prof. Dr. Nicole Frankenberg-Dinkel Second Supervisor: Prof. Dr. Eckhard Hofmann

2 Biosynthese von Phycoerythrobilin und seine Anknüpfung an Phycobiliproteine Dissertation zur Erlangung des Grades eines Doktors der Naturwissenschaften der Fakultät für Biologie und Biotechnologie an der Internationalen Graduiertenschule Biowissenschaften der Ruhr-Universität Bochum angefertigt am Lehrstuhl Biologie der Mikrooragnismen, in der Arbeitsgruppe Physiologie der Mikroorganismen eingereicht von Andrea Wilma Ursula Busch aus Greifswald Bochum April, 2011 Referentin: Prof. Dr. Nicole Frankenberg-Dinkel Korreferent: Prof. Dr. Eckhard Hofmann

3 Acknowledgments First I want to thank Prof. Dr. Nicole Frankenberg-Dinkel for the opportunity to work on this interesting project, for her continuous support and the numerous personal and professional development opportunities that were provided to me. Secondly, I would like to thank Prof. Dr. Eckhard Hofmann for volunteering as the second supervisor, his cooperation in crystallization projects and for his helpful guidance. Next I want to thank all cooperation partners: Dr. E. Reijerse and Prof. Dr. W. Lubitz from the MPI, Mülheim for EPR spectroscopy, Jun.Prof. L. Leichert for mass spectrometry and Prof. Dr. C. Hermann and K. Kock for ITC analyses. Recognition goes to Dr. Shih-Long Tu for his help in establishing a new assay system, his hospitality during my stay in his lab and helpful suggestions. My thanks go to the Research School (Ruhr-University Bochum) for funding and to the DAAD for supporting a one month exchange at the Academia Sinica, Taipei, Taiwan. I also want to thank all people at the chair for Microbial Biology, the entire group of Physiology of Microorganisms for their support and a stimulating working atmosphaere as well as Hanno Boeddinghaus for IT assistance and Britta Schubert for technical support. Special acknowledgments go to Dr. Jessica Wiethaus for cooperation, helpful discussions and suggestions on the project. Last I want to thank my friends and family for their encouragement.

4 Ι List of Abbreviations Ι Ι List of Abbreviations Å Angström APC Allophycocyanin ATP Adenosintriphosphate BV Biliverdin C Grad Celsius Da Dalton DFT Density functional theory 15,16-DHBV 15,16-Dihydrobiliverdin EPR Electron paramagnetic resonance Fd Ferredoxin FDBR Ferredoxin-dependent bilin reductase Fig. Figure FNR Ferredoxin:NADP + -reductase FRET Fluorescence resonance energy transfer FTIR Fourier transformation infra-red spectroscopy GST Glutathion-S-Transferase HO Heme oxygenase HPLC High performance liquid chromatography HY2 Phytochromobilin synthase MM Molecular mechanics NADPH Nicotinamide adenine dinucleotide phosphate NADP + Nicotinamide adenine dinucleotide phosphate (oxidized) NMR Nuclear magnetic resonance PBP Phycobiliprotein PBS Phycobilisome PebA 15,16-Dihydrobiliverdin:ferredoxin oxidoreductase PebB Phycoerythrobilin:ferredoxin oxidoreductase Pcb Prochlorophyte chlorophyll-binding PCB Phycocyanobilin PcyA Phycocyanobilin:ferredoxin oxidoreductase PEB Phycoerythrobilin PEG Polyethylene glycol PUB Phycurobilin PVB Phycoviolobilin PΦB Phytochromobilin PS Photosystem Phy Phytochrome QM Quantum mechanics RMSD Root mean square deviation UV/vis Ultraviolet/visible-spectroscopy WT Wild-type

5 ΙΙ Table of Contents ΙΙ ΙΙ Table of Contents Ι List of Abbreviations ΙΙ Table of Contents 1 Introduction Photosynthesis Light-Harvesting Light-Harvesting in Cyanobacteria Phycobilisomes and Phycobiliproteins Phycobilins Phycobilin Attachment Phycobilin Biosynthesis Ferredoxin-Dependent Bilin Reductases (FDBRs) Phytochromobilin Biosynthesis Phycocyanobilin Biosynthesis Phycoerythrobilin Biosynthesis Objectives of the Work 19 2 Radical Mechanism of Cyanophage Phycoerythrobilin Synthase (PebS) 21 3 Structural and Mechanistic Insights into the Ferredoxin-Mediated Two Electron Reduction of Bilins 35 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 52 5 CpeS Is a Lyase Specific for Attachment of 3Z-PEB to Cys82 of β-phycoerythrin from Prochlorococcus marinus MED Discussion Ferredoxin-Dependent Reduction of Bilins FDBRs Act via a Radical Mechanism What Determines the Regiospecificity of Different FDBRs? Bilin-Channeling: From Heme to Biliproteins FDBRs Distribution and Evolution What Function Does the Occurrence of FDBR Sequences Imply? Evolutionary Path of FDBR Sequences 90 7 Summary Summary Zusammenfassung 95 8 References 97 9 Appendix Synthesis of Putative FDBR Sequences, Cloning and Expression Allocation of Personal Contribution to Publications/Manuscripts in Percent Conference Contributions Curriculum Vitae Statement of Authorship 121 Ι ΙΙ

6 1 Introduction 1 1 Introduction 1.1 Photosynthesis Photosynthesis, deduced from the greek words for light (phōs) and composition (sýnthesis), describes a process in which light energy is converted into chemical energy in form of carbon hydrates. Photosynthesis evolved about 3.5 billion years ago. The first organisms to utilize light for energy conservation were anoxygenic. Oxygenic photosynthesis emerged roughly 2.5 billion years ago (Rye and Holland 1998; Xiong and Bauer 2002). With the development of oxygenic photosynthesis the oxygen levels on earth rose allowing an aerobic metabolism to evolve. This process was mainly driven by the photosynthetic activity of cyanobacteria (Bekker et al., 2004; Summons et al., 1999). Light is used as an energy source in all three kingdoms of life, eukarya, bacteria and archaea. Bacterial photosynthesis has been reported for six phyla: oxygenic cyanobacteria and anoxygenic proteobacteria (purple bacteria), green sulfur bacteria (Chlorobi), green filamentous bacteria (Chloroflexi), heliobacteria and the newly discovered acidobacteria (Blankenship, 2010; Bryant et al., 2007; Raymond, 2008). The evolution of photosynthesis in bacteria is complex and is thought to have largely occurred via horizontal gene transfer (Raymond et al., 2002). Eukaryotic photosynthesis is based on bacterial systems, originating from endosymbiosis of a cyanobacterial-like cell, which in photosynthetic eukaryotes, like plants and algae, forms the chloroplast (Douglas and Raven, 2003; Kutschera and Niklas, 2005; Margulis, 1992; Raven and Allen, 2003). In the light reactions of oxygenic photosynthesis, the pigments in the two reaction centers are excited, which results in electron flow via a gradient down an electron transport chain that leads to the reduction of NADP + to NADPH. The electron flow also creates a proton gradient across the thylakoid membrane that drives ATP synthesis. The photoinduced charge separation takes place in the reaction centers and light-harvesting antennae are employed to funnel excitation energy to the reaction centers. These light-harvesting structures widen the range of utilizable wavelengths and thereby allow the organism to efficiently adapt to the available spectrum (Blankenship, 2001; Green and Parson, 2003; Neilson and Durnford, 2010).

7 1 Introduction Light-Harvesting Light-harvesting and transfer of excitation energy is realized by pigments. Pigments are utilized by the antenna complexes and the reaction centers and can be distinguished into primary photosynthetic pigments and accessory pigments. Accessory pigments harvest light and transfer the excitation energy to the primary pigments located in the reaction centers. These can be excited by either transfer of energy from the accessory pigments or by direct excitation from the sunlight. Antenna systems are present in almost all photosynthetic organisms. The diversity of structures and pigment compositions found suggests that the antenna systems have evolved independently contrary to the evolution of the photosynthetic reaction centers (Blankenship, 2010). Brightly colored antenna complexes called phycobilisomes (PBS) are present in red algae, cyanobacteria and glaucophyta. Being attached to the outer thylakoid membrane they are predominantly associated with photosystem ΙΙ (PSΙΙ) (Glazer, 1977; Grossman et al., 1993; Steiner et al., 2003) Light-Harvesting in Cyanobacteria Cyanobacteria are unicellular Gram-negative species that are found in marine and limnic environments and even on soil or rock. About one half of global primary production occurs in the oceans and a large portion of this can be attributed to cyanobacterial photosynthetic activity of the two main marine genera Synechococcus and Prochlorococcus (Urbach, Robertson et al. 1992; Li 1994; (Garcia-Pichel et al., 2003; Scanlan et al., 2009). In the oceans the light availability depends on the depth. Further down in the water column less light of longer wavelengths is transmitted. While red light only penetrates surface areas, green light can penetrate down to about 100 m and blue light at least 200 m (Davis, 1991). Marine species are very well adapted to this light availability. Cyanobacteria possess two photosystems and special light-harvesting antennae. Two main types of antennae are known in cyanobacteria, the large PBS and the prochlorophyte chlorophyll binding (Pcb)-antenna (Fig. 1). The Pcb-antenna is present in the marine genus Prochlorococcus while other cyanobacteria employ PBS for light-harvesting (MacColl, 1998; Partensky et al., 1999; Ting et al., 2002).

8 1 Introduction 3 Fig. 1: Different light-harvesting strategies in Synechococcus and Prochlorococcus. Synechococcus uses phycobilisomes (PBS) as antennae. They are attached to the outer thylakoid membrane to funnel energy to the reaction centers (green). The core is made up by allophycocyanin (APC) and the rods contain phycocyanin (PC), which can also incorporate phycoerythrin (PE). Prochlorococcus on the other hand harvests light via its Pcb-antenna. Prochlorococci are the smallest known organisms to perform oxygenic photosynthesis (Chisholm et al., 1988). They have a very reduced genome, a small cell size of 0.5 µm 0.7 µm, can reach cell densities up to 10 5 cells per milliliter and mainly populate oligotrophic open waters (Partensky et al., 1999). Prochlorococci have lost PBS and evolved the unusual Pcb-antenna for photosynthetic light-harvesting as a result of niche differention. This membrane integral light-harvesting structure contains special divinyl chlorophylls (Chl), the socalled Chl a 2 and Chl b 2. These pigments are derivatives of Chl a and Chl b that contain an additional vinyl group at the C8 carbon of the B-ring in the tetrapyrrole macrocycle (Chisholm et al., 1992; Goericke and Repeta, 1992). Prochlorococcus occurs in two ecotypes. High-light adapted strains are found in the high irradiated nutrient depleted upper layer of the ocean while low-light strains colonize the nutrient rich depths with low-light intensities (Moore et al., 1998). In low-light adapted strains Pcb-antenna for both photosystems have been found whereas in high-light adapted strains only a Pcb-antenna for PSΙΙ is present (Bibby et al., 2003). This reduced antenna proves advantageous in the deep depth of the

9 1 Introduction 4 oceans due to an absorption range in the blue spectrum (430 nm 490 nm) (Ting et al., 2002). The Pcb-antenna system does not seem to be related to other lightharvesting structures like the light-harvesting complexes (LHCs) of plants, it rather evolved from CP43, a Chl a antenna protein of PSΙΙ (La Roche et al., 1996). Besides the Pcb-antenna, Prochlorococcus contains remnants of its ancestor PBS in form of the phycobiliprotein (PBP) phycoerythrin (PE) located at the thylakoid membrane, whose function is still unknown (Steglich, Mullineaux et al. 2003; Steglich, Frankenberg-Dinkel et al. 2005; Hess, Partensky et al. 1996; Hess, Steglich et al. 1999) The PBS containing Synechococci are unicellular cyanobacteria found widespread in the oceanic environment where they inhabit equatorial and polar waters as well as coastal and open waters (Waterbury et al., 1979). Their ability to adapt to a multitude of environmental conditions is resembled by their large diversity (Ahlgren and Rocap, 2006; Penno et al., 2006). The coccoid cells vary from a size of 0.8 µm to 1.5 µm and can reach densities of up to 200,000 cells per milliliter. Aside from that, they are also found in fresh water environments (Callieri, 2002). Synechococcus strains can be distinguished by the composition of accessory pigments, their ability to chromatically adapt and other characteristics like motility, nitrate utilization and cell cycle behavior (Binder and Chisholm, 1995; Toledo et al., 1999). Synechococcus serves as a model organism to study PBS assembly, cyanobacterial circadian clock, cyanobacterial evolution and marine diversity (Dong and Golden, 2008; Dong et al., 2010; Markson and O'Shea, 2009; Palenik et al., 2009; Schluchter et al., 2010). Synechocccus and Prochlorococcus both inhabit different parts of the ocean but also coexist in open ocean regions. Their growth and evolution is highly influenced by cyanophages, which are viruses that can infect them (Clokie and Mann, 2006; Mann, 2003; Suttle, 2007). Some cyanophages are thought to solely infect Prochlorococcus or Synechoccous, while others could be shown to crossinfect both clades (Sullivan et al., 2005; Sullivan et al., 2010; Sullivan et al., 2003; Wilson, 1993). Recent genome sequencing of several marine cyanophages identified a number of genes of cyanobacterial origin (Lindell et al., 2004; Mann et al., 2003; Sullivan et al., 2010). Besides genes involved in carbon metabolism, phosphate stress and those with other functions, many of them are related to photosynthesis. A number of these genes could be shown to encode functional

10 1 Introduction 5 proteins (Dammeyer et al., 2008a; Lindell et al., 2005). The possible advantage for marine cyanophages to carry photosynthetic genes could be to prevent photoinhibition and to keep photosynthesis rates high during infection thereby increasing phage fitness (Bragg and Chisholm, 2008; Hellweger, 2009). Among the photosynthetic phage genes some code for proteins involved in biosynthesis and assembly of PBP. 1.3 Phycobilisomes and Phycobiliproteins Phycobilisomes are phycobiliprotein containing light-harvesting complexes that allow light-capturing in the green gap where chlorophyll absorbs poorly. PBS are large rod-shaped protein structures attached to the outer thylakoid membrane (Fig. 1) that are formed in an ordered process of self assembly (Fig. 2) (Grossman et al., 1993; Sidler, 2004; Tandeau de Marsac, 2003). Phycobilisomal proteins can comprise about 50 % of the cell s protein and often account for the strong coloration of the organism. In general, PBS can structurally be divided into the core and the surrounding rods. PBSs consist of different chromophoric PBPs that directly participate in light-harvesting and non-chromophoric linker-proteins. Linker-polypeptides are present in the core and the rods to connect the PBPs to form this large protein scaffold and also influence its absorbance properties (Grossman et al., 1993; Neilson and Durnford, 2010). PBPs can be subdivided into four main classes (Table 1). The core is mainly made up by allophycocyanin (APC) while the rods can be composed of the PBPs phycoerythrin (PE), phycocyanin (PC) and phycoerythrocyanin (PEC). Table 1: Phycobiliproteins and their absorbance characteristics, adapted from (Colyer et al., 2005) Phycobiliprotein Absorbance range Phycobiliprotein color PE a 540 nm nm orange-red PEC 570 nm nm purple PC 610 nm nm blue APC 650 nm nm blue-green a PE, phycoerythrin; PEC, phycoerythrocyanin; PC, phycocyanin; APC, allophycocyanin Most phycobiliproteins form heterodimeric complexes of α- and β-subunits, referred to as the biliprotein (αβ)- monomer (Fig. 2). These monomers are associated to higher oligomeric states to form trimers (αβ) 3 or hexamers (αβ) 6. The central APC core exists as a trimer (αβ) 3 and is in direct contact with the thylakoid

11 1 Introduction 6 membrane surface. About six rods are radiating from the core. The PBS rods are composed of up to six stacked disks. Each disk is formed by trimers (αβ) 3 or hexamers (αβ) 6 of PC, PEC or PE (Grossman et al., 1993; Samsonoff and MacColl, 2001; Tandeau de Marsac, 2003). PC hexameric disks are found in the core proximal region and PE is found in the more distal regions to create an energy gradient from PE (via PEC) to PC, then to the APC core and finally to the reaction center (Table 1). Fig. 2: Oligomeric states of phycobiliproteins in phycobilisomes. Phycobiliproteins form α/βheterodimers ( monomer ). The monomers can oligomerize to (αβ) 3 trimeric structures and (αβ) 6 hexameric structures. A) Scheme of subunit oligomerization. B) The (αβ)-monomer of phycoerythrin (Gloeobacter violaceus) is shown with the α-subunit (blue) and the β-subunit (lightblue) carrying the chromophores PEB (red) and PUB (pink). A C-PC (αβ) 3 -trimer is shown in the middle (Thermosynechocccus elongatus) and a C-PC (αβ) 6 hexamer (Synechococcus elongatus) is on the right. The light absorption properties of PBP depend on their prosthetic group, the phycobilin. Cyanobacterial PBP α- and β-subunits each carry one or up to three chromophores covalently bound via a thioether bond to a conserved cysteine; (MacColl and Guard-Friar, 1987; Sidler, 2004). The chromophores found in PBS are phycocyanobilin (PCB, A max 620 nm), phycoerythrobilin (PEB,

12 1 Introduction 7 A max 540 nm), phycourobilin (PUB, A max 500 nm) and phycoviolobilin (PVB, A max 590 nm). The attachment can occur via a single or double thioether linkage at the A-ring of the bilin or at the A- and D-ring, respectively (Fig. 3) (Scheer and Zhao, 2008). PBPs differ in the nature, quantity and site of attachment of their chromophoric groups. APC usually carries one PCB at each subunit whereas the chromophore composition of the other PBP strongly varies. However, PC typically contains PCB whereas PE usually carries PEB (Sidler, 2004; Tandeau de Marsac, 2003). Fig. 3: Phycobiliprotein chromophoric groups. Phycocyanobilin (PCB) is the precursor of protein-bound PCB and phycoviolobilin (PVB), whereas phycoerythrobilin (PEB) is the precursor of protein-bound PEB and phycourobilin (PUB). The attachment usually occurs at the A-ring via a thioether bond on a conserved cysteine residue to the phycobiliprotein. PEB and PUB can also be bound at two sites in the protein via their A- and D-ring. The PBP composition within the rods of PBS depends on the organism and the light quality it is exposed to. Major PBPs are PE and PC. The latter is always present in the rods whereas not all PBS-containing organisms possess PE. The rods of Synechococcus contain PC and in most cases also PE (Six et al., 2007). PE is the main PBP in open-ocean Synechococcus strains. Two different PE s (PEΙ and PEΙΙ) are present. Synechoccus sp. WH8020 possesses PEΙ and PEΙΙ with five and six bilins, respectively. While PEB is present in both forms PEΙΙ additionally contains up to four PUB chromophores which are fewer or absent in the less abundant PEΙ (Ong and Glazer, 1991). The possession of PEB and

13 1 Introduction 8 especially PUB allows the cell to capture light efficiently in a marine environment where blue and green light is dominant (Wood, 1985). This pigment composition is typical for marine Synechococci but rather unusal compared to other cyanobacteria because of the large number of bilins per subunit. In addition, these strains possess PEB and/or PUB chromophores at linker subunits of the PBS (Wilbanks and Glazer, 1993b). So far, chromophorylated rod linker subunits have only been described for red algae as well (Glazer, 1977). Only marine Synechococcus and Synechocystis strains are known to have two constitutively expressed PEs (Wilbanks and Glazer, 1993a). In most cyanobacteria only the APC adjacent PC is constitutively expressed. In a process called complementary chromatic adapation cyanobacteria can adapt their PBP and/or chromophore composition depending on the incoming wavelengths. Some cyanobacterial species like Fremyella diplosiphon are able to adjust their PC to PE ratios depending on the prevalent light (Kehoe and Gutu, 2006). Synechococcus strain WH8020 for example is able to increase its PUB/PEB ratio in blue light (Palenik, 2001) Phycobilins Phycobilins are open-chain tetrapyrrole molecules. They are composed of four pyrrole rings named A, B, C and D linked via methine bridges (Fig. 4). The open ring structure of bilins allows them to adopt various conformations from linear to a porphyrin-like structure depending on the protein environment and solvent. Their light absorption properties and therefore the strong coloration are based on their conjugated π-electron system. Phycobilins differ in their alternating bonding pattern and substituents on ring A and D. The length of the conjugated system determines the energy required to excite the system. If for example a reduction on a double bond shortens the length of the conjugated system, more excitation energy is necessary which results in blue shifted absorption maxima (Falk, 1989). Bound to a protein, phycobilins mainly participate in light-sensing and lightharvesting. Light-sensing is mediated by a class of photoreceptors called phytochromes. These biliprotein receptors covalently bind a bilin chromophore and convert the perceived light signal into a photochemical reaction (Auldridge and Forest, 2011; Lamparter, 2004; Rockwell and Lagarias, 2010). In contrast, in the

14 1 Introduction 9 PBP the light energy absorbed by the bilin is funneled to the photosynthetic reaction centers. Fig. 4: Tetrapyrroles. Structure of a tetrapyrrole with numbering and notation of rings. 1.4 Phycobilin Attachment Both, light-sensing phytochromes and the light-harvesting PBP use covalently attached bilins. Although in both biliproteins the bilin is bound via a conserved cysteine residue the mechanism of attachment is different (Li and Lagarias, 1992; Rockwell and Lagarias, 2010; Scheer and Zhao, 2008). In general, apoproteinchromophore assembly requires accurate attachment of the chromophore to ensure presence of the proper bilin, correct stereochemistry of the bilin linkage and attachment to the right protein residue. Upon attachment of the bilin new asymmetric C-atoms are generated. Most chromophores are attached to the cysteine residue by a single thioether bond at the C3 1 position of the A-ring which results in either the R- or S- configuration (Fig. 3/5) (Lagarias et al., 1988; Scheer and Zhao, 2008). Fig. 5: Chromophore attachment at the A-ring, adapted from (Scheer and Zhao, 2008). Chromophores bind differently to the apoprotein. Attachment at the A-ring of the chromophore creates new chiral C-atoms (in grey) in either R- or S-configuration. Phycoerythrobilin (PEB) and phycocyanobilin (PCB) are attached via the C3 1 carbon to the apoprotein. Phycourobilin (PUB) and phycoviolobilin (PVB) undergo an additional isomerization.

15 1 Introduction 10 Cyanobacterial PBP assembly can either be autocatalytically or supported by specialized enzymes, the phyccobiliprotein lyases (referred to as lyases or bilin lyases in the following). Autocatalytic attachment of bilins is rare and solely described for a core linker protein (Zhao et al., 2005a). Spontaneous attachment can occur in vitro forming improper phycobiliprotein adducts (Arciero et al., 1988; Fairchild and Glazer, 1994). However, correct attachment in vivo is ensured by bilin lyases. Bilin lyases constitute a large heterogenic family of enzymes. One can distinguish between different types of lyases in cyanobacteria: E/F-type lyases, T- type lyases and S/U-type lyases (Table 2) (Biswas et al., 2010; Scheer and Zhao, 2008; Schluchter et al., 2010). Table 2: Phycobiliprotein lyases and their substrates. Different possible binding sites of phycobilins in certain phycobiliproteins mediated by different types of phycobiliprotein lyases. Type Lyase Attachment Site Bilin CpcE/CpcF (Synechococcus sp. PCC 7002) 1 Cys 84 - αpc a PCB E/F - type PecE/PecF (Mastigocladus laminosus) 2 Cys 84 - αpec PCB (PVB) RpcG (Synechococcus sp. WH 8102) 3 Cys 84 - αpc PEB (PUB) CpcT (Synechococcus sp. PCC 7002) 4 Cys βpc PCB T - type CpcT (Nostoc sp. PCC 7120) 5 Cys155 - βpc Cys βpec PCB S/(U)-type CpcS-ΙΙΙ b (Nostoc sp. PCC 7120) 6 Cys 84 - βpc Cys 84 - βpec Cys 84 -αapc Cys 84 -βapc PCB CpcS-Ι/CpcU (Synechococcus sp. PCC 7002) 7 Cys 82 -βpc Cys 81 -αapc Cys 81 -βapc PCB a PE, phycoerythrin; PC, phycocyanin; PEC, phycoerythrocyanin; APC, allophycocyanin; PEB, phycoerythrobilin; PCB, phycocyanobilin b In original publications referred to as CpeS and CpcS1 but renamed based on activity and phylogenetic evidence (Schluchter et al., 2010). 1 (Fairchild et al., 1992), 2 (Zhao et al., 2005b), 3 (Blot et al., 2009), 4 (Shen et al., 2006), 5 (Zhao et al., 2007b), 6 (Zhao et al., 2006a; Zhao et al., 2007a), 7 (Saunee et al., 2008; Shen et al., 2008) The common feature of E/F-type lyases seems to be the presence of HEAT-repeat motifs and a short region of high sequence similarity (Schluchter et al., 2010; Wilbanks and Glazer, 1993a). They act either as heterodimers like CpcE/F from Synechococcus sp. PCC 7002 and PecE/F from Mastigocladus laminosus, or as a monomeric fusion protein of E- and F-subunit like RpcG from Synechococcus sp.

