Chemical Transformations Encoded by a Streptomyces coelicolor Gene Cluster with an Unusual GTP Cyclohydrolase

Item Type text; Electronic Dissertation

Authors Spoonamore, James Edward

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 11/10/2021 06:44:52

Link to Item http://hdl.handle.net/10150/194825 CHEMICAL TRANSFORMATIONS ENCODED BY A GENE CLUSTER IN STREPTOMYCES COELICOLOR CONTAINING AN UNUSUAL GTP CYCLOHYDROLASE

by

James Edward Spoonamore

______

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF BIOCHEMISTRY AND MOLECULAR BIOPHYSICS

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2008 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by James Edward Spoonamore entitled Chemical Transformations Encoded by a Gene Cluster in Streptomyces coelicolor Containing an Unusual GTP Cyclohydrolase and recommend that it be accepted as fulfilling the dissertation requirement for the

Degree of Doctor of Philosophy

______Date: April 16, 2008 Vahe Bandarian

______Date:April 16, 2008 Indraneel Ghosh

______Date: April 16, 2008 Nancy Horton

______Date:April 16, 2008 John Osterhout

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: April 16, 2008 Dissertation Director: Vahe Bandarian 3

STATEMENT BY THE AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED: James E Spoonamore 4

TABLE OF CONTENTS

LIST OF FIGURES ...... 6

ABSTRACT ...... 7

CHAPTER I: INTRODUCTION ...... 8

Explanation of the problem and its context ...... 8

A review of literature ...... 11 Riboflavin biosynthesis and GTP Cyclohydrolase II ...... 11 The discovery of vitamins and riboflavin ...... 11 Elucidating the biosynthetic pathway of riboflavin ...... 13 The riboflavin gene cluster ...... 29 G TP cyclohydrolases in Streptomyces coelicolor A 3(2) ...... 30 Detailed background about GCH II ...... 40

An explanation of the Dissertation Format ...... 45

CHAPTER II: PRESENT STUDY...... 47

SCO 6655 is a non­canonical GTP cyclohydrolase II that catalyzes the transformation of GTP to 2­amino­5­formylamino­6­ribosylamino­4(3H)­ pyrimidinone 5'­phosphate (Appendix A) ...... 47

A single amino acid change in the active site of SCO 1441, SCO 2687, and SCO 6655 is sufficient to interchange their activities between canonical and non­canonical (Appendix B) ...... 48

SCO 6654, a creatininase homolog, catalyzes the transformation of FAPy to APy (Appendix C) ...... 49

SCO 6650, a homolog of 6­pyruvoyltetrahydrobiopterin synthase, is a T­fold protein that appears to bind a pyrimidine ring containing substrate (Appendix D) ...... 49

REFERENCES ...... 51 5

TABLE OF CONTENTS ­ Continued

APPENDIX A: SCO 6655 IS A NON­CANONICAL GTP CYCLOHYDROLASE II THAT CATALYZES THE TRANSFORMATION OF GTP TO 2­AMINO­5­FORMYLAMINO­6­RIBOSYLAMINO­4(3H)­ PYRIMIDINONE 5'­PHOSPHATE ...... 63

APPENDIX B: A SINGLE AMINO ACID CHANGE IN THE ACTIVE SITE OF SCO 1441, SCO 2687, AND SCO 6655 IS SUFFICIENT TO INTERCHANGE THEIR ACTIVITIES BETWEEN CANONICAL AND NON­CANONICAL ...... 90

APPENDIX C: SCO 6654, A CREATININASE HOMOLOG, CATALYZES THE TRANSFORMATION OF FAPY TO APY ...... 114

APPENDIX D: SCO 6650, A HOMOLOG OF 6­ PYRUVOYLTETRAHYDROPTERIN SYNTHASE, IS A T­FOLD PROTEIN THAT APPEARS TO BIND A PYRIMIDINE RING CONTAINING SUBSTRATE ...... 148 6

LIST OF FIGURES

FIGURE 1.1, Riboflavin biosynthetic pathway ...... 16 FIGURE 1.2, Chemical structure of GTP and riboflavin ...... 18 FIGURE 1.3, Diacetyl derivation assay ...... 27 FIGURE 1.4, Predicted GCH II containing transcripts of S. coelicolor A3(2) ...... 34 FIGURE 1.5, Proposed products of the GCH II containing gene clusters ...... 39 FIGURE 1.6, Mechanism of GCH II from the literature ...... 44 7

ABSTRACT

Bacterial secondary metabolite biosynthetic pathways are frequently encoded in gene clusters. Genomic sequence information allows the identification of likely biosynthetic clusters based on sequence homology to known proteins. Biochemical characterization of suspected biosynthetic enzymes affords the discovery of pathways which may never be identified by traditional screening approaches. In the work presented here, I, in some cases in collaboration with others, characterize the three intragenomic

GTP cyclohydrolase II (GCH II) homologs from Streptomyces coelicolor A3(2) and show that one catalyzes a related but distinct reaction from the other two. The basis for the altered activity is investigated and speaks to the chemical mechanism of not only the unusual enzyme but also to all GCH II enzymes. Further, I investigate two other enzymes found in the same gene cluster as the unusual GCH II. Using biochemical techniques, I show that the product of the unusual GCH II is used as a substrate by a creatinine amidohydrolase homolog. Using structural biology, I show that the third enzyme, a 6­ pyruvoyltetrahydropterin synthase (PTPS), can not catalyze the PTPS reaction but is capable of binding a pterin substrate. Finally, I propose that the cluster from S. coelicolor containing the unusual GCH II encodes enzymes for a novel pathway to produce a pterin. 8

CHAPTER I: INTRODUCTION Explanation of the problem and its context Traditionally, bacterial secondary metabolites, such as antibiotics, have been discovered through serendipity by screening fractioned culture media to isolate and identify an active species. Once identified, the biosynthetic route can often be at least partially discerned through judicious use of labeled precursors. Generation of industrially useful over­producing strains are dependent upon successive rounds of random genetic mutation followed by screening. Identifying the metabolite, characterizing its synthesis, and improving the yield of the wild strain each requires an enormous effort and a pinch of luck. For several classes of secondary metabolites, such as polyketides or non­ribosomal peptides, the enzymatic toolkit of bacteria is fairly modular and has been studied sufficiently such that one can reasonably predict the number and kind of proteins, and thus genes, that are needed to effect the chemical transformations from the primary metabolite(s) to the secondary metabolite.

Improvements in DNA sequencing technologies in the last twenty years have led to an explosion in the quantity of primary sequence data available. The genomes of many model laboratory and other organisms are known. As of early 2008, the National Center for Biotechnology Information makes available 574 bacterial, 49 archaeal, and 22 eukaryotic genomes through its website along with many more partially completed genomes. Each of these genomes has been annotated, largely automatically, by using comparisons to known gene sequences. The availability of genomic information, 9

especially in organisms that may produce useful secondary metabolites, enables the discovery of new secondary metabolite genes and pathways, and thus new secondary metabolites via a non­traditional approach. By looking directly at genes that are likely to perform interesting chemical transformations, one is afforded a method which relies neither on the organism being able to produce enough metabolite to give a signal during screening nor on the frequently complex regulation of secondary metabolism. This dissertation is an example of what can be discovered using this bioinformatic approach.

Streptomycyes are a rich source of diverse and useful secondary metabolites. They maintain, for bacteria, unusually large genomes and complex lifestyles. Streptomyces exist as single cells, colonies, and spores. During sporulation, a colony cannibalizes itself in a controlled manner to support the reproductive effort. The S. coelicolor genome contains about 8,000 genes, a number akin to a primitive eukaryote such as yeast. Some

Streptomycetes are known to produce members of a class of secondary metabolites known as deazapurines. Deazapurines play roles as diverse as modified tRNA bases and have been evaluated for their potential as antimicrobials, antifungals, enterotoxins, etc.

While several of these metabolites have been described in the literature, the biosynthetic routes are generally obscure. Radiotracer studies have suggested that GTP cyclohydrolases may be involved in their production. GTP cyclohydrolases have been found in two forms; GCH I is involved in pterin synthesis and GCH II is involved in riboflavin synthesis. In the hope of identifying biosynthetic pathways related to deazapurines, the S. coelicolor genome was examined for genes which encoded GTP 10

cyclohydrolases. GCH II catalyzes the first biosynthetic step in the production of riboflavin, which involves removal of carbon­8 of GTP, a chemical step that would be needed to synthesize a deazapurine secondary metabolite from a guanine ring.

This dissertation describes the biochemical characterization of the three intragenomic GTP cyclohydrolase II homologs from S. coelicolor. In collaboration with others, I show that one of the three enzymes produces a related but distinct product than the other two. I show that simple changes in the enzyme active site explain the basis for the altered activity. I describe the activity of an enzyme encoded adjacent to the unusual

GCH II which uses the unusual product as a substrate. Finally, a structural model of a third enzyme in the cluster has been generated using X­ray crystallography in collaboration with others. 11

A review of literature Riboflavin biosynthesis and GTP cyclohydrolase II

The biosynthetic pathway for riboflavin has remained elusive until perhaps, the last twenty years. By reviewing the literature on the biosynthesis of riboflavin, in which

GTP cyclohydrolase II (GCH II) catalyzes the first committed step, I hope to show the sort of secondary metabolite that GCH II is known to help manufacture, the typical way that such pathways have been elucidated, and review what has been learned about the mechanism of GCH II along the way. Furthermore, by reviewing what is known about riboflavin biosynthesis, it will provide a frame of reference whereby similarities and differences between the relatively well studied riboflavin biosynthesis enzymes and the three gene clusters in S. coelicolor which harbor a GCH II can be noted.

The discovery of vitamins and riboflavin

The concept of a vitamin, a trace substance that is necessary for human health, took a long time to develop. The power of certain foodstuffs to cure 'diseases' such as beri­beri (thiamine deficiency) and scurvy (ascorbic acid deficiency) has been understood to some extent since at least the 1600's. The Scottish naval surgeon James Lind conducted what is considered the earliest modern clinical trial which showed that citrus fruit could cure scurvy (1). The causative agent of the cure contained within the fresh citrus fruit, however, remained mysterious. The concept of micro­nutrients was not formulated until much later. During the 1817 Franco Prussian war lengthy siege of Paris 12

malnutrition, especially among children, was a serious problem. Inavailability of fresh milk for children spurred failed attempts to generate an artificial substitute with the materials on hand. The French chemist Jean­Baptiste Dumas was involved in the attempts to make baby formula and observed that unknown micro­nutrients in milk and eggs, which were neither sugars, fats, nor proteins, were required for human health (2).

The conception of micro­nutrients would remain a theory until a century later when systematic studies using purified proteins, fats, and sugars unequivocally showed that dietary factors soluble in ether are required for growth of rats (3). The number and kind of A (fat soluble) and B (water soluble) vitamins remained somewhat confused due to a lack of nomenclature (4) as well as technical problems such as the inability to isolate individual components from the vitamin B mixtures through the first half of the 20th century.

Simultaneously with the discovery of vitamins, two enzyme cofactors, FAD and

FMN were discovered and their relationship to nutrients elucidated. First, a “yellow respiratory enzyme” (NADPH oxidase) was isolated from yeast (5). The following year it was shown that the yellow coloring in eggwhite and milk, 'Ovoflavin' and 'Lactoflavin', were spectrally indistinguishable from the yellow substance derived from the yeast enzyme (6, 7) and, further, the enzyme was inactive without the yellow dye (8). The link to nutrition was made by showing that purified lactoflavin was both spectrally and

nutritionally identical to vitamin B2 (9). The striking similarity of yellow fluorescent group to the alloxan rings synthesized and described in 1891 by Kuehling (10) was noted 13

(11) and by 1935, the full chemical synthesis of riboflavin was described (12). In just a few years, the yellow protein had been purified, crystallized, its molecular weight determined, and the ratio of one flavin group per molecule protein established (13). The discovery of a related cofactor, flavin adenosine dinucleotide (FAD), was soon to follow.

Working with the kidney amino acid deaminase described by Krebs (14), Warburg and

Christian identified and characterized the second flavin based enzyme cofactor (FAD) and went on to show that the cofactor was widespread. FAD could be found in biological niches as diverse as horse liver and yeast (15). Further, they showed that FAD was in some cases bound irreversibly to its protein carrier and in other cases could be readily

extracted (15). It is remarkable that the discovery of riboflavin (vitamin B2) and the link to enzymatic cofactors was established in the astonishingly short period of 6 years.

Elucidating the biosynthetic pathway of riboflavin

The current state of thought on the biosynthetic pathway to riboflavin is shown in

FIGURE 1.1. Several basic approaches have been used over the years to tease out the steps in the biosynthesis of riboflavin: radiotracer studies, genetic screens, and biochemical studies. To date, no enzyme has been shown to specifically catalyze a dephosphorylation step in the biosynthesis (FIGURE 1.1, chemical step 3).

From a historical standpoint, the similarity of the C ring of riboflavin to uridine and xanthine made either of them intuitively attractive candidates as riboflavin precursors

(FIGURE 1.2). The first clue into the biosynthetic precursor of riboflavin came in 1952 14

when MacLaren reported that supplementing purines in the growth media of a flavogenic yeast, Eremothecium ashbyii, increases riboflavin production (16). MacLaren's data showed that xanthine was the strongest promoter of riboflavin biosynthesis but that guanine and adenine also promoted production. Similar observations were reported by other labs (17­23). 15

FIGURE 1.1 Riboflavin biosynthetic pathway. Each compound is given a Roman numeral while each chemical transformation is given an Arabic numeral. Compound and enzyme names are listed in the key below the pathway. 16 17

FIGURE 1.2 The chemical structures of guanosine 5'­triphosphate with numbering system and riboflavin with rings denoted A, B, and C. 18 19

Radiotracer studies were used to further elucidate both the primary metabolites and their during incorporation into riboflavin. Support for a purine serving as a building block became very strong. Ashbya gosypii cultures grown in the presence of labeled formate and glycine lead to a riboflavin molecule whose B and C rings are labeled in the same way in which they would normally be incorporated into a purine (24). That is, the source of the B and C ring atoms is consistent with derivation from a purine that had incorporated the radiotracer labels thru the de novo purine biosynthetic pathway. E. ashbyii cultures similarly incorporated labeled formate and acetic acid into the B and C rings of riboflavin in the pattern consistent with a purine precursor (25, 26).

Radiotracer studies using purines and pyrimidines also pointed toward a purine being a primary metabolite used for riboflavin biosynthesis. C. flareri cultures grown with supplemental labeled guanine or adenine showed that adenine label was weakly incorporated while guanine was strongly incorporated into riboflavin (27). E. ashbyii cultures fed with 14C labeled adenine showed that 94% of the carbons remained in the riboflavin ring system and that the specific activities of purine carbons­2, ­4, ­5, and ­6 were incorporated into the riboflavin ring C carbons without dilution but that carbon­8 was lost (18, 28). It thus appeared that a purine ring system is not only the source of the atoms but is transformed directly into the riboflavin rings B and C with the loss of carbon­8. The same result was found with 14C and 15N labeled xanthine, all nitrogens and carbons of the purine ring except for carbon­8 were incorporated directly into riboflavin rings B and C (29). In resting Escherichia coli it was found that guanine labeled at 20

carbon­6 was incorporated into the equivalent postition of the riboflavin C ring (30).

Feeding and supplemental studies suffer from several unknowns. Different organisms may preferentially absorb or uptake different compounds. Differential labeling from one compound versus another may be due to transport, uptake, or degradation rather than to the biosynthetic route, per se. Additionally, the purine bases guanine and adenine are interchanged through xanthine, so a basal­level of "cross­talk" between all three bases would be expected. Evidence that guanine, in particular, rather than xanthine or adenine, was the purine source of riboflavin was shown in a bacterial species which lacks the ability to transform either xanthine or guanine to adenine (31) and in a yeast mutant which lacked the purine interchange pathway (32).

Radiotracer studies also shed light on the metabolites that are used to form the A and B rings and ribityl side chain of riboflavin. Cultures of A. gosypii grown in the presence of radiolabeled glucose and acetate incorporated label into both the A ring and ribityl side chain of riboflavin (33, 34). Tracer studies with E. ashbyii showed that the ring A and ribityl side chain are derived from the pentose phosphate pool (35). The radiolabel tracer studies in the 1950s and 1960s, in summary, found that a guanine base sans carbon­8 serves as the primary metabolite precursor of rings B and C of riboflavin and that ring A and the ribityl side chain are derived from the pentose phosphate pool.

The first intermediate to riboflavin to be detected was 6,7­dimethyl­8­ribityl­ lumazine (DMRL) (FIGURE 1.1, compound VII), whose stability and spectral properties made it easy to detect. When extracts of E. ashbyii, a natural riboflavin overproducer, 21

were subjected to paper chromatography, in addition to riboflavin, there was a green fluorescent substance. IR spectroscopy of this compound was consistent with the structure of DMRL, a pyrimidine ring connected to a sugar. Further, the green fluorescent compound could be chemically converted with four carbon donors like diacetyl or acetoin to riboflavin (36­38). Similarly, cultures of A. gosypii supplemented with 14C formate produced labeled DMRL concomitant with labeled riboflavin.

