University of Groningen

Microbial production of thioether-stabilized peptides Kuipers, Anneke

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2010

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Kuipers, A. (2010). Microbial production of thioether-stabilized peptides. s.n.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 23-09-2021

Microbial production of thioether-stabilized peptides

Anneke Kuipers

Cover: Amino acid sequence of the thioether-stabilized angiotensin-(1-7) analogue.

Printing: Ridderprint, Ridderkerk, the Netherlands.

The studies described in this thesis were performed at the Biomade Technology Foundation, Nijenborgh 4, 9747 AG Groningen.

RIJKSUNIVERSITEIT GRONINGEN

Microbial production of thioether-stabilized peptides

Proefschrift

ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op vrijdag 10 september 2010 om 16.15 uur

door

Anneke Kuipers

geboren op 29 april 1967 te Groningen Promotor : Prof. dr. O. P. Kuipers

Copromotor : Dr. G. N. Moll

Beoordelingscommissie : Prof. dr. A. J. M. Driessen : Prof. dr. J. Kok : Prof. dr. M. Kleerebezem

ISBN: 978-90-367-4421-8 (boekversie) ISBN: 978-90-367-4422-5 (digitaal)

Contents

Chapter 1 Introduction 9

Chapter 2 NisT, the transporter of the lantibiotic nisin, can 21 transport fully modified, dehydrated and unmodified prenisin and fusions of the leader peptide with non- lantibiotic peptides

Chapter 3 Post-translational modification of the therapeutic 35 peptides by NisB, the dehydratase of the lantibiotic nisin

Chapter 4 Mechanistic dissection of the enzyme complexes 49 involved in the biosynthesis of lacticin 3147 and nisin

Chapter 5 Sec-mediated transport of post-translationally 63 dehydrated peptides in Lactococcus lactis

Chapter 6 Translocation of a thioether-bridged azurin peptide 77 fragment via the Sec pathway in Lactococcus lactis

Chapter 7 Angiotensin-(1-7) with thioether-bridge: an 83 angiotensin-converting enzyme-resistant, potent Angiotensin-(1-7) analogue

Chapter 8 Summary, general discussion and perspectives 95

Chapter 9 References 101

Samenvatting 119

Nawoord 123

List of 125 publications

Introduction

The idea to exploit lantibiotic biosynthesis enzymes for the stabilization of peptide hormones by introducing thioether bridges was already brought up in a Nature paper from 1988. In that same article Schnell et al. introduced the name lantibiotics for antimicrobial peptides that contain lanthionines (183). The first description of a lanthionine was in a publication of Horn and co-workers, who isolated a thioether- cross-linked amino acid from sodium carbonate treated wool (63). The name lanthionine (Latin, lana = wool) was introduced and represents two alanine residues coupled via a thioether linkage. Thioether bridges are more stable than disulfide linkages and peptide bonds (205). Before 1988 several lantibiotics, among which nisin, subtilin, epidermin and Pep5 (1, 52, 53, 72), were discovered and in 1971 the structure of the lantibiotic nisin was elucidated by chemical degradation (Fig. 1). This study also revealed that all lanthionines and methyllanthionines are composed of one D-amino acid coupled by a mono-sulfide linkage to one L-amino acid. Hence their synthesis takes place stereospecifically. The thioether bridge pattern in the peptide nisin was subsequently confirmed by nuclear magnetic resonance (NMR) spectroscopy (109, 210, 214).

Figure 1. The lantibiotic nisin (Gross and Morrell 1971).

Nisin is a pentacyclic peptide. The first ring is a lanthionine (Ala-S-Ala) and the remaining four are methyllanthionines (Abu-S-Ala), of which the last two are intertwined. Furthermore, the peptide contains 3 dehydroresidues and 21 unmodified amino acids. These (methyl)lanthionines give lantibiotics their unique features like Chapter 1 thermostability, proteolytic resistance and most (methyl)lanthionines are essential for high antimicrobial activity. The assumption that enzymes were involved in the stereospecific introduction of these thioether bridges in peptides, urged the search for the genes implicated in the biosynthesis of lantibiotics. In 1993 the gene cluster involved in the nisin biosynthesis was unraveled (90, 212). In 1999 a project was initiated for the starting Biomade Technology Foundation. This project aimed at the utilization of the nisin modification enzymes for stabilization of nonnatural substrate peptides. New thioether-bridged peptides may be novel antibiotics, which are of great interest because of the increase in resistance to multiple antibiotics, or may be stabilized peptide drugs. By stabilization, these therapeutic peptides are less sensitive to proteolytic breakdown and accordingly need less frequent administration and/or in a lower dose. In addition, stabilization may allow oral and pulmonary delivery. These delivery ways are more patient-friendly than injection. Furthermore, the structural constraint resulting from the introduction of thioether bridges may enhance the receptor specificity and/or the efficacy of the receptor interaction, thus enhancing the therapeutic potential. Research at BiOMaDe demonstrated the feasibility of these concepts and resulted in several publications that partly contributed to this thesis. This thesis discusses the dissection and exploitation of the nisin synthetase and the lacticin 3147 synthetase complexes and their utilization for stabilization of nonnatural substrate peptides. Furthermore, this thesis deals with the examination of the different transport routes that can be used for translocation of the modified peptides across the membrane of the producing bacteria.

Lantibiotics Lantibiotics are ribosomally produced peptides, containing posttranslationally modified amino acids, mainly (methyl)lanthionines and dehydroresidues. Lantibiotics are predominantly produced by Gram-positive bacteria and are principally effective against Gram-positive bacteria. The antimicrobial activity can be effected by disruption of the membrane of the target organism by pore formation (85). Lantibiotics may use lipid II as a docking molecule to efficiently generate hybrid pores (13, 17). They may also inhibit cell growth by binding to lipid II and displacing this precursor of cell wall synthesis (56). Besides, they may inhibit the germination of spores of the species Bacillus and Clostridium (200). All the above-mentioned modes of action are valid for the lantibiotic nisin. Nisin was the first lantibiotic described in literature (166) and is the most studied lantibiotic. It is produced by different Lactococcus lactis strains which are designated as GRAS (generally recognized as save) organisms. Already in 1969, nisin was approved for use as a food preservative (36). Nisin has a broad activity spectrum against Gram-positive bacteria, amongst others against strains of Staphylococcus , Streptococcus, Micrococcus, Lactobacillus, Bacillus, Listeria and Clostridium (199),

10 Introduction and has antimicrobial activity in the nanomolar range (38). These features make the search for novel nisin variants by genetic engineering an interesting approach in the battle against multiple resistant pathogens (49, 88, 159). By their stability, high activity and virtual absence of resistance development, lantibiotics are promising candidates for biomedical application (113). For example, like nisin, lacticin 3147 can be used for prevention and/or treatment of bacterial mastititis and MRSA (27). Nisin, subtilin, epidermin, Pep5 and some similar lantibiotics were first designated as type A lantibiotics, which are rod-shaped, flexible with an elongated structure and which mainly act by forming pores in the bacterial membrane (70). Type B lantibiotics (e.g. cinnamycin, duramycin and ancovencin) were discerned as having a higher degree of cyclization resulting in structures that are more globular and being devoid of pore-forming activity. Nowadays, more than 60 different lantibiotics have been discovered (10) and no less than 15 different posttranslational modifications have been described (224). After the finding of many new lantibiotics the old type A and type B classification became blurred and it was suggested in a scheme by Pag and Sahl to use three groups for classification (143, 224). This new classification places all lantibiotics in one of the three classes on the basis of the biosynthesis machinery used for maturation of the peptide (class I and class II) or the absence of antibiotic activity (class III). Moreover, in class III lantibiotics, LanM maturation enzymes may act via a mechanism distinct from that of class II LanM enzymes.

The lantibiotics biosynthesis machineries The genes involved in lantibiotic synthesis are genetically arranged in gene clusters. These gene clusters can be organized on a transposon (nisin), on the chromosome (subtilin) or on a plasmid (epidermin). Genes on these clusters have been designated the generic locus symbol lan (37). Besides gene products required for the biosynthesis of the peptides, also proteins are encoded which are needed for the processing ( lanP ), translocation ( lanT ), self-protection ( lanI, lanEFG ) and regulation (lanRK ). Per type, many of these proteins encoded in the different gene clusters show amino acid homology, which indicates that indeed they have similar functions (90, 153, 190). The lantibiotics nisin, epidermin and Pep5 belong to the class I lantibiotics and their gene clusters were characterized as depicted in figure 2.

11 Chapter 1

Figure 2. Gene clusters of class I lantibiotics . Genes involved in the modification ( lanB and lanC ) and transport ( lanT ) of the peptide ( lanA ) are illustrated with filled arrows. Promoters are indicated by wedges (21).

In class I lantibiotics, reviewed by Willey and van der Donk, the prepeptide LanA is modified by two distinct enzymes, LanB and LanC. The LanA prepeptide contains a leader sequence that is thought to be necessary for targeting the propeptide part to the modifying - , processing - and translocating enzymes. LanB dehydrates the serines and threonines in the propeptide part of LanA, and LanC couples these dehydrated residues regio- and stereoselectively to cysteines to form respectively lanthionines and methyllanthionines (Fig. 3).

Figure 3. Introduction of an intramolecular thioether bridge by lantibiotic enzymes . LanB dehydrates serine (1) (or threonine) and LanC couples the formed dehydroalanine (2) (or dehydrobutyrine) stereoselectively to a cysteine (3), thus forming one DL-lanthionine (4).

12 Introduction

After translocation of the modified peptide via an ABC transporter LanT, the leader part is in most class I lantibiotic systems removed by a protease LanP, releasing the active lantibiotic (224). In class II lantibiotics, only one enzyme is responsible for dehydration and cyclization of the propeptide LanA. These bifunctional LanM enzymes bear no homology with the LanB enzymes. However, the C-terminal part of these LanM enzymes has low sequence homology with the LanC enzymes, including three zinc- coordinating amino acids (146, 190). Knockouts of one of these zinc ligands completely abolished the cyclase activity of NisC or LctM (106, 149). Another dissimilarity to class I lantibiotics is the dual functionality of LanT. Before translocation of the modified peptide, the peptide is intracellularly processed by the conserved N-terminal protease part of LanT (143, 224). Prototypes of class II lantibiotics are mersacidin and lacticin 481. The second class also comprises the two- component lantibiotics, like lacticin 3147 (168). The two prepeptides LanA α and LanA β are each separately modified by the two enzymes, respectively LtnM1 and LtnM2. After modification, both peptides are processed and translocated by one LtnT enzyme. The gene clusters of the group II lantibiotics are depicted in Figure 4. The gene cluster of lacticin 3147 contains an additional post-translational modification enzyme, LtnJ. This enzyme converts some dehydroalanines in the prepeptides Ltn α and Ltn β to D-alanines (169).

Figure 4 . Gene clusters of class II lantibiotics . Genes involved in the modification ( lanM ) ( lanJ ) and transport ( lanT ) of the peptide ( lanA ) are illustrated with filled arrows. Promoters (known ones) are indicated by wedges (60, 21).

The third class of lantibiotics contains (methyl)lanthionine-containing peptides are mainly devoid of antimicrobial activity. Instead, they have other -for instance morphogenetic- features that may be beneficial to the producing cells. Three lantibiotics in this group are known by now: SapB (79), SapT (80) and AmfS (207). SapB and SapT are believed to be biosurfactants that may have a positive effect on the surface of arial hyphae of the producer strains. Furthermore, the LanM enzymes involved in the biosynthesis of SapB and AmfS have homology with the C-terminal part of other LanM enzymes except for the zinc ligands, which are missing.

13 Chapter 1

Engineering of lantibiotics With the elucidation of gene clusters involved in the biosynthetic pathways of lantibiotics, the next challenge became genetic engineering of lantibiotics. The existence of natural variants among lantibiotics (i.e. nisin A/nisin Z, epidermin/gallidermin) and the high homology between certain lantibiotics (i.e. nisin/subtilin, mutacin II/lacticin 481) suggests that the identity of amino acids present at certain locations is flexible. Generation of mutant lantibiotics with enhanced biological activity or improved physical properties therefore seems promising. In fact, by site directed mutagenesis of the structural genes and the development of expression systems many lantibiotic variants were designed and produced in vivo (Fig. 5AB).

Figure 5A. Some early mutants of some class I lantibiotics created by site directed mutagenesis. Black circles indicate amino acid differences between natural variants. Grey circles indicate mutations (adapted from Cotter 2005a).

14 Introduction

The most engineered lantibiotic is nisin. In 1992 the first nisin mutants were reported (88) and these mutants were followed by many other nisin mutants, which have been reviewed (92, 113). The alteration of residues that take part in formation of the third ring of nisin by the substitution T13C resulted in reduced antimicrobial activity of the nisin mutant. Also the substitution S3T, changing ring A of nisin from a lanthionine in a methyllanthionine, led to a dramatic loss of bioactivity. The mutation T2S resulted in an interesting mutant that displayed a two-fold higher antimicrobial activity against two target organisms (92). Some hinge region mutants had antimicrobial activity against Gram-negative species and furthermore by altering the charge of the nisin lantibiotic, solubility could be improved (233).

Figure 5B. Some class II mutants obtained by site directed mutagenesis. Grey circles for the lantibiotics mutacin II, cinnamycin and mersacidin indicate mutations (adapted from Cotter 2005a). For lacticin 3147, comprising LtnA1 and LtnA2, grey circles represent essential residues. Continuous lines indicate essential domains/amino acids, dashed lines indicate domains that are for the most part variable (adapted from Cotter 2005a).

15 Chapter 1

In addition, other lantibiotics were altered by site-directed mutagenesis. The subtilin E4I substitution displayed a 57-fold improvement in stability and had 3-4 fold the specific activity in suppression of bacterial spore outgrowth (110). Interesting gallidermin mutations were the substitutions L6V, A12L and Dhb14P in the mature peptide. The L6V gallidermin variant had an increased antimicrobial activity, whereas the A12L and Dhb14P variants resulted in a remarkable resistance against proteolytic breakdown (141). The first introduced novel thioether bridge in a lantibiotic reported was for Pep 5. By substitution of A19C, a methyllanthionine was introduced in the peptide, which was formed between the Dhb on position 16 and the introduced cysteine at position 19. This mutant exhibited an increase in proteolytic stability against chymotrypsin and Lys-C. However, the novel thioether bridge had a negative effect on the antimicrobial activity of Pep 5 (8). Also in the class II lantibiotics, comprising mutacin II (24), mersacidin (196) and cinnamycin (221) new variants were made by site-directed mutagenesis. A systematic mutant analysis by alanine scanning of the two-peptide lantibiotic lacticin 3147 revealed the areas within the peptide that are amenable to changes and areas that are essential for the production. None of the mutants displayed an antimicrobial activity higher than that of the wild type producer (30). More recently mutagenesis and screening were accelerated by genetic randomization of specific amino acid coding sites within lantibiotic genes. By random mutagenesis and NNK scanning of nukacin ISK-1, a bank of nukacin ISK-1 variants was generated to identify the positional importance of individual residues responsible for antimicrobial activity (63). Furthermore, by random mutagenesis of mersacidin, 80 mutants were made that produce mature mersacidin at good levels and novel variants were obtained with improved overall bioactivity, such as F3W (4). In addition, novel variants of nisin with improved bioactivity were found by random mutagenesis. Nisin ring A mutants I4K/S5F/L6I and I4K/L6I showed enhanced activity against some target strains (159) even as mutations M21V, N20P and K22T in the hinge region (49). Novel lantibiotics with enhanced bioactivity may be lethal for the producer itself. For the nisin producer this was circumvented by using a production system without the presence of NisP. Without removal of the leader, there is no antimicrobial activity. After production, the leader can be removed by trypsin. Another approach is using an in vitro modification system. The lantibiotics lacticin 481 and the two peptide lantibiotic haloduracin were both modified successfully by incubation of the precursor peptide with the LanM enzymes in vitro (121, 226). In addition, the dehydrated precursor of nisin was successfully cyclized by incubation with NisC in vitro (106). Overall, the biosynthetic system used for the biosynthesis of lantibiotics seems to have a remarkable flexibility.

16 Introduction

The application of these enzymes for the modification and production of modified peptides that are entirely different in size and sequence from their native substrates is subject of this thesis.