16 1 Introduction 11 WH8102. E/F-type lyases seem to be specific for attachment of PCB or PEB to the Cys 84 binding site in the subunit of different PBPs. In addition, some E/F-lyases have an additional isomerase activity, which leads to the formation of PUB upon attachment of PEB or PVB upon attachment of PCB. PUB and PVB are both isomeric forms of PEB and PCB and can only be found covalently bound to their PBP but have never been isolated in their free form (Blot et al., 2009; Fairchild et al., 1992; Jung et al., 1995; Zhao et al., 2000; Zhao et al., 2002; Zhao et al., 2006b; Zhao et al., 2005b). CpcT and CpeT are paralogous T-type lyases (Schluchter et al., 2010; Shen et al., 2006). CpcT attaches PCB specifically to Cys 153 at the β-subunit of PC and acts as a monomer. Phylogenetic analyses of cpet loci suggested its involvement in PE maturation at the equivalent Cys 153 position of β-pe. S/U-type lyases comprise CpeS, CpcS, CpcU and CpcV (Biswas et al., 2010; Saunee et al., 2008; Shen et al., 2008; Zhao et al., 2006a). Heterodimerization has been shown but some members are also active as monomers. They are involved in PEC, PC and APC formation and target the Cys 82 of α- as well as β-subunits. So far, only PCB-transferring CpcS lyases are characterized. CpcSs are divided in three groups: CpcS-Ι, CpcS-ΙΙ and CpcS-ΙΙΙ. The only bilin lyase structure available is that of a CpcS-ΙΙΙ from T. elongatus BP1, which crystallized as a dimer (Schluchter et al., 2010; Shen et al., 2008). CpcS-ΙΙΙ from Anabaena sp. PCC 7120 (Nostoc sp. PCC 7120) is responsible for attachment of PCB to all subunits of APC as well as to the β-subunits of PC and PEC (Zhao et al., 2007a). The function of CpeV and CpeS has not been elucidated yet, but due to their occurrence in PE containing cyanobacteria a PEB transferring role has been proposed (Schluchter et al., 2010). The different lyases greatly differ in their specificity. Some like the T-type lyases seem to be highly specific for attaching one chromophore to one specific residue in a certain phycobiliprotein. Others like S-type lyases appear to be more universal serving various sites in different PBPs with more than one type of chromophore. In addition to their isomerase and lyase activity, lyases may also be involved in removal of the bilin chromophore (Dolganov and Grossman, 1999; Fairchild et al., 1992; Schluchter et al., 2010).

17 1 Introduction Phycobilin Biosynthesis The biosynthesis of phycobilins starts with the cleavage of the heme tetrapyrrole macrocycle catalyzed by heme oxygenase (HO) (Fig. 6). In an oxygen-dependent reaction, heme serves as a cofactor and substrate itself, activating oxygen for its own degradation. In total, the reaction requires three dioxygen molecules and seven electrons (Frankenberg-Dinkel, 2004; Montellano, 2000). In photosynthetic organisms, the reduction equivalents are retrieved from ferredoxin (Fd) but alternative electron donors like ascorbate can be employed to yield the green colored open chain tetrapyrrole BV ΙΧα (Gisk et al., 2010; Rhie and Beale, 1992). During the reaction, iron, carbon monoxide (CO) and water are released. Fig. 6: Biliverdin synthesis. Heme oxygenase catalyzes the ring opening of heme yielding the open chain tetrapyrrole biliverdin ΙΧα (BV ΙΧα). The reaction requires oxygen and electron equivalents thereby releasing iron and carbon monoxide. The ring opening can occur on different positions but most common is cleavage at the α-position yielding BV ΙΧα. If the ring opening occurs at the β-, γ-, or δ-position, BV ΙΧβ, BV ΙΧγ and BV ΙΧδ are formed. Formation of BV ΙΧα is a widespread reaction occurring in eukaryotes and prokaryotes. The fate of BV ΙΧα depends on the organism and its specific requirements (Mantle, 2002). Cleavage of cyclic tetrapyrroles serves different purposes. In mammals, heme cleavage and further reduction of BV ΙΧα to bilirubin by NAD(P)H-dependent biliverdin reductase is a degradation process to dispose of heme (Kapitulnik, 2004; Kutty and Maines, 1981; Tenhunen et al., 1970). Heme cleavage is a process of detoxification, iron reutilization and occurs during transport- and excretion processes. The byproduct of heme cleavage, CO, is an important signaling molecule in eukaryotes (Liu et al., 2010; Ryter et al., 2002). In some pathogenic bacteria this process can be used to

18 1 Introduction 13 gratify the cells demand for iron and is crucial for pathogenicity (Skaar et al., 2004; Wilks and Schmitt, 1998; Zhu et al., 2000). While a role of BV ΙΧα as an antioxidant has been discussed, the main role of BV ΙΧα in cyanobacteria is its precursor function for the light-harvesting PBP chromophores (Stocker et al., 1990). Here, BV ΙΧα is the source material for further reduction reactions to yield bilins with different absorption properties. A family of bilin-biosynthetic enzymes, the ferredoxin-dependent bilin reductases (FDBRs), are imperative for chromophore production (Frankenberg et al., 2001). 1.6 Ferredoxin-Dependent Bilin Reductases (FDBRs) FDBRs have been identified in plants, cyanobacteria and cyanophage (Dammeyer et al., 2008a; Frankenberg et al., 2001; Kohchi et al., 2001). They show relatively low sequence homology but several conserved residues are present. FDBRs do not require metal or organic cofactors and catalyze two- and four-electron reductions to yield the phycobilins (Fig. 7). Like the HO reaction, bilin reduction requires reduction equivalents, which are provided by the small redox protein Fd that is directly reduced by PSΙ (Beale and Cornejo, 1991b; Frankenberg et al., 2001). The FDBR family consists of five members: phytochromobilin synthase (HY2), phycocyanobilin:ferredoxin oxidoreductase (PcyA), 15,16- dihydrobiliverdin:ferredoxin oxidoreductase (PebA), phycoerythrobilin:ferredoxin oxidoreductase (PebB) and phycoerythrobilin synthase (PebS). FDBRs were suggested to act via a radical mechanism. Substrate radical intermediates could already be verified for two members of the family, HY2 and PcyA (Tu et al., 2008; Tu et al., 2007).

19 1 Introduction 14 Fig. 7: Biosynthesis of phycobilins. BV ΙΧα is the substrate of ferredoxin-dependent bilin reductases (FDBRs). Phytochromobilin synthase (HY2) performs a two-electron reduction to yield the plant phytochrome chromophore 3Z-phytochromobilin (3Z-PΦB). PcyA reduces BV ΙΧα to 3Zphycocyanobilin (3Z-PCB) with formation of the intermediate 18 1,18 2 -dihydrobiliverdin. 15,16- dihydrobiliverdin:ferredoxin oxidoreductase (PebA) catalyzes reduction of BV ΙΧα to 15,16- dihydrobiliverdin (15,16-DHBV), which is the substrate of phycoerythrobilin:ferredoxin oxidoreductase (PebB) to form 3Z-phycoerythrobilin (3Z-PEB). The same reaction is catalyzed by phycoerythrobilin synthase (PebS) in a formal four-electron reduction yielding 3Z-PEB. P represents the propionate side chains.

20 1 Introduction Phytochromobilin Biosynthesis In plants, the enzyme responsible for production of phytochromobilin (PΦB) is the nucleus encoded phytochromobilin synthase (HY2, EC ) (Kohchi et al., 2001; McDowell and Lagarias, 2001; Parks and Quail, 1991; Terry et al., 1995; Terry et al., 1993). It is transported into the chloroplast where it synthesizes the plant phytochrome chromophore PΦB (Fig. 7) that is then transferred into the cytosol for holo-phytochrome assembly (Kohchi et al., 2001). The ~29 kda large protein has been solely described for plants but there is indication of PΦB formation in non-photosynthetic organisms like yeast (Wu and Lagarias, 1996). HY2 catalyzes the two-electron reduction of the 2,3,3 1,3 2 -diene system on the A- ring of BV ΙΧα. This reduction of BV ΙΧα results in two possible stereoisomeric ethylidene groups, the 3Z- or 3E-isomer (Fig. 8). The reaction product is thought to be the Z- isomer of PΦB (3Z-PΦB) whereas the formation of holophytochrome Fig. 8. Stereoisomers. The reduction at the 2,3,3 1,3 2 -diene system of BV ΙΧα, 18 1,18 2 -DHBV or 15,16-DHBV results in the formation of an ethylidene-group, which can either be the 3E- or 3Zisomeric form Phycocyanobilin Biosynthesis seems to require 3E-PΦB. The presence of an isomerase was suggested but 3E-PΦB is also formed non-enzymatically (McDowell and Lagarias, 2001; Terry et al., 1995; Terry et al., 1993). Biochemical analyses suggested an aspartate (Asp 256 ) and a water molecule to act as proton donors in the HY2 reaction (Tu et al., 2008). PCB synthesis is catalyzed by phycocyanobilin:ferredoxin oxidoreductase (PcyA, ~ 28 kda, EC ) in a formal four-electron reduction (Fig. 7) (Frankenberg and Lagarias, 2003). PcyA s substrate is BV ΙΧα and the first reduction occurs at the D-ring vinyl group. The intermediate 18 1,18 2 -dihydrobiliverdin (18 1,18 2 -DHBV) is subject to a second two-electron reduction at the A-ring 2,3,3 1,3 2 -diene system forming the final product. The structure of PcyA has been solved for Synechocystis sp. PCC 6803 and Nostoc sp. PCC 7120 (Hagiwara et al., 2006a; Hagiwara et al., 2006b; Tu et al., 2007). PcyA is a single domain enzyme that exhibits an α/β/α-

21 1 Introduction 16 sandwich fold structure in which the substrate BV ΙΧα is positioned between the central β-sheet and C-terminal α-helices in a cyclic conformation (Fig. 9A). Fig. 9: Structure of PcyA from Synechocystis sp. PCC 6803, adapted from (Hagiwara et al., 2006a). A) Overall fold. The β sheets are colored yellow, α-helices in red, loop regions in blue and BV ΙΧα in green. B) Active site residues with bound BV ΙΧα. The catalytically important residues Asp 105 and His 88 coordinate the substrate from the β-sheet site. Glu 76 that is involved in D-ring reduction is positioned close to the same. His 88, Tyr 212 and Gln 216 are located near the A-ring. The substrate BV ΙΧα is placed in the binding pocket in a way that the D- and A- rings are buried inside whereas the propionate side chains on the B- and C-ring are solvent exposed. Binding of BV ΙΧα is stabilized via hydrogen bonds and salt bridges. Upon binding the imidazole side chain of His 88 forms hydrogen bonds with the carbonyl oxygen (O1) at the A-ring of the substrate and the Asp 105 coordinates the D-ring (Fig. 9B). When no substrate is bound Asp 105 and His 88 would be present as a carboxylate-imidazolium ion pair. Biochemical and structural analyses of PcyA and variants depict the following picture of the underlying mechanism of PCB synthesis (Hagiwara et al., 2010; Hagiwara et al., 2006a; Kohler et al., 2010; Stoll et al., 2009; Tu et al., 2007; Tu et al., 2006). The neutral bound BV ΙΧα will be found in lactam-lactim tautomeric states (Fig. 10A). In the major form the A-ring adopts the lactim tautomer and the D-ring is found in the lactim or lactam form comprising a bidentate interaction with

22 1 Introduction 17 Asp 105. In addition, the protonated His 88 can transfer a proton onto the Asp 105 carboxylate side chain, which in turn protonates the O19 on the D-ring generating a BV ΙΧα cation. In this scenario Asp 105 can form a carboxylate cation pair with the protonated BV ΙΧα (Fig. 10B). The positive charge allows subsequent reduction by Fd forming a neutral radical. A complex proton relay system is present in PcyA which reprotonates the His 88 via water, lysine and arginine residues that form a proton shuttle from the bulk solvent. The second proton is donated by Glu 76. Further electron transfer and rearrangements finally result in the formation of the intermediate 18 1,18 2 -DHBV. 18 1,18 2 -DHBV is not released but Fig. 10: Tautomers. A) Lactam and lactim tautomerism. B) Tautomeric forms of BV ΙΧα bound to PcyA. The substrate can adopt the major bidentate neutral conformation or a minor axial ion-pair conformation. The positive charge of the minor conformation on the B-ring can also be delocalized on the other pyrrole rings and on His 88. The different states are interconvertible as depicted by double arrows (Kohler et al., 2010) Phycoerythrobilin Biosynthesis rather is subject to the second reduction step catalyzed by PcyA. The direct proton donating residues in the first reduction of BV ΙΧα yielding 18 1,18 2 -DHBV are Asp 105 and Glu 76. His 88 is the key residue in PcyA reaction since it participates in both, A- and D- ring reduction, evident by the inability of a H88Q variant to promote BV ΙΧα as well as 18 1,18 2 -DHBV reduction and its favorable position to the O1 of the A-ring and O19 of the D- ring. For A-ring reduction, an axial water, Asp 105 and His 88 are critical. Two different pathways exist for PEB biosynthesis (Fig. 7). In cyanobacteria, two FDBRs each catalyze a two-electron reduction to form PEB from BV ΙΧα (Frankenberg et al., 2001). 15,16-DHBV:ferredoxin oxidoreductase (PebA, ~ 27 kda, EC ) catalyzes the reduction of the 15,16-methine bridge of the BV ΙΧα substrate. Its product is directly transferred to PEB:ferredoxin

23 1 Introduction 18 oxidoreductase (PebB, ~ 29 kda, EC ) in a process of metabolic channeling to serve as its substrate (Dammeyer and Frankenberg-Dinkel, 2006). Further reduction occurs at the 2,3,3 1,3 2 -diene structure at the A-ring of 15,16-DHBV yielding 3Z-PEB. This A-ring reduction is the most common in all known FDBRs and is only absent in PebA. Besides the cyanobacterial two-enzyme system of PEB synthesis, another four-electron reducing FDBR was identified in the genome of Prochlorococcus infecting cyanophage P-SSM2 (Sullivan et al., 2005). First identified as a PebA homolog, PEB synthase (PebS, ~ 26 kda, EC ) was shown to catalayze reduction of BV ΙΧα to PEB with 15,16-DHBV as the intermediate (Dammeyer et al., 2008a). The structure of PebS has been solved and revealed high similarity to PcyA (Dammeyer et al., 2008b). Analogous to PcyA, PebS is a globular single domain enzyme displaying an α/β/α-sandwich fold (Fig. 11A). Fig. 11: PebS_P-SSM2 crystal structure with bound BV ΙΧα, adapted from (Dammeyer et al., 2008b). A) Chain C of SePebS-BV (2VCK). Loops are colored blue, β-sheets in yellow and α- helices in red. B) Positioning of the substrate in the active site. BV ΙΧα is shown in green with Asp 105 and Asn 88 coordinating BV ΙΧα from the β-sheet site and Asp 206 from the α-helical site. The central water molecule and polar contacts are shown in red.

24 1 Introduction 19 A central seven-stranded antiparallel β-sheet is embedded between two α-helices on one side (H1/2) and three on the other side (H3/4/5). The substrate BV ΙΧα is located in the binding pocket formed by the β-sheet (S1-S7) and by helices H3 and H4. A long flexible loop is present between S6 and S7, which closes up upon substrate binding. Differences to PcyA lay in this flexible loop that forms two short helices in PcyA and a shorter helix H3. Like in PcyA, the propionate side chains of the substrate are solvent exposed whereas the A- and D-ring is buried in the binding pocket. From the β-sheet-side BV ΙΧα is coordinated by Asn 88 (S5) and Asp 105 (S5) and from the distal side Asp 206 is closely positioned to the substrate. Asp 206 together with Pro 207 form a linker between helices H3 and H4. Asp 206 shows high flexibility evident by different binding modes. A central water molecule is present forming a hydrogen bond network with all pyrrole nitrogens of the substrate and Asp 206 with BV ΙΧα showing a planar conformation (Fig. 11B). Asp 206 can also be positioned further inside replacing the central water molecule. Additionally Asp 206 can be rotated outside the active site causing the BV ΙΧα to adopt a more helical conformation where the A-ring is partially located above the D-ring. This helical conformation was suggested to be a likely conformation for protonation at the O19 because the oxygen at this position could be involved in hydrogen bonding with Asp 105, which could facilitate following electron transfer. An important finding is that no proton relay system will be present in PebS as proposed for PcyA. This possible proton relay would be disrupted by Asn 88, which essentially has a similar positioning as His 88 in PcyA. This could explain why PebS lacks D-ring reduction capacity (Dammeyer et al., 2008b; Hagiwara et al., 2006b; Tu et al., 2007). 1.7 Objectives of the Work Cyanobacterial PEB synthesis proceeds via a two enzyme system whereas cyanophages carry the information for one enzyme catalyzing the equivalent reaction. This work was undertaken to elucidate the underlying reaction mechanisms and structural requirements of PEB biosynthesis. Biochemical analyses of PebA, PebB and PebS and their variants as well as their crystal structures should give information about the catalytically important residues. Since the substrate coordinating residues of PebS are conserved in

25 1 Introduction 20 PebA and PebB the aim was to clarify if these residues serve the same purpose in all three enzymes. By that, the differences between the PEB biosynthesis enzymes that lead to their intrinsic regiospecificities should be investigated. Comparison of these results should give further information about common mechanisms of bilin reduction in the FDBR family. To improve the understanding of FDBR mechanism, function, distribution and evolution, the spread of FDBR like sequences and the properties of their gene products should be analyzed.

26 2 Radical Mechanism of Cyanophage Phycoerythrobilin Synthase (PebS) 21 2 Radical Mechanism of Cyanophage Phycoerythrobilin Synthase (PebS) Busch, A.W., Reijerse, E.J., Lubitz, W., Hofmann, E., Frankenberg-Dinkel, N. Biochemical Journal 433(3): (2011)

27 Biochem. J. (2011) 433, (Printed in Great Britain) doi: /bj Radical mechanism of cyanophage phycoerythrobilin synthase (PebS) Andrea W. U. BUSCH*, Edward J. REIJERSE, Wolfgang LUBITZ, Eckhard HOFMANN and Nicole FRANKENBERG-DINKEL* 1 *Physiology of Microorganisms, Faculty of Biology and Biotechnology, Ruhr-University Bochum, Bochum, Germany, Max-Planck-Institute for Bioinorganic Chemistry, Mülheim an der Ruhr, Germany, and Biophysics, Faculty of Biology and Biotechnology, Ruhr-University Bochum, Bochum, Germany PEB (phycoerythrobilin) is a pink-coloured open-chain tetrapyrrole molecule found in the cyanobacterial light-harvesting phycobilisome. Within the phycobilisome, PEB is covalently bound via thioether bonds to conserved cysteine residues of the phycobiliprotein subunits. In cyanobacteria, biosynthesis of PEB proceeds via two subsequent two-electron reductions catalysed by the FDBRs (ferredoxin-dependent bilin reductases) PebA and PebB starting from the open-chain tetrapyrrole biliverdin IXα. A new member of the FDBR family has been identified in the genome of a marine cyanophage. In contrast with the cyanobacterial enzymes, PebS (PEB synthase) from cyanophages INTRODUCTION Cyanobacteria form a large and diverse group of photoautotrophic bacteria that contribute significantly to the global primary production. They efficiently harvest light via their antenna complexes, the phycobilisomes [1]. One major chromophore incorporated into the phycobilisomes is PEB (phycoerythrobilin), a pink-coloured open-chain tetrapyrrole molecule (phycobilin). The biosynthesis of PEB starts with the oxidative cleavage of haem by haem oxygenase to yield BV (biliverdin IXα) [2]. Further reduction of BV is then catalysed by FDBRs [Fd (ferredoxin)-dependent bilin reductases] [3]. As the name implies, electrons are provided by the one-electron transferring redox protein Fd. FDBRs are a class of radical enzymes that do not possess any metal or organic co-factors [3 5]. They catalyse not only the biosynthesis of PEB, but also of the phycobilins PCB (phycocyanobilin) and P B (phytochromobilin), the chromophore of plant phytochromes (Figure 1) [3,6]. FDBRs can be subdivided into two classes depending on the number of electrons transferred to the substrate. One class catalyses two-electron reductions, the other four-electron reductions. In plants, P B synthesis proceeds via a two-electron reduction at the A-ring position catalysed by P B synthase (HY2) [7]. In cyanobacteria, PEB synthesis requires two sequential twoelectron reductions, 15,16-DHBV (15,16-dihydrobiliverdin):Fd oxidoreductase (PebA) reduces the substrate BV at the 15,16- methine bridge of the D-ring to the intermediate 15,16-DHBV, which is further reduced in a second two-electron reduction step to the final product PEB by PEB:Fd oxidoreductase (PebB) [8]. In contrast, PCB synthesis is catalysed by PCB:Fd oxidoreductase (PcyA) in a formal four-electron reduction combining 18 1, DHBV:Fd oxidoreductase activity and PCB:Fd oxidoreductase activity. Here, the D-ring exovinyl reduction yielding the intermediate 18 1,18 2 -DHBV precedes the A-ring reduction yielding the final product PCB (Figure 1) [4]. These reduction combines both two-electron reductions for PEB synthesis. In the present study we show that PebS acts via a substrate radical mechanism and that two conserved aspartate residues at position 105 and 206 are critical for stereospecific substrate protonation and conversion. On the basis of the crystal structures of both PebS mutants and presented biochemical and biophysical data, a mechanism for biliverdin IXα conversion to PEB is postulated and discussed with respect to other FDBR family members. Key words: bilin reductase, cyanophage, EPR, open-chain tetrapyrrole, phycoerythrobilin synthase, radical. steps were shown to proceed via bilin radical intermediates and similar results were obtained for the Arabidopsis thaliana FDBR HY2 [5,9,10]. A second FDBR able to catalyse a four-electron reduction was discovered in the genome of the Prochlorococcusinfecting cyanophage P-SSM2 [11]. First annotated as a PebA, the enzyme is able to catalyse the synthesis of PEB with BV as the substrate in a formal four-electron reduction with 15,16- DHBV as the intermediate. Due to its novel properties, it was named PebS (PEB synthase) [12]. PebS combines the activities of cyanobacterial PebA and PebB in one enzyme, which has so far not been found in cyanobacteria [12]. PebS is a single globular protein with an α-β-α sandwich fold and high structural similarity to PcyA, the best studied member of the FDBR family. The structure revealed amino acid residues likely to be involved in substrate conversion and possible proton transfer [13]. These residues comprise two aspartate residues at position 105 and 206 that are also highly conserved in the whole FDBR family [9]. The Asp 105 residue has been suggested to be involved in initial protonation of the BV substrate to facilitate subsequent electron transfer from Fd [13]. In the present paper we provide evidence that PebS-catalysed PEB synthesis indeed proceeds via a radical mechanism and that both aspartate residues are important for stereospecific substrate protonation and conversion. EXPERIMENTAL Materials All chemicals were American Chemical Society grade or better unless specified otherwise. All assay components were purchased from Sigma. Glutathione Sepharose TM 4FF, PreScission Protease and expression vector pgex-6p-3 were obtained from GE Healthcare. Alternatively, Protino Glutathione Agarose 4B from Macherey-Nagel was used. HPLC-grade acetone, acetonitrile, formic acid and spectroanalytical grade glycerol were obtained Biochemical Journal Abbreviations used: BV, biliverdin IXα; DVBV, dihydrobiliverdin, Fd, ferredoxin; Fd 7002, Fd from Synechococcus sp. PCC 7002; Fd P SSM2, Fd from cyanophage P-SSM2; FDBR, ferredoxin-dependent bilin reductase; FNR, ferredoxin:nadp + oxidoreductase; PCB, phycocyanobilin; PEB, phycoerythrobilin; PebS, PEB synthase; PEG, poly(ethylene glycol); P B, phytochromobilin; WT, wild-type. 1 To whom correspondence should be addressed ( nicole.frankenberg@rub.de). The structural co-ordinates reported will appear in the PDB under accession code 2 X 9I for PebS_D105N and 2 X 9J for PebS_D206N. c The Authors Journal compilation c 2011 Biochemical Society