Moreover, both DMRL and riboflavin had equivalent specific activities suggesting that

DMRL is a precursor to riboflavin (39). Feeding experiments with 14C labeled guanine generated both labeled DMRL and riboflavin in several riboflavin producing organisms

(40). Finally, cell extracts from bacteria (E. coli) and plants (cabbage, lettuce, radish leaf, spinach, wheat) were shown to convert labeled DMRL to riboflavin demonstrating intermediacy of DMRL to the pathway (41, 42). The source of the remaining four carbon atoms that separated DMRL from riboflavin remained elusive (41, 43­45). Ironically, it was shown later that DMRL is both the donor and acceptor of the four carbons through a dismutation reaction (42, 46).

Riboflavin synthase (FIGURE 1.1, enzyme 7) was the first riboflavin biosynthetic enzyme to be purified and characterized. Riboflavin synthase from a yeast, E. coli and

A. gosypii were shown to use two molecules of DMRL to produce riboflavin (47, 48).

Riboflavin synthase has two distinct binding sites for the donating and accepting DRML

(48­51). The DMRL donating the four carbons is released from the enzyme as 5­amino­ 22

6­ribitylamino­2,4(1H,3H)­pyrimidinedione (FIGURE 1.1, compound IV) (52).

Riboflavin synthase catalyzes the dismutation reaction only with dephosphorylated

DMRL; dimethyl lumazines with other sugar side chains or even ethyl or propyl ribityl lumazines are not substrates for the enzyme (53, 54). Currently, the riboflavin synthases of 9 organisms have been characterized structurally.

Concurrent with the biochemical work to identify intermediates, genetic experiments were taking place to identify genes related to riboflavin biosysnthesis. As early as 1946 a mutant of Neruospora crassa was found which was a temperature sensitive riboflavin auxotroph (55). Somewhat later a second riboflavin biosynthesis deficient variant was identified, but no intermediates could be detected accumulating in either strain (56). Mutagenesis screens in Bacillus subtilis and Saccromyces cerevisiea led to identification of a number of riboflavin deficient mutants (57). Similar studies with E. coli soon followed (58). For historical reasons and because of their industrial import, a substantial amount of the primary literature for the B. subtilis and E. coli pathways is either in Russian (available in English reviews) or contained in patents in various languages. Since the the pathways of all organisms are nearly identical, I will focus on reviewing the yeast literature.

The S. cerevisiea haploid strain S288C was subjected to random mutagenesis with ethylmethanesulfonate (EMS) during growth on media with supplemental riboflavin with subsequent screening to identify riboflavin deficient mutants (59). Using cross­feeding experiments, no pairs of the 26 identified mutants were shown to support mutual growth. 23

However, 2 strains accumulated a green fluorescent compound in their media which was identified as DMRL based on thin layer chromatography and UV­VIS spectroscopy (60).

Two other mutant strains were shown to build up an amino pyrimidine or APy (FIGURE

1.1, compound II) based on spectroscopic (IR, MS, UV) studies and chemical means (60­

62). One of these two mutant strains could synthesize both DMRL and riboflavin upon supplementing the media with diacetyl (60). These results concurred with observations made a decade earlier which showed that APy is a metabolite in E. ashbyii but which could not directly link APy with riboflvain production (63). The 26 mutant strains were classified into six complementation groups and four biochemical groups that accumulated: no intermediate, APy, 5­amino­6­ribitylamino­2,4(1H,3H)­pyrimidinedione

(two complementation groups) (FIGURE 1.1, compound IV), or DMRL (64­66). Further, two of the genes were shown to catalyze the two earliest steps in the biosynthesis (67).

Because of the previous work on riboflavin synthase, it was straightforward to show that mutants which accumulated DMRL lacked riboflavin synthase activity while the other 5 groups did not (68). These studies established that six genes are involved in the production of four intermediates leading to riboflavin. Two key features of the riboflavin biosynthesis pathway remained unresolved. One, the source of the ribityl side chain was not known. It could arise from the starting guanosine nucleoside or could be added to a

6­hydroxy­2,4,5­triaminopyrimidine (APy with no sugar) derived from any form of guanine base (69). Second, the actual form of the four carbons used to transform 5­ 24

amino­6­ribitylamino­2,4(1H,3H)­pyrimidinedione to DMRL was unknown.

The serendipitous discovery of a new enzyme, GCH II (FIGURE 1.1, chemical step 1), soon suggested the answer to one of these open questions. An enzyme which catalyzed the transformation of GTP to dihydroneopterin triphosphate with loss of carbon­8 as formate, GTP cyclohydrolase, was noted to be very large and to conduct a very complex transformation. During the course of ensuring that preparations of GTP cyclohydrolase from E. coli were actually a single enzyme, investigators discovered a second enzyme, GCH II, which was also capable releasing carbon­8 of GTP as formate

2+ (70). GCH II is a 44 kDa, Mg dependent (Km app 14 M) enzyme that converts GTP to

APy, formate and pyrophosphate. GDP, GMP, guanosine, dGTP, ATP, ITP and XTP are not substrates. The biochemistry of this enzyme suggested strongly that GCH II is the first enzyme in the biosynthesis of riboflavin. While the discovery of GCH II set the stage for the discovery of the next two enzymes in riboflavin biosynthesis, it was nearly another twenty years before the position of GCH II in riboflavin biosynthesis was directly confirmed using recombinant techniques. A plasmid containing GCH II could rescue E. coli riboflavin deficient mutants generated with EMS mutagenesis that lacked GCH II activity (71). Recombinant GCH II opened up the door to biophysical, biochemical, kinetic and structural studies of the enzyme. This more intimate GCH II literature will be introduced later in order to focus on the missing pieces in riboflavin biosynthesis; the transformation of APy to DMRL. 25

Using GCH II to make [U14C]­APy, Burrows and Brown looked for deamination and/or reduction of the ribose to ribityl by assaying E. coli extracts via reaction with butanedione (72). The butanedione derivative would be dimethyl­lumazine, dimethyl­ ribityl­pterin, or DMRL in the case of deamination, reduction, or both, respectively.

Figure 1.3 shows the assay strategy more clearly. Both activities were found and two enzymes were partially purified. The first enzyme was a deaminase which had an apparent molecular weight of 80 kDa, a pH optimum of 9.1 and could use APy as its substrate (FIGURE 1.1, chemical step 2). The second was a pyridine nucleotide dependent (NADPH or NADH) reductase with an apparent molecular weight of 40 kDa, that utilized the product of the deaminase only after dephosphorylation (FIGURE 1.1, chemical step 4). Using the same approach, but with the organism A. gossypii, it was also shown that both enzyme activities existed (73). Complementation experiments in E. coli, to confirm the biochemical work by showing cloned deaminase/reductase could work in the pathway in vivo, would wait until 1997 (74). 26

FIGURE 1.3 Diacetyl derivation assay to detect APy deamination, reduction or both used by Burrows and Brown (72). If APy were deaminated then dimethyllumazine (DML) would be detected. Note that the ribose­phosphate is liberated during the derivitization.

If the sugar of APy were reduced then dimethylribitylpterin­phosphate would be produced

(DMRPt(P)). If both reactions occured, dimethylribytillumazine phosphate would be produced (DMRL(P)). 27 28

The final step in the biosynthesis of riboflavin is catalyzed by riboflavin synthase

(FIGURE 1.1 chemical step 7). Studies using purified riboflavin synthase from B. subtilis showed that it was, in fact, a huge protein found in two forms; a heavy form (a homotrimer of 'alpha' chains and sixty molecules of 'beta' chains) or a light form of just the 'alpha' chain homotrimer (76, 77). The 'alpha' homotrimer and 'beta' chains could be separated using gradient ultacentrifugation of purified heavy form enzyme. Riboflavin deficient mutant strains which were 'beta' knockouts maintained riboflavin synthase activity and accumulated DMRL when supplemented with diacetyl, so it was guessed that the 'beta' chains of riboflavin synthase may, in fact, be an enzyme involved in the biosythesis of DMRL. Purified 'beta' chains were shown to synthesize stoichiometric amounts of DMRL from 5­amino­6­ribitylamino­2,4(1H,3H)­pyrimidinedione (FIGURE

1.2, compound IV) using a semipurified extract of B. subtilis to generate the as yet unknown four carbon donor derived from labeled pentose phosphate (78).

The last major unknown step in riboflavin biosynthesis was the source of the four carbons used to make DMRL and the enzyme that catalyzes its generation. From the early radiotracer work it was known that the four carbons were derived from the pentose phosphate pool. More recent work had shown that carbons 1, 2, 3, and 5 of pentose­ phosphate were used though the particulars were still unclear (79­82). In 1990 it was reported that a magnesium dependent enzyme (FIGURE 1.1, chemical step 5, DHBP 29

synthase) from Candida guilliermondii had been purified 600­fold and shown to use D­ ribulose 5­phosphate (FIGURE 1.1, compound V) as its substrate to generate L­3,4­ dihydroxy­2­butanone 4­phosphate (FIGURE 1.1, comound VI), a compound that could serve as the 4 carbon donor (83­85). The DHBP synthase of E. coli was cloned and characterized in 1992 (86). The description of DHBP and its synthase has closed the book on the biosynthetic pathway to riboflavin from primary metabolites with the exception of one step (FIGURE 1.2, chemical step 3).

The riboflavin gene cluster

In B. subtilis, the genes encoding for all of the riboflavin biosynthetic enzymes are found in a cluster comprised of a single transcript of ribGBAH, a gene organization that is conserved in many gram­positive and gram­negative bacteria (87). Since we just reviewed literature indicating that there were six genes found in yeast, there do not appear to be enough genes to encode for all of the enzymes required for riboflavin biosynthesis.

Two of the proteins in B. subtillis are bi­functional or dual domain proteins. The ribG encodes a bi­functional riboflavin deaminase/reductase, ribB encodes riboflavin synthase, ribA encode a bi­functional GCH II / DHBP synthase, and ribH encodes lumazine synthase. Riboflavin biosynthesis using bi­functional GCH II / DHBP synthase proteins can be found in a wide range of organisms including plants and many bacteria

(88, 89). It may be the case that these two enzymes benefit from highly coupled co­ expression since they can be the rate limiting step in riboflavin biosynthesis (90). 30

Riboflavin biosythesis is regulated at many levels. Control of riboflavin biosynthesis in B. subtilis has been shown to work through repression not by riboflavin directly, but rather through the cofactor FMN (91). An interesting aside is that the leading sequence of the riboflavin biosynthesis cluster encodes an approximately 300 base pair conserved sequence, a riboswitch, which, in the presense of FMN, forms a structure which prevents translation of the RNA (92­94). There are several examples of riboswitches in the literature, occasionally RNA encoding biosynthesis genes can interact with the product of the genes' synthesis to suppress their own translation (95, 96).

G TP cyclohydrolases in Streptomyces coelicolor A 3(2)

Actinomycetes including Streptomyces have proven to be one of the most exploitable sources of useful secondary metabolites such as antibiotics. More than 50 antibiotics have been identified from various Streptomyces; the suffix mycin is used to indicate that an antibiotic was identified from a Streptomyces. Genes of secondary metabolite biosynthesis and frequently, in the case of antibiotics, the resistance to the secondary metabolite are often clustered near one another and transcribed as one or more polycistronic transcripts. The polyketide antibiotic actinorhodin biosynthesis genes are found in a gene cluster in S. coelicolor A3(2) (97, 98). The antibiotic prodiginine is also found in a gene cluster in S. coelicolor A3(2) (99­101). The genes for biosynthesis of and resistance to tetracenomycin C are found in a gene cluster in Streptomyces glaucesens

(102). The genes for the biosynthesis of the peptidic antibiotic bialaphos are found in 31

gene cluster in both Streptomyces hygroscopicus and Streptomyces viridochromogenes

(103). The list of genetically mapped biosynthetic gene clusters could go on.

To date, genome sequencing projects have determined the DNA sequences of several Streptomyces including Streptomyces coelicolor A3(2) (104), Streptomyces avermitilis (105), and Streptomyces scabies (the S. scabies genome is available from the

Sanger Institute through a collaboration with the U.S. Department of Agriculture but has not been published in a journal). Several more Streptomyces species are being sequenced at the Broad Institute, Sanger Institute, Genoscope, and the University of Tokyo. Full metabolic models can now be constructed with at least some experimentally confirmed predictions (106). The genomic information coupled with clever manipulations are hoped to greatly extend the efficiency and utility of these organisms for industrial uses (107,

108). The genomic information of Streptomyces and other organisms has also revealed that there are many more metabolic gene clusters than known metabolites, which has led to efforts to screen and exploit these hitherto unknown but predicted capabilities using technologies such as phage display or bioinformatic comparisons to known systems (109­

111).

GTP cyclohydrolase activity is key to the biosynthesis of a number of metabolites; riboflavin, pterins (such as folic acid), toxoflavin, and some pyrrolopyrimidine nucleoside antibiotics. In the biosynthesis of riboflavin, the first required step, opening of the purine ring of GTP with loss of carbon­8, is catalyzed by GCH II. An activity that was first found serendipitously while purifying GTP cyclohydrolase I, an enzyme involved in the 32

biosynthesis of the pterin ring of folic acid which also releases carbon­8 of GTP (70).

Pterins other than folic acid are similarly derived from GTP (112, 113). Streptomyces also produce pyrrolopyrimidine (deazapurine) antibiotics such as toyocamycin, tubercidin, and sangivamycin (114). Though the biosynthesis of these nucleoside antibiotics is not well understood, they too appear to be derived from GTP with the first step of their synthesis being the loss of carbon­8 (115). The antibiotic and poison, toxoflavin (aza­pteridine), is derived from GTP with loss of carbon­8 as well (116).

A BLAST search of the S. coelicolor genome with the protein sequence of GCH I and GCH II from E. coli identifies four genes, SCO 3403, SCO 1441, SCO 2687, and

SCO 6655. SCO 3403 is a homolog of GCH I and is found near three other genes with homology to enzymes involved in the biosynthesis of folate; SCO 3398, SCO 3400, SCO

3401. SCO 1441 appears to be dual functional GCH II / DHBP synthase and is found near riboflavin biosynthetic enzymes including a riboflavin synthase and lumazine synthase (SCO 1443 and SCO 1440). The other two GCH II genes, SCO 2687 and SCO

6655 are found in unusual gene contexts and their role is unknown. The predicted transcripts from S. coelicolor which contain a GCH II are shown along with the predicted homology of the accompanying proteins (FIGURE 1.4). 33

FIGURE 1.4 Predicted GTP cyclohydrolase II containing transcripts of S. coelicolor

A3(2). 34 35

While there are no reports of S. coelicolor making toxoflavin, the genes near SCO

2687 are consistent with a gene cluster involved in toxoflavin biosythesis. Toxoflavin was first discovered in the 1930's as the source of human toxicity due to bacterial contamination in processed coconuts and coconut oil in Java (117, 118). Toxoflavin seems to circumvent the cytochrome system to generate hydrogen peroxide and has been shown to be non­toxic to anaerobically grown bacteria (119). Once ingested, toxoflavin might be actively transported in the small intenstines by the same uptake system which captures riboflavin (120). Toxoflavin is also a phytotoxin produced by the plant pathogen

Burkulderia glumae, an agricultural pest that causes rice grain rot and rice seedling rot.

Toxoflavin is made from GTP using a Mg++ dependent enzyme which releases carbon­8 and generates GMP concurrent with toxoflavin synthesis (116). Toxoflavin biosynthesis in B. glumae is conducted by genes in a biosynthetic cluster which are transcribed as the tox operon (121). The toxBCDE transcript encodes a GCH II (toxB) and a riboflavin deaminase (toxE) (121). These enzymes encode the first two steps of riboflavin biosynthesis implying that toxoflavin biosynthesis diverges from riboflavin biosynthesis after these steps. Toxoflavin has also been detected in the enterobacterium

Photorhabdus luminescens which maintains a three gene toxoflavin cluster that houses a

GCH II (phu2475) (122). The actinobacteria, such as Streptomyces, appear to have evolved shortly after the oxygenation of the atmosphere by photosynthetic plants about

440 million years ago, perhaps to take advantage of decaying plant matter as a nutrient 36

source (123, 124). Toxoflavin may serve Streptomyces as a phytotoxin to hasten the demise of their food source, an insecticide to protect the colony from predation by worms or springtails, an antibacterial to hinder microbial competition, or some combination of all three. The presumed toxoflavin cluster in S. coelicolor is annotated as containing a

GCH II (SCO 2687), a riboflavin deaminase (SCO 2688), a protein of unknown function

(SCO 2691), a membrane protein (SCO 2689) and a membrane transport protein

(SCO2690). The protein of unknown function could be involved in the latter poorly understood synthetic step of generating the azapteridine ring while the membrane proteins could be involved in protecting the organism from the toxin and/or transporting it into the environment. It is not uncommon for antibiotic biosynthetic clusters to contain self protective or anti­suicide genes (125) and the membrane proteins may play this role.