Production and secretion In this thesis the modification and transport enzymes used in the biosynthesis of nisin (NisBTC) and lacticin 3147 (LtnTM2) were applied to investigate the feasibility to introduce thioether bridges in nonlantibiotic peptides. Both systems, NisBTC and LanTM2 were derived from Lactococcus lactis strains. Accordingly, the first approach to produce therapeutic peptides with thioether bridges made use of Lactococcus lactis NZ9000. Lactis NZ9000 is a plasmid-free and prophage-cured L. lactis MG1363 strain with nisRK integrated on the chromosome (93). The two- component regulatory system NisRK, in which NisR is a response regulator and NisK is a histidine kinase, is involved in the autoregulation of nisin biosynthesis. The fully maturated nisin induces via NisRK activation of the Pnis promoter, which controls transcription of the nisABTC genes (40, 91). These components led to the development of the well known NICE, Nisin Controlled Gene Expression, system. Nowadays this system is widely and successfully used for gene expression in Gram- positive bacteria, including bacterial genera other than Lactococcus (44, 76). In our lab we developed a two-plasmid system in which the two plasmids are compatible with each other for the expression and translocation across the membrane of modified peptides (77, 157)(Fig. 6).

Figure 6. Nisin inducible two-plasmid system for the production of modified peptides by L. lactis

17 Chapter 1

The genes encoding the enzymes NisBTC or LtnTM2 were cloned behind the nisin inducible Pnis promoter on a pIL-derived plasmid (192). This plasmid replicates bidirectionally and is appropriate for expression of larger proteins. The encoding sequence for the substrate that has to be modified was fused to the C-terminus of the leader-encoding sequences of NisA or LtnA2 under control of the nisin-inducible promoter. The expression plasmid used for this purpose is a high copy rolling-circle- replicating plasmid, a pNZ8048 derived plasmid (93). When possible, translocation occurred via NisT or via LtnT, respectively, and the modified peptide was harvested from the medium. L. lactis is a suitable producer-strain for peptides in vivo . An advantage of L. lactis as a producer is the absence of production of lipopolysaccharides or proteases like occurring in E. coli and B. subtilis, respectively. Extracellular production of peptides simplifies purification methods, especially in the case of L. lactis , which secretes only a very low number of proteins in the culture media. Moreover, it has been shown that the production level of secreted proteins reached mostly a higher level than that of proteins that were produced intracellularly (101). L. lactis NZ9000 harbors a wide range of enzymes (peptidases, housekeeping proteases) committed to intracellular proteolysis. On the contrary, it possesses only one extracellular housekeeping protease, HtrA (152). In L. lactis, the conserved Sec pathway is successfully used for translocation of homologous and heterologous proteins. These proteins are preceded by a Sec signal sequence that targets the proteins to the Sec pathway. During translocation across the membrane, this signal sequence is removed and the mature protein is integrated in the membrane, anchored to the cell wall or released into the medium. The Sec pathway translocates unfolded proteins across the membrane (45, 59). In this thesis, besides the dedicated lantibiotic transporters NisT and LtnT, also the Sec pathway is examined for transport of modified peptides. Another translocation pathway, which might be of interest, is the Tat (twin arginine translocation) pathway. The Tat pathway translocates folded proteins across the membrane and may consequently be an ideal route for transport of the more bulky (lanthionine-containing) peptides. E. coli and B. subtilis have both a well studied Tat export system (97). However, L. lactis lacks the Tat pathway, which might be a disadvantage for heterologues expression. When a Tat pathway can be introduced there will be no competition with homologues substrates.

Outline of this thesis As already mentioned above, the introduction of thioether bridges in peptides can have a tremendous impact on the stability of the peptide. Moreover, thioether- bridge-imposed peptide structures can improve the pharmacodynamic properties of peptides. Examples of improved therapeutic peptide variants with thioether bridges are enkephalin (155) and somatostatin (138). Both had increased stability and improved pharmacodynamic properties. These improved peptides with thioether

18 Introduction bridges were chemically synthesized. Importantly, thioether peptides produced via lantibiotic enzymes contain only one isomer, which is a significant advantage. A biologically introduced thioether bond bridges a D-amino acid to an L amino acid, whereas chemically induced thioether formation can lead to several stereo isomers (i.e. DL, LL, LD and DD). In the case of engineering more than one thioether bridge in one peptide, regiospecificty of the lantibiotic enzymes can have an additional advantage. For more complex peptides with intertwined or multiple rings and for larger polypeptides the biological production may dramatically reduce the cost and time of synthesis compared to chemical synthesis. Dehydroresidues can also have several valuable properties. For instance, they may play a role in inhibition of biological processes (131, 132) or they can function as attachment sites for further chemical modifications. Therefore, the introduction of lanthionines as well as the introduction of dehydroresidues in nonlantibiotic peptides, exploiting the lantibiotic enzymes NisBC and LtnM2, have a huge potential. The feasibility of engineering these residues in a broad range of peptides will be outlined in this thesis. Chapter 2 focuses on the applicability of the NisT transporter for export of nonlantibiotic peptides and the dissection of the NisBTC enzyme complex. By mass spectrometry, this chapter shows that NisT transports dehydrated NisA prepeptides in the presence of NisB and the absence of NisC. In the absence of NisB and NisC, the unmodified prepeptide NisA is transported.These findings demonstrate that NisT can function independently and that NisB can function without the presence of NisC. Furthermore, it is proven for the first time that NisB can dehydrate and NisT can transport peptides unrelated to nisin when preceded by the nisin leader, like variants of angiotensin, vasopressin and enkephalin. Progress in further exploiting the NisBTC enzymes for posttranslational modification of therapeutic peptides is presented in Chapter 3. The development of the two-plasmid system has a huge beneficial impact on the production level of modified peptides and makes analysis of these new peptide variants more straightforward. This chapter demonstrates that NisB has a wide substrate specificity. Furthermore, it demonstrates for the first time that NisC can cyclize nonlantibiotic peptides and that NisT can transport these novel thioether-bridged peptide variants. In conclusion, the NisBTC enzyme complex can successfully be used for the synthesis of stabilized potential therapeutic peptides. Chapter 4 reports the dissection and utilization of the LtnTM2 part of the lacticin 3147 sythetase complex for modification and transport of nonnatural substrate peptides. Class II lantibiotic synthetase systems may be essential tools for the production of more globular peptides with a higher degree of cyclization. Although the LtnTM2 enzymes appear successful in modification and transport of nonnatural substrate peptides, it is not clear whether the substrate specificity of LtnT and LtnM were as broad as, respectively, NisT and NisBC. Analysis is hampered by lack of

19 Chapter 1 secretion of a number of peptides. Whether this is caused by improper processing by LtnT or by blocked LtnT-dependent translocation is still unknown. In Chapters 5 and 6, the well known Sec pathway is studied as a possible alternative secretion route for posttranslationally modified peptides. First, it is demonstrated that even though the nisin leader is preceded by a signal sequence up to 44 amino acids, the NisA peptide is still modified by NisB and NisC. Thioether bridged pronisin is too large for translocation via the Sec pathway, but the Sec pathway successfully translocates dehydrated peptides and the thioether bridged peptide fragment of azurin. These data reveal once more that the nisin synthetase complex can completely be dissected and that the enzymes NisB and NisT can function independently. Taken together the Sec pathway might be a successful alternative for the secretion of modified peptides. The impact of this thesis is well illustrated in Chapter 7, which describes the development and therapeutic potential of thioether-bridged angiotensin-(1-7). Further perspectives and results are discussed and summarized in Chapter 8.

20 NisT, the transporter of the lantibiotic nisin, can transport fully modified, dehydrated and unmodified prenisin and fusions of the leader peptide with non-lantibiotic peptides

Abstract Lantibiotics are lanthionine-containing peptide antibiotics. Nisin, encoded by nisA , is a pentacyclic lantibiotic produced by some Lactococcus lactis strains. Its thioether rings are posttranslationally introduced by a membrane-bound enzyme complex. This complex is composed of three enzymes: NisB which dehydrates serines and threonines, NisC which couples these dehydrated residues to cysteines thus forming thioether rings and the transporter NisT. We followed the activity of various combinations of the nisin enzymes by measuring export of secreted peptides using antibodies against the leader peptide and mass spectroscopy for detection. L. lactis expressing the nisABTC genes efficiently produced fully posttranslationally modified prenisin. Strikingly, L. lactis expressing the nisBT genes could produce dehydrated prenisin without thioether rings and a dehydrated form of a non-lantibiotic peptide. In the absence of the biosynthetic NisBC enzymes, the NisT transporter was capable of excreting unmodified prenisin and fusions of the leader peptide with non-lantibiotic peptides. Our data show that NisT specifies a broad spectrum (poly)peptide transporter that can function either in conjunction with or independently from the biosynthetic genes. NisT secretes both unmodified- and partially or fully posttranslationally modified forms of prenisin and non-lantibiotic peptides. These results open the way for efficient production of a wide range of peptides with increased stability or novel bioactivities.

J Biol Chem. 2004 ; 279 :22176-82 Chapter 2

Introduction A wide spectrum of biological functions, such as hormone, growth factor, enzyme inhibitor, antigen, antibiotic and ionophore can be found among peptides. Cyclization of peptides has been shown to be a valuable method to obtain biostable analogs. Furthermore, by conformational constraints enhanced or modulated receptor interaction can be obtained (108). Thioether rings can contribute to enhanced peptide stability, enhanced resistance against proteolytic degradation (8, 138, 212) and modulation of receptor interaction (138). Lantibiotics are bacterial peptides with intramolecular thioether bridges (9). They owe their name to their antibiotic activities and the presence of lanthionine residues. Lanthionines are thioether containing amino acids. A variety of activities has been demonstrated for lantibiotics e.g. autoinduction of lantibiotic synthesis (91), permeabilization of target membranes (42, 127, 128, 129, 171), inhibition of cell wall synthesis (222), lipid II binding (13), inhibition of phospholipase A 2 (115), modulation of autolytic enzymes (7) and of an angiotensin-converting enzyme (74). These activities all depend on the presence of thioether rings. By controlling the lanthionine- synthesizing enzyme complex, one might envisage the possibility to introduce thioether rings at any peptide position. However, at present only one new thioether ring has been synthesized in a lantibiotic (8). Most interestingly, in vitro activity of the lanthionine synthesizing enzyme LctM has recently been demonstrated (226). The best known lantibiotic is nisin, which is produced by some Lactococcus lactis strains. Nisin is widely applied as a food preservative (35) because of its antimicrobial activity. It displays a variety of antibiotic activities against many Gram- positive bacteria (171, 215). Breukink et al. (13) found that nisin interacts with lipid II of the target cell, and already at very low concentrations this complex permeabilizes the membranes for small ions and solutes. Nisin is composed of four methyllanthionines, one lanthionine, two dehydroalanines, one dehydrobutyrine and twenty six unmodified amino acids (52, 90). At position 33 mostly a dehydroalanine is present but in some cases an unmodified serine can be found (72). The above-mentioned uncommon residues are posttranslationally produced by intracellular membrane-associated enzyme complexes (75, 189). The enzyme NisB dehydrates serines and threonines; NisC is responsible for thioether bridge formation by coupling the dehydroresidues to cysteines, and NisT exports fully modified prenisin. The extracellular serine protease NisP cleaves off the N-terminal 22-23 amino acid leader peptide (with or without initiating methionine), whereupon the active nisin is released (Fig. 1).

22 Dissection of the nisin modification machinery

Figure. 1. Posttranslational modifications of prenisin . 1: Serines and threonines of unmodified prenisin are dehydrated by NisB. 2: The resulting dehydrated prenisin contains dehydroalanines (Dha) and dehydrobutyrines (Dhb). 3: NisC forms thioether rings between dehydrated residues and cysteines resulting in fully modified prenisin. 4: NisT transports the fully modified prenisin and NisP cleaves off the leader peptide, liberating nisin.

23 Chapter 2

Here we show that the nisin-precursor transporter NisT, in different combinations with NisB and NisC, is able to transport a wide variety of modified and unmodified peptides. This now opens the way to the biotechnological production of modified peptides with novel bioactivities and improved stability. Moreover, we show that the processing enzyme, NisP, requires lanthionine formation of the propeptide for proper functioning.

Materials and Methods Leader peptide . Synthetic leader peptide without initiating methionine was purchased from Synpep, Dublin, CA, US. Bacterially produced leader peptide was purified by binding to teflon beads, a C18 column and elution with an acetonitrile gradient. Cleavage of leader peptide from fully modified prenisin resulted from a 15 min incubation at 37 °C with 1 mg/mL trypsin. Anti-leader peptide antibodies. Polyclonal antibodies were raised in rabbits against the peptide H 2N-STKDFNLDLVSVSKKDC-CONH 2 coupled via the cysteine to keyhole limpet haemocyanin. Samples for Western blotting were prepared as follows. Ten mL (20 mL for samples from cells with pLP1vp and pLP1ang) of bacterial culture supernatant was precipitated with 10% trichloroacetic acid (TCA), kept on ice for 2 h, pelleted by centrifugation at 18514 g during 30 min at 4 ºC, washed with 10% TCA and with acetone and vacuum dried. Pellets were dissolved in 50 µL (for TP9703, NZ9700 and PA1001 containing pBMDL5) or 20 µL sample buffer and applied on a gel. Bacterial strains and plasmids . Strains and plasmids are listed in Table 1. L. lactis strains PA1001 and TP9703 were prepared from NZ9000 and NZ9700 respectively using the pOri gene replacement system (98). In TP9703, the nisP start codon of NZ9700 was replaced by a Not I restriction site. Molecular cloning . Nisin gene(s) (combinations) were amplified from chromosomal DNA of L. lactis NZ9700. DNA amplification was carried out using Expand High Fidelity Polymerase (Roche, Mannheim, Germany) or Pfu polymerase (Invitrogen, CA) in case of insertion or deletion via round-PCR. Plasmid DNA was isolated using the Roche kit. DNA was restricted using restriction enzymes from New England BioLabs Inc. Ligation was carried out with T4 DNA ligase (Roche). DNA fragments were isolated from agarose gel using the Zymoclean gel DNA recovery kit (Orange, CA) or from a PCR mix by using the Roche PCR purification kit. For intermediate cloning steps pGEM-T (Promega) was used. Transformation of Escherichia coli (E. coli DB3.1 (ccdB mutant); E. coli DH5alpha and E. coli TOP10, all obtained from Invitrogen) was carried out using established procedures (170). Electrotransformation of L. lactis was carried out as previously described (61) using a Bio-Rad gene pulser (Biorad, Richmond, CA). Nucleotide sequence analyses were performed by BaseClear (Leiden, NL).