28 470 A. W. U. Busch and others Figure 1 Bilin biosynthesis pathway The substrate BV is reduced in two- and four-electron reduction steps by the FDBRs. Synthesis of the plant chromophore P B requires a two-electron reduction by HY2. In cyanobacteria, PebA reduces BV to 15,16-DHBV which is further reduced to PEB catalysed by PebB. Phage-derived PebS catalyses the four-electron reduction from BV to PEB with 15,16-DHBV as the intermediate. PcyA also performs a formal four-electron reduction to PCB with 18 1,18 2 -DHBV as the intermediate. from J.T. Baker Inc. Sep-Pak Light cartridges were obtained from Waters. BV was obtained from Frontier Scientific, Carnforth, Lancashire, U.K. Protein expression, purification and site-directed mutagenesis PebS from cyanophage P-SSM2, Fd 7002 (Fd from Synechococcus sp. PCC 7002) or Fd P SSM2 (Fd from cyanophage P-SSM2) were recombinantly expressed and purified as described previously [12]. Purified Fd was dialysed against 25 mm Tes/KOH (ph 8.0), 100 mm KCl and 15 % glycerol and stored at 20 C. The concentration was determined as described in [12]. All site-directed mutants of PebS were generated in pgex_pebs_p- SSM2 [12] using the QuikChange site-directed mutagenesis kit (Stratagene) using the following primers (shown is only the forward primer, the reverse primer is the complement, introduced base pair changes are underlined): D105N, 5 -CTTGTTTTGGT- ATGAACCTGATGAAGTTTAGTG-3 ; D105E, 5 -CTTGTTT- TGGTATGGAACTGATGAAGTTTAGTG-3 ; D206N, 5 -CT- TATATGACTGAACTTAATCCTGTTAGAG-3 ; and D206E, 5 -CTTATATGACTGAACTTGAACCTGTTAGAGG-3. PebS mutants were expressed and purified following the method used for the WT (wild-type) protein. Preparation of bilins Preparative production of chromophores was performed under anaerobic conditions as described earlier with the following modifications [5]. FNR (Fd:NADP + oxidoreductase) from Synechococcus sp. PCC7002 and Fd P SSM2 or Fd 7002 were used in varying concentrations. PebS concentrations ranged from μm. The assay was carried out at 20 C and substrate was added sequentially in excess amounts. The reaction was started with an NADPH-regenerating system with final concentrations of 2.2 mm glucose 6-phosphate, 27 μm NADP + and 0.37 units/ml glucose-6-phosphate dehydrogenase. The reaction was stopped and the products purified according to [8]. 15,16-DHBV was prepared with PebA from Synechococcus sp. WH8020 [8] or respective PebS mutants unable to perform 15,16-DHBV reduction. 15,16-DHBV was then purified according to the previously published protocol [8]. Product formation was verified via HPLC. Spectroscopic analysis of bilin reductase activity Anaerobic bilin reductase assays were performed utilizing an Agilent 8453 UV visible spectrophotometer as described previously [5] with the following modifications. Assay conditions c The Authors Journal compilation c 2011 Biochemical Society

29 Radical mechanism of PebS 471 consisted of 25 mm Tes/KOH (ph 8.0), 100 mm KCl, 0.01 μm FNR from Synechococcus sp. PCC7002 and 1 μm Fd P SSM2, 10 μm BSA,10μM BV,10μM purified PebS WT or mutant, 50 units/ml glucose oxidase, 100 mm glucose and 5 μm catalase. To initiate catalysis, 100 μl of NADPH-regenerating system was added. The final concentration of NADPH-regenerating system contained 3.25 mm glucose 6-phosphate, 41 μm NADP + and 0.55 unit/ml glucose-6 phosphate dehydrogenase. Reaction mixtures were incubated at 17 C for 10 min (determined to be within the linear range of PebS activity). A spectrum was taken every 30 s for 10 min. Crude bilins were extracted with a Sep- Pak Light C 18 cartridge and subsequently evaporated to dryness using a SpeedVac concentrator. HPLC analysis was performed as described previously [4]. The concentration of protein and bilin was determined as described previously [8]. Freeze-quench EPR measurements For EPR measurements, the anaerobic assay described above was used with 4-fold increased concentrations of PebS, BV, FNR and Fd, and 15% spectro-analytical grade glycerol was included in the reaction mixtures. The total volume was 3 ml. At various times after addition of 200 μm NADPH, 200 μl aliquots were withdrawn from reaction mixtures, transferred to 4 mm quartz EPR tubes, and immediately frozen in liquid nitrogen. Continuous-wave EPR studies of PebS were performed using a Bruker Elexsys E500 CW X-band EPR spectrometer. EPR spectra were acquired in a standard TE102 resonator at 40 K using a microwave frequency of 9.43 GHz, a power of 20 μwandafield modulation amplitude of 10 G. The temperature of the sample was maintained using an Oxford ESR900 liquid helium flow cryostat and an Oxford ITC 503 temperature controller. Crystallization of PebS mutants in complex with BV Protein was concentrated to 7 12 mg/ml in 25 mm Tes/KOH (ph 8.0) and 100 mm KCl. Crystallization conditions were screened by the sitting drop vapour diffusion method using the Classic, Cryo, PEG and JSCG+ Suites (Qiagen), applying 200/ 100 nl and 100/100 nl mixtures of the protein solution/reservoir solution incubated at 18 C in the dark. Initial hits were refined using the hanging drop technique. For the D105N mutant, final crystals diffracting to 2.2 Å (1 Å = 0.1 nm) were obtained using a reservoir containing 170 mm (NH 4 ) 2 SO 4 and 20% PEG [poly(ethylene glycol)] Crystals were briefly soaked in mother liquor supplemented with 10 % PEG 400 before transfer into liquid N 2. Crystals for the D206N mutant were grown using a reservoir containing 170 mm (NH 4 ) 2 SO 4,25% PEG 8000 and 15% glycerol. Crystals were directly transferred into liquid N 2 and diffracted to 1.85 Å. Data collection and structure determination Oscillation data of the D105N mutant were collected at 100 K at the ESRF (European Synchrotron Radiation Source, Grenoble, France) on beamline ID23.2. Data of D206N mutant crystals were collected at 100 K at the SLS (Swiss Light Source, Villigen, Switzerland) on beamline X10SA. All data were processed and scaled using the XDS package [14]. Data statistics are given in Supplementary Table S1 (at Both structures were solved by molecular replacement in Molrep using the structure of PebS with bound BV (PDB code 2VGR) as the search model. The models were improved by iterating manual rebuilding in Coots [15] and refinement using Refmac [16] for the initial cycle, and phenix.refine [17] for the final cycles. Model and refinement statistics are given in Supplementary Table S1. Non-crystallographic symmetry restraints were used throughout refinement and were only released for variable parts of the structures. RESULTS Anaerobic bilin reduction In order to get a deeper insight into the PebS reaction mechanism, the anaerobic assay system described for other FDBRs was employed. In contrast with an aerobic assay, this system allows the detection of possible bilin radical intermediates during the enzymatic reduction of BV and 15,16-DHBV by PebS and mutants. The PebS reaction was started by the addition of reducing equivalents and resulted in a fast increase of absorbance at 460 nm and nm. Concomitantly, a decrease of BV absorption at 680 nm was observed (Figure 2A, WT). Absorption at the indicated wavelengths is probably due to the formation of bilin radical intermediates as shown for other FDBRs [5,9,10]. Although the absorption of the putative radical intermediate(s) decreased during the course of the reaction, formation of the product PEB was observed at 540 nm. Furthermore, the semireduced intermediate 15,16-DHBV becomes temporarily visible at 600 nm but disappears as the reaction proceeds (Figure 2A). The final product of the PebS reaction is 3Z-PEB as confirmed by HPLC analysis (Figure 2B). Asp 105 and Asp 206 are critical protonating residues On the basis of the PebS crystal structure, the two conserved amino acid residues Asp 105 and Asp 206 were proposed to be crucial for PebS function, as their carboxy groups are in hydrogen bonding distance to the substrate [13], and both residues are conserved in many members of the FDBR family. We therefore generated PebS mutants in which the carboxy group was changed to the corresponding amide (D105N and D206N), or in which the aspartate residue was replaced by the more extended glutamate residue (D105E and D206E). In our anaerobic assay, incubation of D105N with BV and an excess of electron equivalents resulted in an absorption increase at 460 nm and nm similar to WT, but with a very slow decay. These results suggest the formation of bilin radical intermediates and their stabilization/accumulation by the mutant (Figure 2A). Subsequent HPLC analyses revealed that this mutant is unable to convert the substrate BV (Figure 2B). Interestingly, the D105E mutant fully retained the ability to catalyse the first reduction at the 15,16- methine bridge, but could not catalyse the second reduction at the A-ring 2,3,3 1,3 2 -diene system, thereby yielding 15,16-DHBV as the final product (Figure 2). Investigation of PebS_D206N with BV as the substrate showed a decrease of BV absorption at 680 nm with subsequent increase and decrease of a possible bilin radical absorption at 460 nm and nm. HPLC analyses of this mutant demonstrated that only the first reduction step was performed. The final product of this reaction with a maximum absorption at 605 nm is 15,16-DHBV (Figure 2). The same product was obtained when Asp 206 was substituted by a glutamic acid residue (results not shown). However, in this mutant the reaction can be pushed further to produce PEB by a 10-fold increase in FNR concentration (Figure 2). c The Authors Journal compilation c 2011 Biochemical Society

30 472 A. W. U. Busch and others Figure 3 EPR measurement of bilin radicals occurring during reactions of PebS with BV EPR spectroscopic measurement of BV complexes of PebS_WT, PebS_D105N and PebS_D206N mutant. Samples were withdrawn before addition of NADPH and 0.5 min, 2.5 min, 4.5 min and 20 min after the reaction was started and were immediately frozen in liquid nitrogen. All spectra are on the same scale and were recorded at 40 K, 20 μw power and 10 Gauss field modulation (see text for further details). Figure 2 Absorption spectra under anaerobic conditions (A) and HPLC analyses (B) of PebS_WT and mutants PebS BV complex (10 μm) was used and absorption spectra were monitored for 10 min every 30 s. The reaction was started by addition of an NADPH-regenerating system. The reaction was stopped and the products analysed via HPLC at 380 nm (solid line) and 560 nm (dashed line). (A) Time course of substrate conversion. BV absorption due to reduction decreases at 680 nm, 15,16-DHBV formation results in an absorption increase at 605 nm, PEB increases with an absorption at 540 nm. Absorption of 460 nm, 740 nm and 760 nm increases, with further decrease for PebS_WT and PebS_D206N/E and PebS_D105E indicated by double arrows, and is stable in the case of PebS_D105N. (B) HPLC analyses were conducted with C 18 reverse-phase columns from Phenomenex. The Ultracarb 5 μ column was used for all mutants and WT, except for PebS_D105E and PebS_D206E, where the Luna 5 μ column was used. Retention times for the two columns slightly differ. The reaction products DHBV and PEB were detected at 560 nm. BV was detected at 380 nm. Products were confirmed by retention times of bilin standards and whole spectrum analysis of elution peaks. Asterisks indicate non-specific degradation products. Catalytic turnover was also tested with the intermediate of the reaction, 15,16-DHBV. As expected, none of the asparagine residue mutants was able to convert 15,16-DHBV, whereas the WT protein catalysed the formation of PEB (indicated by an absorption increase at 540 nm; results not shown). Interestingly, UV visible spectra of the catalytic turnover of these mutants with 15,16-DHBV as the substrate showed no indication of absorbance that could point to the formation of a 15,16-DHBV radical intermediate (results not shown). In this regard, it has to be noted that externally applied 15,16-DHBV to D206N revealed a 60 nm spectral shift to the shorter wavelength of the UV visible spectrum as compared with the enzymatically produced and enzyme bound 15,16-DHBV (see Supplementary Figure S1 at Bilin radical intermediates during PebS reaction Since the performed anaerobic assays suggested the formation of bilin radical intermediates and their accumulation in the D105N mutant, additional anaerobic bilin reductase assays were performed using an excess of NADPH as the reducing agent and 40 μm of the PebS BV complex. The reactions were monitored via UV visible spectroscopy (see Supplementary Figure S2 at and EPR spectroscopy (Figure 3). For the latter, aliquots were taken at different time points and flash frozen for further EPR measurements. Paramagnetic species could be detected for PebS_WT and the mutants D105N and D206N, indicated by an isotropic EPR signal at g 2 with a first derivative peak to peak linewidth of approx. 15 G (Figure 3). The PebS_WT protein incubated with BV showed a fast increase and decay of the detectable radical with the strongest signal being observed after 0.5 min. The PebS_D105N mutant generated a relatively stable paramagnetic intermediate as shown by an increasing EPR signal with no significant decrease (Figure 3, middle panel). The PebS_D206N mutant showed a similar behaviour of radical signal increase and decrease in the course of the reaction as observed for the WT reaction. At the end of the reaction at 20 min, a weak signal was still detectable either due to weak radical stabilization or an uncompleted reaction (Figure 3, right-hand panel). Absorption changes at the long and short wavelength ( 460 nm, nm), which do not correspond to the substrate (BV) or product (DHBV, PEB) absorption maxima followed similar kinetics as the EPR signal (Figure 4). For comparison, the time point with the highest EPR intensity has been assigned as 100 %. Since the EPR linewidth of the radical signal was invariant over the course of the experiment, the EPR c The Authors Journal compilation c 2011 Biochemical Society

31 Radical mechanism of PebS 473 Figure 4 Correlation of EPR signal and absorption at 760 nm EPR signal intensity (black bar) was calculated from the first derivative peak to peak amplitude and compared with the absorption change at 760 nm (white bar) during the reaction with BV. The time point of strongest EPR signal was set to 100 %. Relative absorbance at 760 nm and EPR signal intensity were normalized to the time point at 0.5 min (PebS_WT) and 2.5 min (PebS_D105N and PebS_D206N). Figure 5 Superposition of the active site of the WT, the D105N (A) and D206N (B) mutant with bound substrate BV The WT is shown in grey and the mutants in red (D105N) and blue (D206N). Only the residues at position 88, 105 and 206 are shown. The hydrogen bonding network is only shown for the mutants. signal intensities were taken as the peak to peak amplitudes of the first derivative signals as presented in Figure 3. Relative EPR signal intensity and absorbance at 760 nm were normalized to the amplitude of the signal at 0.5 min for the PebS_WT reaction and at 2.5 min for the reaction of the two mutants. As shown in Figure 4, the EPR intensity followed very similar kinetics as the absorbance change at 760 nm. Therefore, absorption at this wavelength can most probably be attributed to the formation of bilin radical intermediates. These results confirm a radical mechanism of BV reduction catalysed by PebS. Unfortunately, similar experiments employing 15,16-DHBV as a substrate to detect paramagnetic species of the second reduction catalysed by PebS were technically not feasable due to the high amount of 15,16-DHBV required for the analysis. Structure of the PebS mutants D105N and D206N In addition to the biochemical evidence that both aspartate residues are critical for catalysis, the respective asparagine mutants were crystallized to obtain structural information to help elucidate the underlying catalytic mechanism. X-ray data of PebS_D105N and PebS_D206N with bound BV have been collected. The structure of PebS_D105N (PDB code 2X9I) was refined at 2.2 Å resolution and PebS_D206N (PDB code 2X9J) at 1.85 Å. Both mutants show the identical overall fold as observed for PebS_WT [RMSD (root mean square deviation) 0.27 Å for approx. 190 Cα atoms for all observed domains] (see Supplementary Figure S3 at bj/433/bj add.htm). In all PebS structures, the substrate BV is bound in a pocket parallel to the central β-sheet with the propionate side chains facing the solvent. As already observed for the WT structures, the amino acid residues located on the proximal β-sheet side define the rigid part of the active site. In contrast, residues on the distal side are more flexible (Figure 5 and Supplementary Figure S4 at Accordingly, residue Asp 105 and the respective Asn 105 mutant superimpose perfectly (Figure 5A and Supplementary Figures S4A and S4B). Asp 206, on the other hand, has been found in three different conformational modes in our PebS_WT structures. In one conformation, Asp 206 faces the solvent ( out ). In the second dominant conformation, Asp 206 co-ordinates the substrate indirectly via a central pyrrole water ( in ). In the third conformation observed in the WT enzyme, Asp 206 directly coordinates the pyrrole nitrogens and effectively displaces the pyrrole water ( deep ). Using this nomenclature, both copies of the D206N mutant in the asymmetric unit are in the in conformation. In the D105N structure, we find four copies in the asymmetric unit, two of which have Asp 206 in the deep conformation and two in the in conformation (see Supplementary Table S2 at In one of the latter we did not find electron density to support the c The Authors Journal compilation c 2011 Biochemical Society

32 474 A. W. U. Busch and others placement of the pyrrole water. In both PebS mutants, BV is found exclusively in the porphyrin-like planar form. DISCUSSION PebS is a phage-derived FDBR involved in the biosynthesis of PEB, one of the main chromophores in cyanobacterial lightharvesting complexes. So far, PebS activity is solely found in the enzyme derived from the marine cyanophage P-SSM2 infecting low-light-adapted Prochlorococus strains [11]. Putative PebS orthologues have been identified in sequenced cyanophage genomes and in metagenome-derived sequences of phage origin only [12,18]. In a first two-electron reduction step, PebS regiospecifically reduces the 15,16-methine bridge of BV forming the intermediate 15,16-DHBV. 15,16-DHBV is further reduced to PEB at the A-ring 2,3,3 1,3 2 -diene structure in a second twoelectron reduction step. Both reduction steps generate two new chiral C-atoms at position C-2 and C-16, both of which are in the R configuration [19]. This second reduction step also formally resembles the second reduction catalysed by PcyA and the reaction catalysed by HY2 as they all target the same structure at the tetrapyrrole s A-ring but use different substrates [15, 16-DHBV (PebS), 18 1,18 2 -DHBV (PcyA) and BV (HY2)]. The sole product of the PebS reaction is 3Z-PEB Improvement of our assay work-up conditions enabled us to ultimately prove that the sole product of the PebS reaction is the 3Z-isomer of PEB (Figure 1). Although we previously had indications from UV visible spectroscopy that this is the case [13], HPLC analyses now clearly show that 3Z-PEB is the final product (Figure 2B). Therefore, the occurrence of the 3Eisomer is simply due to experimental artefacts such as assay work-up conditions at slightly higher temperatures [3]. Further studies from our laboratory furthermore suggest similar results for PcyA, where the 3Z-isomer of PCB is the sole product (A.W.U. Busch and N. Frankenberg-Dinkel, unpublished results). Earlier results from phytochromobilin synthase from oats and other FDBRs from Cyanidium caldariorum also confirmed the production of the 3Z-isomer [20 22]. Therefore, we propose that all FDBRs only produce the 3Z-isomer. PebS acts via a radical mechanism Since one common feature of all FDBRs is the lack of any metal or organic cofactors, the presence of bilin radical intermediates was proposed [3]. Radical species could already be confirmed for PcyA [5] and HY2 [9]. In the present study, we show that PebS also acts via bilin radical intermediates which makes them a specific feature of this family of enzymes [5,9]. Employing the anaerobic assay system in combination with UV visible spectroscopy, these radical species can be observed by their distinct absorption bands in the long and short wavelength range. During the course of the PebS reaction, the observed signal emerges from different radical species that cannot be distinguished here since the reaction probably proceeds via a concerted proton-coupled electron transfer. Asp 105 and Asp 206 are both proton donating residues in the PebS reaction A strong accumulation of radical intermediate was observed for the mutant D105N. Since this mutant is unable to catalyse the conversion of BV, it is catalytically stalled in one of the protonations of the first reduction. The observed radical is therefore most likely one of the two radicals occurring in the first two-electron reduction of BV to 15,16-DHBV. This finding is in agreement with studies on PcyA where the homologous amino acid exchange leads to a loss of activity and also to BV radical stabilization [23]. In contrast, D206N is still able to convert BV to the intermediate 15,16-DHBV and therefore is only essential for A-ring reduction. This is consistent with observations for plant P B synthase (HY2) which catalyses the two-electron reduction of the BV A-ring to P B. For HY2, the Asp 206 homologue Asp 256 has been found to be essential for BV reduction as well [9]. Although PcyA also contains an aspartate residue at this position, its mutation to asparagine only has a minimal effect on activity [23]. Therefore, it is not surprising that this residue is found in the out position in the PcyA crystal structure [24]. This conserved residue seems to play a special role in both A-ring reductions of HY2 and PebS but not in PcyA. In this regard, PcyA is the only member of the FDBR family where this amino acid residue is not strictly conserved (see Supplementary Figure S5 at Alternative residues are found in PcyA of Prochlorococcus marinus MED4 (glutamic acid) and the cyanophage P-SSM4 (lysine). Both enzymes were shown to be catalytically active and produce PCB [12,25]. Another fundamental difference between the PcyA and PebS structures is the existence of a continuous proton uptake channel (proton relay) leading to a conserved His 88 residue in PcyA. This channel plays a central role in the PcyA reaction scheme and has been proposed to be important for reprotonation of Asp 105 [10]. Therefore, although we find certain structural features to be conserved within the whole FDBR family, no common catalytic mechanism can be proposed. Proposed catalytic mechanism of PebS On the basis of our biochemical, biophysical and structural data we propose the following catalytic mechanism for PebS (Figure 6A). The bound bilins will always be present as a mixture of tautomers (lactim lactam) regarding the location of the protons. We envisage the first event of the reaction being a protonation of a mono-lactim BV (in the ACδ configuration; using the nomenclature in [26]) through the pyrrole water (H 2 O). The resulting monohydrobiliverdin cation will take up an electron from Fd and generate a neutral monohydrobiliverdin radical (ACDδ). Residue Asp 105 will be important for the next concerted proton-electron transfer and also for directing the proton stereospecifically to C-16 to generate the R configuration at this carbon atom. We envisage that the latter might be possible through a Asp 105 -mediated stereospecific tautomerization (indicated by a black asterisk in Figure 6A) to yield 15,16- DHBV. We expect that PebA will work in a similar fashion with Asp 105 being the critical proton-donating residue. This scheme would also suggest that a D105N mutant accumulates a neutral monohydrobiliverdin radical. The nature of this radical will be further investigated by future high-field EPR measurements. Asp 206 has been found to be flexible (ranging from an out to a deep in conformation in direct contact with the BV) and to be essential for the second reduction step of PebS. Including this information in our mechanism we propose an initial protonation of a 15,16-DHBV monolactim (αcd) from Asp 206 to generate a trihydrobiliverdin cation, concurrent electron transfer from Fd will result in a neutral radical (αbcd). Next would be the stereospecific protonation of C-2. Again, Asp 105 couldplaya crucial role in this step. The still deprotonated Asp 105 could c The Authors Journal compilation c 2011 Biochemical Society

33 Radical mechanism of PebS 475 Figure 6 PebS-catalysed reduction of BV (A) Proposed mechanism. Black stars indicate stereospecific reductions. See text for details. (B) Illustration of the used nomenclature of protonation sites in the substrate BV. The pyrrole nitrogens are denoted A, B, C and D, whereas the two carbonyl-oxygens on the A- and D-ring are denoted α and δ. Nomenclature was adapted from [26]. catalyse a stereospecific tautomerization, yielding a neutral lactam radical (BCD). The final step could be a concerted proton-electron transfer followed by a final tautomerization. The exact nature of which is still unknown but could possibly again involve Asp 206 and water molecules. We would envisage that the homologous aspartic acid residues Asp 107 and Asp 231 of PebB from Synechococcus sp. WH8020 are also crucial for the PebB reaction. However, the initial protonation state of Asp 107 in PebB is currently unknown and is the subject of current investigation in our laboratory. One has to note that due to the reduced extent of the conjugated π-system, the reaction intermediate 15,16-DHBV will adopt a different conformation in the active site than the BV found in our ground state structures. In addition, the D105E mutant, which retains the catalytically critical carboxy group, can still catalyse the first, but not the second, reduction. The longer side chain changes the geometry of the active site, thereby hindering the productive positioning of the reaction intermediate 15,16- DHBV. Clearly more work is needed to clarify the structural changes involved in these intermediates, both by analysing the two glutamate mutants, and the accumulated radicals of PebS. In summary, the following points would be fundamental for the proposed mechanism: (i) electron transfer from Fd is directly coupled to proton transfer from water or the protein; (ii) tautomerization of the different protonation states of the substrate occurs constantly; (iii) stereospecific reduction in both steps is enforced by Asp 105 ; and (iv) flexibility of Asp 206 facilitates water release and re-protonation. AUTHOR CONTRIBUTION AndreaBuschandNicoleFrankenberg-Dinkel designedthestudy; AndreaBuschperformed all the experiments; Andrea Busch and Edward Reijerse performed the EPR experiment; Andrea Busch, Edward Reijerse and Wolfgang Lubitz analysed the EPR data; Eckhard Hofmann solved the crystal structures; Andrea Busch, Eckhard Hofmann and Nicole Frankenberg-Dinkel analysed all biochemical data; Andrea Busch, Eckhard Hofmann and Nicole Frankenberg-Dinkel wrote the paper. ACKNOWLEDGEMENTS We thank the beamline staff at the ESRF and the SLS and colleagues from the Max-Planck Institute for Physiology (Dortmund, Germany) for help during data collection. Thanks are due to Dr Shih-Long Tu for his help in setting up the anaerobic assay system in our laboratory and to Dr Jessica Wiethaus for helpful discussion. FUNDING This work was financially supported by the SFB 480 (Teilprojekt C6 and C8) from the Deutsche Forschungsgemeinschaft (to N.F.D. and E.H.). A.B. received a PhD fellowship from the Ruhr-University Bochum Research School. REFERENCES 1 Glazer, A. N. (1985) Light harvesting by phycobilisomes. Ann. Rev. Biophys. Biophys. Chem. 14, Cornejo, J., Willows, R. D. and Beale, S. I. (1998) Phytobilin biosynthesis: cloning and expression of a gene encoding soluble ferredoxin-dependent heme oxygenase from Synechocystis sp. PCC Plant J. 15, c The Authors Journal compilation c 2011 Biochemical Society