The gene context of the third GCH II in S. coelicolor is unusual. GCH II has only been directly linked with two biosynthetic processes; the production of riboflavin and toxoflavin. The unusual cluster of genes is conserved in both the S. avermitilis and S. scabies genomes. The predicted transcript (FIGURE 1.4) encodes seven proteins: SCO

6655 (GCH II), SCO 6654 (creatinine amidohydrolase), SCO 6653 (membrane protein),

SCO 6652 (SAM methyl transferase), SCO 6651 (glycosyl transferase), SCO 6650 (6­ pyruvoyl­teterahydrobiopterin synthase, PTPS), and SCO 6649/6648 (serine dehydrogenase). The last gene is dual numbered because it is annotated as two genes in the S. coelicolor genome. By comparison to other Streptomyces genomes, it seems clear that a single protein with homology to serine dehydrogenase is encoded here and that an 37

unfortunately placed sequencing error led to one gene being mis­annotated as two in S. coelicolor (A. Hsieh, personal communication). This constellation of genes is novel and would seem to encode for a biosynthetic pathway of unknown function. As outlined above, S. coelicolor has genes in typical contexts for producing folic acid, riboflavin, and toxoflavin. The novel gene cluster could be a separate and redundant system which produces one of these compounds or perhaps it is involved in making one of the deazapurine antibiotics whose biosynthesis is poorly understood. The work presented in this dissertation goes into some depth about three of the proteins from this unusual cluster, while the remaining three enzymes, SCO 6652, SCO 6651, and SCO 6649/6648 remain untouched. The likely products of each of the GCH II containing gene clusters is shown in Figure 1.4. The SCO 1441 cluster is consistent with riboflavin production, the

SCO 2687 cluster is consistent with toxoflavin production, and the SCO 6655 cluster generates an unknown product which could be a pterin, nucleoside antibiotic, or an as yet described product. 38

FIGURE 1.5 The likely and possible products of the GCH II containing gene clusters found in Streptomyces coelicolor A3(2). 39 40

Detailed background about GCH II

Much of the work presented in Appendices A and B revolve around the chemistry of GCH II. GCH II is involved in the first step of the biosynthesis of riboflavin from

GTP. The products of its reaction are APy, pyrophosphate, formate and a small amount of GMP. GCH II was first purified and characterized from E. coli (70) but has since been characterized from several other organisms (71, 88, 126, 127). There are two notable differences observed between the GCH II enzymes of different species. In some species

GCH II is found as a dual domain protein fused with a DHBP synthase while in others it is monofunctional. Among the monofunctional enzymes oligomerization states seem to vary; GCH II from E. coli is a dimer while that from P. guilliermondii is a tetramer (127).

The best studied GCH II is that from E. coli.

GCH II is a zinc metalloenzyme with a conserved CX2GX7CXC binding motif

(128). The enzyme purifies with 1 molar equivalent of zinc and loss of any one of the conserved cysteine residues causes loss of the zinc. With no zinc, the enzyme can not produce APy but is still able to catalyze the transformation of GTP to GMP, implying the metal center is only used for the ring hydrolysis. Overall, the results with recombinant protein agree well with the first studies on purified GCH II save that Kaiser et al (128) reported that the recombinant E. coli enzyme could catalyze both ring opening and pyrophosphate release with dGTP as a substrate, an observation that contradicts the first report that no ring opened product was observed with dGTP (70).

NMR studies using recombinant E. coli GCH II have shown that a solvent oxygen 41

is incorporated into both the APY product and GMP while the pyrophosphate maintains the oxygens of the original GTP triphosphate (129). Based on this information, a covalent enzyme­GMP intermediate has been proposed though there is no direct evidence for one.

Recent steady state kinetic experiments conducted anaerobically at 37 ºC found that the enzyme deviated from Michaelis­Menten behavior suggesting some cooperativity between the monomers. One of the products, pyrophosphate, is a competitive inhibitor

with a Ki of 24 M.

Presteady state techniques have also been applied to E. coli GCH II (130). Single turnover, quenched­flow analysis was unable to detect any intermediates implying that the rate limiting step must be early in the reaction sequence. Stopped­flow analysis was, likewise, unable to detect any intermediates. Since the expectation is that two hydrolysis steps must occur to open the purine ring to release carbon­8 as formate, the authors prepared FAPy triphosphate and showed that it could be used as a substrate to make APy but much more slowly than GTP, meaning FAPy­triphosphate is not kinetically competent. So the summary of the kinetic work with GCH II is that a step prior to ring opening is rate limiting.

In addition to kinetic work, an X­ray crystallographic model of E. coli GCH II is also available (131). Two models of GCH II, one with no ligand and another with a non­ hydrolyzable substrate analog bound (GMPcPP), have been reported. A mechanism for

GCH II (FIGURE 1.5) was proposed based on the structural models in which the first step is a nucleophilic attack by Arg128 at the ­phosphate of GTP to generate a covalent 42

enzyme­GMP complex with the release of pyrophosphate. Premature hydrolysis of the

GMP­enzyme at this step would lead to generation of GMP. Next, a zinc activated water, in concert with Tyr105, attacks carbon­8 of GTP through a tetrahedral intermediate which collapses to form FAPy­enzyme. A second zinc activated water then attacks leading to formation of enzyme­APy and release of formate. The final step is the attack of a solvent water on the enzyme­APy guanidino­phosphate to release APy and regenerate the protein.

Data presented in Appendices A and B cast strong doubt upon elements of this proposed mechanism, particularly with regards to the participation of the active site tyrosine playing a role in the first ring opening hydrolytic attack on the purine. 43

FIGURE 1.6 Proposed mechanism of GCH II based on previous biochemical and structural studies. 44 45

Explanation of dissertation format

As described in the University of Arizona Graduate College Manual for

Electronic Theses and Dissertations and in accord with the policies of the Department of

Biochemistry and Molecular Biophysics, I present my work in two chapters and 4 appendices. Each appendix is a reprint of a previously published work or a manuscript to be submitted for publication in a peer reviewed journal. Chapter I introduces the problem and reviews the literature. Chapter II describes the present study, where I summarize each of the manuscripts.

My contributions to each manuscript are summarized here.

Appendix A: Experiments behind Figures 2, 3, 4, and 7 were solely my work. In

Table 1, specific activities and turn over numbers are my work, while the zinc contents were determined by an external analytical lab. The NMR experiments presented in

Figure 5 and Tables 2, 3, and 4 were conducted and analyzed by Neil Jacobsen and Vahe

Bandarian using samples that I prepared. Annie Dahlgran conducted the initial cloning, overexpression, and purification work of the wildtype enzymes. Both her and my own preparations of wildtype enzymes were used to generate data for this manuscript while the point mutant work was my own.

Appendix B: Experiments behind Tables 1, 2, and 3, with the exception of the zinc content in Table 2, were solely my work. Experiments that produced Figures 2 and 3 are solely my work. The structural alignment of GCH I and GCH II presented in Figure 46

5C is my work. Annie Dahlgran conducted the initial cloning, overexpression, and purification work of the wildtype enzymes. Both her and my own preparations of wildtype enzymes were used to generate data for this manuscript. I produced 17 of the 24 variant overexpression plasmids while the remaining seven were made under my supervision by Joon S. Kim and Jesus Hernandez. Of the various variant protein preparations, 21 of the 24 used in these experiments were my work while the remaining three were prepared by Joon S. Kim or Ivan Ogloblin under my supervision.

Appendix C: All experiments in the manuscript are solely my work. Annie

Dahlgran cloned SCO 6654 and purified one of the wild­type SCO 6654 enzyme preparations used for these experiments.

Appendix D: The bioinformatic analysis was solely my work. The structural work was highly collaborative. Some of the protein preparations used prepared by me while others were prepared by Annie Dahlgran and Vahe Bandarian. Crystallization and freezing conditions were found by myself and Vahe Bandarian. X­ray data collection and initial phasing were done by Annie Heroux. The structural model was refined by Dr. Sue

Roberts. 47

CHAPTER II: PRESENT STUDY

The methods, results and conclusions of this study are presented in the papers appended to this dissertation. What follows is a summary of the most important findings from each one.

SCO 6655 is a non­canonical GTP cyclohydrolase II that catalyzes the transformation of GTP to 2­amino­5­formylamino­6­ribosylamino­4(3H)­ pyrimidinone 5'­phosphate (Appendix A)

Summary

Three intragenomic homologs of GCH II are found in disparate regions of the

Streptomyces coelicolor A3(2) genome. While each of these proteins has a high (>40%) sequence identity to E. coli GCH II, two of them have canonical GCH II activity while the third produces a related but distinct product. SCO 6655 generates 2­amino­5­ formylamino­6­ribosylamino­4(3H)­pyrimidinone 5'­phosphate (FAPy) rather than 2,5­ diamino­6­ribosylamino­4(3H)­pyrimidinone 5'­phosphate (APy). The identity of the unusual product was confirmed by UV­VIS spectroscopy, mass spectrometry, and nuclear magnetic resonance. Surprisingly, introduction of a single conserved tyrosine into the active site of SCO 6655 is sufficient to transform its activity into a canonical GCH II.

This suggests that the previously proposed mechanism for GCH II should be modified to 48

indicate that the active site tyrosine plays an important role only in the second hydrolysis reaction of GCH II.

A single amino acid change in the active site of SCO 1441, SCO 2687, and SCO 6655 is sufficient to interchange their activities between canonical and non­canonical

(Appendix B)

Summary

The non­canonical GCH II, SCO 6655, has only a few amino acid changes in highly conserved residues compared to canonical GCH II enzymes. Only one of these amino acid changes alters the enzyme catalytic activity. All three of the intragenomic

GCH II homologs in Streptomyces coelicolor A3(2) can be switched to produce either the

APy or FAPy with this single amino acid change. Using site directed mutagenesis, it is shown that only two amino acids efficiently support canonical GCH II activity in this position (Tyr and His). Histidine is also found in the active site GCH I. Structural overlays of the active sites of both enzymes show remarkable similarity in zinc and

Tyr/His residue arrangement relative to the GTP substrate. Together, these results indicate a remarkably short mutational pathway is possible to evolve the different activities observed among the GCH II homologs of S. coelicolor A3(2). Furthermore,

GCH II and GCH I have converged on the same mechanistic strategy to release carbon­8 49

of GTP.

SCO 6654, a creatininase homolog, catalyzes the transformation of FAPy to APy

(Appendix C)

Summary

SCO 6654, an enzyme whose gene is adjacent to the unusual GCH II (SCO 6655), uses FAPy as a substrate and generates APy. SCO 6654 is homologous to creatininase, an enzyme which uses a dimetallic zinc­manganese center to catalyze the reversible hydrolysis of creatinine to creatine. Site directed mutagenisis, metals analysis, and metals dependence all suggests that SCO 6654 also uses a dimetallic zinc­manganese metal center for catalysis.

SCO 6650, a homolog of 6­pyruvoyltetrahydrobiopterin synthase, is a T­fold protein that appears to bind a pyrimidine ring containing substrate (Appendix D)

Summary

SCO 6650 is a homolog of 6­pyruvoyltetrahydrobiopterin synthase (PTPS). A X­ ray crystallographic model of SCO 6650 reveals a hexameric tunnel­fold protein. The structure of SCO 6650 overlays with the structure of rat PTPS and reveals that residues 50

known to bind the substrate pyrimidine ring in PTPS are arranged nearly identically in

SCO 6650. Residues known to be important for catalysis in PTPS are, by contrast, not conserved. This indicates that SCO 6650 likely binds a substrate with a pyrimidine ring though its catalytic activity remains cryptic. 51

REFERENCES

1. Lind, J. A. (1753) A treatise of the scurvy. Sands, Murray and Cochran, Edinburgh, Scottland, The James Lind Library, http://www.jameslindlibrary.org /trial_records/17th_18th_Century/lind/lind_kp.html (accessed April 1, 2008)

2. McCollum, E. V. (1953) Who discovered vitamins?, Science 118, 632.

3. McCollum, E. and Davis, M. (1913) The necessity of certain lipins in the diet during growth, J. Biol. Chem. 15, 167­175.

4. Dutcher, R. A. (1928) Vitamin B terminology, Science 68, 206­209.

5. Warburg, O. and Christian, W. (1932) Ein zweites sauerstoffuebertragendes Ferment und sein Absorptionsspektrum, Naturwiss. 20, 688.

6. Kuhn, R., Gyoergy, P. and Wagner­Jauregg, T. (1933) Ueber Ovoflavin, den Farbstoff des Eiklars, Ber. Chem. Ges. 66, 576­580.

7. Kuhn, R., Gyoergy, P. and Wagner­Juaregg, T. (1933) Ueber Lactoflavin, den Farbstoff der Molke, Ber. Chem. Ges. 66, 1034­1038.

8. Warburg, O. and Christian, W. (1933) Ueber das gelbe Ferment und seine Wirkungen, Biochem. Z. 266, 377­411.

9. Gyoergy, P., Kuhn, R. and Wagner­Jauregg, T. (1933) Das Vitamin B2, Naturwiss. 21, 560­561.

10. Kuehling, O. (1891) Ueber Azine der Hansaeuregruppe, Ber. Chem. Ges. 24, 2363­2369.

11. Stern, K. and Holiday, E. (1934) Zur Konstitution des Photo­flavins; Versuche in der Alloxazin­Reihe, Ber. Chem. Ges. 67, 1104­1106.

12. Kuhn, R., Reinemund, K., Kaltschmitt, H., Stroebele, R. and Trischmann, H. (1935) Synthetisches 6.7­Dimethyl­9­d­ribo­flavin, Naturwiss. 23, 260.

13. Kekwick, R. A. and Pedersen, K. O. (1936) Some physico­chemical characteristics of the yellow respiratory enzyme, Biochem. J. 30, 2201­2205. 52

14. Krebs, H. A. (1935) Metabolism of amino­acids: Deamination of amino­acids, Biochem. J. 29, 1620­1644.

15. Warburg, O. and Christian, W. (1938) Isolierung der prosthetischen Gruppe der d­ Aminosaeureoxydase, Biochem. Z. 298, 150­168.

16. MacLaren, J. A. (1952) The effects of certain purines and pyrimidines upon the production of riboflavin by Eremothecium ashbyii, J. Bacteriol. 63, 233­241.

17. Goodwin, T. W. and Pendlington, S. (1954) Studies on the biosynthesis of riboflavin; nitrogen metabolism and flavinogenesis in Eremothecium ashbyii, Biochem. J. 57, 631­641.

18. McNutt, W. S. (1954) The direct contribution of adenine to the biogenesis of riboflavin by Eremothecium ashbyii, J. Biol. Chem. 210, 511­519.

19. Brown, E. G., Goodwin, T.W. and Pendlington, S. (1955) Studies on the biosynthesis of riboflavin. 2. Further observations on nitrogen metabolism and flavinogenesis in Eremothecium ashbyii, Biochem. J. 61, 37­46.

20. Mehta, S. U., Mattoo, A.K. and Modi, V. V. (1972) Ribitol and flavinogenesis in Eremothecium ashbyii, Biochem. J. 130, 159­166.

21. Brown, E. G., Goodwin, T.W. and Jones, O. T. (1958) Studies on the biosynthesis of riboflavin. 4. Purine metabolism and riboflavin synthesis in Eremothecium ashbyii, Biochem. J. 68, 40­49.

22. Osman, H. G. and Soliman, M. H. (1960) Biosynthesis of riboflavin by Eremothecium ashbyii. II. Effect of some amino acids. Ureido acids, hydantoins and some other nitrogenous compounds on the growth and Vitamin B2 production, Biochem. Z. 333, 351­360.

23. Goodwin, T. W. and McEvoy, D. (1959) Studies on the biosynthesis of riboflavin. 5. General factors controlling flavinogenesis in the yeast Candida flareri, Biochem. J. 71, 742­748.

24. Plaut, G. W. (1954) Biosynthesis of riboflavin. I. Incorporation of C14­labeled compounds into rings B and C, J. Biol. Chem. 208, 513­520.

25. Klungsöyr, L. (1954) The Biosynthesis of Riboflavin in Eremothecium ahbyii, Acta. Chem. Scand. 8, 723­727. 53

26. Klungsöyr, L. (1954) The Biosynthesis of Riboflavin in Eremothecium ashbyii, Acta. Chem. Scand. 8, 1292.

27. Audley, B. G., Goodwin, T.W. and McEvoy D (1959) Flavogenesis in Candida flareri; the role of urea and purines, Bioc. J. 72, 8P.

28. McNutt, W. S. J. (1956) The incorporation of the pyrimidine ring of adenine into the isoalloxazine ring of riboflavin, J. Biol. Chem. 219, 365­373.

29. McNutt, W. S. (1961) The Incorporation of the four nitrogen atoms of purines into the pyrimidine and pyrazine rings of riboflavin, J. Am. Chem. Soc. 83, 2303­2307.