24 Dissection of the nisin modification machinery

Table 1. Lactococcus lactis strains and plasmids strains and plasmids characteristics Reference NZ9700 nisABTCIPREFG, nisin producer (93) NZ9800 Derived from NZ9700 by disruption of nisA (90) NZ9743 Derived from NZ9700 by disruption of nisT (154) NZ9000 nisRK (93) PA1001 ΔacmA (18), ΔhtrA (152), derived from NZ9000 This study TP9703 ΔnisP , derived from NZ9700 This study

pNZ8048 derived (93) plasmids pNGnisT nisT , Em-r or Cm-r This study pNGnisTP nisTP , Em-r or Cm-r This study pNGnisP nisP , Cm-r This study pNZnisA-E3 nisA , Em-r This study pLP1 nisin’s leader, Cm-r This study pLP1ang an angiotensin 1-7 variant (NRSYICP) behind the nisin leader, Cm-r This study pLP1vp a vasopressin variant (SYFQNCPRG) behind the nisin leader, Cm-r This study pNG-enkT an enkephalin variant (YTGFC) behind nisA, Em-r This study pBMDL5 nisABTC in a gateway plasmid. An inverted repeat is present This study between nisA and nisB as on the chromosome of NZ9700, Em-r pBMDL8b nisABT (S-6P, P-2L nisA ) in a gateway plasmid. An inverted repeat This study is present between nisA and nisB as on the chromosome of NZ9700, Em-r pNGnisBT nisBT , Em-r or Cm-r This study Em-r: erythromycin resistance gene Cm-r: chloramphenicol resistance gene

Culturing. L. lactis was grown in M17 broth (198) supplemented with 0.5% glucose (MG17) with or without chloramphenicol (5 µg/mL) and/or erythromycin (5 µg/mL). E. coli was cultured in TY medium with or without ampicillin (100 µg/mL) or erythromycin (100 µg/mL). Preceding mass spectrometry, cells were cultured and samples were prepared as follows. Overnight cultures of L. lactis grown in MG17 broth were diluted 1/100. At optical density at 660 nm of 0.4, cells were centrifuged and the medium was replaced by minimal medium (66) with or without 1/1000 volume of filtered (0.4 µm) overnight L. lactis NZ9700 culture medium containing nisin. Incubation was continued for 4 h or overnight after which mass spectrometry samples were prepared. In the case of cells containing pNGnisBT, 50 mL medium was subjected to TCA precipitation prior to further analysis. Growth inhibition experiments were performed as described previously (136), but in the absence of Tween. Mass spectrometry. Samples were obtained by ziptip purification (C18 ziptip, Millipore). Ziptips were wetted and equilibrated with 50% acetonitrile followed by demineralized water. Then peptides were bound and washed with demineralized water, eluted with a solution of 0.1% trifluoroacetic acid (TFA) with 30- or 50% acetonitrile, vacuum dried and stored at –20 °C until analysis. The dried ziptip eluent was

25 Chapter 2

resuspended in 50% acetonitrile containing 0.1% (v/v) TFA and 1 µL was applied to the target. Subsequently, 1 µL of matrix (10 mg/mL alpha-cyano-4-hydroxycinnamic acid completely dissolved by mildly heating and vortexing in 50% acetonitrile containing 0.1% (v/v) TFA) was added to the target and allowed to dry. Mass spectra were recorded with a Bruker Biflex III MALDI-time-of-flight mass spectrometer. In order to maintain high sensitivity, an external calibration was applied. Ethanethiol treatment (137) was applied to confirm posttranslational modification. Measurement of leader peptidase specificity. NisP-mediated cleavage of peptides was measured by MALDI-TOF MS. Log phase L. lactis strains were induced during 4 h or overnight. In indicated cases NisP-mediated cleavage was measured after pH-induced ring closure. Ring closure in dehydrated prenisin was achieved by 1 h incubation at pH 8.0, which was followed by readjustment of the pH to 4.3. Treatment of peptide containing supernatant with cells of NZ9000 containing pNGnisTP was performed during 4 h, followed by MALDI-TOF MS analysis. Induction assay . The presence of nisin was also tested using the sensitive GusA assay (91) that monitors the capacity of nisin to induce the nisin promoter.

Results Detection of the nisin leader peptide. We first investigated whether the nisin leader peptide would remain present in the culture medium of a nisin-producing L. lactis NZ9700 strain. Peptides isolated from the supernatant of this strain reacted with anti-leader peptide antibodies (Fig. 2, lane 3). A Coomassie-stained gel of a C18 column elution fraction showed the presence of peptide (Fig. 3A) with a mass of 2350.9 Da (MALDI-TOF MS). An identical peptide (Fig. 3B) was observed when ziptip-treated supernatant from a nisin-producing L. lactis NZ9700 strain, grown overnight in minimal medium (66), was analyzed via mass spectrometry. This mass corresponds to the nisin leader peptide without the initiating methionine. In addition, peaks with masses corresponding to the nisin leader peptide with the initiating methionine (2482.2 Da) and of nisin (3354.2 Da) were detected (Fig. 3B). In control incubations with L. lactis strain NZ9000, which does not produce nisin, no peaks were observed. Synthetic leader peptide of residue 2-23 gave a mass peak identical to the peak we assigned to the cleaved leader peptide without the initiating methionine. These data for the first time demonstrated the presence of intact nisin leader peptide in the culture medium following secretion and processing.

26 Dissection of the nisin modification machinery

Figure 2 . Extracellular peptide detection by anti-leader peptide antibodies . L. lactis cells were induced with nisin (+) or not induced (-), incubated for 4 h (NZ9000 containing pNGnisT and pLP1vp) or overnight, and subjected to TCA precipitation and Western blotting. Lanes 1, 2: strain TP9703; lane 3 strain NZ9700; lanes 4, 5: strain NZ9000 containing pNGnisT and pNGenkT; lanes 6, 7: strain NZ9000 containing pNGnisT and pLP1vp; lanes 8, 9: strain NZ9000 containing pNGnisT and pLP1ang; lanes 10, 11: strain NZ9000 containing pNGnisT and pNZnisA-E3; lane 12: strain NZ9000 containing pBMDL8b; and lanes 13, 14: strain PA1001 containing pBMDL5. Each experiment was repeated at least three times with similar results .

Figure 3. Isolation of nisin leader peptide from culture medium .Fig. 3A: Leader peptide isolated from L. lactis NZ9700 culture medium by binding to teflon beads and a C18 column was subjected to gel electrophoresis and stained with Coomassie. Fig. 3B: Detection of the nisin leader peptide by MALDI-TOF MS. The supernatant of overnight L. lactis NZ9700 grown on minimal medium was ziptip-treated followed by MALDI-TOF MS analysis. Expected average masses (M+H +) are for the nisin leader peptide residues 2-23: 2352.6 Da, nisin leader peptide residues 1-23: 2483.8 Da, nisin: 3355.2 Da. The experiment shown is a typical result that was repeated more than ten times with identical results.

27 Chapter 2

Thioether ring formation by NisBTC. In order to investigate whether NisBTC are sufficient for thioether ring formation we cloned the nisABTC genes. Both uninduced (Fig. 2, lane 13) and induced PA1001 (Fig 2, lane 14) containing pBMDL5 produced prenisin. Apparently, when no inducing nisin was added, some transcription still occurred; this has previously also been reported for wild-type nisin producers (91). Subsequently, we analyzed the prenisin peptides. The uninduced strain PA1001 containing pBMDL5 produced two peptides with masses corresponding to dehydrated prenisin or fully modified prenisin with and without the initiating methionine (5817.4 and 5686.5 Da) and a third peptide corresponding to unmodified prenisin (5833.9 Da). Similar peptide masses were observed after induction of PA1001 containing pBMDL5 (Table 2), but the mass peak of dehydrated or fully modified prenisin (5688.9 Da) was more pronounced.

Table 2 . Modification and export of peptides in L. lactis . Cells were induced and grown 4 h (NZ9000 containing pNGnisT and pNGang, NZ9000 containing pNGnisT and pLP1vp), or overnight in minimal medium followed by TCA precipitation and/or direct ziptip treatment of the supernatant and MALDI-TOF MS. Experiments were repeated at least three times with similar result. Theoretical values are average masses in Da (M+H +).

L. lactis plasmids Peptide Observed Theoretical mass Theoretical mass strain mass (Da) with Methionine1 w/o Methionine1 PA1001 pBMDL5 Fully modified prenisin 5688.9 5820.0 5688.8 (nisABTC ) 5820.3

Unmodified prenisin 5836.3 5964.0 5832.8 NZ9000 pBMDL8b Fully dehydrated S-6P, 5711.2 5846.1 5714.9 (nisABT ) P-2L prenisin

Unmodified prenisin 5857.0 5990.1 5858.9 NZ9000 pNGnisBT Leader peptide fused to 3186.5 3317.8 3186.6 pLP1ang angiotensin: 3317.9

Dehydrated 3169.2 3299.8 3168.6 3302.9 Dehydrated, ethanethiol 3235.6 3362.8 3231.6 addition

Unmodified, (ethanethiol) 3188.6 3317.8 3186.6 NZ9000 pNGnisT Unmodified prenisin 5833.2 5964.0 5832.8 pNZnisA-E3 NZ9000 pNGnisT Unmodified prenisin C- 6403.2 6535.6 6404.4 pNGenkT terminally fused to enkephalin NZ9000 pNGnisT Leader peptide fused to 3405.2 3537.0 3405.8 pLP1vp vasopressin NZ9000 pNGnisT Leader peptide fused to 3186.6 3317.8 3186.6 pLP1ang angiotensin

28 Dissection of the nisin modification machinery

The trypsin-treated supernatant of induced PA1001 containing pBMDL5 showed a growth inhibiting activity comparable to that of NZ9700, indicating the presence of fully modified prenisin. Uninduced cells showed a much lower activity. In addition, trypsin-treated culture medium of uninduced and nisin-induced PA1001 cells containing pBMDL5 gained in induction capacity as measured with the GusA assay (91). The production of fully modified prenisin by PA1001 containing pBMDL5 was confirmed by overlaying agar plate cells with NZ9000 (Fig. 4A) and with NZ9000 containing pNGnisTP (Fig. 4B). As expected, NZ9700 (position 2) produced clear halos with both overlays, whereas PA1001 (position 1) did not form a halo. By contrast, PA1001 containing pBMDL5 (position 3) and TP9703 (position 4) only produced halos when overlayed with NZ9000 containing pNGnisTP (Fig. 4B). Consistent with the lower activity of trypsin-treated TP9703 supernatant, the halos produced by TP9703 were much smaller than of PA1001 cells containing pBMDL5. These data demonstrate that both induced- and -to a lesser extent- uninduced PA1001 cells containing pBMDL5 produce fully modified prenisin, thus showing that the NisBTC enzymes are sufficient for the thioether ring formation.

Figure 4 . Leader peptidase activity in NisP producing cells . PA1001 (position 1), NZ9700 (2), PA1001 containing pBMDL5 (3) and TP9703 (4) cells were grown overnight on agar plates that contained no antibiotic next to inducing amounts of nisin. Subsequently, the cells were overlayed with log phase-grown sensitive strain NZ9000 (A) and NZ9000 containing pNGnisTP (B) cells and further grown for one more night. The size of the halo indicates the presence of active nisin processed by NisP. The experiment was repeated three times in duplicate with similar result.

Export of dehydrated prenisin via NisBT. Next we investigated the functionality of NisBT in the absence of NisC. Strikingly, NZ9000 cells containing pBMDL8B produced prenisin (Fig. 2, lane 12). Mass spectrometry analysis of the supernatant of these cells with plasmid-encoded NisABT, (the nisA gene of this construct having two leader peptide mutations: S –6 P and P –2 L) demonstrated the production of dehydrated prenisin (5711.2 Da) and unmodified prenisin (5857.0 Da) (Table 2). This

29 Chapter 2

shows that NisBT can act independently of NisC and also demonstrates that the NisT transport activity is not strictly coupled to full posttranslational modification of prenisin. Modification and transport of a non-lantibiotic peptide. In order to investigate whether fusions of the leader with a non-lantibiotic peptide could be modified by NisB and transported by NisT, peptide production by NZ9000 containing pNGnisBT and pLP1ang, encoding a fusion of the leader peptide with NRSYICP, was investigated. Both unmodified (3317.9 and 3186.5 Da) and dehydrated fusions of leader peptide with angiotensin 1-7 (3302.9 and 3169.2 Da) with and without methionine1, were observed (Table 2). In order to confirm the observed dehydration, prior to analysis the peptide samples were treated with ethanethiol, which reacts with dehydroresidues. Indeed, after ethanethiol treatment, the peptide with the mass of dehydrated leader- angiotensin 1-7 had disappeared whereas a peptide corresponding to ethanethiol- modified dehydrated peptide appeared (3235.6 Da). As expected, ethanethiol treatment did not alter the mass of the non-dehydrated peptide (3188.6 Da). These data clearly prove that the fusion of leader peptide with angiotensin 1-7 was dehydrated and transported by NisBT. NisT has a broad substrate specificity. We subsequently determined whether NisT in the absence of NisBC, is capable of transporting various unmodified fusion peptides containing the nisin leader peptide. Experiments were performed with NZ9000 cells containing two plasmids, one coding for NisT and the second for a leader peptide fusion. Export of unmodified prenisin (Fig. 2, lane 11), unmodified prenisin with a C-terminally fused enkephalin peptide (Fig. 2, lane 5), a fusion of the leader peptide with a vasopressin variant (Fig. 2, lane 7) and a fusion of the leader peptide with an angiotensin variant (Fig. 2, lane 9) was measured using anti-leader peptide antibodies. Mass spectra clearly demonstrated export of unmodified prenisin without the initiating methionine (5833.2 Da), of a fusion peptide of unmodified prenisin with a C-terminal enkephalin variant (6403.2 Da), of a fusion of nisin leader peptide with vasopressin (3405.2 Da) and of a fusion peptide of the nisin leader peptide with angiotensin 1-7 (3186.6 Da), (Table 2). Control experiments without pNGnisT and with strain NZ9743 (154) containing disrupted nisT showed no detectable levels of secreted peptide in the culture medium. Furthermore, in the cell fraction of induced NZ9743 cells an antibody-reactive peptide was detected (data not shown). Taken together the data clearly demonstrate that NisT can act independently of the other lantibiotic enzymes and further show that the substrate specificity of NisT is much wider than only the fully modified prenisin. NisP is specific for thioether ring containing prenisin. We measured which leader peptide-containing peptides could be cleaved by the leader peptidase. The leader peptide could neither be cleaved from the leader peptide-angiotensin fusion nor from the unmodified or dehydrated prenisin (Table 3). However, after keeping the

30 Dissection of the nisin modification machinery

dehydrated prenisin 1 h at pH 8.0, the leader peptide could be cleaved off. At pH 8.0 thioether rings can be closed spontaneously (19, 139) after which the peptide apparently had become substrate for the leader peptidase. The pH 8.0- and NisP- treated dehydrated prenisin had however no antimicrobial activity (data not shown), which indicates that no fully modified prenisin had been formed. Fully modified prenisin itself was also cleaved by the leader peptidase, and a control experiment showed that, in the absence of leader peptidase, the pH 8.0 treatment alone did not liberate the leader peptide (Table 3).

Table 3. The leader peptidase is specific for thioether ring containing prenisin. Cleavage of peptides by NisP was measured by MALDI-TOF MS. NisP was expressed by the peptide producing cells or cells expressing NisP were added or –in control experiments- no NisP was present. By incubating one hour at pH 8.0 thioether ring closure in dehydrated prenisin was induced. Theoretical values are average masses in Da (M+H +). L. lactis Plasmid Post- Peptide Observed Theoretical Theoretical strain treatment mass (Da) mass after mass w/o (M+H +) cleavage cleavage NZ9000 pLP1vp + - leader peptide- - 2352.6 3405.8 pNGnisTP vasopressin 1072.2

NZ9000 pLP1ang + - leader peptide- 3187.1 2352.6 3186.6 pNGnisTP angiotensin 853.0

NZ9000 pNZnisA-E3 - unmodified 5836.3 2352.6 5832.8 + pNGnisTP prenisin 3499.2 NZ9000 pBMDL8b NZ9000/pNG dehydrated S-6P, 5714.9 2378.7 5714.9 (nisABT ) nisTP cells P-2L prenisin 3355.2

NZ9000 pBMDL8b pH 8.0 and dehydrated S-6P, 2377.1 2378.7 5714.9 (nisABT ) NZ9000/pNG P-2L prenisin with 3349.8 3355.2 nisTP cells 1 or more thioether rings

PA1001 pBMDL5 pH 8.0 fully modified 5686.3 2352.6 5688.8 (nisABTC ) prenisin 3355.2

PA1001 pBMDL5 NZ9000/pNG leader peptide and 2351.9 2352.6 5688.8 (nisABTC ) nisTP cells fully modified nisin 3353.9 3355.2

In order to discriminate between NisT-dependent- and -independent NisP activity, a plasmid containing the nisP gene was constructed. An antimicrobial activity assay involving fully modified prenisin and mass spectrometry analysis confirmed that the NZ9000 cells containing pNGnisP expressed active NisP as they were able to cleave the leader peptide from externally added fully modified prenisin (data not shown). Hence, NisP can act independently of other lantibiotic enzymes.