34 476 A. W. U. Busch and others 3 Frankenberg, N., Mukougawa, K., Kohchi, T. and Lagarias, J. C. (2001) Functional genomic analysis of the HY2 family of ferredoxin-dependent bilin reductases from oxygenic photosynthetic organisms. Plant Cell 13, Frankenberg, N. and Lagarias, J. C. (2003) Phycocyanobilin:ferredoxin oxidoreductase of Anabaena sp. PCC Biochemical and spectroscopic. J. Biol. Chem. 278, Tu, S. L., Gunn, A., Toney, M. D., Britt, R. D. and Lagarias, J. C. (2004) Biliverdin reduction by cyanobacterial phycocyanobilin:ferredoxin oxidoreductase (PcyA) proceeds via linear tetrapyrrole radical intermediates. J. Am. Chem. Soc. 126, Dammeyer, T. and Frankenberg-Dinkel, N. (2008) Function and distribution of bilin biosynthesis enzymes in photosynthetic organisms. Photochem. Photobiol. Sci. 7, Kohchi, T., Mukougawa, K., Frankenberg, N., Masuda, M., Yokota, A. and Lagarias, J. C. (2001) The Arabidopsis HY2 gene encodes phytochromobilin synthase, a ferredoxin-dependent biliverdin reductase. Plant Cell 13, Dammeyer, T. and Frankenberg-Dinkel, N. (2006) Insights into phycoerythrobilin biosynthesis point toward metabolic channeling. J. Biol. Chem. 281, Tu, S. L., Chen, H. C. and Ku, L. W. (2008) Mechanistic studies of the phytochromobilin synthase HY2 from Arabidopsis.J.Biol.Chem.283, Tu, S. L., Rockwell, N. C., Lagarias, J. C. and Fisher, A. J. (2007) Insight into the radical mechanism of phycocyanobilin-ferredoxin oxidoreductase (PcyA) revealed by X-ray crystallography and biochemical measurements. Biochemistry 46, Sullivan, M. B., Coleman, M. L., Weigele, P., Rohwer, F. and Chisholm, S. W. (2005) Three Prochlorococcus cyanophage genomes: signature features and ecological interpretations. PLoS Biol. 3, e Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W. and Frankenberg-Dinkel, N. (2008) Efficient phage-mediated pigment biosynthesis in oceanic cyanobacteria. Curr. Biol. 18, Dammeyer, T., Hofmann, E. and Frankenberg-Dinkel, N. (2008) Phycoerythrobilin synthase (PebS) of a marine virus: crystal structures of the biliverdin complex and the substrate-free form. J. Biol. Chem. 283, Kabsch, W. (1993) Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26, Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr. Sect. D Biol. Crystallogr. 60, Murshudov, G. N., Vagin, A. A. and Dodson, E. J. (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. Sect. D Biol. Crystallogr. 53, Adams, P. D., Grosse-Kunstleve, R. W., Hung, L.-W., Ioerger, T. R., McCoy, A. J., Moriarty, N. W., Read, R. J., Sacchettini, J. C., Sauter, N. K. and Terwilliger, T. C. (2002) PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr. Sect. D Biol. Crystallogr. 58, Sullivan, M. B., Huang, K. H., Ignacio-Espinoza, J. C., Berlin, A. M., Kelly, L., Weigele, P. R., Defrancesco, A. S., Kern, S. E., Thompson, L. R., Young, S. et al. (2010) Genomic analysis of oceanic cyanobacterial myoviruses compared with T4-like myoviruses from diverse hosts and environments. Environ. Microbiol. 12, Gossauer, A. and Weller, J. P. (1978) Synthesis of bile pigments. 9. Chemical total synthesis of (+)-(2R,16R)- and (+)-(2S,16R)-phycoerythrobilin dimethyl ester. J. Am. Chem. Soc. 100, Beale, S. I. and Cornejo, J. (1984) Enzymic transformation of biliverdin to phycocyanobilin by extracts of the unicellular red alga Cyanidium caldarium.plant Physiol. 76, Beale, S. I. and Cornejo, J. (1991) Biosynthesis of phycobilins. 3(Z)-Phycoerythrobilin and 3(Z)-phycocyanobilin are intermediates in the formation of 3(E)-phycocyanobilin from biliverdin IXa. J. Biol. Chem. 266, McDowell, M. T. and Lagarias, J. C. (2001) Purification and biochemical properties of phytochromobilin synthase from etiolated oat seedlings. Plant Physiol. 126, Tu, S. L., Sughrue, W., Britt, R. D. and Lagarias, J. C. (2006) A conserved histidine-aspartate pair is required for exovinyl reduction of biliverdin by a cyanobacterial phycocyanobilin:ferredoxin oxidoreductase. J. Biol. Chem. 281, Hagiwara, Y., Sugishima, M., Takahashi, Y. and Fukuyama, K. (2006) Crystal structure of phycocyanobilin:ferredoxin oxidoreductase in complex with biliverdin IXa, a key enzyme in the biosynthesis of phycocyanobilin. Proc. Natl. Acad. Sci. U.S.A. 103, Dammeyer, T., Michaelsen, K. and Frankenberg-Dinkel, N. (2007) Biosynthesis of open-chain tetrapyrroles in Prochlorococcus marinus. FEMS Microbiol. Lett. 271, Stoll, S., Gunn, A., Brynda, M., Sughrue, W., Kohler, A. C., Ozarowski, A., Fisher, A. J., Lagarias, J. C. and Britt, R. D. (2009) Structure of the biliverdin radical intermediate in phycocyanobilin:ferredoxin oxidoreductase identified by high-field EPR and DFT. J. Am. Chem. Soc. 131, Received 5 October 2010/1 November 2010; accepted 4 November 2010 Published as BJ Immediate Publication 4 November 2010, doi: /bj c The Authors Journal compilation c 2011 Biochemical Society

35 Biochem. J. (2011) 433, (Printed in Great Britain) doi: /bj SUPPLEMENTARY ONLINE DATA Radical mechanism of cyanophage phycoerythrobilin synthase (PebS) Andrea W. U. BUSCH*, Edward J. REIJERSE, Wolfgang LUBITZ, Eckhard HOFMANN and Nicole FRANKENBERG-DINKEL* 1 *Physiology of Microorganisms, Faculty of Biology and Biotechnology, Ruhr-University Bochum, Bochum, Germany, Max-Planck-Institute for Bioinorganic Chemistry, Mülheim an der Ruhr, Germany, and Biophysics, Faculty of Biology and Biotechnology, Ruhr-University Bochum, Bochum, Germany Figure S1 Binding of DHBV to PebS_WT and variants Protein and DHBV were incubated in equimolar amounts on ice and the spectrum was taken (A C). Equimolar amounts of BV and D206N were used in the anaerobic assay. At the end of the reaction, a spectrum was taken with the bound product DHBV (D). Figure S2 BV reduction time courses monitored by absorption spectroscopy prior to EPR analysis Shown are UV visible spectra which correspond to the EPR spectra shown in Figure 3 of the main text. Samples for EPR measurements were withdrawn at 0, 0.5, 2.5, 4.5 and 20 min. 1 To whom correspondence should be addressed ( nicole.frankenberg@rub.de). The structural co-ordinates reported will appear in the Protein Data Bank under accession code 2X9I for PebS_D105N and 2X9J for PebS_D206N. c The Authors Journal compilation c 2011 Biochemical Society

36 A. W. U. Busch and others Table S1 Data collection and refinement statistics Data in parentheses represent values in the highest resolution bin. For the definition of R meas and R mrgd F see [1]. R free is calculated from 5% of data in thin shells omitted from refinement. The Ramachadran plot was calculated with MolProbity [2]. Data set PebS_D105N BV PebS_D206N BV Figure S3 PebS overlay Overlay of PebS_WT (grey), PebS_D105N (red) and PebS_D206N (blue) with substrate BV. Beamline SLS PXII ESRF ID23.2 Resolution (Å) ( ) ( ) Cell parameters a,b,c (Å) 52.92, 70.14, , 70.04, α,β,γ ( ) 109.4, 109.2, , 70.74, Spacegroup P1 P21 Wavelength Completeness (%) 96.6 (96.6) 99.8 (99.7) Multiplicity 2.7 (2.8) 5.9 (5.9) Average I/σ I 10.6 (3.7) 16.2 (4.3) R sym (%) 7.1 (29.2) 7.5 (45.2) R meas (%) 8.8 (36.6) 8.3 (49.6) R mrgd F (%) 11.6 (46.0) 8.6 (39.0) Refinement PDB code 2X9I 2X9J Resolution (Å) ( ) ( ) R cryst (%) 17.6 (21.0) (26.1) R free (%) 21.2 (n/a) 28.5 (n/a) Average B-factor Protein Biliverdin IXα Water SO Number of atoms Protein Biliverdin IXα Water/SO 4 349/10 429/- Number of side-chains with alternate 0 2 conformations Ramachandran plot Favoured 781 (96.7%) 413 (96.9 %) Allowed 25 (3.1%) 13 (3.1%) Disallowed 2 (0.2%) 0 (0%) Table S2 Active site conformations Conformational modes of Asp 206 or Asn 206 in the different crystal structures of PebS and PebS variants. Protein Chain Asp/Asn 206 Pyrrole H 2 O BV conformation PebS 5A In Yes Planar 5B In Yes Planar 5C In Yes Planar 5D Deep No Planar 8A Out No Helical 8B In Yes Planar 8C Out No Helical 8D In Yes Planar D105N A In Yes Planar B Deep No Planar C Deep No Planar D In No Planar D206N A In Yes Planar B In Yes Planar c The Authors Journal compilation c 2011 Biochemical Society

37 Radical mechanism of PebS Figure S4 Substrate binding pockets Stereoview of the active site of PebS structures with bound BV. Shown are PebS_WT (PDB code 2VCK, chain C) (A), PebS_D105N BV, chain A (B) and chain B (C) and PebS_D206N BV, chain A (C). BV (green) is orientated with the D-ring left and the A-ring right. BV and active site residues (salmon) in the 3.6 Å sphere of BV are represented as stick models. Ordered water molecules in the active site are shown as red spheres. Figures were prepared using PyMOL (DeLano Scientific; c The Authors Journal compilation c 2011 Biochemical Society

38 A. W. U. Busch and others Figure S5 ClustalW-based structural alignment of length sequences of various members of the FDBR family Sequences were selected for which biochemical and/or structural information is available [3 9]. Secondary structure assignment was based on PDB code 2VCK for PebS and PDB code 2D1E [7] for PcyA. The Figure was generated using ESPript and amino acids are marked by the strict function set to 0.66 [10]. ARATH, Arabidopsis thaliana; HY2, phytochromobilin synthase; PebS, phycoerythrobilin synthase; PebA, 15,16-dihydrobiliverdin:ferredoxin oxidoreductase; PcyA, phycocyanobilin: ferredoxin oxidoreductase; P-SSM2, cyanophage P-SSM2; P-SSM4, cyanophage P-SSM4; FREDI, Fremyella diplosiphon; Syn_WH8020, Synechococcus sp. WH8020; Nostoc, Nostoc sp. PCC7120; Syn_PCC6803, Synechocystis sp. PCC6803. c The Authors Journal compilation c 2011 Biochemical Society

39 Radical mechanism of PebS REFERENCES 1 Diederichs, K. and Karplus, P. A. (1997) Improved R-factors for diffraction data analysis in macromolecular crystallography. Nat. Struct. Biol. 4, Davis, I. W., Leaver-Fay, A., Chen, V. B., Block, J. N., Kapral, G. J., Wang, X., Murray, L. W., Arendall, 3rd, W. B., Snoeyink, J., Richardson, J. S. and Richardson, D. C. (2007) MolProbity: all-atom contacts and structure validation for proteins and nucleic acids. Nucleic Acids Res. 35, W375 W383 3 Dammeyer, T., Michaelsen, K. and Frankenberg-Dinkel, N. (2007) Biosynthesis of open-chain tetrapyrroles in Prochlorococcus marinus. FEMS Microbiol. Lett. 271, Dammeyer, T. and Frankenberg-Dinkel, N. (2006) Insights into phycoerythrobilin biosynthesis point toward metabolic channeling. J. Biol. Chem. 281, Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W. and Frankenberg-Dinkel, N. (2008) Efficient phage-mediated pigment biosynthesis in oceanic cyanobacteria. Curr. Biol. 18, Dammeyer, T., Hofmann, E. and Frankenberg-Dinkel, N. (2008) Phycoerythrobilin synthase (PebS) of a marine virus: crystal structures of the biliverdin complex and the substrate-free form. J. Biol. Chem. 283, Hagiwara, Y., Sugishima, M., Takahashi, Y. and Fukuyama, K. (2006) Crystal structure of phycocyanobilin:ferredoxin oxidoreductase in complex with biliverdin IXα, a key enzyme in the biosynthesis of phycocyanobilin. Proc. Natl. Acad. Sci. U.S.A. 103, Tu, S. L., Rockwell, N. C., Lagarias, J. C. and Fisher, A. J. (2007) Insight into the radical mechanism of phycocyanobilin-ferredoxin oxidoreductase (PcyA) revealed by X-ray crystallography and biochemical measurements. Biochemistry 46, Tu, S. L., Chen, H. C. and Ku, L. W. (2008) Mechanistic studies of the hytochromobilin synthase HY2 from Arabidopsis. J. Biol. Chem. 283, Gouet, P., Robert, X. and Courcelle, E. (2003) ESPript/ENDscript: extracting and rendering sequence and 3D information from atomic structures of proteins. Nucleic Acids Res. 31, Received 5 October 2010/1 November 2010; accepted 4 November 2010 Published as BJ Immediate Publication 4 November 2010, doi: /bj c The Authors Journal compilation c 2011 Biochemical Society

40 3 Structural and Mechanistic Insights into the Ferredoxin-Mediated 35 Two Electron Reduction of Bilins 3 Structural and Mechanistic Insights into the Ferredoxin- Mediated Two Electron Reduction of Bilins Busch, A.W., Reijerse, E.J., Lubitz, W., Frankenberg-Dinkel, N., Hofmann, E. (Manuscript submitted to Journal of Biological Chemistry)

41 STRUCTURAL AND MECHANISTIC INSIGHTS INTO THE FERREDOXIN-MEDIATED TWO ELECTRON REDUCTION OF BILINS Andrea W. U. Busch 1, Edward J. Reijerse 3, Wolfgang Lubitz 3, Nicole Frankenberg-Dinkel 1* and Eckhard Hofmann 2* From Physiology of Microorganisms 1, Biophysics 2, Department of Biology and Biotechnology, Ruhr- University Bochum, Bochum, Germany, 3 Max-Planck-Institute for Bioinorganic Chemistry, Mülheim an der Ruhr, Germany Running head: Molecular basis of PEB synthesis * Address correspondence to: Eckhard Hofmann, Ruhr-University Bochum, Proteincrystallography, Department of Biophysics, Bochum, Germany. Fax: (0) ; eckhard.hofmann@bph.ruhr-uni-bochum.de or Nicole Frankenberg-Dinkel, Ruhr-University Bochum, Physiology of Microorganisms, Bochum. Fax: (0) ; nicole.frankenberg@rub.de Phycoerythrobilin (PEB) is one of the major open-chain tetrapyrrole molecules found in cyanobacterial light-harvesting phycobiliproteins. In these organisms, two enzymes of the ferredoxin-dependent bilin reductase family work in tandem to reduce biliverdin IXα to PEB. In contrast, a single cyanophage-encoded enzyme of the same family has been identified to catalyze the identical reaction. Using UV-Vis and EPR spectroscopy we investigated the two individual cyanobacterial enzymes PebA and PebB showing that the two subsequent reactions catalyzed by the phage enzyme PebS are clearly dissected in the cyanobacterial versions. While a highly conserved aspartate residue is critical for both reductions, a second conserved aspartate is only involved in the A- ring reduction of the tetrapyrrole in PebB and PebS. The crystal structure of PebA from Synechococcus sp. WH8020 in complex with its substrate BV at 1.55 Å resolution revealed further insight into the understanding of enzyme evolution and function. Based on the structure it becomes obvious that in addition to the importance of certain catalytic residues, the shape of the active site and consequently the binding of the substrate highly determines the catalytic properties. Ferredoxin-dependent bilin reductases (FDBRs 1 ) are a class of enzymes involved in reducing the heme metabolite biliverdin IXα (BV) to form several individual open-chain tetrapyrroles (phycobilins) used for light-perception or lightharvesting in plants and cyanobacteria (1). FDBRs 1 are distinct from BV reductases in mammals or cyanobacteria which are mainly involved in catabolic degradation of BV to bilirubin (2,3). Currently several members of the FDBR family are known. The first cloned member was phytochromobilin (PΦB) synthase (HY2) from Arabidopsis thaliana, producing PΦB, the chromophore of the photoreceptor phytochrome (4). Based on the amino acid sequence of PΦB synthase, additional members of the FDBR family were identified (5,6). Among those are enzymes, which likewise catalyze a two-electron reduction but also two members that catalyze a formal fourelectron reduction. The most common target for reduction within the FDBR family is the 2,3,3 1,3 2 - diene system of the A-ring (Figure 1). However, only PΦB synthase acts on the substrate BV directly thereby producing PΦB. All other A-ring reductions target intermediates of four-electron reductions leading to the cyanobacterial pigments phycocyanobilin (PCB) and phycoerythrobilin (PEB). Specifically, the second reduction performed by PcyA (PCB:ferredoxin oxidoreductase) targets the A-ring of 18 1, dihydrobiliverdin (DHBV), an isolatable intermediate in the reduction of BV to PCB (7). The recently identified cyanophage PEB synthase (PebS) on the other hand utilizes the intermediate 15, 16-DHBV to produce PEB (6,8). The identical A-ring reduction is also performed by the twoelectron reducing PebB (PEB:ferredoxin oxidoreductase). This enzyme acts in tandem with PebA (15,16-DHBV:ferredoxin oxidoreductase), which reduces BV at the C15-C16 double bond to

42 produce 15,16-DHBV (5,9). Both enzymes are proposed to function in close contact and metabolic channeling of 15,16-DHBV has been postulated (9). In contrast to that, PebS realizes a perfect metabolic channeling of the same intermediate by combining the two activities of PebA and PebB in one enzyme. While the similarity between PebS and PebA is rather low (23%), they do however still serve as a great paradigm of enzyme evolution and function. We previously presented the crystal structure and biochemical analysis of cyanophage PebS (8,10). Like PcyA, the first crystallized member of the FDBR family (11,12), PebS shows an α/β/αsandwich fold, with a central substrate binding site parallel to the plane of the sheet (8). The binding of the substrate BV is rather flexible which was reflected in the different binding modes observed. Concurrent with these different binding modes is also the flexibility of aspartate D206, one of two aspartate residues (D105 and D206) shown to be critical for the reaction. While D206 seems to be important for the A-ring reduction activity of PebS, D105 is proposed to be involved in both enzymatic steps which both proceed via a tetrapyrrole radical intermediate (10). Interestingly, both residues are highly conserved within the whole FDBR family (Figure S1) and a common function has been discussed (5,10,13). This current study has been undertaken to understand the structural and mechanistic details discriminating between the two cyanobacterial and the phage enzyme. While PebS is able to reduce the intermediate 15,16-DHBV further to the final product, the PebA reaction is terminated at this point. PebB on the other hand has a distinct substrate specificity for the intermediate 15, 16- DHBV. In an interdisciplinary approach we combined X-ray crystallography with biochemical and biophysical characterization of wild type and mutant proteins to get a deeper understanding of the evolution of these three enzymes. EXPERIMENTAL PROCEDURES Reagents Unless otherwise specified, all chemical reagents were ACS grade or better. Glucose-6-phosphate dehydrogenase, NADP +, catalase, glucose, glucose-6-phosphate, spectrophotometric grade glycerol, trifluoro acetic acid, 4-methylmorpholine and aminoacids were purchased from Sigma-Aldrich. HPLC grade formic acid, acetone and acetonitrile were purchased from J.T. Baker. Biliverdin IXα (BV) was obtained from Frontier Scientific. Production and purification of recombinant proteins PebA and PebB from Synechococcus sp. WH8020 were expressed, purified and the concentration determined as described previously (9). The purified proteins contained eight additional N-terminal residues from the expression vector. Selenomethionine (SeMet) labeled PebA (SePebA) was produced using M9 minimal medium. 15 min prior to induction with isopropylβ-thiogalactoside 100 mg/l lysine, phenylalanine, threonine, 50 mg/l isoleucine, valine, leucine and 60 mg/l selenomethionine were added to the culture. Expression and purification for labeled protein followed the same procedure as for unlabeled PebA with the exception that all buffers contained 5 mm dithiothreitol. Ferredoxin from cyanophage P-SSM2 and ferredoxin:nadp+ oxidoreductase (FNR) from Synechococcus sp. WH 7002 were prepared as described earlier (10). All protein variants were generated by sitedirected mutagenesis from pgex-6p-1_pebb Syn and pgex-6p-1_pebb Syn (5) using the QuikChange site-directed mutagenesis kit (Stratagene). The used primers are listed in Table S1. Anaerobic bilin reductase assay, electron paramagnetic resonance and HPLC measurements Bilin reductase activity assays, electron paramagnetic resonance (EPR) measurements, HPLC analyses and preparative production of 15,16-DHBV were performed under anaerobic conditions as described (10). To start the reaction, an NADPH-regenerating system was used (5). Only the reaction of the PebA_D84N variant was started with an excess of NADPH (10 electron equivalents, 50 µm final concentration) instead of the NADPH-regenerating system to reduce unspecific reduction of accumulating radical species. For PebB, the following modifications were used: The assay was performed at 20 C with 0.1 µm FNR and 25 Units catalase. After 2 min 10 µm of PebB_WT or PebB variant was added to 2

43 the reaction mixture. The reaction was stopped after a total of 10 min. EPR measurements were performed as described in (10). The following modifications were applied only for the PebB reaction: The reaction was performed at 20 C with 37.5 Units catalase, µm FNR and 4 µm ferredoxin using PebA_WT and BV as the substrate. After 2 min PebB was added to the PebA-DHBV complex at a 1:1 ratio. Crystallization Crystallization conditions were screened by the sitting drop vapor diffusion method utilizing the Cryos, PEG and PACT Suites (Qiagen) applying 200/100 nl and 100/100 nl mixtures of the protein solution (10-20 mg/ml) / reservoir solution incubated at 18 C in the dark. Conditions were further optimized with the hanging drop vapor diffusion method. Substrate BV was added in two-fold excess. PebA crystallized at protein concentrations between mg/ml in 0.1 M HEPES, ph 7 and 28 % PEG Structure determination and refinement Oscillation data of SeMet-labelled protein crystals were collected at 100K at the Swiss Light Source (SLS) on beamline X10SA. Data were processed using XDS (14). 5% of the data were randomly assigned as test set. Data statistics are given in Table 1. The Matthews coefficient was estimated to 2.1 Å 3 /Da for one molecule per asymmetric unit. Phases were determined from a 2 Å dataset collected at the Se-K-edge. AutoSharp readily located the two Se-atoms in the asymmetric unit. Arp/Warp was used to autotrace PebA in the resulting map. This model was then used with Arp/Warp to rebuild the model with the second 1.55 Å PebA dataset. The resulting model was improved using alternating cycles of manual correction in COOT (15) and automatic refinement in PHENIX (16). The first six residues from the expression vector, residues 128 and 129, and the last three residues were not modeled due to missing density. The model has been deposited at the Protein Data Bank under accession number 3X9O. RESULTS All members of the FDBR family are radical enzymes 3 Cyanobacteria use the dual enzyme system PebA and PebB to produce the phycobilin PEB. Although both enzymes have in parts already been studied biochemically (9), we reinvestigated the enzymatic conversions under anaerobic conditions. This is essential to detect and stabilize possible tetrapyrrole radical intermediates as shown for other members of the FDBR family, including the recently investigated cyanophage PebS (10,13,17). In contrast to the aerobic reduction of BV to 15,16-DHBV by PebA, which showed a decrease of absorption at 690 nm and a concomitant increase at 590 nm (9), two additional absorption maxima are observed under anaerobic conditions. In the course of the reaction an increase and further decrease at ~ 440 nm and ~ 750 nm was monitored (Fig. 2A, upper panel). Absorbance at these wavelengths has been attributed to bilin radical intermediates (10,13,17). EPR studies confirmed this assumption. For PebA, the strongest EPR signal was observed 1 min after start of the reaction (Fig. 2 B), when no product was yet detectable via HPLC (data not shown). The appearance of this signal followed similar kinetics as the absorption at ~ 440 nm and ~ 750 nm. In addition, the decrease of radical signal was accompanied by an increase of product formation indicating that the observed radical is a BV radical. For the first time we monitored the PebB reaction which uses 15,16-DHBV as a substrate. In order to facilitate proper 15,16-DHBV delivery, a completed PebA reaction was employed. This was preferred over external supply of 15,16-DHBV because metabolic channeling of 15,16-DHBV from PebA to PebB has been postulated and might be involved in proper delivery and subsequent binding of 15,16-DHBV (9). This is underlined by the observation that externally supplied 15,16- DHBV to the PebS_D206N variant binds differently than the intermediate produced by itself (10). Due to the lack of an extinction coefficient and the instability of 15,16-DHBV, this approach provides an easy and accurate way to supply PebB with equimolar amounts of substrate. First, a regular PebA reaction employing a PebA-BV complex was used. After 2 min almost all BV was converted to 15,16-DHBV which remained bound to PebA (PebA-DHBV). Subsequently, PebB_WT