30. Howells, D. J. and Plaut, G. W. (1965) Biosynthesis of riboflavine by a purine­ requiring mutant strain of Escherichia coli, Biochem. J. 94, 755­759.

31. Baugh, C. M. and Krumdieck, C. L. (1969) Biosynthesis of riboflavine in Corynebacterium species: the purine precursor, J. Bacteriol. 98, 1114­1119.

32. Bacher, A. and Lingens, F. (1969) The structure of the purine precursor in riboflavin biosynthesis, Angew. Chem. Int. Ed. Engl. 8, 371­372.

33. Plaut, G. W. (1954) Biosynthesis of riboflavin. II. Incorporation of C14­labeled compounds into ring A, J. Biol. Chem. 211, 111­116.

34. Plaut, G. W. and Broberg, P. L. (1956) Biosynthesis of riboflavin. III. Incorporation of C14­labeled compounds into the ribityl side chain, J. Biol. Chem. 219, 131­138.

35. Ali, S. N. and al­Khalidi, U. A. (1966) The precursors of the xylene ring in riboflavine, Biochem. J. 98, 182­188.

36. Masuda, T. (1956) Isolation of a green fluorescent substance produced by Eremothecium ashbyii, Pharm. Bull. 4, 71­72.

37. Masuda, T. (1957) Application of chromatography. XXXII. Biosynthesis of riboflavin by Eremothecium ashbyii, Pharm. Bull. 5, 136­141.

38. Masuda, T. (1957) Application of chromatography. XXXI. Structure of a green fluorescent substance produced by Eremothecium ashbyii, Pharm. Bull. 5, 28­30. 54

39. Maley, G. F. and Plaut, G. W. (1959) The isolation, synthesis, and metabolic properties of 6, 7­dimethyl­8­ribityllumazine, J. Biol. Chem. 234, 641­647.

40. Korte, F., Aldag, H., Ludwig, G., Paulus, W. and Storiko, K. (1958) Zur Biosynthese des Riboflavins, Just. Lieb. An. Chem. 619, 70­79.

41. Katagiri, H., Takeda, I. and Imai, K. (1958) Synthesis of riboflavin by microorganisms. III. On the role of G compound, J. Vitaminol. 4, 211­216.

42. Mitsuda, H., Kawai, F. and Moritaka, S. (1961) Biogenesis of riboflavin in green leaves. I. Confirmation of enzymatic reaction for synthesis of riboflavin from 6,7­ dimethyl­8­ribityllumazine, J. Vitaminol. 7, 128­136.

43. Birch, A. and Moye, C. (1957) Studies in Relation to Biosynthesis. Part X. A Synthesis of lumichrome from non­benzenoid precursers, J. Chem. Soc., 412­414.

44. Birch, A. and Moye, C. (1958) Studies in relation to biosynthesis. Part XVI. The synthesis of lumiflavin from non­benzenoid precursors, J. Chem. Soc., 2622­2624.

45. Cresswell, R. and Wood, H. (1959) Chemical Studies of the biosynthesis of riboflavin, Proc. Chem. Soc., 387­388.

46. Goodwin, T. W. and Horton, A. A. (1961) Biosynthesis of riboflavin in cell­free systems, Nature 191, 772­774.

47. Plaut, G. W. (1960) Studies on the stoichiometry of the enzymic conversion of 6,7­dimethyl­8­ribityllumazine to riboflavin, J. Biol. Chem. 235, PC41­2.

48. Plaut, G. W. (1963) Studies on the nature of the enzymic conversion of 6,7­ dimethyl­8­ribityllumazine to riboflavin., J Biol Chem 238, 2225­2243.

49. Mitsuda, H., Kawai, F., Suzuki, Y. et al. (1970) Biogenesis of riboflavin in green leaves. VII. Isolation and characterization of spinach riboflavin synthetase., J Vitaminol (Kyoto) 16, 285­292.

50. Harvey, R. A. and Plaut, G. W. (1966) Riboflavin synthetase from yeast. Properties of complexes of the enzyme with lumazine derivatives and riboflavin., J Biol Chem 241, 2120­2136.

51. Suzuki, Y. and Mitsuda, H. (1971) A second­order kinetic reaction catalyzed by riboflavin synthetase from a high­riboflavinogenic Eremothecium ashbyii., 55

Biochim Biophys Acta 242, 500­503.

52. Wacker, H., Harvey, R. A., Winestock, C. H. and Plaut, G. W., (1964) 4­(1'­D­ ribitylamino)­5­amino­2,6­dihydroxypyrimidine, the second product of the riboflavin synthetase reaction, J. Biol. Chem. 239, 3493­3497.

53. Harzer, G., Rokos, H., Otto, M.K., Bacher, A. and Ghisla, S. (1978) Biosynthesis of riboflavin. 6,7­Dimethyl­8­ribityllumazine 5'­phosphate is not a substrate for riboflavin synthase., Biochim. Biophys. Acta. 540, 48­54.

54. Winestock, C. and Plaut, G. (1961) Synthesis and Properties of Certain Substituted Lumazines, J. Org. Chem. 26, 4456­4462.

55. Mitchell, H. and Houlahan, M. (1946) Neurospora. IV. A temperature­sensitive riboflavinless mutant, Am. J. Bot. 33, 31­35.

56. Garnjobst, L. and Tatum, E. (1956) A temperature­independent riboflavin­ requiring mutant of Neurospora crassa, Am. J. Bot. 43, 149­157.

57. Bresler S. E., Cherepenko E. I. and Perumov D. A. (1971) Study of the riboflavin operon in Bacillus subtilis Part 3 Isolation and properties of mutants with the complex regulatory genotype, Genetika 7, 123.

58. Bandrin, S. V., Beburov, M.I., Rabinovich, P.M. and Stepanov, A. I. (1979) [Riboflavin auxotrophs of Escherichia coli], Genetika 15, 2063­2065.

59. Oltmanns, O. and Lingens, F. (1967) [Isolation of riboflavin­deficient mutants of Saccharomyces cerevisiae], Z. Naturforsch. B 22, 751­754.

60. Lingens, F., Oltmanns, O. and Bacher, A. (1967) [On intermediates of riboflavin biosynthesis in Saccharomyces cerevisiae], Z. Naturforsch. B 22, 755­758.

61. Bacher, A. and Lingens, F. (1968) Accumulation of 2,5­Diamino­6­hydroxy­4­ ribitylaminopyrimidine in a riboflavin­deficient mutant of Saccharomyces cerevisiae, Angew. Chem. Internat. Edit. 7, 219­220.

62. Bacher, A. and Lingens, F. (1970) Biosynthesis of riboflavin. Formation of 2,5­ diamino­6­hydroxy­4­(1'­D­ribitylamino)pyrimidine in a riboflavin auxotroph, J. Biol. Chem. 245, 4647­4652.

63. Goodwin, T. and Treble, D. (1957) The detection of 4:5­Diaminouracil as a 56

metabolite of Eremothecium ashbyii, Biochem. J. 67, 10P.

64. Oltmanns, O., Bacher, A., Lingens, F. and Zimmermann, F. K. (1969) Biochemical and genetic classification of riboflavine deficient mutants of Saccharomyces cerevisiae, Mol. Gen. Genet. 105, 306­313.

65. Oltmanns, O. (1971) Riboflavinless mutants of Saccharomyces cerevisiae: the problem of the rib 6 ­gene, Mol. Gen. Genet. 111, 300­302.

66. Bacher, A., Baur, R., Oltmanns, O. and Lingens, F. (1969) Biosynthesis of riboflavin. Mutants accumulating 6­hydroxy­2,4,5­triaminopyrimidine., FEBS Lett. 5, 316­318.

67. Oltmanns, O. and Bacher, A. (1972) Biosynthesis of riboflavine in Saccharomyces cerevisiae: the role of genes rib 1 and rib 7, J. Bacteriol. 110, 818­822.

68. Baur, R., Bacher, A. and Lingens, F. (1972) The structural gene for riboflavin synthetase in Saccharomyces cerevisiae, FEBS Lett. 23, 215­216.

69. Bacher, A. and Lingens, F. (1971) Biosynthesis of riboflavin. Formation of 6­ hydroxy­2,4,5­triaminopyrimidine in rib 7 mutants of Saccharomyces cerevisiae, J. Biol. Chem. 246, 7018­7022.

70. Foor, F. and Brown, G. M. (1975) Purification and properties of guanosine triphosphate cyclohydrolase II from Escherichia coli, J. Biol. Chem. 250, 3545­ 3551.

71. Richter, G., Ritz, H., Katzenmeier, G., Volk, R., Kohnle, A., Lottspeich, F., Allendorf, D. and Bacher, A. (1993) Biosynthesis of riboflavin: cloning, sequencing, mapping, and expression of the gene coding for GTP cyclohydrolase II in Escherichia coli, J. Bacteriol. 175, 4045­4051.

72. Burrows, R. B. and Brown, G. M. (1978) Presence in Escherichia coli of a deaminase and a reductase involved in biosynthesis of riboflavin, J. Bacteriol. 136, 657­667.

73. Hollander, I. and Brown, G. M. (1979) Biosynthesis of riboflavin: reductase and deaminase of Ashbya gossypii, Biochem. Biophys. Res. Commun. 89, 759­763.

74. Richter, G., Fischer, M., Krieger, C., Eberhardt, S., Lüttgen, H., Gerstenschläger, I. 57

and Bacher, A. (1997) Biosynthesis of riboflavin: characterization of the bifunctional deaminase­reductase of Escherichia coli and Bacillus subtilis, J. Bacteriol. 179, 2022­2028. 75. Fischer, M. and Bacher, A. (2005) Biosynthesis of flavocoenzymes, Nat. Prod. Rep. 22, 324­350.

76. Bacher, A. and Mailänder, B. (1978) Biosynthesis of riboflavin in Bacillus subtilis: function and genetic control of the riboflavin synthase complex, J. Bacteriol. 134, 476­482.

77. Bacher, A., Baur, R., Eggers, U., Harders, H.D., Otto, M.K. and Schnepple, H. (1980) Riboflavin synthases of Bacillus subtilis. Purification and properties, J. Biol. Chem. 255, 632­637.

78. Neuberger, G. and Bacher, A. (1986) Biosynthesis of riboflavin. Enzymatic formation of 6,7­dimethyl­8­ribityllumazine by heavy riboflavin synthase from Bacillus subtilis, Biochem. Biophys. Res. Commun. 139, 1111­1116.

79. Bacher, A., Le Van, Q., Keller, P.J. and Floss, H. G. (1983) Biosynthesis of riboflavin. Incorporation of 13C­labeled precursors into the xylene ring, J. Biol. Chem. 258, 13431­13437.

80. Nielsen, P., Neuberger, G., Fujii, I., Bown, D.H., Keller, P.J., Floss, H.G. and Bacher, A. (1986) Biosynthesis of riboflavin. Enzymatic formation of 6,7­ dimethyl­8­ribityllumazine from pentose phosphates, J. Biol. Chem. 261, 3661­ 3669.

81. Bacher, A., Van, Q., Keller, P. and Floss, H. (1985) Biosynthesis of riboflavin. Incorporation of multiply 13C­labeled precursors into the xylene ring, J. Am. Chem. Soc. 107, 6380­6385.

82. Floss, H., Van, Q., Keller, P. and Bacher, A. (1983) Biosynthesis of riboflavin. An unusual rearrangement in the formation of 6,7­dimethyl­8­ribityllumazine, J. Am. Chem. Soc. 105, 2493­2494.

83. Volk, R. and Bacher, A. (1990) Studies on the 4­carbon precursor in the biosynthesis of riboflavin. Purification and properties of L­3,4­dihydroxy­2­ butanone­4­phosphate synthase, J. Biol. Chem. 265, 19479­19485.

84. Volk, R. and Bacher, A. (1991) Biosynthesis of riboflavin. Studies on the mechanism of L­3,4­dihydroxy­2­butanone 4­phosphate synthase, J. Biol. Chem. 58

266, 20610­20618.

85. Volk, R. and Bacher, A. (1988) Biosynthesis of riboflavin. The structure of the four­carbon precursor, J. Am. Chem. Soc. 110, 3651­3652.

86. Richter, G., Volk, R., Krieger, C., Lahm, H.W., Röthlisberger, U. and Bacher, A. (1992) Biosynthesis of riboflavin: cloning, sequencing, and expression of the gene coding for 3,4­dihydroxy­2­butanone 4­phosphate synthase of Escherichia coli, J. Bacteriol. 174, 4050­4056.

87. Perkins, B. and Pero, J. (2002) Bacillus subtilis and its closest relatives: from genes to cells. (Sonenshein, A., Ed.), ASM Press, Washington, DC.

88. Herz, S., Eberhardt, S. and Bacher, A. (2000) Biosynthesis of riboflavin in plants. The ribA gene of Arabidopsis thaliana specifies a bifunctional GTP cyclohydrolase II/3,4­dihydroxy­2­butanone 4­phosphate synthase, Phytochemistry 53, 723­731.

89. Fassbinder, F., Kist, M. and Bereswill, S. (2000) Structural and functional analysis of the riboflavin synthesis genes encoding GTP cyclohydrolase II (ribA), DHBP synthase (ribBA), riboflavin synthase (ribC), and riboflavin deaminase/reductase (ribD) from Helicobacter pylori strain P1, FEMS Microbiol. Lett. 191, 191­197.

90. Humbelin, M., Greisser, V., Keller, T., Schurter, W., Maiker, M., Hohmann, H., Ritz, H., Richter, G., Bacher, A. and van Loon, A. (1999) GTP cyclohydrolase II and 3,4­dihydroxy­2­butanone 4­phosphate synthase are rate­limiting enxymes in riboflavin synthesis of an industrial Bacillus subtilis strain used for riboflavin production, J. Ind. Micro. Biotech. 22, 1­7.

91. Bacher, A., Eggers, U. and Lingens, F. (1973) Genetic control of riboflavin synthetase in Bacillus subtilis, Arch. Mikrobiol. 89, 73­77.

92. Gelfand, M. S., Mironov, A.A., Jomantas, J., Kozlov, Y.I. and Perumov, D. A. (1999) A conserved RNA structure element involved in the regulation of bacterial riboflavin synthesis genes, Trends. Genet. 15, 439­442.

93. Mironov, A. S., Gusarov, I., Rafikov, R., Lopez, L.E., Shatalin, K., Kreneva, R.A., Perumov, D.A. and Nudler, E. (2002) Sensing small molecules by nascent RNA: a mechanism to control transcription in bacteria., Cell 111, 747­756.

94. Kreneva, R. A. and Perumov, D. A. (1990) Genetic mapping of regulatory 59

mutations of Bacillus subtilis riboflavin operon, Mol. Gen. Genet. 222, 467­469.

95. Vitreschak, A. G., Rodionov, D.A., Mironov, A.A. and Gelfand, M. S. (2004) Riboswitches: the oldest mechanism for the regulation of gene expression?, Trends. Genet. 20, 44­50.

96. Nudler, E. and Mironov, A. S. (2004) The riboswitch control of bacterial metabolism, Trends. Biochem. Sci. 29, 11­17.

97. Malpartida, F. and Hopwood, D. A. (1984) Molecular cloning of the whole biosynthetic pathway of a Streptomyces antibiotic and its expression in a heterologous host, Nature 309, 462­464.

98. Malpartida, F. and Hopwood, D. A. (1986) Physical and genetic characterisation of the gene cluster for the antibiotic actinorhodin in Streptomyces coelicolor A3(2), Mol. Gen. Genet. 205, 66­73.

99. Malpartida, F., Niemi, J., Navarrete, R. and Hopwood, D. A. (1990) Cloning and expression in a heterologous host of the complete set of genes for biosynthesis of the Streptomyces coelicolor antibiotic undecylprodigiosin, Gene 93, 91­99.

100. Feitelson, J. S. and Hopwood, D. A. (1983) Cloning of a Streptomyces gene for an O­methyltransferase involved in antibiotic biosynthesis, Mol. Gen. Genet. 190, 394­398.

101. Cerdeño, A. M., Bibb, M. J. and Challis, G. L. (2001) Analysis of the prodiginine biosynthesis gene cluster of Streptomyces coelicolor A3(2): new mechanisms for chain initiation and termination in modular multienzymes, Chem. Biol. 8, 817­829.

102. Motamedi, H. and Hutchinson, C. R. (1987) Cloning and heterologous expression of a gene cluster for the biosynthesis of tetracenomycin C, the anthracycline antitumor antibiotic of Streptomyces glaucescens, Proc. Natl. Acad. Sci. U. S. A. 84, 4445­4449.

103. Hara, O., Murakami, T., Imai, S., Anzai, H., Itoh, R., Kumada, Y., Takano, E., Satoh, E., Satoh, A., Nagaoka, K. et al, (1991) The bialaphos biosynthetic genes of Streptomyces viridochromogenes: cloning, heterospecific expression, and comparison with the genes of Streptomyces hygroscopicus, J. Gen. Microbiol. 137, 351­359.