31 Chapter 2

Discussion The leader peptide of nisin may fulfill several functions. First, prior to export it may have a role in the posttranslational modification and recognition events (183, 226). Secondly, it is needed for recognition by the transport system, and third, it keeps the lantibiotic in an inactive state until maturation has taken place (212). Here, we demonstrated that the nisin leader peptide accumulates in the bacterial culture medium of nisin producing L. lactis cells. The subtilin leader peptide has been shown to act as a translocation signal in B. subtilis (65) and in E. coli (148). Export of alkaline phosphatase of E. coli when fused with the subtilin leader peptide seemed to be enhanced in the presence of a transporter that is encoded within the subtilin operon (65). When the leader peptidase of subtilin is inhibited by phenylmethyl-sulfonyl fluoride, accumulation of fully modified presubtilin, subtilin and a series of degradation products in the medium has been observed (193). Mutagenesis studies of the leader peptide of various lantibiotics revealed that some of the leader peptide residues are essential for export and possibly for interaction with the modifying enzymes (25, 135, 213). In contrast to the present work, it has also been suggested that fully modified prenisin is the only form of nisin recognized by the transporter (83). A membrane-associated enzyme complex of NisBTC has been reported to be responsible for dehydration of the serine and threonine residues of prenisin, the thioether ring formation by cross-linking of dehydroresidues to cysteines, and the final export step (189). In nisin, Ser29 is not dehydrated whereas Ser33 sometimes escapes dehydration. Overexpression of NisB results in a more frequent dehydration of Ser33 (72). Here we report that the genes nisABTC are sufficient for production and export of fully modified prenisin. Strikingly, we also demonstrate that NisB and NisT suffice to export dehydroresidue-containing peptides. This implies that ring formation is not a prerequisite for export. Dehydrated prenisin is, however, not a substrate for NisP which indicates that further modification is needed before NisP recognizes the prenisin as substrate. Incubation of the dehydrated form of prenisin at pH 8.0 results in spontaneous ring closure (19, 139). Under those conditions, the peptide becomes a substrate of NisP and the leader peptide can be cleaved off. L. lactis NZ9000 with NisABT but without NisC indeed produced the dehydrated prenisin with the dehydroalanines and dehydrobutyrines present and without the ring closure. The serine and threonine residues in the leader peptide are never dehydrated as confirmed by mass spectrometry (Fig. 3B). NisBT-expressing cells produced the dehydrated prenisin with eight dehydrated serine and threonine residues; they also produced a fusion of leader peptide with dehydrated angiotensin 1-7 (Table 2). This implies that bacteria that only contain LanBT (homologs of NisBT) or LanT and the equivalent LanM part (the N-terminal dehydration domain of LanM) can produce peptides with dehydroresidues. Therefore, two more amino acids are in principle available as building blocks for the

32 Dissection of the nisin modification machinery

synthesis of novel (poly)peptides with desired properties. Dehydroresidues have been reported to be essential for the activity of some bio-active (poly)peptides (111, 132, 186, 188, 203, 219). Although the mechanism of NisB action is not known, statistical studies on the variability of the flanking regions of the eight dehydratable serine and threonine residues in nisin and related lantibiotics suggest that NisB, in contrast to the specific NisP, is equipped with a broad substrate specificity. Engineering of dehydroresidues in nisin and Pep5 has been demonstrated (8, 88). Therefore, a wide variety of peptides with dehydroresidues might be produced and exported via NisBT. In this context it is interesting to note that prePep5 fragments with dehydrated residues of Pep5 are exported (125) when the pepC gene is largely deleted. However, those studies all concern original lantibiotics and it would be of interest to introduce such residues in peptides that are normally not modified. Lantibiotic transporters are generally considered to export only specific lantibiotics synthesized by the gene products encoded in the same operon structure. Also some nisin-subtilin chimeras are exported (20) while nisin Z export can be directed by the subtilin leader peptide (90). Those studies, however, all pointed at a rather narrow substrate specificity. Here we demonstrate that the specificity of the NisT transporter is much wider than originally anticipated. Various unrelated peptides, in either a modified form (in the presence of NisB) or an unmodified form (in the absence of NisB) can be secreted provided that they are fused to the leader peptide. Previous attempts to demonstrate functionally active NisP upon overproduction in L. lactis have not been conclusive (144). Here we show that, in the absence of other lantibiotic enzymes, NisP can be functionally expressed in L. lactis . The enzyme shows a clear leader peptidase activity on fully modified prenisin. These data furthermore demonstrate that the NisP activity is not coupled to the transport step by NisT. This agrees with observations that extracellularly added fully modified prenisin is processed by L. lactis NZ9800 (144). Remarkably, neither the dehydrated prenisin nor the leader peptide fusions were cleaved by NisP. This strongly suggests that the leader peptidase is specific for thioether ring-containing prenisin. In this respect, the dehydrated prenisin became a substrate for NisP after pH-induced ring closure which is very suggestive of region- and stereospecific closure of one or more rings. Using model peptides non-enzymatic, stereospecific ring closure has been shown for ring B (19, 139) and ring E of nisin, whereas region specificity and a three to one stereo preference was shown for ring A of subtilin (19). These data indicate that production of dehydroresidue-containing peptides may be followed by extracellular specific ring closure, e.g. at pH 8.0. Thioether rings are essential for most lantibiotic activities. Opening of ring A or C (127) or replacement of a thioether ring by a disulfide bridge and reducing it (216) causes a severe loss of activity. In addition, the rings can protect (poly)peptides against proteolytic degradation (8, 212) and their presence may modulate the activity of

33 Chapter 2

peptides (138). Active lanthionine analogs of somatostatin and enkephalin have been synthesized chemically (138, 155), but this involved elaborate methods that more easily could be performed by a fermentative route. The transport by NisT of medically relevant therapeutic peptides like enkephalin, vasopressin and dehydrated angiotensin (variants) compel further research on modifying such peptides by the lantibiotic enzymes. Summarizing, we have shown that NisBT is sufficient to dehydrate and export the dehydrated non-lantibiotic angiotensin 1-7, the dehydrated prenisin, and that the NisT transporter is equipped with a wide substrate specificity, transporting various peptides provided they are fused to the nisin leader peptide. Production of peptides via NisBT provides an adequate system to study the substrate specificity of NisB, and may enable the synthesis and export of a wide variety of peptides with dehydroresidues. This process can then be followed by extracellular stereospecific ring closure to avoid possible export incompatibilities of bulky thioether ring containing peptides.

Acknowledgements : Karin Scholtmeijer, Hans Hektor and George T. Robillard are gratefully acknowledged for help and helpful discussions.

Footnotes. The abbreviations used are: MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight; MS, mass spectrometry.

34

Post-translational modification of the therapeutic peptides by NisB, the dehydratase of the lantibiotic nisin

Abstract Post-translationally introduced dehydroamino acids often play an important role in the activity and receptor specificity of biologically active peptides. In addition, a dehydroamino acid can be coupled to a cysteine to yield a cyclized peptide with increased biostability and resistance against proteolytic degradation and/or modified specificity. The lantibiotic nisin is an antimicrobial peptide produced by Lactococcus lactis . Its post-translational enzymatic modification involves NisB-mediated dehydration of serines and threonines, and NisC-catalyzed coupling of cysteines to dehydroresidues, followed by NisT-mediated secretion. Here, we demonstrate that a L. lactis strain containing the nisBTC genes effectively dehydrates and secretes a wide range of medically relevant non-lantibiotic peptides among which variants of adrenocorticotropic hormone, vasopressin, an inhibitor of tripeptidyl peptidase II, enkephalin, luteinizing hormone-releasing hormone, angiotensin and erythropoietin. For most of these peptides ring formation was demonstrated. These data show that lantibiotic enzymes can be applied for the modification of peptides, thereby enabling the biotechnological production of dehydroresidue-containing and/or thioether-bridged therapeutic peptides with enhanced stability and/or modulated activities.

Biochemistry. 2005 ; 44: 12827-34 . Chapter 3

Introduction The presence of unusual dehydroamino acids in peptides can have a large effect on the biological activity. For instance, a synthesized pentapeptide containing a dehydroalanine acts as a potent inhibitor of a spider venom peptide epimerase (132). Dehydroalanines at position 5 in nisin and subtilin are responsible for the inhibition of the outgrowth of bacterial spores by reacting with sulfhydryl groups of membrane components (131). Besides their role in inhibitors, dehydroalanines can increase the efficiency by which (poly)peptides transmit signals by interacting with receptor or acceptor molecules; the synthesis of a dehydroalanine-containing neurokinin A receptor antagonist resulted in a more rigid and potent peptide (111). Dehydroamino acids are also important for the activity of - amongst others - thiostrepton, nosiheptide and berninamycin (144). Dehydroresidues can furthermore be versatile starting points for the synthesis of unnatural amino acids, attachment sites for further modification or serve as sites for peptide cyclization. Intramolecular coupling of a dehydroresidue to a cysteine has proven to be a valuable method to obtain biostable analogs with resistance against proteolytic degradation (8, 212) or modulated receptor interaction (73, 108, 138). A well-known group of cyclized peptides is formed by the lantibiotics: antimicrobial peptides that contain the thioether amino acids lanthionine and/or methyllanthionine (9). A variety of lantibiotic activities is known, which all depend on the presence of thioether rings (21, 120, 172, 227). The best studied lantibiotic is nisin, a widely applied food preservative that is produced by a number of Lactococcus lactis strains (13). Nisin biosynthesis involves the activity of four enzymes (90, 212). NisB dehydrates serines and threonines in the nisin propeptide, after which the formed dehydroresidues are stereo- and regiospecifically coupled to cysteines by NisC. The ABC transporter NisT then exports the fully modified prenisin, whereupon the extracellular peptidase NisP cleaves off the leader peptide, to liberate active nisin that contains four methyllanthionines, one lanthionine, two dehydroalanines and one dehydrobutyrine (52). Chemical synthesis of dehydroamino acids in peptides (26, 139, 182, 194) is costly, and chemical cyclization methods are cumbersome and primarily lack regio- and stereoselectivity. Chemical cyclization methods may also result in oligomerization (81). On the other hand, fermentative methods in which peptides are enzymatically modified would allow for a controlled formation of the desired product. In the past, only a few dehydroresidues in lantibiotics have been engineered (8, 13, 88) and only one new thioether ring in an existing lantibiotic has been generated (8). Here, we demonstrate that dehydroamino acids can be introduced in a broad range of therapeutic peptides by exploiting the bacterial serine/threonine dehydratase, NisB.

36 NisB dehydrates therapeutic peptides

Materials and methods Bacterial strains and plasmids . The strains and plasmids used in this study are listed in Table 1. Peptides were encoded on pNZ8048-derived plasmids. Peptides were co-expressed with pIL253-based plasmids containing the nisBTC genes, except in combination with the peptide ACTH, for which the nisBTC genes were located on a pNZ8048-derived plasmid. Angiotensin(1-7) was encoded on the pILNisBTC plasmid. Molecular cloning. Nisin genes or combinations were amplified from chromosomal DNA of L. lactis NZ9700, using Expand High Fidelity Polymerase (Roche, Mannheim, Germany) or Pfu polymerase (Invitrogen, CA). Plasmid DNA was isolated using the QIAGEN purification kit (Qiagen). DNA was restricted using restriction enzymes from New England BioLabs Inc. Ligation was carried out with T4 DNA ligase (Roche). DNA fragments were isolated from agarose gel using the Zymoclean gel DNA recovery kit (Orange, CA) or from a PCR mix using the Roche PCR purification kit. Electrotransformation of L. lactis was carried out as previously described (61) using a Bio-Rad gene pulser (Biorad, Richmond, CA). Nucleotide sequence analyses were performed by BaseClear (Leiden, NL). Growth conditions . L. lactis was grown at 30°C in MG17 broth, as described previously (25). Sample preparation for mass spectrometry or Western blot analysis was carried out as follows: overnight cultures of L. lactis were transferred to fresh

MG17 medium. At an OD 660 nm of 0.4, cells were pelleted (2,057 x g, 5 min, Eppendorf 5810 R) and medium was replaced by minimal medium, adapted from Jensen and Hammer (66). Induction was carried out by adding 1/1000 volume of filtered (0.45 m) medium from an overnight-grown culture of the nisin-producer L. lactis NZ9700. Incubation was continued overnight. Sample preparation. Samples were purified from the medium fraction by ziptip purification (C18 ziptip, Millipore) (86). Peptides from larger volumes were precipitated with 10% trichloroacetic acid (TCA) and kept on ice for 2 h. The sample was then pelleted by centrifugation at 18,514 x g during 30 min at 4ºC, washed with acetone and vacuum dried. In the case of EPO, the medium fraction was freeze-dried (Labconco), desalted using PD-10 columns (Amersham Biosciences), and again freeze-dried. Dehydration was confirmed by ethanethiol treatment (137). A total of 40 µL of an ethanethiol mixture (80 µL of ethanol, 65 µL 5 M NaOH, 60 µL ethanethiol, and 400 µL MQ) was added to vacuum-dried peptide and incubated at 50°C for 2.5 h. The reaction was stopped by adding 10 µL of acetic acid. CDAP (1-cyano-4- dimethylaminopyridinium tetrafluoroborate) was used to react with free cysteine residues. Vacuum-dried sample was resuspended in 9 µL 25 mM citrate buffer at pH 3.0, and reduced with 1 µL Tris[2-carboxyethyl]phosphine (TCEP). After a 10 min incubation at room temperature, 2 µL of CDAP was added, followed by 15 min of incubation at room temperature. As positive controls two chemically synthesized peptides (CRYTDPKPHIRLRIK and MSTKDFNLDLVSVSKKDS-GASPRITRICK)

37 Chapter 3

were used, both containing one cysteine. For tryptic digestion, TCA-precipitated LHRH was dissolved in 50 mM Tris at pH 6.8 (0.1 mg LHRH/mL), 200 µL of which was incubated with 20 µL of trypsin (0.01 mg/mL) in the same buffer at 37°C for 2 h. Peptide analysis . Mass spectra were recorded with a Bruker Biflex III matrix- assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometer. To maintain high sensitivity, an external calibration was applied. Trypsin-digested LHRH was separated on a reversed-phase high-performance liquid chromatography (HPLC) column (Alltima, C18, 5 µm, Alltech Chromatography). Separation was carried out at 1.0 mL/min, using acetonitrile (ACN) as a solvent in a gradient from 10 to 90%. Peak fractions (detection at 280 nm using a diode-array detector) were collected and analyzed by MALDI-TOF mass spectrometry. N-terminal amino acid sequence analysis of LHRH was carried out by Eurosequence (Groningen, NL) using a Procise 494 sequencing system equipped with a 140C Microgradient System and a 785A Absorbance Detector (Applied Biosystems, Foster City, CA). Samples (1000 pmol) of LHRH were solubilized in 50 % (v/v) acetic acid, and sequencing was performed using procedures and chemicals supplied by the manufacturer. Purified LHRH was modified by thiol addition, peroxidation using trifluoroperacetic acid, and a second thiol addition step (124). Western blot analysis. Polyclonal anti-leader peptide antibodies were raised in rabbits against the peptide H 2N-STKDFNLDLVSVSKKDC-CONH 2 coupled via the cysteine to keyhole limpet haemocyanin. TCA-precipitated peptides from 10 mL cultures were dissolved in 20 µL sample buffer and applied on a Tricine SDS-PAGE gel. Peptides were transferred to PVDF Western blotting membrane (Roche) using a Trans-Blot SD semi-dry transfer cell (Bio-Rad). The membrane was blocked with 2% skim milk (Oxoid) in TBST (10 mM Tris-HCl, pH 8.0/150 mM NaCl/0.05% Tween- 20) and 100 mM EDTA for 18 h at 4°C and washed twice with TBST for 10 min. The membrane was incubated with anti-leader antibody (1:500) in TBST + 0.2% skim milk for 1 h at 25°C, washed three times (10 min each) with 0.2% skim milk in TBST, and incubated with anti-rabbit IgG antibody conjugated to alkaline phosphatase (Sigma) in TBST + 0.2% skim milk for 1 h at 25°C (1:5000 dilution). The membrane was washed twice (10 min each) with TBST, followed by six times washing (10 min each) in alkaline phosphatase buffer (100 mM Tris-HCl, pH 9.5/100 mM NaCl/5 mM MgCl 2). Finally, the membrane was incubated with 10 mL of alkaline phosphatase, containing BCIP/NBT (5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium) until bands appeared.

Results and discussion We have recently shown that the nisin transporter NisT of L. lactis excretes non-lantibiotic peptides when fused to the nisin leader peptide. We here examined a range of medically relevant peptides with respect to their translocation and

38 NisB dehydrates therapeutic peptides

dehydration via NisBTC. For that purpose, the nisin propeptide was partly or entirely replaced by a non-lantibiotic peptide. In addition, serine/threonine and cysteine mutants of bioactive peptides were made, aiming at dehydration or thioether bridge formation. Peptide production and export was examined using L. lactis NZ9000 cells, containing one plasmid carrying the nisBTC genes and another one carrying the code for the fusion peptide (Table 1).