44 was added leading to an immediate transfer and binding of 15,16-DHBV to PebB (PebB-DHBV) as observed by a shift of 15,16-DHBV absorbance from 590 nm (PebA-DHBV) to ~ 605 nm (PebB- DHBV) (Fig. 3A, Fig. S2). Almost simultaneously a strong increase at 682 nm was observed, most likely representing the formation of radical intermediates which were again confirmed by EPR measurements (Fig. 3B). A small EPR signal was still observed at 2 min due to a not fully completed PebA reaction. Addition of PebB_WT resulted in an increased radical signal, which disappeared with the completion of the reaction at ten minutes. At this time point the formation of PebB-bound 3Z-PEB (PebB-PEB) was detected at 548 nm (Fig. S2). This is in agreement with the PebS reaction where the final product is also 3Z-PEB (10). Herewith we prove that all different members of FDBRs act via radical intermediates. Identification of critical residues for catalytic activity Central to the proposed reaction mechanism of PebS are two aspartate residues D105 and D206, both involved in interactions with the pyrrole nitrogens upon substrate binding. Both are shown to be essential for the complete reduction of BV to PEB by PebS and are highly conserved throughout the family of FDBRs (Fig. S1). To study the role of the corresponding residues in PebA and PebB (Table 1), several variants were characterized in anaerobic bilin reductase assays. The resulting products were analyzed by HPLC and UV-Vis spectroscopy. If aspartate at position 84 in PebA is changed into an asparagine, no conversion of BV is observed. Interestingly, the D84N variant appears to stabilize a radical intermediate as indicated by a fast increase and very low decrease of an absorption maximum at 740 nm (Fig. 2A). If the carboxylic side chain is retained, but exchanged by a longer side chain which is expected to sterically interfere with substrate binding, the resultant PebA variant D84E is still able to convert BV to 15,16-DHBV at a similar efficiency (Fig. 2A). However, the produced 15,16-DHBV is significantly more unstable (data not shown) suggesting that D84 is also involved in stabilization of the enzymeproduct complex. This catalytic behaviour is in agreement with data of the PebS_D105N/E 4 variants which showed the same properties for the first reduction (10). Similar results were obtained for the asparagine variant of the homologous amino acid residue in PebB (PebB_D107N). A stabilization of a radical intermediate was observed both in UV-Vis (increase of absorption at 670 nm) and EPR spectroscopy (Fig. 3). Consequently, no product formation was detected (data not shown). Interestingly, retaining the carboxylic side chain of this residue in a glutamate variant (PebB_D107E) did not rescue the activity indicating that space constraints might be crucial for reduction. However, this variant is still able to bind 15,16- DHBV in a manner similar to the WT protein (Fig. S2) and trace amounts of product PEB were monitored by HPLC (data not shown). If D205 is changed into an asparagine, PebA retains its activity. The D205N variant converted its substrate BV to 15,16-DHBV (Fig. 2A). In contrast, the homologous variant of PebB (PebB_D231N) showed a complete loss of PEB formation. An increase of absorption at 666 nm with no further decay again suggested the stabilization of a radical intermediate (Fig. 3A). The observed EPR signal increased over time with highest signal intensity at 10 min (Fig. S3). PebB_D231E on the other hand showed only slightly decreased activity with significant product formation. The presented data clearly demonstrate that the individual activities of PebA and PebB with their dependency on important catalytic residues are combined in the phage enzyme PebS. Specifically, D105 is important for both reductions while D206, although conserved, is only critical for A-ring reduction (10). Overall structure of PebA In order to determine the structural differences between PebS and PebA, which only catalyzes the first reduction of BV to 15,16-DHBV, the structure of selenomethionine labeled PebA with BV IXα was solved by the single wavelength anomalous dispersion method and was refined at 1.55 Å resolution (Table 2). Six residues of the linker peptide used in the expression construct and the last three residues of the mature protein were not modeled due to missing electron density. In addition, the two

45 residues N124 and D129 could not be placed for the same reason. Clear density for the substrate allowed to unambiguously place BV into the binding site. Identification of the correct orientation was possible due to the asymmetric vinyl substituents of the A- and D-ring. The overall structure consists of a central sevenstranded antiparallel β-sheet, which is from both sides flanked by a total of six α-helices (Fig. 4). This topology represents the typical α/β/α - sandwich found also for PcyA and PebS (8,11). The binding pocket for BV is on one side formed by the rigid β-sheet. On the other side the substrate is bound by the loop (termed lid in the following) between helix H3 and strand S7 and by helices H5 and H6. While the binding pocket is lined mostly with apolar and aromatic residues, coordination of the polar pyrrol nitrogens and the carbonyl oxygens is facilitated by the highly conserved residues D84 and N88, respectively (Fig. 5). In addition, six water molecules are resolved inside the binding niche. The BV propionate side chains are coordinated by salt bridges with three positively charged protein side chains (K93, R134, R150). BV is bound in PebA in a roofed conformation, with both A- and D-ring tilted 40 out of the plane and the A-ring buried deeper inside the pocket. The catalytic residue D84 is in an identical orientation in PebA as compared to D105 in PebS and is therefore not the structural reason for the roofed BV binding. The other conserved aspartate D205 which is of catalytic importance for PebS and has been shown to adopt different conformations in PebS is rotated away from the active site in PebA. It is the first of three residues connecting H5 and H6 (termed Dloop in the following). This residue is not involved in direct ligand interaction which is consistent with its non-catalytic role in PebA. DISCUSSION All FDBRs share the same overall fold The structure of PebA represents the first structural view of an FDBR catalyzing a twoelectron reduction. Together with the different structures available of the four-electron reduction enzymes PcyA and PebS we are now able to make 5 a clearer discrimination between structural features preserved throughout the family and features with distinct differences, which might be key elements in controlling the stereoselective reactions. All structurally solved FDBRs share the common seven-stranded β-sheet, which forms a mostly hydrophobic basis of the substrate binding site. A prominent feature of the binding pocket is the central polar centering pin, formed by D105 (PebS numbering used throughout the discussion for simplicity reasons) and the adjacent N88 on strands S5 and S4, respectively (Fig. S5). Only for HY2, D105 is replaced by asparagine, but its function is preserved by the change of the neighboring N88 to aspartate, restoring the placement of a functional carboxylic group at this position (13). The binding pocket is always formed by two long helices (PebA:H5-H6, PebS:H3-H4, PcyA:H7-H8) connected by the D- loop. The position of these helices is similar in all structures, but variations of the linker geometry occur between enzymes and upon substrate binding. The largest structural variations are located in the lid between β-strands S6 and S7 in response to substrate binding in both PebS and PcyA (Fig. 4) (8,18). In addition, large differences in length and folding of the lid region exist between the different FDBRs. In PebA and PcyA helical elements are found, which are missing completely in PebS (Fig. S4). Both the changes upon substrate binding and differences in structure make this area the most promising target for interaction of PebA with both ferredoxin (as proposed by (19)) and PebB. Structural flexibility of the active site is required for two consecutive reductions within one FDBR Based on the structures and on the sequence alignment (Fig. S1), the binding pockets of all FDBRs are lined with hydrophobic and aromatic residues, making van-der-waals and π-stacking interactions the dominating factor for substrate binding. Very few residues in the active site are chemically able to participate in substrate protonation and are found to be highly conserved amongst FDBRs. Also the general binding mode with the A-ring buried deeper seems to be identical throughout the FDBR family, regardless of the site of reduction (Fig. S5).

46 Considering the similarities between PebS and PebA, the question arises, which structural features of PebA are responsible for the termination of the BV reduction at the intermediate 15,16-DHBV, while it is directly processed to PEB in PebS. The binding pocket of PebA forces BV into a strained, roofed conformation (Fig. 5). This conformation results from steric restraints imposed by size variation of apolar protein side chains. For example, substitution of methionine M212 by phenylalanine in combination with the substitution of isoleucine I86 by valine in PebA forces the rotation of the D- ring of BV (Fig. S5 and S6). In contrast, in PebS the pocket is much larger, leaving room for the tilting of the D-ring after reduction of the C15-C16 double bond to form the nonplanar product 15,16- DHBV (Fig. S6). It has been proposed that rearrangement of 15,16-DHBV in PebS is required for stereospecific reduction of the A-ring. Even though a structure of PebS with the bound intermediate 15,16-DHBV is still missing, the observed binding flexibility for BV supports such a rearrangement (8). In contrast, very little conformational variations for a variety of substrates have been observed in PcyA (11,18,20). The intermediate 18 1,18 2 -DHBV of the PcyA reduction still retains the conjugated π-system of BV, which extends over all four tetrapyrrole rings. Thereby no large structural changes have to be accomodated during the two successive reductions (Fig. S5 and S6). D206 is critical for A-ring reduction Based on our and other biochemical and structural data it is obvious that the conserved D105 residue is highly important for bilin reduction in PcyA, PebS, PebA and PebB. The only exception so far seems the above mentioned HY2 protein where double exchange of an Asp/Asn pair occurs. D105 is well positioned to facilitate stereospecific protonation to both C-2 and C-16 to generate the R configuration at these carbon atoms. In contrast to the general importance of D105, D206, although conserved, is only important for A-ring reduction of BV (in HY2) and 15,16-DHBV (in PebS and PebB). For PebS we previously postulated a role of D206 as a proton shuttle to supply the protons for the second reduction. In PebA this proton shuttle is not required since only a two-electron reduction is catalyzed, representing the first reduction in PebS. In both cases the active site seems to be able to supply the two protons necessary for the reaction either from D105 and/or water molecule(s). For the second reduction PebS requires additional delivery of protons from the surrounding medium as the intermediate stays tightly bound in the active site. D206 is proposed to fulfill this function. In contrast, it is hypothesized that PebB binds 15,16-DHBV into its protonated active site analogous to BV binding to PebA and PebS alleviating the need for an additional proton shuttle. We would expect at least residual function of the PebB_D231N variant if this residue is only relevant for proton delivery in PebB. Therefore, an additional function of D206 in substrate coordination is required to fully explain our data. We believe that D206 will be involved in direct coordination of 15,16-DHBV, thereby affecting local reactivity of the substrate to support stereospecific protonation by D105. In turn, these data suggest a similar function of D206 also for PebS. Full understanding of these processes will require structures of PebA, PebB and PebS with bound 15, 16-DHBV as well as high field EPR characterization of the substrate radicals. These experiments are currently underway in our laboratories. 6

47 FOOTNOTES 1 Abbreviations used: BV, biliverdin IXα; DHBV, dihydrobiliverdin; EPR, electron paramagnetic resonance; FDBR, ferredoxin-dependent bilin reductase; FNR, ferredoxin:nadp + oxidoreductase; PCB, phycocyanobilin; PEB, phycoerythrobilin; PebS, PEB synthase; PΦB, phytochromobilin; SeMet, selenomethionine; WT, wild-type. SUPPLEMENTAL INFORMATION One Table and six supplemental Figures are available online. ACKNOWLEDGEMENT This work was financially supported by the SFB 480 (Teilprojekt C6 and C8) of the Deutsche Forschungsgemeinschaft (to N.F.D. and E.H.). A.B. received a PhD fellowship from the Ruhr-University Bochum Research School. Thanks are also due to the beamline staff at Swiss Light Source (Villigen, CH) and the European Synchroton Radiation Facility (Grenoble, F). 7

48 REFERENCES 1. Dammeyer, T., and Frankenberg-Dinkel, N. (2008) Photochem Photobiol Sci 7, Kutty, R. K., and Maines, M. D. (1981) J Biol Chem 256, Schluchter, W. M., and Glazer, A. N. (1997) J Biol Chem 272, Kohchi, T., Mukougawa, K., Frankenberg, N., Masuda, M., Yokota, A., and Lagarias, J. C. (2001) Plant Cell 13, Frankenberg, N., Mukougawa, K., Kohchi, T., and Lagarias, J. C. (2001) Plant Cell 13, Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W., and Frankenberg-Dinkel, N. (2008) Curr Biol 18, Frankenberg, N., and Lagarias, J. C. (2003) J Biol Chem 278, Dammeyer, T., Hofmann, E., and Frankenberg-Dinkel, N. (2008) J Biol Chem 283, Dammeyer, T., and Frankenberg-Dinkel, N. (2006) J Biol Chem 281, Busch, A. W., Reijerse, E. J., Lubitz, W., Hofmann, E., and Frankenberg-Dinkel, N. (2011) Biochem J 433, Hagiwara, Y., Sugishima, M., Takahashi, Y., and Fukuyama, K. (2006) Proc Natl Acad Sci U S A 103, Tu, S. L., Rockwell, N. C., Lagarias, J. C., and Fisher, A. J. (2007) Biochemistry 46, Tu, S. L., Chen, H. C., and Ku, L. W. (2008) J Biol Chem 283, Kabsch, W. (2010) Acta Crystallogr D Biol Crystallogr 66, Emsley, P., and Cowtan, K. (2004) Acta Crystallogr D Biol Crystallogr 60, Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd, J. J., Hung, L.-W., Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R. J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010) Acta Crystallogr D 66, Tu, S. L., Gunn, A., Toney, M. D., Britt, R. D., and Lagarias, J. C. (2004) J Am Chem Soc 126, Hagiwara, Y., Sugishima, M., Takahashi, Y., and Fukuyama, K. (2006) FEBS Lett 580, Chiu, F. Y., Chen, Y. R., and Tu, S. L. (2010) J Biol Chem 285, Hagiwara, Y., Sugishima, M., Khawn, H., Kinoshita, H., Inomata, K., Shang, L., Lagarias, J. C., Takahashi, Y., and Fukuyama, K. (2010) J Biol Chem 285,

49 FIGURE LEGENDS Figure 1. Reactions catalyzed by the FDBR family. Highlighted in the box is the most common reduction site within the family, the A-ring 2,3,3 1,3 2 -diene system. The plant enzyme phytochromobilin synthase (HY2) catalyzes the reduction of biliverdin (BV) to 3Z-PФB; cyanobacterial PEB:ferredoxin oxidoreductase (PebB) and cyanophage phycoerythrobilin synthase (PebS) the reduction of 15,16- dihydrobiliverdin (DHBV) to 3Z-phycoerythrobilin (3Z-PEB), and PcyA the reduction of 18 1, dihydrobiliverdin to 3Z-phycocyanobilin (3Z-PCB). The reduction sites are highlighted by the circles. Figure 2. Reaction of PebA_WT and variants monitored by UV-Vis and EPR spectroscopy. A. Anaerobic reduction of BV by PebA and variants was monitored via UV-Vis spectroscopy for 10 min, except for variant D84E. Spectra were taken every 30 sec. Possible radical absorptions (~440 and ~750 nm) are indicated by an asterisk, development of the spectra over time is indicated by arrows. The WT reaction contained a four-fold excess of enzyme as described in (10) and represents the sample used for the EPR measurement shown in B. The reaction of the D84N variant employed NADPH instead of the regenerating system as described in the experimental procedures. B. EPR measurements were performed as described previously with samples taken from the reaction at indicated time points. All spectra are on the same scale and were recorded at T=40 K, a microwave power of 20 μw and 1 mt field modulation (see (10) for further experimental details). Figure 3. Reaction of PebB_WT and variants monitored by UV-Vis and EPR spectroscopy. A. Anaerobic reduction of BV by PebA (spectra shown in grey) was monitored via UV-Vis spectroscopy. At the end of the reaction when all BV was reduced to 15,16-DHBV, PebB was added and the reaction monitored for additional 8 min (spectra shown in black). Spectra were taken every 30 sec. Possible radical absorption is indicated by an asterisk, development of the spectra over time is indicated by arrows. B. EPR measurements for the WT have been performed as described previously (10) and under experimental procedures with samples taken from the reaction at indicated time points. All spectra are on the same scale and were recorded at T=40 K, a microwave power of 20 μw and 1 mt field modulation (see text and (10) for further experimental details). Figure 4. Overall structure of PebA. Shown is the protein backbone in cartoon representation, rainbow colored from blue (N-terminus, labeled N) to red (C-terminus, labeled C). The bound substrate BV and the two catalytic residues D84 and D205 are shown as stick models. Secondary structure elements are labeled. Figure 5. Stereoview of the active site of PebA. Shown is a cartoon representation of both PebA (green) and PebS (yellow, PDB code 2VCK, chain C), superposed based on the Cα-atoms of the central β-sheet. The two catalytic aspartates and the substrate are shown as sticks and are labeled. Water molecules present in the PebA structure are shown as red spheres. 9

50 Table 1. Catalytic activities of protein variants of the two conserved aspartate residues. Homolog residues in each FDBR are shown in the same row. PebS_D105 PebA_D84 PebB_D107 D105N - D84N - D107N - D105E - D84E 15,16-DHBV D107E - PebS_D206 PebA_D205 PebB_D231 D206N 15,16-DHBV D205N 15,16-DHBV D231N - D206E 15,16- DHBV/PEB D231E PEB -, no activity observed; DHBV, dihydrobiliverdin; PEB, phycoerythrobilin (as products of the reaction). 10

51 Table 2. Data Collection and Refinement Statistics a Data set peba-10 peba-19 Beamline SLS X10SA SLS X10SA Resolution (Å) ( ) ( ) Cell parameters (a,b,c; Å) (α,ß,γ; degrees) 42.25, 39.28, , , , 39.35, Spacegroup P2 1 P2 1 Wavelength Completeness (%) 99.3 (92.9) 97.7 (93.8) Multiplicity 7.6 (6.1) 4.3 (3.4) Average I/σI 24.5 (6.7) RRsym (%) 6.3 (24.7) RRmeas (%) b 6.8 (26.9) 11.3 (4.8) 7.4 (22.0) 8.3 (25.8) RRmrgd-F (%) b 5.6 (21.5) 8.6 (28.2) Phasing Phasing Power 1.13 overall 3.45 at 5.5 Å >1 above 2.4Å Refinement Resolution (Å) RRcryst(%) 17.9 RRfree(%) d 22.5 Number of atoms Protein 8179 Biliverdin 43 Water 322 Average Bfactor Protein 23.8 (Biliverdin) Root mean square deviation from ideality Bonds (Å) Angles (degrees) a Data in parentheses represent values in the highest resolution bin. b For definition of c RRmeas and R mrgd-f see Diederichs and Karplus (1997). FOM calculated after density modification and threefold averaging in d RESOLVE. Rfree calculated from 5% of data omitted from refinement. 11

52 FIGURE 1

53 FIGURE 2

54 FIGURE 3

55 FIGURE 4

56 FIGURE 5

57 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 52 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) Busch, A.W., Frankenberg-Dinkel, N., Hofmann, E. (Manuscript in preparation)

58 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 53 Abstract Phycoerythrobilin synthase (PebS) is an enzyme catalyzing the four-electron reduction of biliverdin ΙΧα (BV ΙΧα) to phycoerythrobilin (PEB). 15,16- dihydrobiliverdin:ferredoxin oxidoreductase (PebA) catalyzes the two-electron reduction of BV ΙΧα to 15,16-didydrobiliverdin (15,16-DHBV), which is also an intermediate in the PebS reaction. Two aspartate residues at position 105 and 206 are known to be crucial for PebS reaction. Exchange of these residues hindered the 15,16-DHBV reduction reaction. In this study we ought to investigate the structural background of these two glutamate variants and the influence of the amino acid exchange on the substrate conformation. Introduction Phycoerythrobilin (PEB) is an open chain tetrapyrrole, also called bilin that functions in light-harvesting in oxygenic photosynthetic bacteria and algae. Bound to a phycobiliprotein it is incorporated into phycobilisomes, the large lightharvesting antenna of cyanobacteria, red algae and glaucophytes (Grossman et al., 1993; MacColl, 1998; Neilson and Durnford, 2010). In cyanobacteria, two ferredoxin-dependent radical enzymes 15,16- dihydrobiliverdin:ferredoxin oxidoreductase (PebA) and PEB:ferredoxin oxidoreductase (PebB) catalyze two subsequent two-electron reductions of biliverdin ΙΧα (BV ΙΧα) to yield the 3Z-isomer of PEB (Fig. 1). The product of PebA, 15,16-dihydrobiliverdin (15,16-DHBV), is the intermediate bilin in the reaction (Dammeyer and Frankenberg-Dinkel, 2006; Frankenberg et al., 2001). In the genome of a marine virus that infects cyanobacteria, a gene encoding phycoerythrobilin synthase (PebS) that catalyzes the four-electron reduction of BV ΙΧα to yield PEB with 15,16-DHBV as the intermediate, has been identified (Fig. 1) (Dammeyer et al., 2008a). The structure of PebS revealed two aspartates in proton donating position to the substrate BV ΙΧα (Dammeyer et al., 2008b). These residues, Asp 105 and Asp 206, were suggested to have proton-donating function by analysis of the in vitro activities of their asparagine variants. The homologous residues are conserved in PebA and PebB. As revealed by the analysis of these residues in PebA and PebB, one aspartate (Asp 105 ) was identified to be important for both reactions. The

59 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 54 second Asp 206 is only crucial for 15,16-DHBV reduction. Both residues are similarly positioned in the crystal structures of PebA and PebS (Busch et al., manuscript submitted; Busch et al., 2011; Dammeyer et al., 2008b). Fig. 1: Phycoerythrobilin biosynthesis. Biliverdin ΙΧα (BV ΙΧα) is reduced by 15,16- dihydrobiliverdin:ferredoxin oxidoreductase (PebA) to 15,16-dihydrobiliverdin (15,16-DHBV). 15,16- DHBV is reduced by PEB:ferredoxin oxidoreductase (PebB) to 3Z-phycoerythrobilin (3Z-PEB). Phycoerythrobilin synthase (PebS) catalyzes the reduction of BV ΙΧα to 3Z-PEB; 15,16-DHBV is an intermediate in the reaction. Structural analyses of asparagine variants of these two residues at position 105 and 206 in PebS (PebS_D105N and PebS_D206N) showed an almost identical overall fold and active site structure compared to the wild-type (WT) protein (Busch et al., 2011). This observation implied that lack of activity in the asparagine variants is due to the lack of a carboxylic side chain at this position. To test if activity can be restored when a carboxylic side chain is present, glutamate variants of these two residues (PebS_D105E, PebS_D206E) have been analyzed (Busch et al., 2011). Due to their carboxylic side chain in the glutamate residues one would expect to yield functional enzymes. In the case of PebS_D206E this assumption was correct. PebS_D206E was able to reduce BV ΙΧα to PEB, although the reaction was much slower than for the WT enzyme. Reaction with PebS_D105E only yielded the intermediate 15,16-DHBV. These observations were explained with the bulkier side chain of glutamate that might influence reactivity by changing the active site structure and/or positioning of the substrate. In addition to the PebS structures, the crystal structure of PebA became available and revealed a very similar overall fold compared to PebS (Busch et al., manuscript submitted). Moreover, the active site structure of both enzymes was extremely alike raising the question of what determines the different activities of PebA and PebS. One striking difference between PebS and PebA was the conformation of the substrate BV ΙΧα in both enzymes. In PebS BV ΙΧα was

60 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 55 shown to adopt a more planar porphyrin-like conformation and lockwasher -like conformation with the A-ring being positioned partially above the D-ring (Fig. 2A- C). In PebA BV ΙΧα adopts a more roof -like structure (Fig.2D). This special conformation of the substrate BV ΙΧα observed in PebA suggested that a similar conformation of its product 15,16-DHBV prevents its further reduction. To elucidate whether a `restricted` conformation of 15,16-DHBV, as observed for PebA bound BV ΙΧα (Fig. 2D), prohibits further reduction in PebS_D105E and PebS_D206E, these variants were subjected to structural analysis. Fig. 2: BV ΙΧα conformation in the active site pocket of A) PebS (2VCK, chain C), B) PebS (2VCK, chain D), C) PebS (2VGR, chain A), D) PebA (2X9O). The substrate BV ΙΧα is shown as stick model in green. The aspartate residues are shown as lines in black. Experimental Procedures Primer Primer used for quick change mutagenesis were PebS(QC) D105E fw, CTT GTT TTG GTA TGG AAC TGA TGA AGT TTA GTG; PebS(QC) D105E rev, CAC TAA ACT TCA TCA GTT CCA TAC CAA AAC AAG; PebS(QC) D206E fw,