104. Bentley, S. D., Chater, K.F., Cerdeño­Tárraga, A., Challis, G.L., Thomson, N.R., 60

James, K.D., Harris, D.E., Quail, M.A., Kieser, H., Harper, D., Bateman, A., Brown, S., Chandra, G., Chen, C.W., Collins, M., Cronin, A., Fraser, A., Goble, A., Hidalgo, J., Hornsby, T., Howarth, S., Huang, C., Kieser, T., Larke, L., Murphy, L., Oliver, K., O'Neil, S., Rabbinowitsch, E., Rajandream, M., Rutherford, K., Rutter, S., Seeger, K., Saunders, D., Sharp, S., Squares, R., Squares, S., Taylor, K., Warren, T., Wietzorrek, A., Woodward, J., Barrell, B.G., Parkhill, J. and Hopwood, D. A. (2002) Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2), Nature 417, 141­147.

105. Ikeda, H., Ishikawa, J., Hanamoto, A., Shinose, M., Kikuchi, H., Shiba, T., Sakaki, Y., Hattori, M. and Omura, S. (2003) Complete genome sequence and comparative analysis of the industrial microorganism Streptomyces avermitilis, Nat. Biotechnol. 21, 526­531.

106. Borodina, I., Krabben, P. and Nielsen, J. (2005) Genome­scale analysis of Streptomyces coelicolor A3(2) metabolism, Genome. Res. 15, 820­829.

107. Chater, K. F. (2006) Streptomyces inside­out: a new perspective on the bacteria that provide us with antibiotics, Philos. Trans. R. Soc. Lond. B Biol. Sci. 361, 761­ 768.

108. Donadio, S., Sosio, M. and Lancini, G. (2002) Impact of the first Streptomyces genome sequence on the discovery and production of bioactive substances, Appl. Microbiol. Biotechnol. 60, 377­380.

109. Gross, H. (2007) Strategies to unravel the function of orphan biosynthesis pathways: recent examples and future prospects, Appl. Microbiol. Biotechnol. 75, 267­277.

110. Hranueli, D., Cullum, J., Basrak, B., Goldstein, P. and Long, P. F. (2005) Plasticity of the streptomyces genome­evolution and engineering of new antibiotics, Curr. Med. Chem. 12, 1697­1704.

111. Yin, J., Straight, P.D., Hrvatin, S., Dorrestein, P.C., Bumpus, S.B., Jao, C., Kelleher, N.L., Kolter, R. and Walsh, C. T. (2007) Genome­wide high­throughput mining of natural­product biosynthetic gene clusters by phage display, Chem. Biol. 14, 303­312.

112. Shiota, T. and Palumbo, M. P. (1965) Enzymatic synthesis of the pteridine moiety of dihydrofolate from guanine nucleotides, J. Biol. Chem. 240, 4449­4453. 61

113. Dalal, F. R. and Gots, J. S. (1965) Requirement of GTP for pteridine synthesis in Salmonella typhimurium and its inhibition by AMP, Biochem. Biophys. Res. Commun. 20, 509­514.

114. Suhadolnik, R. (1970) in Nucleoside Antibiotics pp298­354. Wiley­Interscience, New York.

115. Elstner, E. F. and Suhadolnik, R. J. (1971) The biosynthesis of the nucleoside antibiotics. IX. Purification and properties of guanosine triphosphate 8­ formylhydrolase that catalyzes production of formic acid from the ureido carbon of guanosine triphosphate, J. Biol. Chem. 246, 6973­6981.

116. Levenberg, B. and Kaczmarek, D. K. (1966) Enzymic release of carbon atom 8 from guanosine triphosphate, an early reaction in the conversion of purines to pteridines, Biochim. Biophys. Acta. 117, 272­275.

117. Van Veen, A. and Mertens, W. (1934) Das Toxoflavin, der gelbe Giftstoff der Bongkrek, Rec. Trav. Chim. 53, 398­404.

118. Van Veen, A. and Mertens, W. (1934) Die Giftstoffe der sogenannten Bongkrek­ vergiftungen auf Java, Rec. Trav. Chim. 53, 257­266.

119. Latuasan, H. E. and Berends, W. (1961) On the origin of the toxicity of toxoflavin, Biochim. Biophys. Acta. 52, 502­508.

120. Yagi, K. (1954) Distribution of riboflavin nucleotides in rat organs influenced by the administration of flavin compounds, J. Biochem. 41, 757­762.

121. Suzuki, F., Sawada, H., Azegami K and Tsushida, K. (2004) Molecular characterization of the tox operon involved in toxoflavin biosynthesis of Burkholderia glumae, J. Gen. Plant Pathol. 70, 97­107.

122. Duchaud, E., Rusniok, C., Frangeul, L., Buchrieser, C., Givaudan, A., Taourit, S., Bocs, S., Boursaux­Eude, C., Chandler, M., Charles, J., Dassa, E., Derose, R., Derzelle, S., Freyssinet, G., Gaudriault, S., Médigue, C., Lanois, A., Powell, K., Siguier, P., Vincent, R., Wingate, V., Zouine, M., Glaser, P., Boemare, N., Danchin, A. and Kunst, F. (2003) The genome sequence of the entomopathogenic bacterium Photorhabdus luminescens, Nat. Biotechnol. 21, 1307­1313.

123. Heckman, D. S., Geiser, D.M., Eidell, B.R., Stauffer, R.L., Kardos, N.L. and Hedges, S. B. (2001) Molecular evidence for the early colonization of land by 62

fungi and plants, Science 293, 1129­1133.

124. Chater, K. F. and Chandra, G. (2006) The evolution of development in Streptomyces analysed by genome comparisons, FEMS Microbiol. Rev. 30, 651­ 672.

125. Martín, M. F. and Liras, P. (1989) Organization and expression of genes involved in the biosynthesis of antibiotics and other secondary metabolites, Annu. Rev. Microbiol. 43, 173­206.

126. Kobayashi, M., Ohara­Nemoto, Y., Kaneko, M., Hayakawa, H., Sekiguchi, M. and Yamamoto, K. (1998) Potential of Escherichia coli GTP cyclohydrolase II for hydrolyzing 8­oxo­dGTP, a mutagenic substrate for DNA synthesis, J. Biol. Chem. 273, 26394­26399.

127. Liauta­Teglivets, O., Hasslacher, M., Boretskii, I.R., Kohlwein, S.D. and Shavlovskii, G. M. (1995) Molecular cloning of the GTP­cyclohydrolase structural gene RIB1 of Pichia guilliermondii involved in riboflavin biosynthesis, Yeast 11, 945­952.

128. Kaiser, J., Schramek, N., Eberhardt, S., Püttmer, S., Schuster, M. and Bacher, A. (2002) Biosynthesis of vitamin B2., Eur. J. Biochem. 269, 5264­5270.

129. Ritz, H., Schramek, N., Bracher, A., Herz, S., Eisenreich, W., Richter, G. and Bacher, A. (2001) Biosynthesis of riboflavin: studies on the mechanism of GTP cyclohydrolase II, J. Biol. Chem. 276, 22273­22277.

130. Schramek, N., Bracher, A., and Bacher, A. (2001) Biosynthesis of riboflavin. Single turnover kinetic analysis of GTP cyclohydrolase II, J. Biol. Chem. 276, 44157­44162.

131. Ren, J., Kotaka, M., Lockyer, M., Lamb, H.K., Hawkins, A.R. and Stammers, D. K. (2005) GTP cyclohydrolase II structure and mechanism, J. Biol. Chem. 280, 36912­36919. 63

APPENDIX A: SCO 6655 IS A NON­CANONICAL GTP CYCLOHYDROLASE THAT CATALYZES THE TRANSFORMATION OF GTP TO 2­AMINO­5­ FORMYLAMINO­6­RIBOSYLAMINO­4(3H)­PYRIMIDINONE 64 65 66

Reproduced with permission from Spoonamore, J. E., Dahlgran, A. L., Jacobsen, N. E., and Bandarian, V. (2006) Evolution of new function in the GTP cyclohydrolase II proteins of Streptomyces coelicolor, Biochemistry 45, 12144 – 12155

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90

APPENDIX B: A SINGLE AMINO ACID CHANGE IN THE ACTIVE SITE OF SCO 1441, SCO 2687, AND SCO 6655 IS SUFFICIENT TO INTERCHANGE THEIR ACTIVITIES BETWEEN CANONICAL AND NON­CANONICAL 91 92 93

Reproduced with permission from Spoonamore, J. E. and Bandarian, V. (2008), Understanding Functional Divergence in Proteins by Studying Intragenomic Homologues. Biochemistry 47, 2592­2600 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

APPENDIX C: SCO 6654, A CREATININASE HOMOLOG, CATALYZES THE TRANSFORMATION OF FAPY TO APY 115

A creatinine amidohydrolase homologue from Streptomyces coelicolor is a

formylamino deformylase which uses a dimetallic zinc manganese center

James E. Spoonamore and Vahe Bandarian*

Department of Biochemistry and Molecular Biophysics

The University of Arizona

Tucson, Arizona, 85261 USA

* Corresponding author: Vahe Bandarian

University of Arizona

Department of Biochemistry and Molecular Biophysics

1041 E. Lowell St.

Tucson, AZ 85721

Tel: 520 626­0389

E­mail: [email protected] 116

SUMMARY

SCO 6654 is a creatinine amidohydrolase homolog found in a gene cluster of unknown function in Streptomyces coelicolor. Like creatinine amidohydrolase, SCO

6654 appears have a dimetallic manganese ­ zinc center, as shown by site directed mutagenesis, metals analysis, and activity assays. We show that SCO 6654 catalyzes the conversion of 2­amino­5­formylamino­6­ribosylamino­4(3H)­pyrimidinone 5'­phosphate to 2,5­diamino­6­ribosylamino­4(3H)­pyrimidinone 5'­phosphate, the canonical GTP cyclohydrolase II product, based on UV­visisble spectroscopy, chromatography and mass spectroscopy. We speculate that this enzyme participates in a biosynthetic pathway that utilizes a purine to generate a pterin.

INTRODUCTION

There are three intragenomic homologs of GTP cyclohydrolase II (GCH II) in the genome of Streptomyces coelicolor. One of these, SCO 6655, catalyzes the conversion of

GTP to a non­canonical product, 2­amino­5­formylamino­6­ribosylamino­4(3H)­ pyrimidinone 5'­phosphate (FAPy) (1). This protein is found among a cluster of genes

(SCO 6655 ­ 6648) whose biological significance has been difficult to glean; however, the presence of a similar arrangement of genes in the five actinomycetes whose genomes have been determined (S. coelicolor, Streptomyces avermitilis, Streptomyces scabies,

Salinispura tropica, and Salinispura arenicola) suggests that the cluster may be important 117

in producing a purine­based secondary metabolite. In addition to the observed

GTP→FAPy activity of SCO6655, support for this notion comes from the X­ray crystal structure of another protein in the cluster, SCO 6650, which appears capable of binding an amino­pyrimidinone substrate but carries out alternate chemistry.

We have undertaken a systematic study of the open reading frames in the cluster to determine its biological function. SCO 6654 is immediately downstream of SCO 6655 and is divergently transcribed. A PSI­BLAST search, with the amino acid sequence of

SCO 6654, identifies it a creatinine amidohydrolase (creatininase, CNNase) (EC

3.5.2.10). Creatininase catalyzes the reversible hydrolysis of creatinine to creatine

(Figure 1A), an activity which was first identified in four strains of soil dwelling microorganisms which were cultured with creatinine as the carbon source (2). A second bacterial system to breakdown creatinine through 1­methylhydantoin using creatinine deaminase (EC 3.5.4.21) also exists and a third bacterial system, through methylguanidine, is speculated (Figure 2) (3). Creatinine catabolic systems are not widespread; a recent study showed five Psuedomonas species were able to use creatinine as a nitrogen source while only one of the five could also use it as a carbon source (4).

Creatininase is part of an inducible three enzyme system found in some bacteria

(Alcaligenes, Arthrobacter, Flavobacterium, Micrococcus, Pseudomonas, and Tissierella)

(3) in concert with creatinase (EC 3.5.3.3) and either sarcosine oxidase (EC 1.5.3.1) or sarcosine dehydrogenase (EC 1.5.99.1) to break down creatinine into glycine and urea (2, 118

5­7). In at least some cases, the three enzymes are found together in a gene cluster (7).

Biochemical and structural studies have revealed that CNNase is a metalloenzyme which uses a dimetallic Zn ­ Mn center (8) where zinc is held more tightly than the manganese. Native CNNase purified from Pseudomonas putida C­83 contains 0.9 zinc per monomer along with traces of other divalent metal ions (Mn2+, Mg2+, and Ca2+) (9).

When the P. putida protein is overexpresed in E. coli, 2 equivalents of Zn2+ per subunit are found, unless the growth medium is supplemented with manganese, in which case it purifies with 1 Zn and 1 Mn per subunit (8). The activity of purified CNNases are stimulated, in vitro, by the addition of supplemental divalents metal ions to the assay buffer system; this is presumably due to replacement of one or both metal ions which were lost during purification. Different results are observed with different preparations;

Co2+, Mg2+, Ni2+, Ca2+, and Mn2+ supplementation have all been reported to give the maximum observable enzymatic activity (9­11). CNNase can be completely inactivated by removal of divalent metals including the zinc by incubation with the chelator EDTA.

Metal free enzyme can be reactivated by reconstituting the metal site with Mn2+ > Co2+ >

Mg2+ > Fe2+ > Zn2+ > Ni2+ > Cu2+ (9).

The subunit composition of CNNases from different organisms varies from dimeric to octameric and the subunits vary in size from 23 ­ 80 kDa (Table 1). Steady state kinetic analysis of various purified CNNases from A. ureafaciens, Alcalignes sp.,

and P. putida have found Km values for creatinine ranging from 8 – 125 mM (7, 10­13). 119

Structural studies by two groups have provided insight into the P. putida CNNase

(8, 14). The structual models show that each of the two metal ions binds a His and a Glu, and an Asp bridges the two metal ions. The metal ion binding residues are highly conserved. The structural studies both show that one of the active site metal ions is Zn2+ however, they differ in the identity of the metal ion in the second site. In one case a Zn2+ was modeled while in the other a Mn2+ is found. A Zn ­ Mn center is thought to be the native form and the structure of this combination including a creatine in the active site shows residues which are important for substrate binding.

In this manuscript, we describe cloning, overexpression and purification of SCO

6654, a CNNase homolog from S. coelicolor. Biochemical studies show that the protein is not a CNNase. Interestingly, the enzyme catalyzes the conversion of the product of

SCO 6655 to 2,5­diamino­6­ribosylamino­4(3H)­pyrimidinone 5'­phosphate (APy), as shown by analyses using UV­visible spectroscopy, high performance liquid chromatography (HPLC) and mass spectometry (MS). Furthermore, site­directed mutagenesis shows that SCO 6654 has a metal center akin to that found in CNNase.

Comparison of the context of the cluster bearing SCO 6654 in S. coelicolor to that of other sequenced actinomyces suggest that SCO 6655/SCO 6654 may work in concert.

EXPERIMENTAL PROCEDURES

Cloning and purification, of SCO 6654 120

The gene encoding SCO 6654 was cloned from Streptomyces coelicolor A3(2) genomic DNA by 25 cycles of amplification using Pfu­Turbo DNA polymerase and forward and reverse primers

5'­GGAATTCCCATGGCATATGAAAGGCCGTGCCGGAATGG­3' and

5'­TTTTAAGCTTATTACCCGCGCTCGCCGTCGG­3'. The annealing temperature was set to 58˚C. The PCR product was cloned, essentially as has been described previously for SCO 6655 (1), into the NdeI/HindIII sites of pET28a or pET29a vectors for expression

of either His6­ or native SCO 6654.

BL21(DE3) E. coli were transformed with the native expression plasmid

(pVBD32) carrying sco6654. An overnight culture in Lenox Broth (LB) from 1 kanamycin resistant colony were was prepared and used to inoculate six 1 L Fernbach flasks containing 34 µg/mL kanamycin and 0.1 mM zinc sulfate. Cells were grown at

37°C to OD600nm~1 prior to induction by addition of 0.1 mM IPTG. Cells were collected 4

h post­induction by centrifugation at 4000 xg for 30 min, frozen in liquid N2, and stored at

–80°C.

For protein purification, frozen cells were suspended in ice cold 20 mM Tris•HCl

(pH 8.0) containing 1 mM PMSF and 0.5 mM DTT. Cells were lysed in an ice bath by sonciation at 50% power for 7 min using a Branson Sonifier with 10 sec bursts followed

20 sec rests. Lysates were cleared by centrifugation at 34,000 xg for 45 minutes. Cleared lysates were applied to a 2.5 x 14.5 cm Q­sepharose column and washed with 20 mM

Tris•HCl (pH 8) containing 0.5 mM DTT (Buffer A). The protein was eluted with a 121

linear gradient of 0 – 0.5 M KCl in Buffer A. SCO 6654 elutes at ~0.4 M KCl. Fractions containing SCO 6654 were identified based on SDS­PAGE, pooled, and concentated with an Amicon pressure cell using a YM10 membrane. The resulting protein was further purified by application to a 2.5 x 60 cm sephacryl S­300 gel filtration column equilibrated with Buffer A. The protein was eluted with Buffer A and detected by SDS­

PAGE analysis. Fractions containing SCO 6654 were pooled, concentrated and stored at

–80°C until needed. SCO 6654 was quantified using a BCA assay kit with BSA as the standard (Pierce).