Table 1. Lactococcus lactis strains and plasmids Characteristics Reference Strains NZ9700 nisABTCIPRKEFG (93) NZ9000 nisRK + (93)

Plasmids pIL253-derived (192) pIL2angBTC (157)

pIL3BTC nisBTC , derived from pIL2angBTC by deletion of Em-r gene and This study introducing stop codon at start ang gene pIL4BTC nisBTC , derived from pIL2angBTC by deletion of Cm-r gene and This study introducing stop codon at start ang gene pNZ8048 derived (93)

pNGnisBT nisBT , Em-r or Cm-r (86) pLPacth leader peptide fused to SYSMECFRWG, Cm-r This study pTPvp1 leader peptide fused to ATFQCAPRG, Cm-r This study pTPvp2 leader peptide fused to AYTQNCPRG, Cm-r This study pLPvp3 leader peptide fused to ITSYFQCTPRC, Cm-r This study pTPtppII leader peptide fused to ITSISRASVA, Em-r This study pTPenkT4 leader peptide fused to ITSISKYTGFC, Em-r This study pLPlhrh leader peptide fused to This study ITSISLCTPGCKTGALMGCNKQHWSYGCRPG, Em-r pLPepo leader peptide fused to This study YASHFGPLGWVCK, Cm-r pILang7BTC leader peptide fused to This study DRVTIHC, Cm-r. Adaptation from pIL2angBTC

Peptides, including mutants of ACTH, vasopressin, an inhibitor of tripeptidylpeptidase II, enkephalin, LHRH, angiotensin(1-7) and EPO were isolated from the medium fraction and identified via immunodetection with anti-leader peptide antibodies (Figure 1). Clear expression was observed when peptides were either directly cloned behind the nisin-leader ( lanes 2-4, 10-11 ) or fused with part of the N- terminal amino acids of nisin ( lanes 5, 7-9). In some cases the latter strategy resulted in a significantly higher production level (data not shown). The absence of peptides in

39 Chapter 3

the medium fractions of NZ9000 with only pIL4BTC or pIL3BTC (Figure 1, lanes 1 and 6, respectively) proved that the detected peptides are encoded on the co- transformed fusion peptide constructs. Without nisin induction no peptides were observed (not shown).

Mw (kDa) 1 2 3 4 5 6 7 8 9 10 11

30.2 22.3

13.7 6.8 3.9

Figure 1 . Extracellular production of non-lantibiotic peptides in L. lactis . Samples were obtained by precipitating the medium fraction of an induced, overnight culture. Peptides were detected via Western blot analysis using anti-leader antibodies. L. lactis strain NZ9000 contained either pIL3BTC (lanes 6-9) or pIL4BTC (lanes 1-5, 10), supplemented with the plasmids containing genes encoding the peptides to be modified. Maturation of angiotensin(1-7) was carried out on a one-plasmid system (pILang7BTC). M, marker; lane 1, pIL4BTC; lane 2, pIL4BTC/pLPacth; lane 3, pIL4BTC/pTPvp1; lane 4, pIL4BTC/pTPvp2; lane 5, pIL4BTC/pLPvp3; lane 6, pIL3BTC; lane 7, pIL3BTC/pTPtpp II; lane 8, pIL3BTC/pTPenkT4; lane 9, pIL3BTC/pLPlhrh, lane 10, pIL4BTC/pLPepo; lane 11, pILang7BTC.

To investigate whether the exported fusion peptides were dehydrated, ziptip- treated culture medium (86) was subjected to Maldi-TOF mass spectrometry (Table 2). Measured values consistently refer to fusion peptides with and without the N-terminal Met of the leader peptide. For clarity, Table 2 only shows values belonging to fusion peptides without the initial Met. The corresponding mass spectra are electronically available as Supporting Information (Figures S1-11). The adrenocorticotropic hormone (ACTH) acts upon the adrenal cortex by stimulating the secretion of glucocorticoids, such as cortisol. An ACTH fragment (1- 10), with a H6C mutation (SYSMECFRWG), was fused to the leader peptide. Mass spectrometry analysis of the produced peptide yielded peaks corresponding to one (3581.8 Da) and no (3600.3 Da) dehydration (Table 2), suggesting dehydration of one of the two serine residues. To establish which one was dehydrated the serines were independently changed into glycines by site-directed mutagenesis and the peptides were maturated and exported via NisBTC. Mass spectrometry analysis of the new ACTH variants (S1G and S3G, respectively) showed that dehydration only occurred

40 NisB dehydrates therapeutic peptides

when the serine at position 3 was left intact (Table 2). These data prove that in the original ACTH (SYSMECFRWG) the serine at position 1 is not dehydrated, but the serine at position 3 is.

Table 2. Dehydration of therapeutic peptides and analogs by NisB, analyzed by Maldi-TOF mass spectrometry. The dehydration of serines/threonines was verified by ethanethiol modification. Amino acids part of the potentially active part of the peptide are underlined. Amino acids in bold correspond to mutations. Calculated values are displayed in italics. Peptide* Sequence Dehydration Mass (M + H +) w/o Met1 (Da) extent Dehydration + ethanethiol Observed Calculated Observed Calculated ACTH(1-10) SYSME CFRWG 1 3581.8 3582.1 3643.9 3644.2 0 3600.0 3600.1 ACTH(S1G) GYSME CFRWG 1 3552.3 3552.0 3613.8 3614.1 0 3570.7 3570.0 ACTH(S3G) SY GME CFRWG 1 - 3552.0 - 3614.1 0 3570.6 3570.0 VP1 AT FQ CA PRG 1 3264.8 3266.7 3329.4 3328.8 0 - 3284.7 VP2 AYTQNCPRG 1 3327.6 3325.7 3390.6 3387.8 0 3345.3 3343.7 VP3 IT SYFQ CT PR C 3 3625.3 3625.2 # 3814.0 3813.3 2 3643.0 3643.2 3769.7 3768.3 1 3662.4 3661.2 3727.2 3723.3 0 - 3679.2 TPP II ITSISRASVA 4 3265.5 3266.7 3514.9 3517.8 3 3283.8 3284.7 3470.9 3472.8 2 3302.4 3302.7 3427.3 3427.8 1 3320.1 3320.7 3383.5 3382.8 0 3338.7 3338.7 EnkT4 ITSISKYTGF C 4 3482.0 3482.0 3734.6 3733.1 3 3500.4 3500.0 3690.1 3688.1 2 3518.5 3518.0 3646.0 3643.1 1 3536.5 3536.0 3601.9 3598.1 0 3554.1 3554.0 Enk.A IAAIARYTGF C 1 3500.3 3502.0 3564.1 3564.0 0 3518.2 3520.0 Enk.B IAAIARYSGF C 1 3488.9 3488.0 3550.1 3549.8 0 3506.8 3506.0 LHRH ITSISLCTPGCKTGA 6 5499.8 5497.4 5871.0 5874.5 LMGCNK 5 - 5515.4 QHWSYG CRPG Ang(1-7) DRV TIH C 1 3157.9 3159.6 3219.5 3221.7 0 3176.0 3177.6 EPO(1-13) YAS HFGPL GWVCK 1 3784.1 3781.3 3843.9 3843.4 0 3799.3 * VP, vasopressin; ACTH, adrenocorticotropic hormone; TPP II, tripeptidylpeptidase II; Enk, enkephalin; LHRH, luteinizing hormone-releasing hormone; Ang, angiotensin; EPO, erythropoietin. # Leader peptide contains a S10L mutation .

41 Chapter 3

Vasopressin is an antidiuretic hormone with the sequence CYFQNCPRG-NH 2 and has a disulfide bond that bridges the two cysteines. We genetically made three fusion peptides of the nisin leader with vasopressin variants, termed VP1 (leader- ATFQCAPRG), VP2 (leader-AYTQNCPRG) and VP3 (leader-ITSYFQCTPRC). For both VP1 and VP2 mass spectrometry showed peaks corresponding to a single dehydration in each peptide. Since both peptides contain one threonine and no serines, the threonines at position two and three, respectively, were dehydrated by NisB. VP3 contains the first two residues of nisin, IT, between the nisin leader and the vasopressin variant. Figure 2a depicts a Maldi-TOF mass spectrum of VP3 after dehydration. The spectra show peak patterns corresponding to three, two and one-time dehydrated VP3, with and without the N-terminal methionine attached. The phosphopentapeptide RAS(P)VA inhibits tripeptidylpeptidase II (TPP II), after processing of the octapeptide precursor VALRAS(P)VA. Chemical replacement of the serine by a dehydroalanine (RA Dha VA) reduced the Ki of this already potent inhibitor 45 times (203). TPP II degrades amongst others cholecystokinin, responsible for satiety feeling, and inhibition of this peptidase may therefore contribute to battling obesity. The peptide termed TPP II (leader peptide fused to ITSISRASVA) was maturated and exported via NisBTC. Mass spectrometry of the secreted peptide fused to the leader peptide resulted in peaks corresponding to one to four times dehydration of the peptide. Leu-enkephalin (YGGFL) and Met-enkephalin (YGGFM) have a morphine-like effect. The potential impact of cyclization by insertion of a thioether bridge in peptides is well demonstrated by the observation that thioether-enkephalin is resistant to enzymatic degradation, whereas non-cyclized Leu-enkephalin had a t1/2 of less than 7 min (195). Furthermore, thioether-enkephalin is in vivo 10,000 times more potent than morphine and 1,000,000 times more potent than enkephalin cyclized via a disulfide bridge (155). A fusion of a prenisin fragment and a peptide containing a pentapeptide enkephalin variant (YTGFC) was termed EnkT4 (leader-ITSISKYTGFC). After maturation and export via NisBTC mass spectrometry revealed mass peaks for EnkT4 corresponding to a stepwise dehydration of all four serines/threonines present. Two enkephalin variants were made to prevent undesired ring formation of the cysteine with the dehydroamino acids upstream of the ultimate threonine. Enk.A (IAAIAKYTGFC) and Enk.B (IAAIAKYSGFC) were both dehydrated, although the extent of dehydration for Enk.B (serine variant) was much lower than for Enk.A. The luteinizing hormone-releasing hormone (LHRH) has the sequence pGluHWSYGLRPG-NH 2. Its functionality with respect to receptor binding and activation of each amino acid is known (187). Recent studies have shown that mono- and dicyclic versions of this hormone can act as potent antagonists (162, 163). The leucine in position seven can be replaced by another uncharged amino acid. In addition

42 NisB dehydrates therapeutic peptides

to a L7C-LHRH mutation, we also changed the pGlu into a glutamine, which can spontaneously form a pGlu group. A plasmid-encoded peptide was composed of the first N-terminal 44 amino acids of prenisin fused to the L7C-LHRH variant (QHWSYGCRPG). Mass spectrometry showed peaks corresponding to five and six times dehydration. Six times dehydration implied that also the serine in the LHRH- coding part was dehydrated. Angiotensin is a cardiovascular peptide, which occurs in several forms. The angiotensin(1-7) variant (DRVYIHP), belonging to the active angiotensinII variants, plays a crucial role in the preservation of the cardiac function (112). A 2-fold mutant of this peptide was made (Y4T and P7C), fused to the leader peptide and subjected to dehydration. The observed mass peaks related to the non-dehydrated and one time dehydrated peptide. EPO (erythropoietin) is a glycoprotein hormone, responsible for the regulation of the production of red blood cells. We selected an EPO peptide mimetic (EMP1) of 13 amino acids, EPO(1-13) (67), to serve as a substrate for dehydration. Aromatic amino acids at positions essential for mimetic action were kept in tact and one serine for dehydration (C3S) was introduced (YASHFGPLGWVCK). Mass spectrometry gave peaks corresponding to a one time dehydrated peptide. To confirm that the observed mass decreases of 18 Da were the result of a dehydration step, the peptides were submitted to ethanethiol treatment. Ethanethiol couples to dehydrated residues and under the applied conditions also to subsequently formed thioether rings. In both cases this results in a mass increase of 62 Da. For all peptides described above dehydration could be confirmed by ethanethiol addition, even for those containing multiple dehydrations (Table 2). As an illustration, Figure 2b demonstrates the mass shift of dehydrated VP3 as a result of ethanethiol addition. Additions for all three dehydrated amino acids could be identified. Cyclization of the generated dehydroresidue with the downstream cysteine by NisC does not result in a mass shift. The applied conditions for ethanethiol treatment do not allow to discriminate between the absence or presence of a thioether ring. However, ethanethiol addition proves that NisB-mediated dehydration has taken place in a wide variety of unrelated therapeutic peptides.

43 Chapter 3

Figure 2. Detection of multiple forms of dehydration of the vasopressin analog VP3 via Maldi- TOF mass spectrometry . ( A). The extent of dehydration is indicated in roman letters, corresponding to one (I), two (II), and three (III) times dehydration. VP3 peaks representing vasopressin including the N-terminal methionine are indicated in italics. ( B) shows the mass shift of 62 Da as a result of ethanethiol addition to VP3 without methionine. The data corresponding to this graph are shown in Table 2.

CDAP was used to examine whether the cysteines present in most of the peptides are accessible or have undergone a thioether linkage with a dehydroresidue. Formation of the isothiocyanate part will result in a mass increase of 25 Da, whereas the absence of a mass shift indicates that the cysteine is involved in ring formation. Thioethers typically do not react with CDAP (van der Donk, personal communication), neither do addition reactions with other residues occur, as long as large excesses of CDAP are avoided (225). Two cysteine-containing, chemically synthesized, peptides were used as positive controls. In addition, peaks corresponding to non-dehydrated peptides served as an internal control. Figure 3a shows the mass spectrum of the Ang(1-7) peptide before and after CDAP treatment. The dehydration peak (3159.7 Da) remains present, whereas the non-dehydration peak (3177.6 Da) is completely replaced by a +25 Da adduct (3202.9 Da). The presence of a minor peak representing dehydrated Ang(1-7) and CDAP (3185.2 Da) shows that most likely a minor fraction of the peptide has escaped ring formation. As a positive control the cysteine-containing peptide NisB2 was used (Figure 3b). A total mass shift with +25 Da showed complete CDAP addition under the conditions used. Analysis of the other therapeutic peptides after treatment with CDAP showed that, when fully dehydrated, no additions were found. This indicates that the cysteine is not available for CDAP and thus points at the presence of a thioether ring. Simultaneously, peaks corresponding to

44 NisB dehydrates therapeutic peptides

non-dehydrated peptides shifted 25 Da, which points at the accessibility of the cysteine for CDAP. This indicates that the cysteine is not involved in ring formation. The peptides for which total CDAP addition was observed and hence lack ring formation are the enkephalin variants starting with IAAIA. No conclusive results were obtained from CDAP-treated EPO.

Figure 3 CDAP addition of Ang(1−7) shown by MALDI−TOF mass spectrometry . Nontreated peptide is indicated by a solid line, and CDAP-treated peptide is presented by a dashed line. Mass shifts of 25 Da are indicated with an arrow. (A) Leader-Ang(1−7) peptide (DRVTIHC). (B) Control peptide NisB2 (CRYTDPKPHIRLRIK).

When the ACTH variants were analyzed under nonreduced conditions, cysteine additions were observed for only the nondehydrated peptide. Additional peaks of +119 Da were observed for the peptides with (3821.5 Da) and without (3689.8 Da) the initial methionine (see also Figure S3 in the Supporting Information). For the mass peaks corresponding to the dehydrated peptide ACTH(S1G) (3552.0 Da), no equivalent peaks, pointing at cysteine addition, were detected. Cysteinylation can only occur when the cysteine of the peptide is not involved in ring formation. When these data are taken together, they are compatible with ring formation in the dehydrated ACTH variants. LHRH was chosen for more detailed studies to confirm the presence of a lanthionine ring. The LHRH fusion peptide was obtained in larger quantities by TCA- precipitation from a 1 L culture, after which LHRH was released from the rest of the fusion peptide by tryptic digestion. Cleavage of the arginine in LHRH was either hampered by the presence of a proline at position P1’ or protected by the introduced lanthionine ring (Figure 4). The trypsin digest was separated by reversed-phase HPLC, where the LHRH peptide fraction eluted at 28% ACN (See Figure S12 in the Supporting Information). The identity of the LHRH peptide in the fraction was confirmed by mass spectrometry (1172.6 Da). N-terminal sequencing was carried out on both the ethanethiol-modified and non-treated LHRH. The results are shown in Figure 4. Ethanethiol modification results in the formation of S-ethylcysteine at

45 Chapter 3

position 4, which confirms that dehydration at position 4 has taken place. The presumed Cys at position 7 could not be determined. It is known that when becoming N-terminally exposed (e.g., during Edman degradation) dehydroamino acids spontaneously deaminate and become unstable, thereby blocking the sequence reaction (124). However, N-terminal sequencing of the nontreated LHRH peptide did not result in a sequence block, which can only mean that a thioether bridge is present.