61 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 56 CTT ATA TGA CTG AAC TTG AAC CTG TTA GAG G; PebS(QC) D206E rev, CCT CTA ACA GGT TCA AGT TCA GTC ATA TAA G. The changed triplet is shown in bold. Protein purification Protein has been purified as described for the WT (Dammeyer et al., 2008b). Crystallization conditions Crystallization conditions of PebS-BV complex were screened by the sitting drop vapor diffusion method using the PEG, Cryos and Classics Suites (Qiagen), applying 200/100-nl and 100/100-nl mixtures of the protein solution (PebS_D105E, 35 mg/ml; PebS_D206E, 23 mg/ml) / reservoir solution incubated at 18 C in the dark. Conditions were further optimized with the hanging drop vapor diffusion method. Final crystals grew in a 10-µl drop of a 2:1 mixture of protein solution (concentrations as above) and reservoir solution within 1 day in 18 % PEG 4000, 10 % isopropanol and 0.1 M HEPES sodium salt at ph 7.5 (PebS_D105E) and within 3 days in 0.2 M calcium acetate, 0.1 M sodium cacodylate at ph 6.7 and PEG 8000 (PebS_D206E). Substrate has been added in two-fold access. PebS_D105E crystals with 15,16-DHBV were obtained as described for the hanging drop vapor diffusion method. Final crystals were harvested in cryoloops and briefly soaked in paraffin oil as the cryoprotectant prior to flash-freezing in liquid nitrogen. Green colored crystals of PebS_D105E with BV ΙΧα were harvested after 1 day (PebS_D105E-BV). Purple crystals of PebS_D105E grown with 15,16-DHBV were harvested after 1 day as well (PebS_D105E-DHBV). Crystals of PebS_D105E with 15,16-DHBV showed slight color change after 3 days when they were harvested due to reoxidation processes of the 15,16-DHBV (PebS_D105E-DHBV ox ). PebS_D206E crystals with BV ΙΧα showed green appearance and were harvested after 3 days (PebS_D206E-BV). Data collection and structure determination Oscillation data of the D105E-1 and D206E variants were collected at the SLS (Swiss Light Source, Villigen, Switzerland) on beamline X10SA and of the D105E-7 and D105E-11 at the at the ESRF (European Synchrotron Radiation Source, Grenoble, France) on beamline ID23.2. All data were processed and scaled using the XDS package (Kabsch, 1993). Data statistics are given in Table 1. Both structures were solved by molecular replacement in Molrep using the structure of PebS with bound BV ΙΧα

62 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 57 (PDB code 2VGR) as the search model. The models were improved by iterating manual rebuilding in COOT (Emsley and Cowtan, 2004) and refinement using PHENIX (Adams et al., 2002). Model and refinement statistics are given in Table 1. Table 1: Data collection PebS_D105E-BV PebS_D105E-DHBV PebS_D105E-DHBV ox PebS_D206E- BV Spacegroup P P P P1 Unitcell (a,b,c,α,β,γ) Wavelength [Å] Resolution limits [Å] ( ) ( ) ( ) ( ) No. of observations (19049) (13033) (17447) (12479) No. of unique reflections (2957) (3422) (6120) (3373) Multiplicity 7.32 (6.44) 4.19 (3.81) 3.98 (2.85) 3.59 (3.70) Completeness 99.8 % (99.0 %) 94.2 % (97.0 %) 98.0 % (81.9 %) 97.8 % (97.1 %) Ι/σΙ (3.29) (3.69) (2.29) 9.29 (3.10) R sym 5.2 % (45.1 %) 5.6 % (47.3 %) 5.0 % (47.2 %) 9.9 % (38.2 %) R meas 5.5 % (49.0 %) 6.3 % (54.2 %) 5.6 % (57.2 %) 11.6 % (44.7 %) R mergedf 6.5 % (44.2 %) 7.7 % (46.5 %) 9.6 % (67.1 %) 13.4 % (43.6 %) a For definition of RmergedF and R meas see (Diederichs and Karplus, 1997). b Data in parentheses refer to the highest resolution bin. Results and Discussion In order to gain structural information about the substrate positioning in the glutamate variants of PebS X-ray data of PebS_D105E with bound BV ΙΧα and 15,16-DHBV and PebS_D206E with bound BV ΙΧα have been collected. The structure of PebS_D105E-BV was refined at 1.55 Å, PebS_D105E-DHBV at 1.48 Å, PebS_D105E-DHBV ox at 1.15 Å and PebS_D206E-BV at 2.3 Å resolution. Structure of PebS_D105E PebS_D105E crystallized with its substrate BV ΙΧα yielded large (up to 1 mm) green colored crystals after ~ 1 day. The fast growth of rather large crystals obtained in a short amount of time provided the possibility of growing crystals with bound 15,16-DHBV. Earlier attempts crystallizing PebS, PebA and PebB with 15,16-DHBV and/or 3Z-PEB failed most likely due to reoxidation of the bilin, which results in an inhomogenous mixture of different bilin enzyme moieties thus

63 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 58 preventing crystallization. The observation that enzyme bound PEB can be converted back to 15,16-DHBV and 15,16-DHBV back to BV ΙΧα respectively was made before and attributed to the presence of oxygen that could serve as the oxidizing agent in the reverse reaction. Full reversion of PebA bound 15,16-DHBV to BV ΙΧα was described to take place 3-6 days on ice (Dammeyer and Frankenberg-Dinkel, 2006). Tests with PebS-PEB complex have been performed under strictly anaerobic conditions. Reoxidation still took place but was slightly slowed down (data not shown). Crystals of PebS_D105E were obtained with 15,16-DHBV after 1 day. Crystal structures were solved for PebS_D105E-DHBV frozen after 1 day while it still retained its purple color indicating bound 15,16-DHBV and for PebS_D105E- DHBV ox after 3 days when reoxidation was already in progress as observed by a slight color change. Although crystals of PebS_D105E with BV ΙΧα appeared green and crystals grown with 15,16-DHBV showed purple color no electron density for a bilin was obtained in any of the structures. As shown in Fig. 3, the overall fold of all three structures revealed an α/β/α sandwich fold with two N- Fig. 3: Overlay of PebS_D105E structure with wildtype PebS. PebS (2VCL, cyan) has been superimposed with PebS_D105E-BV (red), PebS_D105E-DHBV (blue) and PebS_D105E-DHBV ox (green). terminal α-helices, a seven stranded antiparallel β-sheet and three C-terminal α-helices. This was identical to the substrate free PebS structure with root mean square deviations (RMSD) of Å for 183 C-α atoms (PebS_D105E-BV), Å for 187 C-α atoms (PebS_D105E- DHBV), Å for 194 C-α atoms (PebS_D105E-DHBV ox ). In the substrate-bound PebS structure Asp 206 is coordinating BV ΙΧα from a short loop at the α-helical site whereas Asn 88 and Asp 105 coordinate from the

64 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 59 β-sheet site (Fig. 4A). The latter two residues (Asn 88 and Glu 105 ) are comparably positioned in the PebS_D105E structures whereas Asp 206 adopts a conformation similar to that found in the substrate free structure of the WT protein (pdb code: 2VCL) with the Asp 206 rotated outside the active site adopting the out position (Fig. 4B/C/D). Due to its longer side chain Glu 105 protrudes further in the binding site where the substrate is replaced by water molecules. Fig. 4: Comparison of active site structure of PebS-BV and PebS_D105E. Active site structure of A) PebS with substrate (2VCK, chain C), B) PebS_D105E-BV, C) PebS_D105E-DHBV, D) PebS_D105E-DHBV ox. The protein is displayed as cartoon. The active site residues Asn 88, Asp/Glu 105, Asp 206 and the substrate BV ΙΧ α (grey) are shown as stick model. Water molecules are shown as red spheres. The fact that no electron density was observed for a bilin in the substrate binding pocket could not be explained so far. The obvious coloration of the crystals (green: PebS_D105E-BV, purple: PebS_D105E-DHBV) implied the presence of bound bilin in the binding pocket as observed for PebS crystals that had BV ΙΧα located in the active site (Busch et al., 2011; Dammeyer et al., 2008b). One explanation could be that the protein crystallized without the substrate that instead stayed in solution and adhered to the outer surface of the crystal. Absorption spectroscopy will be conducted utilizing a microspectrophotometer to examine the absorption properties of these crystals, which will allow a distinction between protein-bound and free substrate.

65 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 60 Structure of PebS_D206E PebS_D206E crystallized with BV ΙΧα in space group P1 with four molecules in the asymmetric unit (ASU). For all four molecules an overall fold almost identical to that of the substrate-bound PebS_WT protein (PDB code: 2 VCK, chain D) was observed with RMSDs of Å for 181 C-α atoms (chain A), Å for 178 C- α atoms (chain B), Å for 181 C-α atoms (chain C) and Å for 180 C-α atoms (chain D). The substrate BV ΙΧα is located in the binding niche between the central β-sheet and the C-terminal α-helices (Fig. 5). In the PebS_WT structure the loop formed by residues was shown to close up upon substrate binding (Dammeyer et al., 2008b). In PebS_D206E-BV the residues could not be conclusively modeled but a fold like that observed for substrate- bound WT protein is anticipated. The low electron density in this region is due to a high flexibility in this region that could in turn point towards a loose binding of BV ΙΧα. However, this explanation is not very likely considering the clear electron density for BV ΙΧα in the pocket and a distinct binding spectrum of the PebS_D206E-BV complex (Busch et al., 2011). The substrate is coordinated by the residues Asp 105 and Asn 88 from the β-sheet side and from Glu 206 from the short Fig. 5: Overlay of substrate-bound PebS_D206E with wild-type PebS. PebS (2VCK, chain D; blue) was superimposed with chain A of PebS_D206E (red). loop at the α-helical site. Glu 206 is only slightly moved further in towards the substrate coordinating the pyrrole nitrogen atoms of BV ΙΧα (2.9 Å, 2.8 Å, 2.7 Å, 2.7 Å hydrogen bonding distance) in a way as observed for the WT (2.9 Å, 2.9 Å, 2.8 Å, 2.9 Å hydrogen bonding distance). The longer side chain of glutamate compared to aspartate is rather compensated by an upward movement of the short loop

66 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 61 comprising Glu 206 (Fig. 6). Also, the propionate side chain at ring C of the substrate is rotated upwards in the WT structure compared to the PebS_D206E where the BV ΙΧα is positioned more planar. This planar conformation of the propionate side chains is observed in the WT structures where a pyrrol water Fig. 6: Overlay of the active site of PebS_D206E and wild-type PebS. Coordination of the substrate BV ΙΧα of PebS_WT (2VCK, chain D; cyan) and PebS_D206E (chain A, salmon) by Asn 88 and Asp 105 from the β-sheet site and by Asp/Glu 206 from the α-helical site is shown. Polar contacts are indicated for PebS_D206E only. The enzyme environment is shown as a cartoon, the substrate BV ΙΧα as a stick model and the substrate coordinating residues Asn 88, Asp 105 and Glu 206 as lines. (central water molecule that is coordinated by all four pyrrole nitrogens) is present (Fig. 2A). The fact that both, the planar conformation of the BV ΙΧα together with its proprionate side chains and the deep position of the Asp 206, are found in WT structures aside from each other suggests that this active site structure is also possible in the WT. There is no obvious indication of dissimlarities that could explain the limited activity of PebS_D206E. Nevertheless, the variant was still active (Busch et al., 2011). The slowed down activity will be affiliated to the bulkier residue. We cannot draw any conclusion by the observed structures regarding the prevalence of an

67 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 62 observed conformation. It is possible that the glutamate side chain shifts the reactive conformation frequency to a less preferred conformation thereby slowing down the reaction. Active site structure of glutamate variants The Asp 105 and Asn 88 that are involved in substrate coordination are comparably positioned in all known PebS structures to date including PebS_D206E and PebS_D105E (Busch et al., 2011). This confirms the earlier observation that the β- sheet site of PebS is rather rigid compared to the small loop between the helices H3 and H4 that includes Asp 206. The Asp 206 residue was found to be very flexible and to adapt three main conformations. Its side chain can be rotated out of the active site pocket (Fig. 2C), move in to coordinate the substrate and the pyrrole water (Fig. 2A) or it occupies the deep conformation where Asp 206 takes over the position of the central water to coordinate the substrate (Fig. 2B) (Dammeyer et al., 2008b). The most conclusive hint regarding a PebA-like substrate conformation was anticipated from the PebS_D105E variant that showed no PEB:ferredoxin oxidoreductase activity. Due to the fact that no electron density of a bilin in the active site pocket of PebS_D105E was observed no conclusions can be drawn in terms of active site substrate and product conformation. Therefore the inability of PebA to reduce its product 15,16-DHBV is still not fully explainable. Recent progress has been made in order to solve the structure for PebA with bound 15,16-DHBV. PebA-BV crystals could be reduced with dithionite to yield purple colored crystals, indicating specific reduction of BV ΙΧα had occurred, in turn suggesting formation of 15,16-DHBV in the binding pocket of the PebA crystal. The specific reduction of BV ΙΧα in the protein crystal by dithionite was also shown for PcyA. Here the radical accumulating variant D105N with bound BV ΙΧα was subjected to High-Field EPR measurement after reduction with ferredoxin in solution, and after redution with dithionite in its crystal form. The two reducing agents produced structurally indistinguishable radical intermediates (Stoll et al., 2009). The same technique will be employed in order to solving the structure of PebS with its product PEB. Product/intermediate-enzyme complex structures

68 4 Structure of Glutamate Variants from Phycoerythrobilin Synthase (PebS) 63 might be able to give conclusive hints in order to explain the `enhanced` activity of PebS that combines PebA and PebB activities.

69 5 CpeS Is a Lyase Specific for Attachment of 3Z-PEB to Cys 82 of 64 β- Phycoerythrin from Prochlorococcus marinus MED4 5 CpeS Is a Lyase Specific for Attachment of 3Z-PEB to Cys82 of β-phycoerythrin from Prochlorococcus marinus MED4. Wiethaus, J., Busch, A.W., Kock, K., Leichert, L.I., Herrmann, C., Frankenberg- Dinkel, N. Journal of Biological Chemistry 285(48): (2010)

70 Supplemental Material can be found at: THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 285, NO. 48, pp , November 26, by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. CpeS Is a Lyase Specific for Attachment of 3Z-PEB to Cys 82 of -phycoerythrin from Prochlorococcus marinus MED4 * S Received for publication, August 5, 2010, and in revised form, September 27, 2010 Published, JBC Papers in Press, September 28, 2010, DOI /jbc.M Jessica Wiethaus, Andrea W. U. Busch, Klaus Kock, Lars I. Leichert, Christian Herrmann, and Nicole Frankenberg-Dinkel 1 From the Department of Physiology of Microorganisms, Faculty of Biology and Biotechnology, Physical Chemistry I, Faculty of Chemistry and Biochemistry, and the Medical Proteome Center, Ruhr-University Bochum, Universitaetsstrasse 150, Bochum, Germany In contrast to the majority of cyanobacteria, the unicellular marine cyanobacterium Prochlorococcus marinus MED4 uses an intrinsic divinyl-chlorophyll-dependent light-harvesting system for photosynthesis. Despite the absence of phycobilisomes, this high-light adapted strain possesses -phycoerythrin (CpeB), an S-type lyase (CpeS), and enzymes for the biosynthesis of phycoerythrobilin (PEB) and phycocyanobilin. Of all linear tetrapyrroles synthesized by Prochlorococcus including their 3Z- and 3E-isomers, CpeS binds both isomers of PEB and its biosynthetic precursor 15,16-dihydrobiliverdin (DHBV). However, dimerization of CpeS is independent of bilins, which are tightly bound in a complex at a ratio of 1:1. Although bilin binding by CpeS is fast, transfer to CpeB is rather slow. CpeS is able to attach 3E-PEB and 3Z-PEB to dimeric CpeB but not DHBV. CpeS transfer of 3Z-PEB exclusively yields correctly bound Cys 82 -PEB, whereas Cys 82 -DHBV is a side product of 3E-PEB transfer. Spontaneous 3E- and 3Z-PEB addition to CpeB is faulty, and products are in both cases Cys 82 -DHBV and likely a PEB bound at Cys 82 in a non-native configuration. Our data indicate that CpeS is specific for 3Z-PEB transfer to Cys 82 of phycoerythrin and essential for the correct configuration of the attachment product. Phycobilisomes are the major light-harvesting complexes of cyanobacteria, rhodo-phytes, and cryptophytes (1, 2). The light-absorbing properties of corresponding phycobiliproteins result from up to four linear tetrapyrroles, known as bilins, covalently bound via thioether bonds to conserved cysteine residues of each and subunit. Bilins are derived from heme, which is converted to the first linear tetrapyrrole biliverdin IX (BV IX ) 2 by heme oxygenases (3). In cyanobacteria, BV IX is further reduced by ferredoxin-dependent bilin reductases to 3Z-phycoerythrobilin (3Z-PEB) or 3Z-phycocyanobilin (3Z- PCB) 3 (4, 5). Subsequent bilin attachment to phycobiliproteins is in most cases supported by members of the three known * This work was supported by the Deutsche Forschungsgemeinschaft and a German-Israeli Foundation Young Investigator grant (to N. F.-D.). S The on-line version of this article (available at contains supplemental Figs. S1 S8. 1 To whom correspondence should be addressed. Tel.: ; Fax: ; nicole.frankenberg@rub.de. 2 The abbreviations used are: BV IX, biliverdin IX ; DHBV, 15,16-dihydrobiliverdin; ITC, isothermal titration calorimetry; PE, phycoerythrin; PEB, phycoerythrobilin; PCB, phycocyanobilin. 3 A. W. U. Busch, unpublished data. phycobiliprotein lyase families (E/F, S/U, and T) (6). These proteins are supposed to guide the reaction in a chaperone-like manner probably by conformational control of the bilin (7, 8). In addition, some lyases act as isomerases and generate phycoviolobilin or phycourobilin upon attachment to the phycobiliprotein (9 11). Whereas PCB, phycourobilin, or phycoviolobilin attaching lyases are well characterized, little is known about lyases specific for attachment of PEB to phycoerythrins (PE). We therefore investigated a putative PEB-transferring lyase encoded by P. marinus MED4 (12). Prochlorococcus strains dominate most oceanic phytoplankton communities (13). They possess a remarkable pigment composition and depend on unique divinyl-chlorophyll antenna complexes instead of phycobilisomes for photosynthetic light harvesting (14, 15), Despite the absence of phycobilisomes Prochlorococcus carries a special PE of unknown function (16 18). Genes associated with phycobiliprotein maturation can be found in highly reduced prochlorococcal as well as in Prochlorococcus-infecting phage genomes indicating the importance of PE for cell fitness (12, 19). Prochlorococcus occurs in two ecotypes adapted to specific niches (20). High-light strains inhabit the upper nutrient-depleted but light-irradiated layer of the ocean, whereas low-light strains colonize depths of up to 200 m, which are nutrient-rich but exposed to a low light intensity. Whereas low light-adapted strains encode and subunits of PE, HL strains like P. marinus MED4 carry only a degenerated form of -PE (14, 21). P. marinus MED4 possesses one of the smallest genomes of photosynthetic organisms but still encodes all components for phycobiliprotein maturation: -PE (CpeB), enzymes for PEB and PCB biosynthesis (i.e. HO1, PebA, PebB, and PcyA) and the S-type lyase CpeS (12, 22). We characterized CpeS from MED4 with respect to bilin specificity and affinity, binding kinetics, and transfer activity. Besides BV IX and DHBV, both the 3Z- and 3E-isomers of PEB and PCB were included in this study to address the so far unstudied stereoselectivity of lyases. Our results demonstrate that CpeS specifically transfers 3Z-PEB to CpeB and ensures correct configuration of the attachment product. The functionality of the last component of the phycobiliprotein maturation machinery in Prochlorococcus is demonstrated underlining the relevance of PE in this unusual cyanobacterium. EXPERIMENTAL PROCEDURES Construction of Expression Plasmid The P. marinus MED4 cpes coding region was PCR-amplified from genomic DNA Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, 2011 NOVEMBER 26, 2010 VOLUME 285 NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37561

71 3Z-PEB Cys 82 Lyase CpeS with primers encompassing recognition sites for EcoRI (5 -gaccatgattagaattcgttgacg-3 ) and HindIII (5 -ccaagcttgcctgcagagaataat-3 ). The PCR product was cloned into expression vector paskiba45 (IBA) for N-terminal fusion of a Strep-tag to CpeS. The resulting plasmid was verified by sequencing. Construction of plasmid pet43a-pe for production of N-terminal NusA His-tagged CpeB was described previously (18). Production and Purification of Recombinant Proteins A culture of BL21(DE3) carrying the respective overexpression plasmid was incubated at 37 C to an A 580 of 0.6. After induction with isopropyl- -D-thiogalactopyranoside (100 M; pet43 derivative) or anhydrotetracycline (200 g/ml; paskiba45 derivative) cells were incubated for 18 h at 17 C, harvested by centrifugation, resuspended in sodium phosphate buffer (60 mm sodium phosphate, 100 mm NaCl, ph 7.5) and disrupted by two passages through a French press cell at 20,000 psi. After separation of cell debris the supernatant was loaded on a Talon metal affinity resin (Clontech; CpeB) or a Strep- Tactin-Sepharose (IBA; CpeS) column. Purification was carried out according to the manufacturer s instructions based on sodium phosphate buffer. If needed, purified CpeB was incubated with thrombin (10 cleavage units/mg CpeB) and dialyzed against cleavage buffer (60 mm sodium phosphate, 150 mm NaCl, 5 mm MgCl 2, 2.5 mm CaCl 2, ph 7.8). Tag-free CpeB was then purified by a second metal affinity chromatography. Purified proteins were dialyzed against buffer, which was in case of CpeB supplemented with 1 mm dithiothreitol. CpeB was additionally dialyzed twice against buffer to remove dithiothreitol. Concentrations of proteins were determined using the calculated molar extinction coefficient 280 (23). Chemical Modification of Cysteine Residues CpeB was incubated with 10 mm iodoacetamide for 30 min and subsequently purified via Nap5 columns (GE Healthcare). SDS-PAGE and Zinc-blot Analysis Protein samples were analyzed using 12.5% or 15% SDS-PAGE (24). Zinc-blot analyses were carried out as described in (25). Spectroscopy UV-visible absorption measurements were done using an Agilent Technologies 8453 spectrophotometer. Fluorescence spectroscopy was performed with an Aminco- Bowman AB2 spectrofluorimeter. Preparation of Bilins BV IX was obtained from Frontier Scientific. 3E-PEB and 3E-PCB were isolated from Porphyridium (3E-PEB) or Spirulina (3E-PCB) cells as described in (26) and references therein. Preparative anaerobic production of bilins was performed at 20 C as described earlier (27) with the following modifications. Ferredoxin-NADP(H)-oxidoreductase from Synechococcus sp. PCC7002 (PetH) and ferredoxin from the cyanophage P-SSM2 or Synechococcus sp. PCC7002 were used. Substrate was added sequentially in excess amounts, and the reaction was started with an NADPH-regenerating system. Bilins were prepared with PebS from cyanophage P-SSM2 (3Z-PEB), PcyA from Nostoc sp. PCC7120 (3Z- PCB), PebA from Synechococcus sp. WH8020 (DHBV), and PebS variants unable to perform DHBV reduction (DHBV) (4, 28, 29). 3 Bilins were purified according to Ref. 4 and analyzed by HPLC. While 3E-PEB and 3E-PCB were obtained to almost 100% purity, 3Z-PEB and DHBV contained less than 6% and 3Z-PCB less than 10% contaminations. Determination of Bilin Concentration Bilins were dissolved in dimethyl sulfoxide, diluted 1:100 in MeOH/5% HCl, and concentrations were estimated using the following molar extinction coefficients in MeOH/5% HCl: ,000 M 1 cm 1 (3E- PEB) (30); ,900 M 1 cm 1 (3E-PCB) (31); and ,800 M 1 cm 1 (BV IX ) (32). Concentrations of 3Z-PEB and DHBV were calculated with the 594 of 3E-PEB and the measured absorption at 571 nm (3Z-PEB) or 564 nm (DHBV) (4). Determination of CpeS Bilin Specificity Assays were performed in sodium phosphate buffer, ph 7.5, at room temperature. Lyase (10 M) and bilin (5 M) were incubated for 10 min before affinity purification of CpeS. Columns were washed with increasing NaCl concentrations (100 mm 1 M) and wash fractions screened for released bilins by absorption spectroscopy. Spectroscopic Determination of Protein to Bilin Ratio in CpeS Bilin Complexes Lyase (15 M) and bilin (30 M) were incubated in sodium phosphate buffer, ph 7.5, for 10 min before affinity purification of CpeS. Absorption spectra of eluted CpeS bilin complexes were obtained in sodium phosphate buffer and after addition of 0.5 volume of acidic urea (8 M, ph 1.5). Protein concentrations in buffer were determined with the calculated 280 of CpeS (23), bilin concentrations in acidic urea with ,800 M 1 cm 1 (PEB) (33), or ,000 M 1 cm 1 (DHBV) (34). Stopped Flow Kinetics of Bilin Binding to CpeS Stopped flow experiments were performed in an SFM-400 apparatus with MOS-200 optics (Bio-Logic) at 20 C in sodium phosphate buffer, ph 7.5. Bilin (2 M) was rapidly mixed with various concentrations of CpeS and the absorption increase at 594 nm (3E-PEB), 590 nm (3Z-PEB), and 607 nm (DHBV) was recorded. Ten time traces were accumulated and averaged each. An exponential equation was fitted to the experimental data yielding the k obs value for each concentration. All experiments were repeated two times. Isothermal titration calorimetry Thermodynamic parameters of CpeS interaction with 3E-PEB at 20 C in sodium-phosphate buffer ph 7.5 were determined using an isothermal titration microcalorimeter (MicroCal Auto-iTC 200, GE Healthcare). The first 0.5 lof3e-peb (300 M) were injected into the temperature-controlled sample cell containing 200 l of CpeS (30 M). The change in heating power was monitored for 180 s until equilibrium was reached before the next 1.8- l injection was started. Background heat generation from dilution of 3E-PEB in buffer was measured to be irrelevant. The data were averaged over two ITC experiments. Size Exclusion Chromatography An GE Healthcare Superdex 75 HR10/300 GL size exclusion column was equilibrated in sodium phosphate buffer at a flow rate of 0.5 ml/min. Standards with known M r (i.e. bovine serum albumin, 66,000; carbonic anhydrase, 29,000; cytochrome c, 12,400; and aprotinin, 6,500) were applied to the column, and their elution volumes were determined spectroscopically. CpeS and tag-free CpeB were chromatographed under identical conditions. Analyses of CpeS Bound Bilins Affinity-purified CpeS bilin complexes were denaturated in acidic urea (8 M, ph 1.5). Subsequently, bilins were purified as described in Ref. 4 and further analyzed on a Luna C18 reversed-phase HPLC column (Phenomenex) with diode array detection (Agilent 1100 series) (29). Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 NUMBER 48 NOVEMBER 26, 2010