The molecular weight of SCO 6654 and SCO 6654 Asp51Ala were estimated by analyzing the retention time for the protein (3 mg) on a Sephacryl S­200HR gel filtration column (1.6 x 60 cm) calibrated with a commercial kit which included thyroglobulin, ferritin, catalase, aldolase, albumin, ovalbumin, chymotrypsinogen A, and ribonuclease A

(molecular weights of 669, 440, 232, 158, 67, 43, 25, 14 kDa, respectively). The column was equilibrated and run at room temperature with 20 mM Tris•HCl (pH 8) containing

0.15 M NaCl at a flow rate of 1 mL/min.

Site­directed mutagenesis of SCO 6654 and purification of variant

The SCO 6654 Asp51Ala variant was prepared from the pET29a plasmid pVBD32 using the Quickchange procedure (Stratagene) with primers

5'­GCGACCGCCACCCTCG­3' and 5'­CGAGGGTGGCGGTCGC­3' and an annealing 122

temperature of 50˚C. Presence of the of the desired mutation was confirmed by sequencing the plasmid (pJS229). The variant was expressed essentially as described for the wildtype; however, the yield of soluble SCO 6654 Asp51Ala was improved by lowering the temperature to 25˚C at induction and collecting the cells 12 h post induction.

The variant was purified as described for the wildtype protein.

UV­visible spectral change measurements

An Agilent 8453 diode array spectrophotometer equipped with an 8 cell changer was used for UV­visible measurements. SCO 6654 assays were conducted in quartz

cuvettes containing 0.8 ml of 0.1 M Tris•SO4 (pH 8.0), 5 mM MnSO4, and 0.5 mM DTT in the presence of varying quantities of SCO 6654. Assays were initiated by addition of

7 µL 10 mM FAPy (purified as described (1)). Changes in the absorbance of the sample were monitored in 15 s intervals. Initial rates were calculated using the initial portion of the progress curves (typically 1­5 min) using a difference extinction coefficient at 299 nm

2300 M­1 cm­1.

Mass spectral analysis of the product of SCO 6654

The identity of the product produced by SCO 6654 was confirmed by mass spectrometry as follows. Enzyme (40 µM) was incubated with 0.35 mM FAPy containing

50 mM Tris•SO4 and 5 mM MnSO4, at room temperature, for 30 min. Control 123

experiments were also carried out in the absence of enzyme. The sample was applied to a

1 mL P­2 column equilibrated with water (prepared in a Pasteur pipet plugged with glass wool), eluted with 0.4 mL water, and then the following ~0.2 mL (8 drops) of eluate was collected and diluted with 0.5 mL water. The sample was analyzed using a Thermo LCQ

Decca XP ESI mass spectrometer by direct infusion at a flow rate of 5 L/min. The nozzle was kept at 300°C and a voltage of 3000 V. The instrument was operated in the negative ion mode and was optimized for the m/z = 380.2 signal due to FAPy. The

MS/MS was perfomed at 25% energy applied 30 msec. As a control, we also carried out the identical process with SCO 2687, which we have previously shown to produce APy from GTP (1). In these experiments, SCO 2687 (70 µM) was incubated with 50 mM

Tris•SO4 (pH 8.0), 0.5 mM MgCl2 and 0.35 mM GTP for 30 min at room temperature and analyzed exactly as described for SCO 6654.

HPLC of wildtype and variant reaction timecourse

The reaction catalyzed by wildtype and Asp51Ala variant of SCO 6654 was also monitored by HPLC. In these expeiments, either SCO 6654 (2 µM) or the Asp51Ala

variant of the protein (25 µM) were incubated with 100 µM FAPy in 25 mM Tris•SO4

(pH8), 5 mM MnSO4 and 1 mM DTT. All incubations were carried out at room temperature. At various times an aliquot (0.1 mL) of the reaction mixture was withdrawn, the protein was removed by filtration through a nano­sep 10K centrifugal device and 124

analyzed by HPLC. The filtrate was applied to a 4.6 x 150 mm Zorbax SAX column using an Agilent 1100 series HPLC equipped with a chilled autosampler and diode array

detector. A gradient of buffer A (0.75 M NH4H2PO4 containing 2% (v/v) acetonitrile) and buffer B (water) at a flow rate of 0.75 ml/min was applied as follows: isocratic 1% A for 5 minutes followed by a linear gradient to 10% A over 20 minutes. At these conditions,

APy and FAPy eluted at 4.3 and 13.9 min, respectively.

Activity of SCO 6654 in the presence of various metal ions

The activity of SCO 6654 in the presence of 5 mM CaCl2, CoCl2, CuCl2, MgSO4,

MnSO4, NiSO4, ZnSO4 was determined. The enzyme (0.5 µM) was incubated with the

metal ion for 10 minutes in 50 mM Tris•SO4. Additionally, enzyme was incubated with

CaCl2, MgSO4, and MnSO4 in 50 mM Tris•SO4 1 mM DTT. FAPy (0.24 mM) was added to the mixture and the reactions were allowed to proceed for 30 minutes prior to removal of the protein with a centrifugal microconcentrator. The resulting samples were analyzed by HPLC as described above. APy and FAPy were quantified by comparison to standard

curves. Km (app) curves were generated with the same method but metal was varied from

0.5 ­ 500 M MnSO4.

Michaelis­Menten analysis of SCO 6654

The Michaelis­Menten parameters for SCO 6654 were determined as follows. 125

SCO 6654 (0.5 µM) in 50 mM Tris•SO4 (pH 8) containing 5 mM MnSO4 and a range of

FAPy (0 ­ 1 mM) were incubated at room temperature for 30 min. Data was fit to the

Michaelis Menton equation using the statistical package R (15).

Creatininase Assay

The Jaffe reaction (16) was used to assay SCO 6654 for creatininase activity essentially as described for clinical sample assay but with one modification (17); SDS was not included as an assay component. The reactions contained 2 mM creatinine, 5 µM

SCO6654 and 5 mM MnSO4 in 0.1 M Tris SO 4 (pH 8). To quantify the product, an aliquot (0.06 mL) of the reaction mixture was mixed with 1 ml Jaffe's reagent and allowed to stand for 30 min. Absorbances 505 nm were measured. Standard curves showed that the assay is linear in the range of 0.07­2.2 mM creatinine.

RESULTS

SCO 6654 is homologous to creatinine amidohydrolase proteins

A PSI­BLAST search with the amino acid sequence of SCO 6654 converges after

4 iterations to reveal, exclusively, protein sequences annotated as CNNase (PFAM02633 and COG 1402). The X­ray crystal structure of P. putida CNNase with bound product

(PDB ID: 1J2T) shows that the zinc and manganese of the active site metal center are each coordinated with a Glu and His side chain and that both metals share a bridging 126

Asp. A pair­wise alignment between SCO 6654 and this protein shows that all of the metal ligands from the P. putida CNNase align with identical residues in SCO 6654

(Figure 3). In contrast to the metal­binding residues, CNNase residues that interact with the bound product in the crystal structure are not conserved in SCO 6654 (8). The alignment and structural context of the conserved and missing residues suggest that while

SCO 6654 may have retained a CNNase­like metal center, it is likely to have lost the ability to bind creatinine and catalyze the formation of creatine.

Expression and purification of SCO 6654

To begin biochemical characterization of the protein, native SCO 6654 was overexpressed and purified by anion exchange chromatography and gel filtration. The

Asp51Ala variant did not express well at 37°C; however, incubation at room temperature after induction with IPTG, followed by a more lengthy induction phase (12 h), led to production of soluble protein. SDS­PAGE of the protein preparations stained with

Coomassie blue show that the preparations are > 90% pure (data not shown).

Oligomerization state of SCO 6654

Gel filtration of purified SCO6654 and SCO 6654 Asp51Ala revealed molecular mass of 327 kDa and 314 kDa, respectively. The predicted monomeric mass for SCO

6654, based on the amino acid sequence, is 28.6 kDa. Therefore, SCO 6654 appears to be 127

either a decamer or dodecamer. In any case, it appears to be a higher order oligomer than those reported for CNNases (Table 1).

Metal analysis of SCO 6654

As shown in Figure 3, SCO 6654 conserves residues which bind the putative Zn ­

Mn bimetallic center in CNNase. To determine if SCO 6654 in fact is a metalloprotein, it was subjected to ICP­OES analysis (Garrat­Callahan, Burlingame, CA). Two separate preparations of SCO 6654 revealed 0.46–0.5 equivalents of zinc bound to the protein.

Other metals were also detected: Ca 0.23 ­ 7, Fe 0.7 ­ 1.3, Mn 0.1 ­ 0.2 equivalents. The

X­ray crystal structure of CNNase reveals that an Asp residue bridges the two metal ions.

Loss of the bridging residue leads to loss of metal ion binding; the Asp51Ala variant of

SCO 6654 purifies with 0.06 equivalent of zinc and 1.5 equivalents of Ca only.

SCO 6654 does not hydrolyze creatinine

To determine if SCO 6654 is indeed a creatinine amidohydrolase, the protein was assayed for the activity with creatinine as substrate. We have been unable to detect any turnover using the colorimetric assay. The absence of CNNase activity and loss of substrate binding residues, coupled with the retention of the zinc site in SCO 6654, suggest that the protein may have alternate substrate specificity. We have shown previously that SCO 6655, which is divergently transcribed from SCO 6654, catalyzes the 128

conversion of GTP to FAPy, which contains a formamide moiety. This is an unusual activity which has only been observed in one other protein, GTP cyclohydrolase III, from

Methanocaldococcus jannaschii (18). To determine if the role of SCO 6654 is hydrolysis of the formamide moiety of FAPy, the enzyme was incubated with FAPy and reactions were monitored by UV­visible spectrophotometry (Figure 4). Indeed, SCO 6654 appears to catalyze the time and enzyme dependent conversion of FAPy to APy (Figure 4A). The difference spectrum of the reaction is consistent with loss of the characteristic 274 nm peak due to FAPy and formation of a molecule with 300 nm absorbance, consistent with the formation of APy. The assay was carried out in the presence of 5 mM Mn2+ which

appears to stimulate activity by the greatest extent (see below). We observe a Km = 0.5 ±

­1 0.02 mM and a kcat = 7.4 ± 0.1 min under these conditions.

To confirm that the product of SCO 6654 is indeed APy, the reactions were analyzed by HPLC (Figure 5). In these experiments, enzyme was incubated with FAPy in the presence of Mn2+ and the conversion of the 13.9 min FAPy peak to a 4.3 min peak was monitored. The reaction produces a peak with both the same retention time and UV­ visible spectra as APy. In contrast to wildtype SCO 6654, the Asp51Ala variant does not catalyze conversion of FAPy to any detectable product. This is consistent with the role of

Asp51 as a key metal ion binding ligand in the dimetallic active site.

To unambiguously identify APy as the product of SCO 6654, the reaction products were analyzed by mass spectrometry (Figure 5). In these experiments, we 129

utilized SCO 2687, a GCH II intragenomic homolog in S. coelicolor, which we have previously shown to catalyze the conversion of GTP to APy (1). Incubation of SCO 6654 with FAPy leads to formation of a product with the identical mass to charge ratio as APy

(m/z = 352.2). In the absence of SCO 6654, no APy is observed (Figure 5A). The identity of APy was further confirmed by MS/MS analysis (Figure 5B). In these experiments, the peaks with m/z = 352.2 from both SCO 6654 and SCO 2687 were fragmented and the resulting patterns were identical. Notably, both show a peak at 139, which corresponds to the base fragment of APy.

Divalent metal ions activate SCO 6654

While active in the conversion of FAPy to APy as purified, SCO 6654 is activated

as much as 40­fold by inclusion of 5 mM MnSO4. To determine if what metal ions can activate the protein, assays were carried out in the presence 5 mM each of Mn2+, Mg2+,

Ni2+, Ca2+, Co2+, Zn2+, Cu2+ (Table 2). Note that the protein as purified contains 0.46–0.5

Zn, 0.23 ­ 7 Ca, 0.7 ­ 1.3 Fe, 0.1 ­ 0.2 Mn equivalents and displays 2%­3% of its maximum observed activity without additional divalent metal ions. Of the divalent metal ions tested, Mn2+ activates the protein the best (40­fold), followed by Mg2+ (10­fold), Ni2+

(10­fold), Ca2+ (2­fold), and Co2+ (2­fold). By contrast Cu2+ and Zn2+ inhibits the activity

. The presence or absence of 1 mM DTT strongly affects the ability of Mn2+, Mg2+, and

Ca2+ to activate the enzyme; without DTT, they inhibit the activity while with DTT, they 130

activate.

DISCUSSION

While creatininase is fairly rare, the enzymatic strategy of using a two metal ion active site is common (19). Some active sites rely on a homodimetallic center while others form a heterodimetallic center; urease (Ni ­ Ni) (20), dipeptidase (Zn ­ [Co, Mo,

Mg, Zn, Mn]) (21­23), dihydropyrimidinase (Zn ­ Zn) (24), dihydroorotase (Zn ­ Zn)

(25), hydantoinase (Zn ­ Zn) (26, 27), phosphotriesterase (Zn ­ Co) (28), and amino peptidase (Zn ­ Co) (29, 30) The dimetallic centers are thought to activate a water for nucleophilic attack on the substrate. Interestingly, activity of these proteins can be reconstituted sometimes with homometallic centers using non­native metal ions: examples include phosphotriesterase (Co2+, Mn2+, Cd2+, Ni2+, Zn2+) (28, 31) and dihydroorotase (Zn2+, Co2+, Mn2+) (32­34). In the case of dipeptidase this variability seems to be a strategy to allow flexibility, various (Zn ­ M) combinations produce enzymes with altered catalytic efficiency toward different substrates (22, 23, 35). The active site of the P. putida CNNase has been reconstituted as a homodimetallic center with Mn2+, Co2+, Mg2+, Fe2+, Zn2+, Ni2+, Cu2+ giving greater to lesser activity in the same order (9).

SCO 6654 appears to use the the same zinc manganese dimetallic center found in

CNNase. Loss of the bridging aspartic acid residue leads to an enzyme that, apparently, 131

cannot form an active metal center. Phosphotriesterase, by contrast, is able to form an active site without its bridging ligand, a carbamyllysine, although the enzyme suffers some loss of activity (31). Urease variants lacking the bridging ligand are inactive and purify without a metal center but can be rescued (36).

We speculate that the His168 of SCO 6654 which aligns with His178 of CNNase plays an active role in catalysis. In CNNase this conserved residue has been proposed to function as a general base catalyst by removing a proton from a water which bridges the metal ions (8, 14) in a manner analogous to members of the urease superfamily (19) such as His254 in phosphotriesterase (32), His320 in urease (37), or His 234 in dihydroorotase

(38).

The biological role of SCO 6654 remains to be established. However, it is rather curious that SCO 6654 transforms FAPy, the product of SCO 6655, to APy the product of a canonical GCH II. FAPy, even as a triphosphate, is not a catalytically competent substrate for GCH II (39) and FAPy does not serve as a substrate for the two canonical intragenomic homologs of GCH II in S. coelicolor (SCO 2687 and SCO 1441) (JE

Spoonamore and V Bandarian, unpublished observations). One possibility is that SCO

6654 could serve to salvage FAPy by transfoming it into a chemical species, which can be utilized in biosynthesis of riboflavin.

A second possibility for the role of SCO 6654 is that it plays an integral role in a pathway that produces a purine­based molecule using the genes encoded in the SCO 6655

­ SCO 6648 cluster. Our working model is that the SCO 6655 / 6654 duo produce APy, 132

which is converted by the other enzymes of the cluster into an as yet unknown pterin. If this were true, one would expect to a SCO 6654 activity present in similar clusters.

Surprisingly, however, it is not. In the Actinomycetes S. coelicolor, S. scabies, and S. avermitilis one finds a similar cluster of genes with both a GCH II and a SCO 6654 type protein; in S. tropica CMB­440 and S. arenicola CNS­205, however, only a GCH II is present. Closer examination of the GCH II homologs in these organisms reveals a startling fact. In cases where a GCH II homolog is present that can produce APy (as judged by our knowledge of the GCH II switch residue (40) ), a SCO 6654 type protein is absent; however, when a non­canonical GCH II is present (GTP→FAPy), a SCO 6654 homolog is present. This tight correlation suggests that it is the coordinated activity of

SCO 6655 / 6654 (GTP→APy) that is required by the cluster. This conclusion supports the notion that SCO 6654 is an integral player in the pathway. As more genomes are sequenced it will become possible to judge the validity of this prediciton.