Figure 4 Amino acid sequence determination of the lanthionine-containing LHRH peptide . The primary sequence of LHRH is shown, in which Leader represents the leader peptide (MSTKDFNLDLVSVSKKDSGASPR). Arrows indicate trypsin-sensitive sites, and the dotted arrow designates the arginine excluded from cleavage. The effective LHRH peptide is highlighted in bold. The N-terminal sequence of both ethanethiol (I) and untreated (II) LHRH is shown. SE-C denotes S- ethylcysteine, and a question mark indicates a nonidentified amino acid. Numbers indicate the position of the amino acids. A schematic picture of thioether-linked LHRH is shown, in which the lanthionine ring (−S−) is formed between the N-terminal Ala (derived from Dha) and the C-terminal Ala (derived from Cys).

Membrane-associated enzyme complexes have been described for nisin- and subtilin-modifying and exporting enzymes (75, 189). Karakas Sen et al. (72) demonstrated that additional NisB in a nisin-producing strain resulted in a higher extent of dehydration in nisin. The latter observation suggests that no strict stoichiometry is required for the enzyme complex to function. Recently, we found that the NisBT-complex remains active in the absence of NisC. Because the angiotensin(1- 7) peptide, used in this study, is completely dissimilar to nisin, dehydration is likely directed by the N-terminal leader peptide. For LctM, Xie et al. found that in the absence of the leader peptide no modification occurred (226). Hence the presence of the leader sequence is an effective means to target peptides to the lantibiotic enzymes. Eventually, removal of the leader peptide is a prerequisite. Recently, we demonstrated that the extracellular leader peptidase NisP is highly specific and only recognizes thioether ring-containing prenisin (86). In addition, cleavage of fully modified prenisin by trypsin has proven to be an efficient method to obtain active

46 NisB dehydrates therapeutic peptides

nisin. Trypsin can be used to liberate therapeutic peptides that are free of trypsin- sensitive cleavage sites (arginine or lysine) or when these sites are protected by thioether rings. Alternatively, the introduction of appropriate cleavage sites would allow the removal of the leader peptide in other fusion peptides. The generation of dehydroresidues can serve as a step towards the biochemical synthesis of thioether linkages in peptides. Monosulfide bridges are more stable than disulfide bridges. They enhance the peptide stability and protect against proteolytic degradation. Furthermore, thioether bridges may generate enhanced activities, as mentioned for thioether enkephalin. Thioether variants of oxytocin (69), somatostatin (138) and RGD peptides (73) with modulated receptor interaction have been described. Using lantibiotic enzymes, a large number of analogs of therapeutic peptides can be easily achieved by varying the amino acids that are not involved in dehydration and/or ring formation. Biological synthesis of thioether peptides allows the generation of cyclopeptide libraries by using semi-randomized primers and might eventually lead to cost-effective production of dehydroresidues and/or thioether ring-containing larger peptides. Altogether, the here presented data indicate that a broad range of therapeutic and other bioactive peptides can serve as a substrate for lantibiotic enzyme-mediated insertion of dehydroresidues or and/thioether rings.

Acknowlegdments We are grateful to Wilfred van der Donk for advice on the use of CDAP, and to Kees Leenhouts, Karin Scholtmeijer and George T. Robillard for helpful discussions.

Footnotes Abbreviations: NisB, nisin dehydratase; NisT, nisin transporter;NisC, nisin cyclase; MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight; CDAP, 1- cyano-4-dimethylaminopyridiniumtetrafluoroborate Supporting Information is is available free of charge via the Internet at http://pubs.acs.org.

47

Mechanistic dissection of the enzyme complexes involved in the biosynthesis of lacticin 3147 and nisin

Abstract The thioether rings in the lantibiotics lacticin 3147 and nisin are post-translationally introduced by dehydration of serines and threonines followed by coupling of these dehydrated residues to cysteines. The prepeptides of the two-component lantibiotic lacticin 3147, LtnA1 and LtnA2, are dehydrated and cyclized by two corresponding bifunctional enzymes, LtnM1 and LtnM2, and are subsequently processed and exported via one bifunctional enzyme LtnT. In the nisin synthetase comlex the enzymes NisB, NisC, NisT and NisP respectively dehydrate, cyclize, export and process prenisin. Here we demonstrate that the combination of LtnM2 and LtnT can modify, process and transport peptides entirely different from LtnA2 and that LtnT can process and transport unmodified LtnA2 and unrelated peptides. Furthermore we demonstrate a higher extent of NisB-mediated dehydration in the absence of thioether rings. Thioether rings apparently inhibited dehydration, which implies alternating action of NisB and NisC. Furthermore, specific –but not all- NisC-cyclized peptides were exported with higher efficiency as a result of their conformation. Taken together, these data provide further insight into the applicability of Lactococcus lactis containing lantibiotic enzymes for the design and production of modified peptides.

Appl Environ Microbiol. 2008 ; 74 :6591-7 Chapter 4

Introduction The lantibiotics nisin and lacticin 3147 are produced by some distinct Lactococcus lactis species and inhibit a broad range of Gram-positive bacteria. Lantibiotics are peptide antibiotics that contain thioether-bridged amino acids: lanthionines and methyllanthionines (23, 118, 171). By binding to lipid II, the essential precursor of bacterial cell wall peptidoglycan, they inhibit cell wall synthesis and form hybrid pores in the membrane of the target cell (13, 15, 16, 56, 222, 223). Two main classes of lantibiotics are discerned. In one class (class I) , comprising nisin (90), dehydration and cyclization are catalyzed by separate LanB and LanC enzymes. The nisin dehydratase NisB dehydrates serines and threonines of the prepeptide NisA. Subsequently, NisC couples the formed dehydroalanine and dehydrobutyrine to cysteines to form lanthionine or methyllanthionine, respectively. The ABC transporter NisT exports modified prenisin out of the cell (83, 153). In the second class, comprising amongst others lacticin 481 (156), and the two- component lantibiotics and lacticin 3147 (168), dehydration and cyclization are performed by bifunctional LanM enzymes (23). During synthesis of lacticin 3147 each of the prepeptides LtnA1 and LtnA2 is modified by a separate modification enzyme, respectively LtnM1 and LtnM2 (119). Lacticin 3147 also contains other modifications resulting from conversion of dehydroalanines to D-alanines catalyzed by the enzyme LtnJ (28, 169). The modified peptides Ltn α and Ltn β are processed and transported by LtnT. Both components are essential for the activity of the lantibiotic. Ltn α resembles mersacidin in its globular shape, and Ltn β has some structural similarity to nisin (Fig. 1) (116, 118). The discovery and development of novel antibiotics, is urgent because of the increase in resistance to multiple antibiotics. Lantibiotic mutants with modulated activity have been described (159). Furthermore the application of lantibiotic enzymes can generate biostable thioether-bridged therapeutic peptides, which are resistant against proteolytic degradation (55, 78). Thioether bridged peptides may also have modulated receptor interaction and extended delivery possibilities. Several studies indicated that lantibiotic-modifying and transporter enzymes are organized in multimeric complexes (75, 226, 134). It was demonstrated before that cells containing all lacticin 3147 biosynthesis enzymes, except LtnM1, still produced Ltn β (119). Here we studied whether in L. lactis either the combination of LtnM2 and LtnT or LtnT alone is functional and whether peptides that are entirely different from LtnA2 are LtnM2 and LtnT substrates.

50 Dissection of lantibiotic enzyme complexes

Figure 1. Nisin (A) and processed, fully modified Ltn β peptide (B). In Ltn β both threonines included in lanthionine ringA and the one threonine included in methyllanthionine ringB are not dehydrated.

Intriguingly, three threonines in the Ltn β peptide (Fig. 1) that are included in the thioether rings are not dehydrated by LtnM2. In contrast NisB dehydrates threonines much better than serines and can also succesfully generate polydehydrobutyrine (157, 158). In addition we therefore studied the dehydration of lacticin A2-derived peptides corresponding to these ring-containing sequences by dissected enzyme complexes of the lacticin 3147 and nisin biosynthesis machineries.

Materials and Methods Bacterial Strains and Plasmids . L. lactis NZ9000 was used for expression of the modification enzymes and peptides. The modification genes and transporter genes, ltnM2T, nisBTC were cloned into pIL253-derived plasmids (192) behind the nisin- inducible promoter. The encoding sequences for the N-terminal lacticin leader LtnA2 or nisin leader were fused to the sequence, encoding the substrate peptide, and placed under control of the nisin inducible promoter of pNZ8084-derived plasmids (93). Strains and plasmids are listed in Table 1. Molecular cloning . The ltnA2 and ltnTM2 genes were amplified from pBAC105, isolated from L. lactis IFPL105 and the nisA and nisBTC genes were amplified from chromosomal DNA of L. lactis NZ9700. PCR reactions were performed with Phusion DNA polymerase (Finnzymes, Finland). Restriction enzymes used for cloning strategies were purchased from New England Biolabs Inc.. Ligation

51 Chapter 4

was carried out with T4 DNA ligase (Roche, Mannheim, Germany). Sequences encoding peptides different from the original propeptides were genetically fused to the leader or propeptide by means of PCR (87). Electrotransformation of L. lactis was carried out as previously described (61) using a Biorad gene pulser (Richmond, CA). Nucleotide sequence analysis was performed by BaseClear (Leiden, NL).

Strain or plasmid Characteristic(s) Source or reference

Strains NZ9700 nisABTCIPRKFEG (90) NZ9000 MG1363 derivative; pepN :: nisRK + (91) IFPL105 pBAC105 (117) Plasmids pIL253-derived (192) pIL3BTC nisBTC cloned behind the Pnis promoter; Cm r (157) pIL3BT nisBT behind the Pnis promoter; Cm r (15) pIL31BTdC Cyclase activity knocked out by mutation C326A in NisC This study pILPTM2 ltnT-ltnM2 cloned behind the Pnis promoter; Em r This study pILPltnT-2 Disruption of ltnM2 by BglII digestion; Em r This study pILPTM2-C818A Substitution by means of PCR; Em r This study pNZ8048-derived (93) pNZang1-7.2 Pnis + encoding leader peptide NisA::NisA(1-17) :: This study IEGR::DKTYICP; Em r pNZang1-7.3 Pnis + encoding leader peptide NisA::NisA(1-17):: This study IEGR::DATYICP; Em r pNZang1-7.8 Pnis + encoding leader peptide NisA::NisA(1-17):: This study IEGR::DTVYCHP; Em r pA25-7.2 Pnis + encoding leader peptide LtnA2::NisA(1-17):: This study IEGR::DKTYICP; Cm r pA25-7.3 Pnis + encoding leader peptide LtnA2::NisA(1-17):: This study IEGR::DATYICP; Cm r pA25-7.8 Pnis + encoding leader peptide LtnA2::NisA(1-17):: This study IEGR::DTVYCHP; Cm r pA24 Pnis + encoding prepeptide LtnA2; Cm r This study pA27 Pnis + encoding leader peptide LtnA2 This study ::TTPATPAISILSAYI; Cm r pA28 Pnis + encoding prepeptide LtnA1 plus leader peptide This study LtnA2::ILSAYISTNTCPTTKCTRAC; Cm r pNZ29 Pnis + encoding leader peptide NisA::ILSAYISTNTCPK; This study Em r pNZ30 Pnis + encoding leader peptide NisA::ILPTTKCTRAC; Em r This study pNZ31 Pnis + encoding leader peptide NisA::ILPTTKCARAA; Em r This study pNZ32 Pnis + encoding leader peptide NisA::ILPTTKATRAA; Em r This study pTP-DSRWARVALID Pnis + encoding leader peptide (157) SQKAAVDKAITDIAEKL NisA::DSRWARVALIDSQKAAVDKAITDIAEKL; Em r pTPtppii Pnis + encoding leader peptide NisA::ITSISRASVA; Em r (77)

Mass spectrometry . Samples were purified from the medium fraction by ziptip purification (C18 ziptip, Millipore) or directly spotted by putting 1 µl of the supernatant on the target. After drying, the spots were washed once with 4 µl MQ to

52 Dissection of lantibiotic enzyme complexes

remove the salts. Subsequently, 1 µl of matrix (10 mg/ml α-cyano-4-hydroxycinnamic acid in 50% acetonitrile containing 0.1% (v/v) TFA) was added to the target and allowed to dry. CDAP (1-cyano-4-dimethylaminopyridinium tetrafluoroborate) was used to react with free cysteine residues. The vacuum-dried sample was resuspended in 9 µl 25 mM citrate buffer, pH 3.0, and reduced with 1 µl Tris[2- carboxyethyl]phosphine (TCEP) (10 mg/ml in MQ). After a 10 min incubation at room temperature 2 µl of CDAP (10 mg/ml in 100 % acetonitril) was added, followed by 15 min. of incubation at room temperature. Mass spectra were recorded with a Voyager DE PRO Maldi-TOF mass spectrometer in the linear mode (Applied Biosystems). In order to maintain high sensitivity, an external calibration was applied. All measurements were performed in at least three independent experiments with identical results. Gel electrophoresis . L. lactis cells were induced with nisin and grown further in minimal medium. Peptides were isolated from the supernatant using disposable SPE Bond Elut C18 cartridges from Varian. After applying the supernatant, the column was washed with MQ containing 0.1 % TFA and peptides were eluted with 100 % methanol containing 0.01 % TFA. Peptides out of 1 ml supernatant were applied on tricine gel (181). Analysis was performed using the Silver Staining Kit of Invitrogen or by Coomassie staining.

Results Functionality of the enzymes LtnM2T . In order to analyze the functionality of the enzymes LtnM2 and LtnT in L. lactis , pILPTM2 was co-expressed with pA24, encoding the natural substrate LtnA2. Maldi-TOF analysis of the supernatant revealed a mass of 2846 Da identical to the expected mass of the 8-fold dehydrated Ltn β peptide, lacking conversions of two dehydroalanines to D-alanines due to the absence of the enzyme LtnJ (169, 28) and which has undergone an N-terminal deamination yielding an α-keto amide (+1 Da) (25) (Fig. S1). Clearly the enzymes LtnM2 and LtnT, in the absence of any other lacticin 3147 enzyme, succesfully modified, processed and translocated the prepeptide LtnA2. LtnM2T dehydrate, process and export peptides comprising angiotensin-(1-7) variants. We assessed the versatility of LtnM2T to modify and produce angiotensin- (1-7) variants and compared the results with an existing angiotensin-producing NisBTC system. Sequences coding for peptides comprising nisin(1-17) and angiotensin-(1-7) variants DKTYICP, DATYICP and DTVYCHP were placed behind the lacticin A2 leader or nisin leader . The encoding plasmids were each co-expressed in L. lactis with respectively pILPTM2 and pIL3BTC. The extent of LtnM2T- or NisBTC-mediated dehydration and translocation of the peptides in the supernatant of induced cultures was analyzed by Maldi-TOF (Table 2, Fig. 2AB, Fig. S2AB) and silver-stained tricine gels (Fig. 3). The amount of peptide produced via the NisBTC

53 Chapter 4

system (Fig. 3, lanes 4-6) was higher than via the LtnM2T system (Fig. 3, lanes 1-3). LtnM2T-exported, dehydrated and processed peptides were clearly detectable by Maldi-TOF (Fig. 2B and Fig. S2AB). All three constructs were evidently dehydrated up to 5-fold by NisB (Table 2, Fig. 2A), whereas LtnM2-mediated dehydration was in two cases up to five-fold and for the peptide encoded by pA25-7.8 even up to 6-fold (fully) (Table 2, Fig. S2AB and Fig. 2B). In all cases 5-fold dehydration was the major product, implying that the effciency of LtnM2- and NisB-mediated dehydration of the peptides did not differ to a large extent.

TABLE 2. Comparison of peptide sequences fused to the nisin leader, modified and secreted by NisBTC with the same peptide sequences fused to the LtnA2 leader, modified, processed and secreted by LtnM2T.