72 Chromophore Transfer Assays Bilin (15 M) was incubated with or without CpeS (30 M) for 5 min. Total binding of bilin to CpeS was confirmed via absorption spectroscopy. In control experiments, affinity purified CpeS bilin was used. CpeB (30 M) was added, and reactions were followed spectroscopically. After 45 min, CpeB was isolated by affinity chromatography and further analyzed. In competition assays, bilin (3.5 M) was added to CpeS, CpeB, or premixed CpeS/CpeB (7 M each), and absorption and fluorescence emission spectra were taken immediately. Analyses of CpeB-bound Bilins Purified CpeB was digested with trypsin and resulting peptides desalted via a C18 Sep-Pak column (Waters) (4). Peptides were separated by HPLC on a Jupiter C4 (300 Å) column (Phenomenex) (35). Isolated chromopeptides were dissolved in 50% acetonitrile, 1% TFA, and analyzed by MALDI-TOF MS and MS/MS using the Ultraflex II Instrument (Bruker Daltonik, Bremen) and a modification of a method described in Ref. 36. Peptides were spotted onto an MTP AnchorChip 800/384 target (Bruker Daltonik) preloaded with -cyano-4-hydroxycinnamic acid matrix according to the manufacturer s instructions. MS and MS/MS spectra were acquired using the flexcontrol 3.0 software (Bruker Daltonik). The instrument was set to positive polarity. Calibration of MS spectra was performed using a peptide mix consisting of bradykinin ( Da), angiotensin II ( Da), Substance P ( Da), bombesin ( Da), ACTH ( Da), and ACTH CLIP ( Da). Peak lists were generated from MS and MS/MS spectra with the help of flexanalysis 3.0 software (Bruker Daltonik). Peak masses from MS data were compared with theoretical masses of peptides containing the covalently linked bilin. MS peaks that matched the theoretical masses were selected as precursor ions for MS/MS analysis to confirm peptide identity. RESULTS Purification of Recombinant CpeS and CpeB P. marinus MED4 possesses an unusual green light-absorbing -PE (CpeB) with one PEB bound at Cys 82 (18). Adjacent to the cpeb gene, the putative S-type lyase CpeS is encoded. Besides CpeS, no other lyase homolog was identified in MED4. To study whether the putative lyase is sufficient for bilin attachment to -PE, corresponding proteins were overproduced in E. coli and purified via affinity chromatography (supplemental Fig. S1). CpeS was obtained to almost 100% purity, CpeB to an estimated purity of 50%. Because CpeB was produced with an N-terminal NusA-tag, impurities most likely represent NusA breakdown products, which were removable by affinity purification of thrombin-cleaved CpeB. However, truncated CpeB was unstable and therefore only employed in small-scale control experiments to rule out an influence of either tag or contaminations on bilin binding assays. In addition, covalent binding of bilins to contaminating proteins was ruled out by zinc-blot analysis (supplemental Fig. S2). CpeS Binds DHBV and Both PEB Isomers Prochlorococcus possesses the functional biosynthetic machinery for production of PEB and PCB (37). Although PEB is the chromophore of -PE, no PCB-binding protein has been identified in MED4 so far. 3Z-PEB Cys 82 Lyase CpeS FIGURE 1. Absorption spectra of CpeS bilin complexes. CpeS was incubated with BV IX, 3E-PCB, 3Z-PCB, DHBV, 3E-PEB, or 3Z-PEB and subsequently affinity purified. Unbound bilins were washed out during chromatography with increasing salt concentrations in buffer. Only DHBV (A), 3E-PEB (B), or 3Z-PEB (C) formed stable complexes with CpeS (solid lines). For comparison, absorption spectra of bilins in buffer are given (dashed lines). AU, absorbance units. Because members of the S/U type lyase family are characterized by a low bilin specificity (6, 8), we tested CpeS for binding of all bilins synthesized by MED4. We included 3Z- and 3E-isomers of PCB and PEB to address the largely unstudied bilin stereoselectivity of lyases. Absorption spectra of BV IX, 3E- PCB, and 3Z-PCB were only slightly influenced by CpeS (data not shown). Mixing of CpeS with DHBV, 3E-PEB, or 3Z-PEB immediately resulted in the formation of bluish, nonfluorescent complexes with increased long wavelength absorption at 607 nm (DHBV), 594 nm (3E-PEB), or 590 nm (3Z-PEB) (Fig. 1), indicating binding of bilins in a stretched conformation. During subsequent affinity purification of lyase-bilin mixtures, washing with increasing salt concentrations lead to release of total BV IX, 3E-PCB, and 3Z-PCB at low salt concentrations (200 mm NaCl), whereas DHBV, 3E-PEB, and 3Z-PEB were retained, even at high salt concentrations (1 M NaCl) (data not shown). Absorption spectra of eluted complexes were comparable with those obtained before purification indicating stable binding of DHBV, 3E-PEB, and 3Z-PEB to CpeS. To investigate whether this stability is due to covalent binding, SDS-PAGE and subsequent zinc-blot analyses were performed (supplemental Fig. S2A). Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, 2011 NOVEMBER 26, 2010 VOLUME 285 NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37563

73 3Z-PEB Cys 82 Lyase CpeS FIGURE 2. CpeS homodimer formation. CpeS was analyzed by size exclusion chromatography on a Superdex 75 column. CpeS was chromatographed in the presence (A) or absence (B) of 3E-PEB. Protein was detected by absorbance at 280 nm (solid lines), and bilin was detected by absorbance at 594 nm (dashed lines). The elution volumes of standards with known molecular weight (MW) were used to determine the oligomerization state of CpeS (C). AU, absorbance units. No zinc-induced red fluorescence was detected indicating a noncovalent interaction. Dimeric CpeS Binds Two Molecules of DHBV, 3E-PEB, or 3Z-PEB S-type lyases can occur in different oligomeric states (38 40). Therefore, we performed size exclusion chromatography of CpeS in the presence and absence of bilins. In the absence of bilins, CpeS eluted as a complex of 47.1 kda (Fig. 2A), which correlates well with the calculated mass of a homodimer (46.6 kda). Elution profiles in the presence of bilins were almost identical with protein and bilin eluting simultaneously as CpeS bilin complexes of 50.8 kda (DHBV), 50.1 kda (3E-PEB), and 50.3 kda (3Z-PEB) (Fig. 2B). These data suggest CpeS dimer formation prior to bilin binding. Affinity purified CpeS bilin complexes were denatured in acidic urea to determine the ratio of CpeS to bound bilin. According to the absorption in sodium phosphate buffer and acidic urea, protein to bilin ratios of 1:0.89 (3E-PEB), 1:0.84 (3Z-PEB), and 1:0.79 (DHBV) were calculated (data not shown) FIGURE 3. Bilin binding to CpeS measured by stopped flow experiments. 2 M bilin was rapidly mixed with increasing concentrations of CpeS. The observed rate constants (k obs ) were obtained from the exponential fit of the absorbance increase at 594 nm (3E-PEB; filled circles), 590 nm (3Z-PEB; squares), and 607 nm (DHBV; circles), respectively, and plotted against the CpeS concentration. TABLE 1 Rate constants and thermodynamic parameters for the interaction of CpeS with bilins The association rate constants (k on ) were determined by stopped flow experiments, binding constant (K a ), enthalpy ( H 0 ), and entropy change ( S 0 ) by ITC. From those values, K d 1/K a and k off k on K d were calculated. ND, not determined. Bilin k on K a K d k off H 0 S 0 M 1 s M 1 M s 1 kcal/mol cal/mol/deg DHBV ND ND ND ND ND 3E-PEB Z-PEB ND ND ND ND ND indicating dimeric CpeS with two bilins bound. Absorption spectra of denatured complexes were essentially the same as those of free bilins in the corresponding solvent and released bilins were identified by HPLC to be unmodified (supplemental Fig. S3). Extinction coefficients of CpeS bilin complexes were determined per bilin as follows: 594, 74,000 M 1 cm 1 (CpeS 3E-PEB); ,000 M 1 cm 1 (CpeS 3Z-PEB); and 607, 87,800 M 1 cm 1 (CpeS DHBV). CpeS Binds DHBV and Both PEB Isomers with Similar Binding Kinetics and High Affinity As CpeS bilin complexes were formed immediately after mixing, binding kinetics of both components were analyzed by stopped flow experiments. CpeS and bilins were rapidly mixed at different ratios and absorption increases at 594 nm (3E-PEB), 590 nm (3Z-PEB), or 607 nm (DHBV) were recorded. A single exponential function was fitted to the experimental data, from which the observed rate constant k obs was extracted (supplemental Fig. S4). The k obs values were plotted against the protein concentration (Fig. 3). Linear regression yielded the association rate constants k on as the slopes of the straight lines with standard errors in the range of 10%. 3E-PEB shows the fasted association rate constant (2 M 1 s 1 ) followed by 3Z-PEB (1.4 M 1 s 1 ) and DHBV (0.9 M 1 s 1 ) (Table 1). However, k on values were in the same range especially when considering possible inexactness in DHBV or 3Z-PEB concentration due to the unknown extinction coefficients. The ordinate intercept values, which correspond to the dissociation rate constant k off, are close to 1 s 1.As this value is small, the standard error is of the same magnitude so that a precise value for k off cannot be obtained. Nevertheless, Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 NUMBER 48 NOVEMBER 26, 2010

74 3Z-PEB Cys 82 Lyase CpeS a value close to 1 s 1 can be confirmed by calculation from the k on value and the K d value obtained from ITC (see below), which yields exactly K d k on k off 1.4 s 1. FIGURE 4. 3E-PEB binding to CpeS measured by ITC. The ITC data from a representative experiment are shown. 3E-PEB was injected into the temperature-controlled sample cell containing CpeS, and the change in heating power was recorded (top panel). By integration, the generated heat was obtained, which was plotted against the molar ratio of bilin and CpeS (bottom panel) together with the fitted one site binding isotherm. From this fit, the values for stoichiometry (N), binding constant (K a ), enthalpy ( H 0 ), and entropy change ( S 0 ) were obtained. FIGURE 5. CpeS-mediated and spontaneous chromophorylation of CpeB. CpeB was incubated with free bilins (A C) or CpeS bilin complexes (D F). Fluorescence emission ( ex 550 nm) was recorded directly after addition of CpeB and further every 5 min over a period of 45 min. In the case of CpeB, incubation with DHBV (A) or CpeS DHBV (D), only the 1 min and 45 min spectra are given. In the case of CpeB incubation with 3E-PEB, an additional 90 min spectra is shown. The course of fluorescence emission changes is indicated by arrows. CpeB incubation with 3E-PEB (B)or3Z-PEB (C) resulted after 10 min in a decrease of fluorescence intensity (broken lines), whereas in the case of CpeS 3E-PEB (E) or CpeS 3Z-PEB (F) fluorescence constantly increased (solid lines). Note that scales differ by a factor of 4. AU, absorbance units. CpeS 3E-PEB complex affinity was measured by ITC. 3Z-PEB and DHBV were not included due to insufficient amounts available. Experiments were repeated two times with an example of a result presented in Fig. 4, and mean values are shown in Table 1. The stoichiometry of CpeS 3E-PEB was n in reasonable agreement with the spectroscopically determined 1:0.89 ratio in the complex. The resulting K d value of the CpeS 3E-PEB complex was 0.7 M, which represents tight binding and which is driven by both favorable enthalpy ( H kcal/mol) as well as entropy ( S cal/mol/deg) changes upon complex formation. The K d determined for the CpeS 3E-PEB complex is in range with values obtained for other lyases (8, 39, 41, 42). CpeS Transfers PEB but Not DHBV to Dimeric CpeB To verify whether CpeS is sufficient for correct bilin addition to -PE, we compared spontaneous and lyase-mediated attachment of bilins to CpeB by recording the fluorescence emission of assembled holo-cpeb. DHBV was neither spontaneous nor under catalysis of CpeS attached to CpeB (Fig. 5, A and D). This was further verified by zinc-blot analysis (supplemental Fig. S2B) and HPLC of CpeB peptides (data not shown). Spontaneous and CpeS-mediated reactions of both PEB isomers with CpeB resulted in formation of fluorescent complexes with covalently bound bilins (Fig. 5 and supplemental Fig. S2B). Reaction products were stronger fluorescent when obtained (i) with 3Z-PEB and (ii) in presence of CpeS. The course of spontaneous and lyase-supported reactions were significantly different. While in the presence of CpeS, fluorescence intensities increased constantly (Fig. 5, E and F), maximal fluorescence was reached in the absence of CpeS after 10 min followed by an intensity decrease (Fig. 5, B and C). CpeS-mediated reactions as well as the spontaneous reaction of 3Z-PEB with CpeB were largely completed within 40 min, whereas addition of free 3E-PEB to CpeB was not completed within 90 min. This was confirmed by following spontaneous addition of PEB to CpeB via absorption spectroscopy (supplemental Fig. S5). Mixing of CpeB with either 3E-PEB or 3Z-PEB resulted instantly in a blue shift of the long wavelength absorption maximum of bilins. Whether this blue shift reflects the formation of a thioether linkage between CpeB and PEB, conformational changes of the bilin or a combination of both is unclear. The blue-shifted maximum decreased while simultaneously a red-shifted absorption maximum increased. In the case of 3Z-PEB, the reaction was completed after 40 min (supplemental Fig. S5, C and D). 3E-PEB addition to CpeB was not completed until at least 2 h (supplemental Fig. S5, A and B). Lyase-mediated transfer reactions were not Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, 2011 NOVEMBER 26, 2010 VOLUME 285 NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37565

75 3Z-PEB Cys 82 Lyase CpeS followed via absorbance spectroscopy due to superposition of CpeS bilin signals with those of holo-cpeb. However, the initial blue-shifted intermediate observed for spontaneous addition of PEB to CpeB was not detected for CpeS-mediated reactions. One might speculate that the fluorescence in the first phase of spontaneous PEB addition to CpeB is ascribed to the blueshifted intermediate. While absorption of this intermediate decreases, fluorescence decreases, too. There was no conversion of 3Z- into 3E-PEB or vice versa under our assay conditions (supplemental Fig. S3). As described above, CpeS is a dimer with two bilins bound. To verify the oligomerization state of the lyase target protein CpeB, size exclusion chromatography with tag-free CpeB was performed (supplemental Fig. S6). After incubation with 3E- PEB for 2 h, CpeB eluted mainly as complex of 45 kda as judged by the absorbance of bound bilin (supplemental Fig. S6A). This correlates well with the calculated mass of a CpeB homodimer (43.4 kda). An additional broad shoulder corresponding to either higher ordered oligomers or aggregates was detected. These CpeB fractions had no bilin bound. In the absence of bilin, the elution profile of CpeB was essentially the same as in the presence of 3E-PEB (supplemental Fig. S6B). However, the amount of higher ordered oligomers or aggregates increased in the absence of bilin. These data suggest interaction of dimeric CpeS with dimeric CpeB. One might speculate that both Cys 82 sites in the CpeB dimer are supplied with PEB before the CpeS CpeB complex falls apart. CpeS Is a 3Z-PEB-transferring Lyase CpeB was affinity purified after 45 min of incubation with either free bilin or CpeS bilin. As shown in Fig. 6 and Table 2, CpeB chromophorylation products differed significantly in their spectroscopic properties. Holo-CpeB generated by incubation with free bilins or with CpeS 3E-PEB exhibited at least two absorption maxima, with one at 610 nm typical for phycobiliprotein-bound DHBV (43, 44). Only CpeS 3Z-PEB-mediated reactions resulted in formation of holo-cpeb with identical spectroscopic properties as native -PE from P. marinus MED4 (18). Bilin types and binding sites in CpeB were determined by analysis of CpeB bilin peptides via HPLC (supplemental Fig. S7) and MALDI-TOF MS and MS/MS (supplemental Fig. S8 and Table 3). CpeS-mediated addition of 3Z-PEB to CpeB resulted in binding of PEB to Cys 82, whereas transfer of 3E-PEB yielded in addition Cys 82 -DHBV. Spontaneous addition of 3E-PEB and 3Z-PEB resulted in formation of Cys 82 -DHBV and of a second product whose identification via MALDI-TOF failed as for unknown reasons no bilin peptides were detectable. Absorption spectra of this product were nearly identical to those of lyase-produced Cys 82 -PEB, but elution times differed significantly. The PEB attachment site of spontaneous bilin addition to CpeB was indirectly narrowed down to Cys 82 by several lines of evidence: (i) binding of PEB to the CpeB tag or to impurities was ruled out by repeating experiments with truncated CpeB (data not shown); (ii) modification of CpeB cysteine residues using iodoacetamide abolished bilin binding (data not shown) (18); (iii) CpeB Cys 82 mutants do not form fluorescent adducts with 3E-PEB in vitro (18). Due to the nearly identical absorbance of bilin peptides 1 and 2, we concluded that both carry PEB at Cys 82. Differences in HPLC retention times of FIGURE 6. Absorbance and fluorescence emission spectra of purified holo-cpeb. Absorbance (solid lines) and fluorescence emission (broken lines; ex 550 nm) spectra of affinity purified holo-cpeb are given. CpeB was reconstituted with the following: 3E-PEB (A), 3Z-PEB (B), CpeS 3E-PEB (C), or CpeS 3Z-PEB (D). AU, absorbance units. TABLE 2 Spectroscopic properties of affinity purified holo-cpeb CpeB was reconstituted with 3E-or3Z-PEB in the presence ( ) or absence ( )of CpeS. Fluorescence emission was recorded after excitation at 550 nm. Values for native -PE from P. marinus MED4 were taken from Ref. 18. Bilin CpeS Absorption maxima Fluorescence emission maximum nm nm 3E-PEB 537, 577, Z-PEB 554, 574, E-PEB 565, Z-PEB Native -PE spontaneously and CpeS-generated Cys 82 -PEB peptides must be due to altered conformation or configuration of the bilin peptide. DISCUSSION S/U-type lyases are classified into five groups (45). Until now, only members of groups CpcS-I, CpcS-III, and CpcU were well Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 NUMBER 48 NOVEMBER 26, 2010

76 TABLE 3 Identification of tryptic holo-cpeb bilin peptides via MALDI-TOF Holo-CpeB was reconstituted with 3E-PEB, 3Z-PEB, CpeS 3E-PEB, or CpeS 3Z-PEB as indicated. Three different tryptic bilin peptides (peak nos. 1, 2, 3) of holo-cpeb were separated by HPLC and identified via MALDI-TOF. Bilin peptide sequences are given with the attached bilin in bold face in brackets. Calculated (Cal) and experimentally determined (Exp) m/z of bilin peptides are given. ND, not determined. Peak HPLC retention time Bilin peptide sequence studied. These lyases attach PCB to Cys 84 of phycocyanin or phycoerythrocyanin and to Cys 82 and Cys 82 of allophycocyanin (38, 39, 46). 4 While the function of CpcV group lyases is enigmatic, S-type lyases of groups CpeS and CpeU were speculated to attach PEB to PE subunits, due to their encoding in gene clusters associated to PE maturation (45). We have now characterized the first CpeS-like lyase from P. marinus MED4 and demonstrated that CpeS is a bilin lyase specific for attachment of 3Z-PEB to Cys 82 of PE. S/U-type lyases exhibit different oligomerization states: heterodimeric, homodimeric, or monomeric (38 40). As P. marinus MED4 CpeS is homodimeric, S-type lyases appear to be mainly dimeric with the exception of Nostoc sp. PCC7120 CpcS (39). However, the assumption of a CpcS monomer is based on missing self-interaction in pulldown assays, which might be due to stable dimer formation prior to experiments. Dimeric CpeS binds two molecules of bilin with similar binding kinetics for 3E-PEB, 3Z-PEB, and DHBV. DHBV is a biosynthetic intermediate transferred from the ferredoxin-dependent bilin reductase PebA to its homolog PebB probably by metabolic channeling (4). PebB, in turn, catalyzes the reduction of DHBV to 3Z-PEB. Considering direct transfer of DHBV from PebA to PebB binding of DHBV by CpeS in vivo seems unlikely. In line with that assumption, CpeS does not attach DHBV to CpeB in vitro. However, DHBV-transferring lyases are at least feasible for cryptophytes with DHBV-containing phycobiliproteins (47). As lyases are also able to detach bilins (8, 41, 48), CpeS might remove incorrectly bound DHBV from CpeB, but this could not yet be verified in preliminary studies (data not shown). Although CpeS binds 3E-PEB, 3Z-PEB, and DHBV tightly, there is no complex formation with BV IX, 3E-PCB, or 3Z-PCB. The common difference between these two groups of bilins is the reduction state of the methine bridge between C15 and C16. CpeS bound bilins have a reduced 15,16 bond and thereby a new chiral center at C16 in the R configuration (49). In contrast, PCB-specific lyase-isomerase PecE/PecF from Cal 4 The consensus numbering of bilin attachment sites is given. m/z Exp min 3E-PEB ND ND ND MSVC(DHBV)LR Z-PEB ND ND ND MSVC(DHBV)LR CpeS 3E-PEB MSVC(PEB)LR MSVC(DHBV)LR CpeS 3Z-PEB MSVC(PEB)LR Z-PEB Cys 82 Lyase CpeS Mastigocladus laminosus binds only bilins with a 15,16 double bond (50). Thus, binding of the correct bilin may be ensured in vivo by a specific interaction of lyases with the 15,16 bond of either PCB or PEB. However, CpcS1 from Nostoc sp. PCC7120 and CpcE/CpcF from Synechococcus sp. PCC7002 bind both PEB and PCB (8, 35, 41). One might argue that these species synthesize only PCB but not PEB, so that discrimination between these two bilins is unnecessary. Further comparing studies of bilin binding by PCB- and PEB-transferring lyases must be carried out to address this interesting hypothesis. Although CpeS binds bilins with a reduced 15,16 bond, only those with a further reduction of the A-ring diene system are transferred to CpeB, namely 3E- and 3Z-PEB. The A-ring diene system reduction of BV IX results in two possible stereoisomeric ethylidene groups at the C3 carbon atom, 3Z and 3E. 3Z-isomers are the biosynthetic products of different ferredoxin-dependent bilin reductases 3 (4, 5, 28, 29, 37, 51). However, 3E-isomers are often byproduct of bilin extraction methods and bilins isolated from phycobiliproteins by methanolysis are mainly 3E-configured. CpeS transfers both 3E-PEB and 3Z-PEB to CpeB, but only 3Z-PEB yields exclusively correctly bound PEB at Cys 82 and holo-cpeb with identical spectroscopic features as native -PE (18). 3E-PEB addition by CpeS is less effective and yields oxidized side products. Therefore, 3Z-PEB is the natural substrate for CpeS-mediated transfer to CpeB. Binding and transfer of 3E- and 3Z-isomers was shown for other lyases as well. In some cases, the functionality of lyases was proven in in vivo reconstitution systems and in in vitro assays. While the in vivo assays provide 3Z-PCB as substrate, the in vitro assays were performed with 3E-PCB. In agreement with the results of this study, CpcS1 and CpcT1 from Nostoc sp. PCC7120 are able to bind and transfer both PCB isomers (8, 35, 52). Although assays with 3Z-PCB result in a correctly bound chromophore, attachment of 3E-PCB by CpcT1 is of low fidelity and leads to partial oxidation to mesobiliverdin (52). Products of CpcS1 mediated 3E- and 3Z-PCB transfer to -phycoerythrocyanin have nearly identical spectroscopic features as native -phycoerythrocyanin, but only 3Z-PCB addition products match extinction coefficients of native -phycoerythrocyanin (39). Synechococcus sp. PCC7002 CpcT transfers 3E-PCB and 3Z-PCB to phycocyanin but only spectroscopic data of the latter product are perfectly in agreement with those of native phycocyanin (53). In conclusion, lyases bind bilins with different stereochemistry of the C3 ethylidene group. Both isomers can be attached to phycobiliproteins but the 3E-isomer is only partially protected from oxidation. Thus 3Z-isomer specificity may be a common feature of all lyases. This is in agreement with 3Z-isomers as the primary product of bilin biosynthesis and points toward direct bilin transfer from the last biosynthetic enzyme to the lyase. Binding of 3Z-PEB to CpeS is fast with an association rate constant of 1.4 M 1 s 1 and in competition assays with CpeB, added bilin was completely bound by the lyase (data not shown). As already speculated for other lyases (8), rapid bilin binding to CpeS may prevent nonenzymatic transfer to CpeB, which results in false addition products. Spontaneous addition of PEB to CpeB is of low yield and fidelity and addition products Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, 2011 NOVEMBER 26, 2010 VOLUME 285 NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37567