One possible product from the genes in this cluster is an as yet described pterin.

Our current thinking on the product of the SCO 6655 ­ 6648 gene cluster is shown in

Figure 7. Based on the similarity of the gene clusters of other organisms, it appears that

SCO 6648 and SCO 6649 have been mis­anotated as two genes when in fact, there is only one gene which is homologous to serine dehydrogenase (A. Hsieh and V. Bandarian, unpublished observation). In this model GTP is tranformed to FAPy by SCO 6655, FAPy is transformed to APy by SCO 6654. The product of SCO 6648/6649 and APy are 133

condensed by the PTPS homolog (SCO 6650) to form a pterin ring. It is hard to speculate specifically on the role of the other enzymes in the gene cluster, which are a glycosyl transferase and a methyl transferase. However, we note that formation of glycosyl pterins is precedented in nature (41, 42). If true, then this cluster represents a novel biosynthetic pathway to pterins. 134

References 1. Spoonamore, J. E., Dahlgran, A. L., Jacobsen, N. E. and Bandarian, V. (2006) Biochemistry 45, 12144­12155 2. Dubos, R. and Miller, B. (1937) J Biol Chem 121, 429­445 3. Wyss, M. and Kaddurah­Daouk, R. (2000) Physiol Rev 80, 1107­1213 4. Tricot, C., Piérard, A. and Stalon, V. (1990) J Gen Microbiol 136, 2307­2317 5. Krebs, H. and Eggleston, L. (1939) Enzymologia 7, 310­320 6. Kaplan, A. and Naugler, D. (1974) Mol Cell Biochem 3, 9­15 7. Nishiya, Y., Toda, A. and Imanaka, T. (1998) Mol Gen Genet 257, 581­586 8. Yoshimoto, T., Tanaka, N., Kanada, N., Inoue, T., Nakajima, Y., Haratake, M., Nakamura, K. T., Xu, Y. and Ito, K. (2004) J Mol Biol 337, 399­416 9. Rikitake, K., Oka, I., Ando, M., Yoshimoto, T. and Tsuru, D. (1979) J Biochem 86, 1109­1117 10. Inouye, Y., Matsuda, Y., Naid, T., Arai, S., Hashimoto, Y., Asano, K., Ozaki, M. and Nakamura, S. (1986) Chem Pharm Bull 34, 269­274 11. Liu, W. and Huang, C. (1988) J Chin Agric Chem Soc 26, 398­405 12. Kaplan, A. and Szabo, L. L. (1974) Mol Cell Biochem 3, 17­25 13. Tsuru, D. (1977) Amino Acid Nucleic Acid 35, 31­37 14. Beuth, B., Niefind, K. and Schomburg, D. (2003) J Mol Biol 332, 287­301 15. R Foundation for Statistical Computing (2005) R: A language and environment for statistical computing, Vienna, Austria 16. Jaffe, M. (1886) Z Physiol Chem 10, 391­400 17. Creatinine – Jaffe’s Method, World Health Organisation Guidelines on Standard Operating Procedures for Clinical Chemistry, http://www.searo.who.int/EN/Section10 /Section17/Section53 /Section481.htm (accessed April 11, 2008) 18. Graham, D. E., Xu, H. and White, R. H. (2002) Biochemistry 41, 15074­15084 19. Holm, L. and Sander, C. (1997) Proteins 28, 72­82 20. Jabri, E., Carr, M. B., Hausinger, R. P. and Karplus, P. A. (1995) Science 268, 998­ 1004 21. Patterson, E. K. (1976) Methods Enzymol 45, 377­386 22. Hayman, S. and Patterson, E. K. (1971) J Biol Chem 246, 660­669 23. Simmonds, S. (1970) Biochemistry 9, 1­9 24. Abendroth, J., Niefind, K. and Schomburg, D. (2002) J Mol Biol 320, 143­156 25. Thoden, J. B., Phillips, G. N. J., Neal, T. M., Raushel, F. M. and Holden, H. M. (2001) Biochemistry 40, 6989­6997 26. Abendroth, J., Niefind, K., May, O., Siemann, M., Syldatk, C. and Schomburg, D. (2002) Biochemistry 41, 8589­8597 135

27. Cheon, Y., Kim, H., Han, K., Abendroth, J., Niefind, K., Schomburg, D., Wang, J. and Kim, Y. (2002) Biochemistry 41, 9410­9417 28. Omburo, G. A., Kuo, J. M., Mullins, L. S. and Raushel, F. M. (1992) J Biol Chem 267, 13278­13283 29. Garner, C. W. J. and Behal, F. J. (1974) Biochemistry 13, 3227­3233 30. Little, G. H., Starnes, W. L. and Behal, F. J. (1976) Methods Enzymol 45, 495­503 31. Kuo, J. M., Chae, M. Y. and Raushel, F. M. (1997) Biochemistry 36, 1982­1988 32. Benning, M. M., Kuo, J. M., Raushel, F. M. and Holden, H. M. (1995) Biochemistry 34, 7973­7978 33. Brown, D. C. and Collins, K. D. (1991) J Biol Chem 266, 1597­1604 34. Huang, D. T., Thomas, M. A. and Christopherson, R. I. (1999) Biochemistry 38, 9964­9970 35. Hayman, S., Gatmaitan, J. S. and Patterson, E. K. (1974) Biochemistry 13, 4486­ 4494 36. Pearson, M. A., Schaller, R. A., Michel, L. O., Karplus, P. A. and Hausinger, R. P. (1998) Biochemistry 37, 6214­6220 37. Park, I. S. and Hausinger, R. P. (1993) Protein Sci 2, 1034­1041 38. Williams, N. K., Manthey, M. K., Hambley, T. W., O'Donoghue, S. I., Keegan, M., Chapman, B. E. and Christopherson, R. I. (1995) Biochemistry 34, 11344­11352 39. Ritz, H., Schramek, N., Bracher, A., Herz, S., Eisenreich, W., Richter, G. and Bacher, A. (2001) J Biol Chem 276, 22273­22277 40. Spoonamore, J. E. and Bandarian, V. (2008) Biochemistry 47, 2592­2600 41. Yamazawa, A., Takeyama, H., Takeda, D. and Matsunaga, T. (1999) Microbiology 145 (Pt 4), 949 ­ 954 42. Wachi, Y., Burgess, J. G., Iwamoto, K., Yamada, N., Nakamura, N., and Matsunaga, T. (1995) Biochem. Biophys. Acta 1244, 165 ­ 168

Acknowledgements

The authors thank Ms. Annie L. Dahlgran for cloning SCO 6654, Dr. John Osterhout for insightful comments, and Dr. Osamu Miyashita for translating Japanese language publications. 136

FIGURE LEGENDS

Figure 1. Chemical reaction catalyzed by (A) creatininase and (B) SCO 6654.

Creatininase catalyzes the hydrolysis of creatinine to creatine. SCO 6654 catalyzes the hydrolysis of FAPy to APy. Stars highlight the nitrogen where the carbon­nitrogen bond is hydrolzyed.

Figure 2. Bacterial creatinine degradation pathways. Two pathways, one through 1­ methylhydantoin and the second through creatine are known to exist. A third through methylguanidine has been described. The dashed box highlights the creatininase pathway where creatinine is hydrolyzed to creatine, creatine is hydrolyzed to sarcosine with the loss of urea, and sarcosine is degraded to glycine with the loss of formaldehyde.

Figure 3. Sequence alignment of SCO 6654 and creatininase from Pseudomonus putida.

Thick dashed boxes highlight the zinc ligands (H036, E183 and D045) while thick solid boxes highlight the manganese (E034, H120, and D045) metal center ligands of creatininase. Note that the bridging metal center residue of is D045. Grey dashed boxes highlight residues shown to interact with creatine in the P. putida creatininase crystallographic model (PDB ID: 1J2T) (8). The thin solid box highlights a conserved histidine which may function as a catalytic base. SCO 6654 has 23% identity and 36% similarity to the P. putida sequence. 137

Figure 4. SCO 6654 uses FAPy as a substrate. (A) Spectral change observed upon incubation of FAPy with SCO 6654 shows a loss of absorbance at 275 nm with a concomitant increase in the aborbance at 299 nm and an isobestic point at 280 nm. Inset shows the rate of the reaction is proportional to the amount of enzyme added (an extinction coefficient of 2300 M­1 cm­1 for the transformation of FAPy to APy was used).

(B) Difference absorbance showing the changes observed in between the first and last spectra from the top panel.

Figure 5. Loss of the conserved bridging ligand of the dimetallic metal center inactivates

SCO 6654. HPLC chromatograms over an analytical SAX column show (A) SCO 6654 digests FAPy (retention time 13.9min) to APy (4.3 min). Samples were run at 5 (bottom),

30 (middle), and 120 (top) minutes into the digestion. (B) The SCO 6654 Asp51Ala variant does not produce a detectable product peak even when present at 12­fold higher concentration than wildtype.

Figure 6. Mass spectra showing that the product of SCO 6654 is APy. (A) MS showing

FAPy incubated with no enzyme as a control (top), GTP digested with SCO 2687 (APy control) (middle), and FAPy digested with SCO 6654 (bottom). Dashed lines highlight the m/z of FAPy (380.2) and APy (352.2). (B) The MS/MS fragmentation pattern of the 138

352 m/z peaks from the SCO 2687 (top) and SCO 6654 (bottom) digestions.

Figure 7. Proposed biosynthesis encoded by gene cluster of unknown function which includes SCO 6654. The box highlights speculative roles for other enzymes in the cluster. 139

TABLES

Table 1. Molecular weights and oligomerization state of CNNases from different organisms Organism Molecular Oligomer Subunit Weight Reference Weight (kDa) (kDa) P. putida 140a 6 23b (9) P. putida (PDB ID: 1Q3K) 170c 6 28.4 (14) P. putida (PDB ID: 1J2T) 172c 6 28.6 (8) Arthrobacter eurofaciens 240a 8 30b (12) Alcaligenes sp. 160a 2 80b (10) Flavobacterium U­188 150a N.D.d N.D.d (13) P. putida var nariensis 210a N.D.d N.D.d (13) aDetermined by gel filtration, bDetermined by SDS­PAGE, cDetermined by multiplying the calculated weight of the monomer based on the amino acid sequence reported in the crystal structure, dNot determined. 140

Table 2. Activation of SCO 6654 by various divalent metals Divalent Metal Relative Activity w / DTT no DTT

MnSO4 100 0

MgSO4 26 0 a NiSO4 N.D. 24

CaCl2 5 0 a CoCl2 N.D. 4 no added metal 3 2 a ZnSO4 N.D. 1 a CuCl2 N.D. 0 aNot determined 141

FIGURE 1 142

FIGURE 2 143

FIGURE 3 144

FIGURE 4 145

FIGURE 5 146

FIGURE 6 147

FIGURE 7 148

APPENDIX D: SCO 6650, A HOMOLOG OF 6­ PYRUVOYLTETRAHYDROPTERIN SYNTHASE, IS A T­FOLD PROTEIN THAT APPEARS TO BIND A PYRIMIDINE RING CONTAINING SUBSTRATE 149

X­ray crystal structure of a 6­pyruvoyltetrahydropterin synthase homolog from

Streptomyces coelicolor

James E. Spoonamore1, Sue A. Roberts1, Annie Heroux2 and Vahe Bandarian1*

1 Department of Biochemistry and Molecular Biophysics, University of Arizona, Tucson,

AZ 85721, USA

2 Biology Department, Brookhaven National Laboratory, Upton, NY 11973, USA

* Corresponding author: Vahe Bandarian

University of Arizona

Department of Biochemistry and Molecular Biophysics

1041 E. Lowell St.

Tucson, AZ 85721

Tel: 520 626­0389

E­mail: [email protected] 150

ABSTRACT

The X­ray crystal structure of the 6­pyruvoyltetrahydropterin synthase (PTPS) from Streptomyces coelicolor, SCO 6650, was solved to 1.5 Å resolution. SCO 6650 is a hexameric T­fold that closely resembles other PTPS proteins. The biological activity of

SCO 6650 is unknown but it is not likely to catalyze the PTPS reaction since it lacks both a required active site zinc metal ion and an essential catalytic triad. SCO 6650, however, maintains residues consistent with binding a pterin­like substrate. 151

1. Introduction

Tetrahydrobiopterin (BH4) is biosynthesized in three steps from GTP by the successive actions of GTP cyclohydrolase I (GCH I), 6­pyruvoyltetrahydrobiopterin synthase (PTPS) and sepiapterin reductase (SR). Biopterin is not commonly found in bacteria, but is an essential cofactor in mammalian nitric oxide synthase and aromatic hydroxylation reactions. While rare in bacteria, glycosylated biopterin analogs have been described in a number of photosynthetic bacteria and roles ranging from photoreception to UV­protection have been proposed (Wachi, et al., 1995, Yamazawa, et al., 1999). The biological roles of these compounds and their structural diversity remain to be established.

PTPS catalyzes two successive Amadori reactions in the conversion of 7,8­ dihydroneopterin triphosphate to 6­pyruvoyltetrahydropterin (Figure 1). The X­ray structure of the protein from rat revealed a “tunnel” fold (T­fold)(Nar et al., 1994), which has also been found for GTP cyclohydrolase I (GCH I)(Nar et al., 1995), dihydroneopterin aldolase / epimerase (DHN aldolase)(Sanders et al., 2004), and urate oxidase(Retailleau et al., 2004) (PDB ID: 1B66, 1GTP, 1RRW, and 1R51, respectively).

The T­fold of PTPS is achieved by a dimer of trimers, which form two stacked torroids.

The active site of PTPS is composed of residues contributed from three subunits (A, A',

B)(Ploom et al., 1999). Three histidine sidechains from one subunit (A) and the vicinal hydroxyl groups of the substrate coordinate an essential divalent zinc metal ion in the 152

active site, whereas an Asp­His duo from an adjacent subunit (B) activate a Cys (A) residue to initiate chemistry. The exo­ and endo­cyclic nitrogens of the pyrimidine ring of the substrate interact with the sidechain carboxylate of a glutamate residue, while the pyrimidine carbonyl oxygen hydrogen bonds with the preceding amino acid backbone amide (A). Additionally, a second glutamic acid side chain (A) forms a hydrogen bond with the C1' hydroxyl and may interact with the N5 of the pterin ring (Figure 2).

The genome sequence of Streptomyces coelicolor (Bentley et al., 2002) revealed the presence of a PTPS homolog (SCO 6650), which is nestled in a cluster of genes that appears to be conserved in all strains of Actinomyces sequenced to date (Vahe and Tony, unpublished observations). In addition to PTPS, the cluster houses a GTP cyclohydrolase

II homolog, which previous studies from our laboratory have shown to catalyze the conversion of GTP to 2­amino­5­formylamino­6­ribosylamino­4(3H)­pyrimidinone

5'­phosphate (FAPy)(Spoonamore et al., 2006). This is an unorthodox arrangement from a number of respects. To our knowledge, Streptomyces do not produce biopterin, therefore, the presence of a PTPS homolog and absence of a recognizable sepiapterin reductase protein seem contradictory. Also, even if this organism produced biopterin, the first step would be carried out by GCH I and not by GCH II, which appears to co­localize with SCO 6650 in Actinomyces.

As part of a study to assign the biological function of this cluster, the PTPS

homolog (SCO 6650) was cloned, overexpressed in E. coli, and purified either as His6­ 153

tagged or the native form. Both the native and His6­tagged protein readily crystallized

and the structure of the Se­Met His6­tagged protein was solved to 1.5 Å. The model reveals that unlike the canonical PTPS proteins, SCO 6650 lacks nearly all the catalytically essential residues, but retains residues that are required for binding of a purine­like substrate.

2. Materials and Methods

2.1 Materials

Restriction endonucleases were from New England Biolabs. Expression vectors pET28a and pET29a, and expression strain BL21(DE3) E. coli were from New England

Biolabs. The pGEM­T Easy kit was from Promega. LB refers to Lenox Broth. Pfu

Turbo and DH5α were from Invitrogen. Purification resins were obtained from GE

Healthcare. All other chemicals were obtained from VWR Scientific and Sigma Aldrich.

2.2 Protein Expression and Purification

The gene encoding the S. coelicolor PTPS (sco6650) was amplified with Primer 1

(5'­GGAATTCCCATGGCATATGTTCAGCATCACCGTCCGCGATCAC­3') and

Primer 2 (5' ­ TTTTAAGCTTATTACAGCGCACGCTCGTAACTCGC ­ 3') from S. coelicolor genomic DNA and sub­cloned into the pGEM T­Easy vector prior to being cloned into the NdeI/HindIII site of pET28a and pET29a overexpression plasmids for the 154

expression of His6­ or native protein, respectively. The sequence of the final construct revealed that a non­conservative mutation had been inadvertently introduced (via Primer

2) in the course of the initial cloning. To correct this error, the gene was amplified from the cloning vector using Primer 1 and Primer 3 (5' ­

GCAAGCTTATTACAGCGCACGCTCGTAACTCGCCCAGGC ­ 3') and re­cloned into the expression vectors. DNA sequencing revealed the correct gene had been cloned.