NisBTC LtnM2T plasmid max. dehydrations plasmid max. dehydrations pNZang1-7.2 5 x pA25-7.2 5 x pNZang1-7.3 5 x pA25-7.3 5 x pNZang1-7.8 5 x pA25-7.8 6 x

A B

%Intensity

Mass (m/z)

Figure 2AB . Dehydration of angiotensin fusion peptides by NisB and LtnM2 . Culture supernatant was analyzed by Maldi-TOF as described in the methods section. (A) Supernatant of L. lactis NZ9000 pIL3BTC pNZang1-7.8. Expected mass of the protonated fully dehydrated fusion peptide (STKDFNLDLVSVSKKDSGASPRITSISLCTPGCKTGALMIEGRDTVYCHP) is 5194 Da. The main peak, 5212 Da, represents the 5-fold dehydrated peptide. (B) Supernatant of L. lactis NZ9000 pILPTM2 pA25-7.8. Expected mass of the processed, protonated and fully dehydrated peptide (ITSISLCTPGCKTGALMIEGRDTVYCHP) is 2861 Da.

54 Dissection of lantibiotic enzyme complexes

The observed masses of the peptides in the supernatant indicated correct LtnT- mediated processing and export. In the absence of the transporter no peptide was detected in the supernatant excluding lysis or other translocation mechanisms. These data clearly prove that LtnM2 is capable of dehydrating peptides that are entirely different from its natural substrate. Importantly, these data also show that LtnT can subsequently remove the leader peptide and translocate the modified peptides in vivo . LtnT stays active in the absence of LtnM2 . We investigated whether LtnT itself, in the absence of the other lacticin 3147 synthetase enzymes, retains activity. When LtnT was co-expressed with LtnA2 or the DTVYCHP-comprising peptide preceded by the LtnA2 leader, processed peptides could be identified in the supernatant (Fig. S3AB). In the absence of the transporter no peptide was detected in the supernatant excluding lysis or other translocation mechanisms. These results indicate that functional dissection of LtnT from the other synthetase enzymes is possible. LtnT can process and translocate LtnA2 and unrelated substrate peptides provided that the LtnA2 leader is present.

1 2 3 4 5 6 7

14.8 kDa

7.8

3.9

Figure 3 . Secretion of modified peptides by LtnT and by NisT . L. lactis cells were induced with nisin and grown overnight in minimal medium. Peptides out of 1 ml supernatant were applied on a tricine gel. The gel was stained using the Silver Staining Kit of Invitrogen. Lane 1 : NZ9000 pILPTM2 pA25-7.2; Lane 2 : NZ9000 pILPTM2 pA25-7.3; Lane 3 : NZ9000 pILPTM2 pA25-7.8; Lane4 : NZ9000 pIL3BTC pNZang1-7.2; Lane 5 : NZ9000 pIL3BTC pNZang1-7.3; Lane 6 : NZ9000 pIL3BTC pNZang1-7.8; Lane 7 : Kaleidoscopic marker of Biorad. In Lanes 6, 7 and 8, clearly dimers were detected likely resulting from high peptide concentrations.

Do thioether rings prevent the dehydration of threonines in the peptide Ltn β? In the natural, post-translationally modified Ltn β peptide two threonines included in the first lanthionine ring and one threonine included in the second methyllanthionine are not dehydrated by LtnM2 (Fig. 1). In order to determine whether dehydration is impaired by the thioether rings installed by LtnM2, an LtnM2 mutant was made in which the cyclase activity was knocked out by exchanging the conserved cysteine at position 818 in LtnM2 (107), resulting in plasmid pILPTM2-C818A. However, when

55 Chapter 4

LtnA2-encoding pA24 was co-expressed with pILPTM2-C818A, secretion of the dehydrated natural substrate LtnA2 could not be detected in the supernatant by Maldi- TOF. In the absence of cyclase activity spontaneous formation of not-original large rings in the LtnA2 peptide can not be excluded and these might hamper the processing or transport by LtnT. The peptide variants comprising angiotensin-(1-7), encoded by the plasmids pA25-7.2 and pA25-7.8 and co-expressed with pILPTM2-C818A, were successfully dehydrated, processed and translocated, indicating that the dehydratase activity of C818A-LtnM2 and the processing and transport activity of LtnT were unaltered (data not shown). Moreover, when truncations of the LtnA2 peptide, TTPATPAISILSAYI (pA27) and ILSAYISTNTCPTTKCTRAC (pA28) were preceded by the lacticin A2 leader- peptide and co-expressed with the enzymes LtnM2 and LtnT, no secreted peptides could be detected, neither processed nor unprocessed. Apparently processing and / or the transport by LtnM2T has substrate-specific limitations. For further studies on the possible influence of thioether rings on dehydration we therefore continued this study by applying the nisin biosynthesis NisBT(C) enzymes. Enhanced dehydration in the absence of NisC . When the truncated variants of the LtnA2 peptide, ILSAYISTNTCPK and ILPTTKCTRAC preceded by the nisin leader and encoded by the plasmids pNZ29 and pNZ30, were co-expressed with pIL3BT (Fig. 4AC) or pIL3BTC (Fig. 4BD) a higher extent of dehydration was observed in the absence of NisC. In this context it seems that the threonine residues that are not dehydrated by LtnM2 can be dehydrated by NisB. To verify whether unmodified threonines in the thioether rings were involved, the C-terminally adapted fragment of LtnA2, ILPTTKCARAA, was fused to the nisin leader (pNZ31). Co- expression of pNZ31 with pIL3BT (Fig. 4E) or pIL3BTC (Fig. 4F) showed again a higher extent of dehydration of the peptide in the absence of NisC. Incubations with TCEP and CDAP and the lack of CDAP modification, confirmed that the peptide ILPTTKCARAA was almost fully cyclized by NisC (Fig. S4AB). Control experiments provided consistent data. The cysteines in the peptide ILPTTKCTRAC were exchanged by alanines leading to ILPTTKARAA encoded by pNZ32. This plasmid was co-expressed with pIL3BTC (Fig. S5A) or pIL3BT (Fig. S5B) which resulted in an identical extent of dehydration. Furthermore, knocking out the cyclase activity by the mutation C326A (107) resulting in the the plasmid pIL31BTdC did not alter the extent of peptide dehydration. Identical dehydration levels were observed when co-expressing with pIL31BTdC or pIL3BT (data not shown). Expression of the activity-deficient NisC mutant was confirmed by Western blotting (data not shown). All these data clearly demonstrate reduced dehydration in the presence of the cyclase NisC, consistent with reduced dehydration of threonines that are included in the thioether rings.

56 Dissection of lantibiotic enzyme complexes

%Intensity

Mass (m/z)

Figure 4ABCDEF . Enhanced dehydration by NisBT compared to NisBTC . Culture supernatant was analyzed by Maldi-TOF. Expected masses of the fully dehydrated and protonated fusion peptides encoded by: pNZ29 (nisin leader::ILSAYISTNTCPK): 3672 Da; pNZ30 (nisin leader::ILPTTKCTRAC): 3487 Da and pNZ31 (nisin leader-ILPTTKCARAA): 3443 Da. (A) pNZ29 co-expressed with pIL3BT, main peak is of peptide 4-fold dehydrated; (B) pNZ29 co-expressed with pIL3BTC, main peak is of peptide 2-fold dehydrated; (C) pNZ30 co-expressed with pIL3BT, main

57 Chapter 4

peak is of peptide 2-fold dehydrated; (D) pNZ30 co-expressed with pIL3BTC, main peak is of peptide 1-fold dehydrated; (E) pNZ31 co-expressed with pIL3BT; (F) pNZ31 co-expressed with pIL3BTC.

A 1 2 3 4 5 6 kDa 14.8

7.8 3.9

Figure 5 . Secretion of peptides via NisBT and NisBTC . L. lactis cells were induced at OD600 = 0.4 with nisin and grown further in minimal medium for 5 hours. Peptides out of 3 ml supernatant were applied on a tricine gel and analyzed by Coomassie staining. Lane A: Kaleidoscopic marker of Biorad. Lane 1 : NZ9000 pIL3BT pNZ29 (nisin leader:: ILSAYISTNTCPK); Lane 2 : NZ9000 pIL3BTC pNZ29; Lane 3 : NZ9000 pIL3BT pNZ30 (nisin leader::ILPTTKCTRAC); Lane4 : NZ9000 pIL3BTC pNZ30; Lane 5 : NZ9000 pIL3BT pNZ31 (nisin leader:: ILPTTKCARAA); Lane 6 : NZ9000 pIL3BTC pNZ31.

NisC-mediated cyclization enhances the production level of some peptides but not of others . We investigated whether or not the presence of NisC might enhance export. The amount of secreted peptide produced via NisBT or NisBTC was compared and analyzed on a Coomassie-stained tricine gel (Fig. 5). No noteworthy difference in secretion level was observed between NisBT- and NisBTC-transported peptides when peptide-encoding plasmid pNZ29 was co-expressed with pIL3BT or pIL3BTC. The secretion level of the peptide encoded by pNZ30 was somewhat higher when co- expressed with pIL3BTC instead of pIL3BT. However, when pNZ31 was co- expressed with pIL3BTC, the secretion level was clearly higher than when co- expressed with pIL3BT. Subsequently we wanted to distinguish a channeling effect of NisC itself from enhanced export of peptides due to the cyclized peptide conformation. When peptides without cysteines hence without potential for thioether ring formation, such as the ones encoded by pNZ31, pTP-DSRWARVALIDSQKAAVDKAITDIAEKL and pTPtppii (Fig. S6), were co-expressed with either pIL3BT or pIL3BTC, no differences in transport could be detected. Nor did the presence of the inactive NisC mutant C326A improve the transport efficiency (data not shown). These observations indicate that in these experiments no channeling effect of NisC for transport of peptides by NisT occurs. Hence differences in export via NisBT compared to NisBTC, observed for the peptides encoded by the plasmids pNZ29 and pNZ30, are likely caused by differences

58 Dissection of lantibiotic enzyme complexes

in conformation between (methyl)lanthionine-containing peptides versus linear peptides with free cysteines.

Discussion We here studied post-translational modifications, processing and transport by (sub)complexes of class I and class II lantibiotic enzymes. Since neither the activity of lacticin 3147 M1 and M2 enzymes, nor the nisin dehydratase activity by NisB, has been reconstituted in vitro , we here performed in vivo studies on the dissection of the enzyme complexes of these two lantibiotics, lacticin 3147 and nisin. It has been reported that mutagenized LtnA1 and LtnA2 are each still modified, processed, and transported and retain antimicrobial activity (29, 30, 48). In all cases, the mutagenized substrates were still highly homologous to their natural counterparts. In the study by Cotter (30), not all of the alanine-substituted peptides could be detected. It is not clear whether the lack of detection of some of the mutagenized peptides is due to impaired processing and/or transport. LtnT is an ABC transporter with a dual function. First, the ABC transporter is a maturation protease; its intracellular proteolytic domain resides in the N-terminal part of the protein. Second, after removal of the leader peptide LtnT translocates the substrate across the membrane (57). In the nisin biosynthesis machinery these processes are carried out by two distinct enzymes. NisT transports prenisin across the membrane, NisP extracellularly removes the leader-peptide (212, 230). Hence, removal of the nisin leader is not a prerequisite for transport of peptides fused to the leader (86). As can be seen in Table S1 there is peptide-dependent variation in processing and export. Modification of the peptide substrates does not seem to be a qualification for processing and transport by LtnT. Clearly more research is needed to establish a clear picture of the substrate specificity of the lacticin 3147 synthetase enzymes and transporter involved. In this study we demonstrated for the first time that LtnM2 can dehydrate peptides unrelated to LtnA2 and that LtnT surprisingly was able to process and translocate these modified peptides. Moreover, LtnT was shown to process and transport substrate peptides in the absence of all other lacticin 3147 synthetase enzymes. Van der Donk and co-workers demonstrated in vitro activity of lacticin 481, whose modification enzyme LctM has homology with LtnM2, independent of LctT (226). An in vivo study showed that in the absence of LctT a fully modified but truncated variant of the lacticin 481 was detected in the supernatant by mass spectrometry (208). This truncated, modified variant seems to be translocated via another transporter. In addition, NukM can modify its substrate in the absence of NukT (134). The latter study also proposed a membrane-located multimeric enzyme complex in which NukT and NukM are associated like the enzymes of the proposed

59 Chapter 4

membrane-associated multimeric lantibiotic-synthesis complex (NisBTC / SpaBTC) (189, 75). We here demonstrate that, like the enzymes NisB, NisC, NisT and NisP which can function independently (86, 87, 106), the combination of LtnM with LtnT and LtnT alone can also function independent of other lacticin 3147 synthetase enzymes. We hypothesized that dehydration and cyclization by the enzymes LanM, LanB and LanC are alternating activities and that thioether rings may inhibit dehydration of residues that are included in the thioether ring. In a previous study by Van der Donk the possibility that cyclization prevented dehydration of specific residues of the lacticin 481 peptide was not excluded (126). For the lacticin 3147 LtnM2 enzyme we have no conclusive data as a result of apparent substrate limitations of LtnT. We therefore compared modification by partially dissected enzyme complexes: NisBT versus NisBTC. All the tested truncations of the LtnA2 peptide had a higher extent of dehydration in the absence of NisC. In the case of export via NisBT, fully dehydrated peptides were observed. The apparent impaired dehydration of residues included in the ring is fully consistent with the possibility that the rings themselves interfere with accessibility of NisB. This points at the occurrence of alternating dehydrating and cyclization activities. For alternating activities it is likely that association between NisB and NisC is needed, consistent with the existence of a multimeric enzyme complex (189). It is tempting to speculate on a model in which the nisin leader might bind close to an interface between NisB and NisC allowing the modifiable propeptide substrate to flip from NisB to NisC and back again during alternating activities. Although no data are available yet, such a model is even easier to envisage for LctM enzymes in which the dehydratase and cyclase catalytic sites are present within one and the same enzyme. Lantibiotic cyclases regio-specifically cyclize thioether rings (160, 235). Because of their high reactivity, dehydroalanines may spontaneously react with cysteines. As a result, regio-specific cyclase activity might be required before unwanted thioether bridges are formed. Therefore alternating activity of dehydratase and cyclase might in some peptides be crucial to ascertain the correct bridging pattern, essential for activity. To provide a complete picture it should be emphasized that alternating dehydratase and cyclase activity may occur, but is not an absolute necessity. For instance, in vitro reconstitution demonstrated that NisC can cyclize fully dehydrated prenisin in the absence of NisB and NisT (106). In these studies dehydration and cylization were uncoupled processes and could therefore not be alternating processes. Many lantibiotics possess dehydrated residues in a thioether ring, like Dha5 and Abu25 in nisin. Clearly, while we do not know whether dehydration of these particular residues takes place before or after ring formation, they are dehydrated. Future studies might investigate the possibility that the conformation

60 Dissection of lantibiotic enzyme complexes

of the peptide substrate itself affects whether, when and where the dehydration and cyclization activities alternate.

Acknowledgment : L. D. Kluskens is supported by the Dutch Technology Foundation, STW project 06927. Kathleen Wood is gratefully acknowledged for reading the manuscript.

Footnotes The abbreviations used in this paper are defined as follows:Dha, dehydroalanine; MALDI-TOF, matrix-assisted laser desorption ionization–time of flight; TFA, trifluoroacetic acid; TCEP, Tris(2-carboxyethyl)phosphine; CDAP, 1-cyano-4- dimethylamino-pyridinium tetrafluoroborate; Cmr, chloramphenicol-resistant; and Emr, erythromycin resistant.

Supplemental material for this article may be found at http://aem.asm.org/.

61

Sec-mediated transport of post-translationally dehydrated peptides in Lactococcus lactis

Abstract Nisin is a lanthionine-containing antimicrobial peptide produced by Lactococcus lactis . Its (methyl)lanthionines are introduced by two posttranslational enzymatic steps involving the dehydratase NisB that dehydrates serine- and threonine residues and the cyclase NisC that couples these dehydroresidues to cysteines yielding thioether bridged amino acids called lanthionines. The prenisin is subsequently exported by the ABC transporter NisT and extracellularly processed by the peptidase NisP. L. lactis expressing the nisBTC genes can modify and secrete a wide range of non-lantibiotic peptides. Here we demonstrate that in the absence of NisT and NisC, the Sec-pathway of L. lactis can be exploited for the secretion of dehydrated variants of therapeutic peptides. Furthermore, posttranslational modifications by NisB and NisC still occur even when the nisin leader is preceded by a Sec signal peptide or a Tat signal peptide of 27 or 44 amino acids long, respectively. However, transport of fully modified prenisin via the Sec pathway is impaired. The extent of NisB-mediated dehydration could be improved by raising the intracellular concentration NisB or by modulating the export efficiency through altering the signal sequence. These data demonstrate that besides the traditional lantibiotic transporter NisT, the Sec-pathway with an established broad substrate range can be utilized for the improved export of lantibiotic enzyme-modified (poly)peptides.