77 3Z-PEB Cys 82 Lyase CpeS are Cys 82 -DHBV and most likely a Cys 82 -PEB with an altered, incorrect configuration or conformation. Indeed, nonenzymatic addition of bilins to other phycobiliproteins was shown to be inefficient and to produce oxidized side products (11, 35, 38, 39, 46, 53). In the case of CpcA, CpcB, PecA, and PecB from Mastigocladus laminosus, PCB adds spontaneously to the correct Cys 84 binding site but in a non-native configuration (54). Whereas bilins in solution are in all Z configuration and all syn-conformation (55 57), phycobiliprotein-bound bilins are mainly Z,Z,Z/anti,syn,anti-configured (58 60). However, upon denaturation or digestion of the phycobiliprotein matrix bilins are released and adopt their Z,Z,Z/syn,syn-,syn-porphyrin-like form again. This is in line with the almost identical absorption of non-enzymatic and CpeS-catalyzed Cys 82 -PEB peptides during HPLC. Differences in retention times might be due to differences in PEB binding to Cys 82. Bilin addition to phycobiliproteins generates two new chiral carbon atoms, C3 and C3, with two possible configurations each. Whereas the configuration of C3 is always R, the configuration of C3 at binding sites Cys 84, Cys 143, and Cys 84 of different phycobiliproteins is R, whereas it is S for binding sites Cys 155 and Cys 50/61 (58, 59, 61 63). We therefore speculate that CpeS-mediated attachment of PEB yields Cys 82 -C3 (R)- PEB, while non-assisted attachment of PEB results in formation of Cys 82 -C3 (S)-PEB. However, verification of the C3 configuration by NMR failed so far, due to insufficient amounts. Current research in our laboratory will address this further to tackle this interesting question. Acknowledgments We thank Claudia Steglich and Wolfgang Hess (Freiburg) for providing P. marinus MED4 chromosomal DNA and Rosa Rosello (Karlsruhe) for providing Porphyridium cells. We also thank Hugo Scheer and Stephan Böhm (Munich) for introduction of methods used in this study. REFERENCES 1. Glazer, A. N. (1985) Annu. Rev. Biophys. Biophys. Chem. 14, Sidler, W. (1994) in The Molecular Biology of Cyanobacteria, pp , Kluwer Academic Publishers, Dordrecht, The Netherlands 3. Frankenberg-Dinkel, N., and Lagarias, J. C. (2003) in The Porphyrin Handbook, pp , Academic Press, Amsterdam, The Netherlands 4. Dammeyer, T., and Frankenberg-Dinkel, N. (2006) J. Biol. Chem. 281, Frankenberg, N., Mukougawa, K., Kohchi, T., and Lagarias, J. C. (2001) Plant Cell 13, Scheer, H., and Zhao, K. H. (2008) Mol. Microbiol. 68, Böhm, S., Endres, S., Scheer, H., and Zhao, K. H. (2007) J. Biol. Chem. 282, Kupka, M., Zhang, J., Fu, W. L., Tu, J. M., Böhm, S., Su, P., Chen, Y., Zhou, M., Scheer, H., and Zhao, K. H. (2009) J. Biol. Chem. 284, Fairchild, C. D., Zhao, J., Zhou, J., Colson, S. E., Bryant, D. A., and Glazer, A. N. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, Blot, N., Wu, X. J., Thomas, J. C., Zhang, J., Garczarek, L., Böhm, S., Tu, J. M., Zhou, M., Plöscher, M., Eichacker, L., Partensky, F., Scheer, H., and Zhao, K. H. (2009) J. Biol. Chem. 284, Zhao, K. H., Deng, M. G., Zheng, M., Zhou, M., Parbel, A., Storf, M., Meyer, M., Strohmann, B., and Scheer, H. (2000) FEBS Lett. 469, Rocap, G., Larimer, F. W., Lamerdin, J., Malfatti, S., Chain, P., Ahlgren, N. A., Arellano, A., Coleman, M., Hauser, L., Hess, W. R., Johnson, Z. I., Land, M., Lindell, D., Post, A. F., Regala, W., Shah, M., Shaw, S. L., Steglich, C., Sullivan, M. B., Ting, C. S., Tolonen, A., Webb, E. A., Zinser, E. R., and Chisholm, S. W. (2003) Nature 424, Partensky, F., Hess, W. R., and Vaulot, D. (1999) Microbiol. Mol. Biol. Rev. 63, Steglich, C., Post, A. F., and Hess, W. R. (2003) Environ. Microbiol. 5, Ting, C. S., Rocap, G., King, J., and Chisholm, S. W. (2002) Trends Microbiol. 10, Hess, W. R., Partensky, F., van der Staay, G. W., Garcia-Fernandez, J. M., Börner, T., and Vaulot, D. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, Steglich, C., Mullineaux, C. W., Teuchner, K., Hess, W. R., and Lokstein, H. (2003) FEBS Lett. 553, Steglich, C., Frankenberg-Dinkel, N., Penno, S., and Hess, W. R. (2005) Environ. Microbiol. 7, Sullivan, M. B., Coleman, M. L., Weigele, P., Rohwer, F., and Chisholm, S. W. (2005) PLoS Biol. 3, e Moore, L. R., Rocap, G., and Chisholm, S. W. (1998) Nature 393, Ting, C. S., Rocap, G., King, J., and Chisholm, S. W. (2001) Microbiology 147, Kettler, G. C., Martiny, A. C., Huang, K., Zucker, J., Coleman, M. L., Rodrigue, S., Chen, F., Lapidus, A., Ferriera, S., Johnson, J., Steglich, C., Church, G. M., Richardson, P., and Chisholm, S. W. (2007) PLoS Genet. 3, e Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, Laemmli, U. K. (1970) Nature 227, Wahleithner, J. A., Li, L. M., and Lagarias, J. C. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, Terry, M. J. (2002) in Heme, Chlorophyll, and Bilins, Methods and Protocols, pp , Humana Press, Totowa, NJ 27. Tu, S. L., Gunn, A., Toney, M. D., Britt, R. D., and Lagarias, J. C. (2004) J. Am. Chem. Soc. 126, Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W., and Frankenberg-Dinkel, N. (2008) Curr. Biol. 18, Frankenberg, N., and Lagarias, J. C. (2003) J. Biol. Chem. 278, Gossauer, A., and Klahr, E. (1979) Chem. Bet. 112, Cole, W. J., Chapman, D. J., and Siegelman, H. W. (1968) Biochemistry 7, Heirwegh, K. P., Blanckaert, N., and Van Hees, G. (1991) Anal. Biochem. 195, Glazer, A. N., and Hixson, C. S. (1975) J. Biol. Chem. 250, Wedemayer, G. J., Kidd, D. G., Wemmer, D. E., and Glazer, A. N. (1992) J. Biol. Chem. 267, Zhao, K. H., Su, P., Tu, J. M., Wang, X., Liu, H., Plöscher, M., Eichacker, L., Yang, B., Zhou, M., and Scheer, H. (2007) Proc. Natl. Acad. Sci. U.S.A. 104, Grzendowski, M., Wolter, M., Riemenschneider, M. J., Knobbe, C. B., Schlegel, U., Meyer, H. E., Reifenberger, G., and Stühler, K. (2010) Neuro. Oncol. 12, Dammeyer, T., Michaelsen, K., and Frankenberg-Dinkel, N. (2007) FEMS Microbiol. Lett. 271, Saunée, N. A., Williams, S. R., Bryant, D. A., and Schluchter, W. M. (2008) J. Biol. Chem. 283, Zhao, K. H., Su, P., Li, J., Tu, J. M., Zhou, M., Bubenzer, C., and Scheer, H. (2006) J. Biol. Chem. 281, Schluchter, W. M., Shen, G., Alvey, R. M., Biswas, A., Saunée, N. A., Williams, S. R., Mille, C. A., and Bryant, D. A. (2010) Adv. Exp. Med. Biol. 675, Fairchild, C. D., and Glazer, A. N. (1994) J. Biol. Chem. 269, Zhao, K. H., Wu, D., Zhou, M., Zhang, L., Böhm, S., Bubenzer, C., and Scheer, H. (2005) Biochemistry 44, Fairchild, C. D., and Glazer, A. N. (1994) J. Biol. Chem. 269, Arciero, D. M., Dallas, J. L., and Glazer, A. N. (1988) J. Biol. Chem. 263, Shen, G., Schluchter, W. M., and Bryant, D. A. (2008) J. Biol. Chem. 283, Biswas, A., Vasquez, Y. M., Dragomani, T. M., Kronfel, M. L., Williams, S. R., Alvey, R. M., Bryant, D. A., and Schluchter, W. M. (2010) Appl. Environ. Microbiol. 76, Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 NUMBER 48 NOVEMBER 26, 2010

78 47. Wemmer, D. E., Wedemayer, G. J., and Glazer, A. N. (1993) J. Biol. Chem. 268, Zhao, K. H., Ran, Y., Li, M., Sun, Y. N., Zhou, M., Storf, M., Kupka, M., Böhm, S., Bubenzer, C., and Scheer, H. (2004) Biochemistry 43, Gossauer, A., and Weller, J. P. (1978) J. Am. Chem. Soc. 100, Storf, M., Parbel, A., Meyer, M., Strohmann, B., Scheer, H., Deng, M. G., Zheng, M., Zhou, M., and Zhao, K. H. (2001) Biochemistry 40, Kohchi, T., Mukougawa, K., Frankenberg, N., Masuda, M., Yokota, A., and Lagarias, J. C. (2001) Plant Cell 13, Zhao, K. H., Zhang, J., Tu, J. M., Böhm, S., Plöscher, M., Eichacker, L., Bubenzer, C., Scheer, H., Wang, X., and Zhou, M. (2007) J. Biol. Chem. 282, Shen, G., Saunée, N. A., Williams, S. R., Gallo, E. F., Schluchter, W. M., and Bryant, D. A. (2006) J. Biol. Chem. 281, Zhao, K. H., Zhu, J. P., Song, B., Zhou, M., Storf, M., Böhm, S., Bubenzer, C., and Scheer, H. (2004) Biochim. Biophys. Acta 1657, Z-PEB Cys 82 Lyase CpeS 55. Scharnagl, C., Köst-Reyes, E., Schneider, S., Köst, H. P., and Scheer, H. (1983) Z. Naturforsch. 38, Huber, R., Schneider, M., Epp, O., Mayr, I., Messerschmidt, A., Pflugrath, J., and Kayser, H. (1987) J. Mol. Biol. 195, Huber, R., Schneider, M., Mayr, I., Müller, R., Deutzmann, R., Suter, F., Zuber, H., Falk, H., and Kayser, H. (1987) J. Mol. Biol. 198, Ficner, R., Lobeck, K., Schmidt, G., and Huber, R. (1992) J. Mol. Biol. 228, Brejc, K., Ficner, R., Huber, R., and Steinbacher, S. (1995) J. Mol. Biol. 249, Adir, N., Vainer, R., and Lerner, N. (2002) Biochim. Biophys. Acta 1556, Stec, B., Troxler, R. F., and Teeter, M. M. (1999) Biophys. J. 76, Duerring, M., Schmidt, G. B., and Huber, R. (1991) J. Mol. Biol. 217, Ritter, S., Hiller, R. G., Wrench, P. M., Welte, W., and Diederichs, K. (1999) J. Struct. Biol. 126, Downloaded from at MEDIZINISCHE EINRICHTUNGE, on June 16, 2011 NOVEMBER 26, 2010 VOLUME 285 NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37569

79 6 Discussion 74 6 Discussion 6.1 Ferredoxin-Dependent Reduction of Bilins The FDBRs are a family of extraordinary enzymes that catalyze the regiospecific reduction of the heme cleavage product BV ΙΧα, or in the case of phycoerythrobilin-ferredoxin oxidoreductase, 15,16-DHBV. The reaction requires electrons, which are provided by the small electron carrier protein ferredoxin. Cyanobacterial FDBRs use plant type Fds that posses an [2Fe-2S] cluster instead of bacterial type Fds that contain an [4Fe-4S] cluster (Otaka and Ooi, 1989). In photosynthetic organisms, reduced Fd is always present. Photoreduced PSΙ reduces Fd on the stromal site of the thylakoid membrane. Ferredoxin-NADP + reductase (FNR) then transfers electrons from Fd to NADP + to form NADPH that serves in various reactions (Alte et al., 2010; Setif, 2001). Other Fd-dependent enzymes include ferredoxin:thioredoxin reductase, nitrite reductase, sulfite reductase and glutamate synthase (Flores et al., 2005; Hase et al., 2006; Knaff and Hirasawa, 1991; Nakayama et al., 2000; van den Heuvel et al., 2004). Contrary to the FDBRs, where electrons are directly transferred to the substrate, all other known Fd-dependent enzymes possess metal or organic cofactors like heme, iron-sulfur clusters or flavins that serve as the initial electron acceptors mediating electron transfer to the substrate. Since Fd is a one-electron carrier reduction in the FDBRs occurs in one-electron steps. As a consequence substrate radicals can form. This was proven for all FDBRs employing electron paramagnetic resonance spectroscopy (EPR) (Tu, Rockwell et al. 2007; Tu, Chen et al. 2008; Busch, Reijerse et al. 2011) FDBRs Act via a Radical Mechanism Chemically challenging reactions are often realized employing a radical mechanism. In a reduction reaction electrons are transferred onto the enzyme or the substrate, which can result in the formation of radicals: chemical species that possess an unpaired electron. Radical enzymes are enzymes that catalyze a reaction with radical intermediate substrate-derived species or/and contain radicals such as the protein-bound glycyl radical (Selmer et al., 2005). Radicals are very reactive and therefore very susceptible to oxygen. For this reason radical

80 6 Discussion 75 enzymes are often found in anaerobic organisms. There exists a few found in both aerobes and anaerobes (Buckel and Golding, 2006) but FDBRs are exceptional since they are exclusively found in aerobic organisms. Aerobes developed antioxidant pathways to protect themselves of oxygen and reactive oxygen species. For example, they usually possess high concentrations of enzymes that are involved in deactivation of these reactive species like superoxide dismutases, catalases and peroxidases (Imlay, 2008). In this regard it should be mentioned that Prochlorococcus has lost most protective systems like catalases and peroxidases, but seems to be protected from oxidative damage by depletion of hydrogen peroxide by the surrounding microbial community (Bernroitner et al., 2009; Morris et al., 2011). Mutational studies of an E.coli transcription factor showed that a bulkier group could block access of oxygen to its iron-sulfur cluster preventing its oxidation (Jervis et al., 2009). Similar mechanisms could ensure activity of radical enzymes like FDBRs in cyanobacteria. In PebS for example, a loop was shown to close up upon binding of the substrate (Dammeyer et al., 2008b). A similar observation was made for PcyA were two α-helices and surrounding side chains move upon binding of BV ΙΧα (Hagiwara et al., 2006b). These structural changes might not only stabilize binding but could also stabilize further radical formation by protecting the substrate. In the process of trying to understand the catalytic mechanism of different FDBRs, protein variants were generated that accumulated radical intermediate species (Busch et al., 2011; Tu et al., 2008; Tu et al., 2006). For example, in all cases were Asp 105 and its homologs was shown to be critical for catalysis (in PcyA, PebA, PebB and PebS) exchange to an asparagine residue at this position led to accumulation of radical intermediate in all four variants. The exact nature of these radicals formed by the radical accumulating variants and those observed for the WT is still unknown. However, accumulation of radical allows further analyses of these species due to higher concentrations, and the likelihood of the presence of just one radical species is high contrary to the assumption that a mixture of radical species is present like in the WT reaction. The radicals formed during the reaction can vary regarding their protonation state and the localization of the unpaired electron. Radical accumulation can occur due to the inability of the variant to donate a proton to the substrate. In this way the reduction is fixed in an intermediate state. On the other hand an initial protonation is likely to occur to

81 6 Discussion 76 enable subsequent electron transfer (Dammeyer et al., 2008b; Krois, 1991), which would mean that the accumulation of radical in the variant shows the ability of the system to accept electrons without initial proton transfer but does not resemble an actual reaction intermediate. The one-electron reduced BV ΙΧα radical intermediate bound to D105N has been analyzed in PcyA employing High-Field EPR analyses, Density Functional Theory (DFT) computational studies, Nuclear Magnetic Resonance (NMR) and crystallization to identify the structure and protonation state of the radical intermediate (Kohler et al., 2010; Stoll et al., 2009). A high structural homogeneity of the bilin radical was implied. The radical appeared to be delocalized over the entire BV ΙΧα molecule rather than localized on one or few C-atoms. A BV ΙΧα one-electron reduced bis-lactim radical protonated at the A- and D-ring was proposed (Fig.12B). In this scenario the D105N variant forms a bis-lactim Fig. 12: Proposed structural models for one electron reduction of wild-type PcyA and PcyA_D105N (Kohler et al., 2010). A) The BVH+/Asp ion pair accepts the initial electron in PcyA_WT to generate a neutral radical intermediate. Glu 76 is well positioned for the second protoncoupled electron transfer. B) In PcyA_D105N, BV ΙΧα is not protonated and Glu 76 is poorly positioned for secondary electron transfer. One electron transfer to enzyme-bound substrate is accompanied by proton transfer from a central water not present in the BV ΙΧα bound WT structure and ejection of a hydroxide ion.

82 6 Discussion 77 hydrate with an axial water not present in the WT-bound BV ΙΧα. BV ΙΧα would be protonated from this axial water molecule with subsequent one-electron reduction generating a bis-lactim neutral radical that cannot undergo further reduction due to an unfavorable position of the second proton donor Glu 76, thus remains catalytically stalled in a radical state. Based on these observations and studies on a H88Q radical accumulating variant together with previous data a bis-lactim radical was proposed for the WT (Fig. 12A). This raises the question if these observations for the reduced PcyA_D105N variant apply for PebS and PebA as well. Asp 105 seems to have similar function in the three enzymes, serving in substrate coordination, as a catalytic residue and in mediating regiospecific protonation in the first and second reduction steps of PCB (PcyA) and PEB synthesis (PebS, PebA/B). The observation that PcyA employs a proton relay system as well as other catalytic residues infers that PcyA evolved a mechanism distinct from that in other FDBRs (Hagiwara et al., 2006a; Tu et al., 2007). Also, the far-red absorption maximum observed for PcyA upon substrate binding that was attributed to protonation, is not observed for PebS, PebA or PebB but initial protonation was postulated to occur in PebS (Busch et al., 2011). The similarities in PebS and PebA allow the assumption that the steps of 15,16-DHBV formation including the radical intermediate states are essentially the same for these two. To prove this, approaches similar to experiments performed for PcyA_D105N are in progress. These include a quantum-mechanics/molecularmechanics (QM/MM) approach and High-Field EPR experiments to understand the underlying mechanism leading to radical accumulation in the PebS_D105N and PebA_D84N variants. This will also allow comparison with radical species formed by PcyA. Besides the exact structures of the radical species involved in the reaction and their protonation states it is also not known whether the reduction steps proceed via stepwise electron-proton or proton-electron transfers, or in a concerted protoncoupled electron transfer reaction. The electrons derive from Fd and are easily distributed over the entire bilin substrate molecule. The origin of the protons needed for the reaction though is not clear. Protons can derive from water, protein residues or from a protonated site in the substrate. To examine this Fouriertransform infrared spectroscopy (FTIR) with deuterated solvent and/or labeled substrate will be employed (Berthomieu and Hienerwadel, 2009). The localization

83 6 Discussion 78 and fate of protons during the reaction is the key to understanding the mechanisms of regiospecificity What Determines the Regiospecificity of Different FDBRs? Each FDBR member possesses different regiospecificity for the three sites that are targeted for reduction, the A-ring 2,3,3 1,3 2 -diene system, the 15,16-methine bridge between ring C and D and the D-ring vinyl group (Fig. 13). Fig. 13: Reduction sites in BVΙΧα and corresponding FDBRs. The reduction sites of FDBRs are shown on BVΙΧα. The most common reaction is the A-ring reduction at the 2,3,3 1,3 2 -diene system catalyzed by HY2, PcyA, PebB and PebS (in red). PebA and PebS can perform a reduction at the 15,16-methine bridge (in green). PcyA is the only member to catalyze a reduction at the D- ring vinyl group (in blue). Despite the different activities, the comparison of FDBR structures reveal a very similar overall fold and active site structure (Busch et al., 2011; Dammeyer et al., 2008b; Hagiwara et al., 2006a; Tu et al., 2007). This raises the question of what determines their specificity. Several residues are conserved in the FDBR family and seem to have similar functions in different members whereas others are exclusive to one family member (Fig. 14). Asp 105 (using PebS numbering) is highly conserved in all members of FDBRs except for HY2. Asp 105 serves in A-ring reduction catalyzed by PebB, PebS and PcyA, in 15,16-reduction catalyzed by PebS and PebA, and likely serves in D- ring reduction catalyzed by PcyA as well. In PcyA Asp 105 was proposed to be important for mediating the formation of the correct R-configuration at the C2

84 6 Discussion 79 carbon during A-ring reduction (Tu et al., 2007). In the formation of PEB it most likely has the same function during A-ring reduction in PebB and PebS, whereas it was proposed to be a direct proton donor in 15,16-reduction of PebA and PebS (Busch et al., 2011). In PcyA His 88 and Glu 76 were identified to act in D-ring reduction (His 88, Glu 76 ) and A-ring reduction (His 88 ) (Tu et al., 2007; Tu et al., 2006). His 88 is an asparagine in PebA and PebS, where it is similarly positioned as in PcyA. However, exchange of this residue to histidine in PebS does not change its activity or regiospecificity (data not shown). Fig. 14: Catalytically important residues in the FDBR reaction. Asp 105 is important for PebS and PcyA reaction. It is also necessary for reduction of the 15,16- methine bridge in PebA and PebS. D-ring reduction in PcyA requires Glu 76, His 88 and Asp 105. His 88 and Asp 105 are also involved in PcyA A-ring reduction. Asp 206 is crucial for A-ring reduction performed by HY2, PebS and PebB. The colors refer to Fig. 13. Asp 206 (using PebS numbering) is strictly conserved in all FDBRs except for two functional PcyAs (Dammeyer et al., 2008a; Dammeyer et al., 2007). In HY2 it was postulated to serve as a proton donating residue in the A-ring reduction (Tu et al., 2008). Although highly conserved, Asp 206 does not have a catalytic function in PcyA or PebA. This is displayed by retained activity of asparagine variants of PcyA and PebA respectively (Busch et al., manuscript submitted; Tu et al., 2006). Nevertheless, in PebA Asp 206 was found similarly positioned as in PebS where it is a proton donor (Dammeyer et al., 2008b). These observations implicate that the neighbouring environment determines if the conserved residues are actually involved in the reaction, possibly by influencing positioning and conformation of the substrate. In PebA an unusual conformation of the substrate was implicated to be the determining factor to prevent the second reaction to occur (Busch et al., manuscript submitted). This conformation is likely induced by certain active site residues that are not present in PebS. Also, amino acid residues outside the active

85 6 Discussion 80 site pocket can have an influence on the enzymes reactivity. For example, RNase A activity relies on an intricate network of hydrogen bonding interactions which can be disrupted by mutation of residues far outside the active site (Doucet et al., 2011). In the case of PebA, subtle differences in the positioning of amino acid residues in the enzyme together with the observed substrate conformation, might work together in tightly regulating regiospecificity. In order to further probe the role of the active site structure and its influence on substrate conformation more structural data need to be accumulated. PebB Fig. 15: PebA-BV crystal reduced with dithionite. Crystals were grown in 0.1 M HEPES, ph 7, 28 % PEG 400 and reduced with dithionite in reservoir solution. PebA-BV complex is colored in green and the reduced crystal in purple indicates PebA-DHBV formation. crystals are available but showed insufficient diffraction. It is quiet possible that the PebB active site structure is very similar to that observed for PebS and PebA. It will have to be determined how the crucial residues are positioned and what conformational differences of the substrates as well as the products can be observed in the three enzymes. A step towards this is the ability to regiospecifically reduce the substrate in the crystal (Fig. 15). 6.2 Bilin-Channeling: From Heme to Biliproteins The first step in BV ΙΧα reduction by FDBRs is the binding of the substrate, which in all cases except for PebB is the product of heme oxygenase BV ΙΧα. Since FDBRs can bind bilins other than their natural substrate (Dammeyer and Frankenberg-Dinkel, 2006; Frankenberg and Lagarias, 2003), direct interaction and delivery of BV ΙΧα from HO is anticipated (Fig. 16). HO from cyanophage P-SSM2 had been set in biacore interaction studies together with PebS. With this method it was possible to detect a weak protein-protein interaction of HO with PebS (data not shown). However, due to high background signals different methods need to be employed to verify this finding. The interaction of HO with FDBR is further supported by the observation that product release by HO is the