Native protein was prepared by transformation of the pET29 expression vector bearing SCO 6650 into E. coli BL21(DE3) cells. Transformants were grown overnight in

Lenox broth containing 34 µg/mL kanamycin prior to inoculation of large scale (1 L) growths containing 34 µg/mL kanamycin. Expression was induced by addition of 0.1 mM

isopropyl β­D­1­thiogalactopyranoside at OD600nm~ 1; cells were harvested after 4 h. For the purification, cell pastes were resuspended into 25 mL of 0.02 M Tris•HCl (pH 8) containing 0.5 mM DTT. The extract was also supplemented with one tablet of Complete

(­EDTA) protease inhibitor tablet from Roche. Cells were lysed by sonication, centrifuged to remove cellular debris. The cleared lysate was applied to a Q­Sepharose

Fast Flow column (2.5 x 14.5 cm) that had been equilibrated with 0.02 M Tris•HCl (pH

8) containing 0.5 mM DTT; the protein was eluted with a linear gradient in the loading buffer of 0 ­ 0.5 M KCl. Fractions that contained SCO 6650, as determined by SDS­

PAGE, were pooled. SCO 6650 eluted at ~0.25 M KCl. Pooled fractions were dialyzed against 4 L 0.02 M Tris•HCl (pH 8.0) containing 0.5 mM DTT with one buffer change. 155

SCO 6650 was concentrated with a 10,000 molecular weight cut­off centrifugal concentrator (YM­10 membrane, Amicon) and flash frozen in liquid nitrogen until needed. Protein stock was quantified using the BCA protein assay kit from Pierce with

BSA as standard.

Selenomethionine­labeled N terminal His6­tagged protein was expressed in M9 minimal media supplemented with selenomethionine as described previously (Doublié,

1997). The protein was purified essentially as described for the native protein above, except an additional gel filtration step was included as follows. Dialyzed and concentrated Q­Sepharose fractions were applied to a Sephacryl S­300 column (2.6 x 60 cm) equilibrated with 0.02 M Tris•HCl (pH 8) containing 0.5 mM DTT. Protein was eluted by the same buffer.

Metal analysis of the selenomethionine­labeled protein by ICP­OES (Garratt­

Callahan) revealed 2.1 equivalents of selenium in each monomer, which is consistent with post­translational removal of the first methionine, resulting in two methionine atoms in each monomer.

2.3 Crystallization and structure solution

Conditions yielding crystals of native and SeMet­containing His6­SCO6650 were found using sparse matrix screening kits Crystal Screen I and II from Hampton Research.

The hanging­drop method was utilized in these experiments. For the native crystals, the 156

optimal crystallization conditions entailed combining 3 µL of a solution containing 1.8 M ammonium sulfate and 0.1 M potassium phosphate (pH 8) with 3 µL of the protein solution, which was typically composed of 16 mg/mL SCO 6650 in a buffer that contained 0.02 M Tris•HCl (pH 8.0) and 0.5 mM DTT. The well contained 0.5 mL of the precipitation buffer. Cubic crystalls grew at room temperature in 4 days. The SeMet histidine­tagged SCO 6650 was crystallized as follows. An aliquot (3 µL) of the protein solution, which typically contained 40 mg/mL protein in 0.02 M Tris•HCl (pH 8.0) and

0.5 mM DTT was combined with 3 µL of a solution containing 2 M NaCl and 10% PEG

6000. The well volume was 0.5 mL. Crystals grew at room temperature in 7 days and were flash frozen using 3.2 M NaCl, 10% PEG 6000 for cryoprotection.

Selenium MAD data were collected at 100 K on an ADSC Quantum­4 CCD detector at beam line X26C at Brookhaven National Laboratory. Data from one crystal were collected at three wavelengths (see Table 1 for statistics) and reduced using

HKL2000 (Otwinowski & Minor, 1997). Data to 2.6 Å resolution were used for phasing.

Nine selenium positions were located and initial phases were calculated using SOLVE

(Terwilliger & Berendzen, 1999). Density modification, phase extension to 1.5 Å resolution, and initial structure building was performed using RESOLVE (Terwilliger,

2003). The structure was refined, including TLS refinement, using REFMAC5

(Murshudov et al., 1997) and rebuilt with COOT (Emsley & Cowtan, 2004). There is residual difference electron density on the faces and empty center of the hexamer along 157

the pseudo threefold axis. Imidazole rings of histidine residues from the His­tag

(sequence = MGSSHHHHHHSSGLVPRGSH) can be seen stacked on Trp25 residues from each chain located at the top and bottom of the central cavity. Except for these stacked histidine residues, the residual electron density could not be interpreted and, since neither the residue number nor chain indentification of these histidine residues can be determined, they are not included in the model.

3. Results and discussion

3.1 Overall Structure of SCO 6650

The asymmetric unit of the His6­tagged protein crystals contains a hexamer, which is the biological unit. The structural model shows SCO 6650 is a T­fold protein composed of a homo­hexamer in the form of two trimeric subunits each composed of a large, helix wrapped 12 stranded anti­parallel beta barrel. The two trimers stack to form the hexamer with overall dimensions of 53 Å along the barrel axis and 55 Å diameter (Figure 3A,

3B). Gel filtration chromatography of native SCO 6650 estimates the molecular weight to be 113 kDa, which corresponds to an octomer (data not shown).

3.2 Comparison of SCO 6650 to PTPS proteins

The two highest scoring hits from a PSI­BLAST (Altschul et al., 1997) search with the amino acid sequence of SCO 6650 are two proteins annotated as antibiotic 158

biosynthesis proteins for gentamycin (gi | 85814033) and fortimycin (gi | 85813928) while all other matches are PTPS proteins. Both antibiotic biosynthesis sequences conserve the

PTPS catalytic triad and zinc coordinating histidines. The search identified the PTPS from Pseudomonas aeruginosa (gi 126031480) (PDB ID: 2OBA) as fourth best match (E score 3 e ­39). A multiple sequence alignment using PTPS sequences from the well described protein from Rattus rattus (gi 4929886), Pseudomonas aeruginosa (gi

126031480), Pyrococcus horikoshii (gi 14590525) (Bagautdinov et al., 2007) (PDB ID:

2DTT), and SCO 6650 (Figure 4). The alignment reveals substantial sequence similarities; however, it is clear that key catalytic residues for PTPS chemistry are absent in SCO 6650.

Structural alignments further re­enforce the overall global similarities. In these comparisons, SCO 6650 monomer or oligomer was superimposed with PTPS structures from P. horikoshii, P. aeruginosa, R. rattus, P. vivax, and P. faciparum (PDB ID: 2DTT,

2OBA, 1B66, 2A0S, and 1Y13, respectively) using the SSM (Krissinel & Henrick, 2004)

(Table 2). The root mean square deviation of ­carbons varies between 1.6 and 2.1 for the oligomers, indicating substantial structural similarities. As is discussed below, the structural alignments reveal that the active site of SCO 6650, as had been suspected from the sequence alignments, is indeed devoid of the residues that would be required for PTPS chemistry. 159

3.3 The SCO 6650 Active Site

The active site of PTPS is formed at the interface of the trimers with contributing amino acids from three monomers (A, A', B) leading to a total of 6 active sites formed at the interface between the two trimers (Ploom et al., 1999). SCO 6650 appears to have a similar arrangement. Using the SSM superpose option within COOT, the rat PTPS (PDB

ID: 1B66) and SCO 6650 hexamers were superimposed. Both the structural overlay and sequence alignment show that SCO 6650 can not catalyze the PTPS reaction since it lacks both the active site zinc metal and the catalytic triad. SCO 6650 does not maintain all three zinc­coordinating histidine residues, which are required for activity (Bürgisser et al., 1995) (Figure 3D). Zinc is not observed in the crystal structure of SCO 6650 nor could it be detected by ICP­OES analysis of protein preparations even when the protein was overexpressed with 100 M supplemental zinc in the growth media. SCO 6650 maintains two of the histidines but the third residue is an alanine. Furthermore, the active site cysteine, which is thought to initiate the chemistry and the two other residues in the catalytic triad (asp and his) to which it belongs, are absent in SCO 6650. (Bürgisser et al., 1995; Ploom et al., 1999; Bürgisser et al., 1994). The Cys­Asp­His triad is replaced with Ala­Gln­Tyr.

3.4 Substrate binding

Comparison of published ligand­bound PTPS structures with that of SCO 6650 160

provides insights into the potential substrate of SCO 6650. In this analysis, the ligand­ bound models of P. horikoshii (Bagautdinov et al., 2007), R. rattus (Ploom et al., 1999),

Plasmodium vivax, and Plasmodium faciparum were utilized (PDB ID: 2DTT, 1B66,

2A0S, 1Y13, respectively). All of these PTPS proteins use a Ser/Thr–Glu motif to bind the pyrimidine ring of the substrate. The carboxylate sidechain of the glutamic acid residue and the backbone amide from the preceding amino acid hydrogen bond with the substrate amino­pyrimidinone ring N3, exo­cyclic amino group, and carbonyl oxygen

(Figure 2). A second glutamic acid side chain hydrogen bonds with the C1' hydroxyl (or

C2' hydroxyl in the case of the trimeric P. horikoshii protein). A backbone carbonyl makes an additional hydrogen bond with the exo­cyclic pyrimidine amide. When any of these proteins is aligned with SCO 6650, the substrate binding residues appear strikingly conserved (Figure 3C). This constellation of residues for binding the amino­pyrimidine ring of pterins has been noted in T­fold proteins such as GCH I and DHN aldolase

(Hennig et al., 1998). Hence, SCO 6650 appears to be another example of a T­fold protein capable of binding an amino­pyrimidine substrate such as a pterin. 161

Acknowledgments The authors wish to thank Annie Dahlgran for cloning SCO 6650. 162

Table 1. Data Collection, Phasing, and Refinement Statistics

SeMet crystal SeMet crystal 1 2

crystal class, space group orthorhombic, P212121

a = 77.5 a =77.4 cell parameters (Å) b = 87.8 b = 88.0 c = 120.9 c = 120.9

Z (molecules / au) 6 temperature (K) 100 peak remote inflection wavelength (Å) .9777 .9600 .9787 .9771 resolution (Å) 1.7 1.68 1.7 1.45 523224 / 545914 / 518212 / 958738 / total/unique refls 165708 175003 165212 147043 completeness (%)a 96 / 71 97 / 78 96 / 77 96 / 92 mean I / σ(I)a 34. / 2.3 36 / 2.3 37. / 2.5 44 / 4.2 a,b Rsym 0.08 / 0.49 0.08 / 0.50 0.08 / 0.45 0.12 / 0.32 FOM (solve) 0.60 at 2.6 A c Rcryst / Rfree 0.20 / 0.22 rmsd bonds (Å) / angles (°) 0.14 / 1.36 % of residues in most favorable / additional allowed 92.5 / 7.5 regions of Ramachandran plot avg B (Å2) 20.1 a b Σ Σ   Σ Σ Overall/outermost shell. Rsym = hkl i Ii(hkl) – / hkl i I(hkl), where

c Σ  is the mean intensity of all symmetry­related reflections Ii(hkl). Rcryst = ( Fobs–Fcalc )/

Σ Fobs. Rfree as for Rcryst, using a random subset of the data (5%) not included in the refinement. 163

Table 2. Comparison of the sequence identities and alpha chain similarity of PTPS homologs to SCO 6650

%Identity Cα RMS deviation homolog PDB ID (%Similarity) Hexamer Monomer P. horikoshii 2DTT 21 (40) 1.55a 1.32 P. aeruginosa 2OBA 20 (32) 2.11 1.64 R. rattus 1B66 20 (31) 1.99 1.54 P. vivax 2A0S 16 (29) 1.89 1.45 P. faciparum 1Y13 15 (30) 1.91 1.41 aThe P. horikoshii biological unit is a trimer, this is the a fit to one of the SCO 6650 torroids 164

Figure Legends

Figure 1. PTPS catalyzes the transformation of 7,8­dihydroneopterin triphosphate to 6­ pyruvoyl­tetrahydropterin

Figure 2. Interactions between PTPS and its substrate, 7,8­dihydroneopterin triphosphate, (from PDB ID: 1B66).

Figure 3. The SCO 6650 hexamer looking across (A) and down (B) the T­fold barrel. The hexamer of SCO 6650 (color) overlain with the hexamer of rat PTPS (grey) highlights similarities in the residues known to be important for substrate binding (C). SCO 6650 (color) overlay with rat PTPS (grey) highlights differences between SCO6650 and the PTPS catalytic triad and zinc ligand residues (D).

Figure 4. A multiple sequence alignment of PTPS homologs from P. horikoshii, P. aeruginosa, R. Rattus, and SCO 6650 generated with Clustal W (Thompson et al., 1994). The boxes in the alignment highlight: PTPS zinc ligand histidines (Zn) one of which is not conserved in SCO 6650; the PTPS catalytic triad residues (Cat) are all different in SCO 6650; SCO 6650 conserves residues important for substrate binding in PTPS (Sub). 165

References Altschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Nucleic Acids Res. 25, 3389­3402. Bagautdinov, B., Sugahara, M. & Kunishima, N. (2007). Acta Cryst. F63, 15­17. Bentley, S. D., Chater, K. F., Cerdeño­Tárraga, A., Challis, G. L., Thomson, N. R., James, K. D., Harris, D. E., Quail, M. A., Kieser, H., Harper, D., Bateman, A., Brown, S., Chandra, G., Chen, C. W., Collins, M., Cronin, A., Fraser, A., Goble, A., Hidalgo, J., Hornsby, T., Howarth, S., Huang, C., Kieser, T., Larke, L., Murphy, L., Oliver, K., O'Neil, S., Rabbinowitsch, E., Rajandream, M., Rutherford, K., Rutter, S., Seeger, K., Saunders, D., Sharp, S., Squares, R., Squares, S., Taylor, K., Warren, T., Wietzorrek, A., Woodward, J., Barrell, B. G., Parkhill, J. & Hopwood, D. A. (2002). Nature 417, 141­147. Bürgisser, D. M., Thöny, B., Redweik, U., Hess, D., Heizmann, C. W., Huber, R. & Nar, H. (1995). J. Mol. Biol. 253, 358­369. Bürgisser, D. M., Thöny, B., Redweik, U., Hunziker, P., Heizmann, C. W. & Blau, N. (1994). Eur. J. Biochem. 219, 497­502. Doublié, S. (1997). Methods Enzymol. 276, 523­530. Emsley, P. & Cowtan, K. (2004). Acta Cryst. D60, 2126­2132. Hennig, M., D'Arcy, A., Hampele, I. C., Page, M. G., Oefner, C. & Dale, G. E. (1998). Nat. Struct. Biol. 5, 357­362. Krissinel, E. & Henrick, K. (2004). Acta Cryst. D60, 2256­­­2268 Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Acta Cryst. D53, 240­255. Nar, H., Huber, R., Heizmann, C. W., Thöny, B. & Bürgisser, D. (1994). EMBO J. 13, 1255­1262. Nar, H., Huber, R., Meining, W., Schmid, C., Weinkauf, S. & Bacher, A. (1995). Structure 3, 459­466. Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307­326 Ploom, T., Thöny, B., Yim, J., Lee, S., Nar, H., Leimbacher, W., Richardson, J., Huber, R. & Auerbach, G. (1999). J. Mol. Biol. 286, 851­860. Retailleau, P., Colloc'h, N., Vivarès, D., Bonneté, F., Castro, B., El­Hajji, M., Mornon, J. P., Monard, G. & Prangé, T. (2004). Acta Cryst. D60, 453­462. Sanders, W. J., Nienaber, V. L., Lerner, C. G., McCall, J. O., Merrick, S. M., Swanson, S. J., Harlan, J. E., Stoll, V. S., Stamper, G. F., Betz, S. F., Condroski, K. R., Meadows, R. P., Severin, J. M., Walter, K. A., Magdalinos, P., Jakob, C. G., Wagner, R. & Beutel, B. A. (2004). J. Med. Chem. 47, 1709­1718. Spoonamore, J. E., Dahlgran, A. L., Jacobsen, N. E. & Bandarian, V. (2006). Biochemistry 45, 12144­12155. Terwilliger, T. C. & Berendzen J. (1999). Acta Cryst. D55, 849­861 Terwilliger, T. C. (2003). Acta Cryst. D59, 38­44. 166

Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). Nucleic Acids Res. 22, 4673­ 4680. Wachi, Y., Burgess, J. G., Iwamoto, K., Yamada, N., Nakamura, N. & Matsunaga, T. (1995). Biochim. Biophys. Acta 1244, 165­168. Yamazawa, A., Takeyama, H., Takeda, D. & Matsunaga, T. (1999). Microbiology 145 ( Pt4), 949­954. 167

Figure 1. 168

Figure 2. 169

Figure 3. 170

Figure 4.