Appl Environ Microbiol. 2006 ; 72 :7626-33 Chapter 5

Introduction Improved variants of therapeutic peptides have a large potential in the pharmaceutical market. An important possibility to make new variants is the introduction of dehydrated amino acids in such peptides. Dehydrated residues can enhance or modulate the biological activity of the peptide. For instance they can play an essential role in peptide activities (144), they can act as inhibitors of biological activities (131, 132) or they may alter and improve signal transmission by interaction with receptor or acceptor molecules (111). Furthermore, dehydroresidues can be starting points for further modifications. A powerful modification is the intramolecular coupling of the dehydroresidue to a cysteine resulting in a (methyl)lanthionine. Cyclization by a thioether ring in the peptide enkephalin led to a drastic improvement of the biostability and the in vivo half-life of the peptide (155, 195). Lack of biostability is one of the major limitations for the successful application of therapeutic peptides. Chemical synthesis of therapeutic peptides with dehydroresidues and/or thioether rings is time-consuming and costly. Biological production of such peptides can reduce this cost effectively. Recently, we have shown that Lactococcus lactis can produce and secrete a large variety of dehydrated non-lantibiotic peptides (77, 157) by making use of the lantibiotic enzyme NisB and the ABC transporter NisT of the nisin modification enzyme complex. Four enzymes are involved in nisin biosynthesis: NisB dehydrates serines and threonines in the nisin propeptide. Subsequently, dehydrated residues are stereo- and regioselectively coupled to cysteines in the nisin propeptide by NisC, a feature difficult to achieve by chemical synthesis. Export of the fully modified prenisin occurs via the ABC transporter NisT. Finally, the extracellular peptidase NisP cleaves off the leader which results in the formation of the active lantibiotic nisin (83, 90, 153, 230). Targeting to and modification by the lantibiotic enzymes depends on the leader peptide (226). Although NisT is capable of secreting a broad range of modified peptides, not all peptides are transported efficiently (157). Furthermore, the substrate selectivety of NisT has not yet been established. Therefore, other peptide or protein transport systems might be of interest for secretion of modified peptides by L. lactis . For the transport of prenisin or non-lantibiotic peptides via NisT, the nisin leader peptide plays an important role. The leader region of type A lantibiotics is highly conserved and does not show the characteristics of the signal peptides used in protein secretion systems (20, 39). Particularly, it has been shown that very short N-terminal extensions of leader peptides do not interfere with the activity of modification enzymes (193, 226). Possibly, an N-terminal extension of the leader sequence with a signal sequence may allow modified peptides to be directed into an alternative protein excretion pathway, thereby increasing the versatility of the modification system for the production of therapeutic proteins.

64 Sec-mediated transport of dehydrated peptides

A well known universally conserved protein translocation system is the Sec system. The Sec system translocates unfolded proteins across the cell membrane via a protein conducting pore formed by the SecYEG complex and a molecular motor, the ATPase, SecA. Secretory proteins are equipped with an N-terminal signal peptide sequence which functions as a targeting and recognition signal (45, 151). Signal peptides are composed of three domains: a positively charged N-domain, a hydrophobic core (H-) domain and a more polar C-domain which contains the cleavage site for signal peptidases. The latter enzymes remove the signal peptide from the translocated protein resulting in the release of the mature protein. The average length of a Sec signal peptide is 28 amino acids (202, 220). Protein translocation has been extensively studied in Escherichia coli and Bacillus subtilis , Gram-negative and Gram-positive bacteria, respectively (217). Homologues of SecA and SecYEG are also present in L. lactis (12, 82). Azide, a known SecA inhibitor (140, 122), blocks Sec- dependent translocation of proteins in L. lactis (59). Since the number and amount of secreted proteins in the medium of L. lactis is relatively low, this organism appears an ideal host for the detection and purification of exported therapeutic peptides. The major secreted protein in L. lactis is Usp45. The signal peptide of Usp45 has been used successfully for the secretion of heterologous proteins (2, 100, 209). The Sec system in L. lactis is able to secrete proteins ranging from low-(< 5 kDa) to high-molecular weight (> 160 kDa) (59, 101). In the present study, we show that the dehydration and cyclization of the nisin propeptide still occurs when the nisin leader is preceded by the Sec signal peptide of

Usp45 (SP Usp45 ) and that the Sec system of L. lactis efficiently secretes dehydrated therapeutic peptides.

Material and Methods Bacterial Strains and Plasmids . The host L. lactis NZ9000 was used for expression of the modification enzymes and peptides. The modification genes nisB and nisC were cloned into pIL253-derived plasmids (192). The encoding sequences for the signal peptides were N-terminally fused to encoding sequence of the nisin leader and of the modifiable peptide, and placed under control of the nisin inducible promoter of pNZ8048-derived plasmids (93). Strains and plasmids are listed in Table 1. Molecular Cloning. The nisBTC genes, the sequence encoding the Sec signal peptide of Usp45 and the nisA gene were amplified from chromosomal DNA of L. lactis NZ9700 using the Phusion DNA polymerase (Finnzymes, Finland). The Tat signal sequence of YwbN (SPYwbN) was amplified from chromosomal DNA of B. subtilis 168. Ligation was carried out with T4 DNA ligase (Roche, Mannheim, Germany). Restriction enzymes used for cloning strategies were purchased from New England Biolabs, Inc.

65 Chapter 5

TABLE 1. Bacterial strains and plasmids Strain or plasmid Characteristics Reference L. lactis strains NZ9700 nisABTCIPRKFEG (90) NZ9000 MG1363 derivative; pepN::nisRK + (91) LL108 MG1363 derivative; repA +, Cm-r (99) B. subtilis strain Bs 168 trpC (94) pIL253-derived (192) plasmids pIL3BTC pILangBTC derivate: Pnis + inverted repeat + nisBTC , Cm-r (157) pIL6BTC pIL3BTC derivate: Pnis + nisBTC , Cm-r This study pILhpB pIL3BTC derivate: Pnis + inverted repeat + nisB , Em-r This study pILB pIL3BTC derivate: Pnis + nisB , Em-r This study pIL3BC pIL3BTC derivate: Pnis + inverted repeat + nisBC , Cm-r This study pILBC pIL3BTC derivate: Pnis + nisBC , Em-r This study pNZ8048-derived (93) plasmids pNZE3 Pnis + encoding NisA prepeptide sequence, Em-r (86) pNG41nisA Pnis + encoding SP Usp45 fused to nisin prepeptide sequence, Cm-r This study

pNG41tppii Pnis + encoding SP Usp45 - Leader peptide - ITSISRASVA, Cm-r This study pNG41epo Pnis + encoding SP Usp45 - Leader peptide - YASHFGPLGWVCK,Cm-r This study pNG41lhrh1 Pnis + encoding SP Usp45 - Leader peptide - QHWSYGCRPG, Cm-r This study pNG41hrh 2 Pnis + encoding SP Usp45 - Leader peptide - QHWSYSLRCG, Cm-r This study pNG41vp1 Pnis + encoding SP Usp45 - Leader peptide - ATFQCAPRG, Cm-r This study

pNG41acth Pnis + encoding SP Usp45 - Leader peptide - SYSMECFRWG, Cm-r This study pNG51nisA Pnis + encoding SP YwbN fused to nisA prepeptide sequence, Cm-r This study

pNG51tppii Pnis + encoding SP YwbN - Leader peptide - ITSISRASVA, Cm-r This study pNG51lhrh2 Pnis + encoding SP YwbN - Leader peptide - QHWSYSLRCG, Cm-r This study pTPtppii Pnis + encoding Leader peptide – therapeutic peptide (77) pTPvp1 pLPacth pLPepo Cm-r, Em-r, resistance to chloramphenicol and erythromycin, respectively. Pnis inducible nisA promoter.

For the construction of pNG41nisA, the start codon of ssUsp45 was cloned into the Nco I site of pNZ8048. Subsequently, the nisA gene was fused at the 3' end of the ssUsp45 . The plasmid pNG41nisA was used as template for construction of the peptide-encoding plasmids. Plasmids were amplified by using primers with a complementary part to the nisin leader and a 5' overhang, encoding the peptide sequences, and one universal primer, annealing downstream the nisA gene of pNG41nisA. This allowed the exchange of the propeptide NisA sequence by another peptide sequence leaving the leader peptide intact. The linear peptide-encoding plasmid was self-ligated. For construction of the plasmids pNG51nisA, encoding the signal sequence of YwbN fused to the prepeptide NisA, and the plasmids pNG51tppii and pNG51lhrh2, the same strategy was used as described above. For the pIL-derived constructs, the following strategy was used. The nisTC genes were removed by restriction of pIL3BTC with Eco RI, followed by ligation, leading to pILhpB. The

66 Sec-mediated transport of dehydrated peptides

hairpin between Pnis and nisB was removed by amplification of pILhpB with an antisense primer containing the ribosome binding site behind the Pnis promoter and a sense primer beginning with the start codon of nisB . The pIL3BC construct was made by amplification on pIL3BTC with an antisense primer annealing behind the stop codon of nisB and a sense primer containing the RBS of nisC followed by self-ligation and transformation. From plasmid pIL3BC the hairpin was removed, by amplification with the primers mentioned above, resulting in the plasmid pILBC. Electrotransformation of L. lactis was carried out as previously described (61) using a Biorad gene pulser (Richmond, CA). Nucleotide sequence analysis was performed by BaseClear (Leiden, NL). Growth conditions . L. lactis was grown in M17 broth (195) supplemented with 0.5% glucose (GM17), or in minimal medium (66, 157) with or without chloramphenicol (5 µg/mL) and/or erythromycin (5 µg/mL). Cultures were grown on minimal medium after induction with nisin prior to sample preparation for mass spectrometry (Maldi-TOF: matrix-assisted laser desorption/ionization time-of-flight mass spectrometry), Western blot analysis (77) or the antimicrobial assay. Mass spectrometry . Samples were purified from the medium fraction by ziptip purification (C18 ziptip, Millipore). Dehydration of the peptides was confirmed by ethanethiol treatment which results in a mass increase of 62 Da upon reaction with the dehydrated amino acid (77). Mass spectra were recorded with a Voyager DE PRO Maldi-TOF mass spectrometer (Applied Biosystems). In order to maintain high sensitivity, an external calibration was applied. Western blot analysis . Polyclonal anti-leader peptide antibodies were raised in rabbits against the peptide H 2N-STKDFNLDLVSVSKKDC-CONH 2 coupled via the cysteine to keyhole limpet haemocyanin (77). Peptides, precipitated with 10% TCA from supernatant, or lysozyme-treated cell lysates were dissolved in sample buffer and applied on a tricine gel (181). Peptides were transferred to PVDF Western blotting membrane (Roche) using a Trans-Blot SD semi-dry transfer cell (Bio-Rad). Antimicrobial assay. Samples for analysis, cell lysates or TCA precipitates of culture supernatant were separated on a tricine gel. After electrophoresis, the gel was incubated twice in a mixture of isopropanol (20%) and acetic acid (10%) and subsequently washed thoroughly several times with demineralized water (11). The gel was overlaid with top agar, containing 1/100 of an overnight culture of the indicator, Lactococcus lactis strain LL108, and 0.01 mg/mL trypsin to release mature nisin, and zones of inhibition were scored by halo formation.

Results Modification by NisB and NisC of nisin prepeptide preceded by non-lantibiotic secretion signals . To investigate whether dehydration by NisB and ring formation in the propeptide by NisC occurs when the nisin leader peptide is preceded by a signal

67 Chapter 5

peptide, the Sec signal peptide (SP Usp45 ) or the Tat signal peptide (SP YwbN ) was N- terminally fused to the prepeptide NisA. The Tat-signal sequence of the YwbN-protein directs the protein to the Tat pathway of Bacillus subtilis (68). The Tat pathway translocates folded proteins across the membrane (151) and it has been suggested that a tertiary structure of the substrate seems to be a prerequisite (34). The Tat pathway might be an interesting candidate for the transport of (poly)peptides which have a somewhat bulky character caused by the intramolecular lanthionine rings. Since L. lactis lacks a Tat pathway, reconstitution of a Tat pathway in L. lactis might be considered. Another possibility is expression of a lantibiotic system in B. subtilis.

The corresponding plasmids pNG41nisA (with SP Usp45 ) and pNG51nisA (with

SP YwbN ), respectively, were co-expressed with and without pILBC, in the absence of NisT, in L. lactis . Samples from cultures with and without co-expression of NisB and NisC were compared by Western blotting (Figs. 1A and 1B) and by an antimicrobial assay (Fig. 1C). Dehydration, catalyzed by NisB and ring formation, catalyzed by NisC, were determined by means of a bioassay that monitors the antimicrobial activity of correctly modified nisin in cell lysates after trypsin-treatment to remove the signal peptide and the leader peptide. Without co-expression of NisB and NisC, a high amount of prepeptide was secreted when the SP Usp45 was used to direct the export (Fig.

1A lane 2). When the prepeptide was preceded by the Tat signal sequence SP YwbN , only a small amount of the NisA prepeptide could be detected in the supernatant (Fig. 1A lane 4), even though no functional Tat pathway is present in L. lactis . The latter is likely due to an inefficient SP YwbN -directed transport via the Sec pathway. When NisB and NisC were co-expressed, no significant level of secreted peptides could be detected in the supernatant when SP Usp45 or SP YwbN was used to direct the secretion (Fig. 1A lanes 3, 5), which excludes lysis as well as export, and neither any antimicrobial activity, after trypsin digestion, could be detected in the supernatant (data not shown). However, after analysis of the cell lysates on a tricine gel using an overlay with the indicator strain together with trypsin, peptides remaining in the cell- lysates showed significant antimicrobial activity. This activity was observed at a position in the gel that corresponds to the full length of SP Usp45 -Leader-NisA of 8.3 kDa (Fig. 1C lane 3) and SP YwbN -Leader-NisA of 10.1 kDa (Fig. 1C lane 5), respectively. Without coexpression of NisB and NisC, no antimicrobial activity was observed (Fig. 1C lanes 2 and 4). Also in the absence of trypsin, in all mentioned cell lysates, there was no antimicrobial activity (data not shown). These data demonstrate that both NisB and NisC are capable of modifying their substrate despite the fact that the leader peptide is preceded by a signal peptide of 27 (SP Usp45 ) or 44 (SP YwbN ) amino acids long. Moreover, the absence of extracellular levels of fully modified prenisin suggests that secretion of the fully modified peptide via the Sec pathway is impaired.

68 Sec-mediated transport of dehydrated peptides

Figure 1 . Modification by NisBC of the NisA propeptide that contains a signal peptide N- terminally of prenisin . L. lactis cells were induced with nisin and grown overnight. TCA precipitates of 1 mL supernatant (A) and of approximately 3.0 × 10 7 lysed cells (B) were analyzed using tricine gels and Western blotting with antileader antibodies. By trypsin treatment of the cell extracts on the tricine gel active nisin could be liberated and detected by an overlay with nisin-sensitive L. lactis . The black areas in the picture represent zones with no growth of the indicator strain (C). Lane 1 : NZ9000; Lane 2 : NZ9000 pNG41nisA; Lane 3 : NZ9000 pNG41nisA pILBC; Lane 4 : NZ9000 pNG51nisA; Lane 5 : NZ9000 pNG51nisA pILBC; Lane 6 : positive control NZ9000 pNZ-E3 pIL3BC. Arrows indicate the full length products of SP Usp45 :Leader:NisA, SP YwbN :Leader:NisA and Leader:NisA.

Sec-mediated transport of NisB-dehydrated peptides . Recently, we have shown that the ABC transporter NisT of L. lactis transports a wide range of dehydrated non- lantibiotic peptides into the culture medium (77). Here, we investigated whether the general Sec-pathway of L. lactis can be exploited for the translocation of dehydrated peptides, thereby increasing the range and amount of modified peptides and proteins that can be produced and exported. For this purpose, plasmids (pNG41peptide), encoding fusions of the signal peptide of Usp45 (SPUsp45 ) with the nisin leader and the peptides of interest, including variants of the inhibitor of tripeptidylpeptidase II (TPPII), the luteinizing hormone-releasing hormones (LHRH1 and LHRH2), the erythropoietin fragment(1-13) (EPO), vasopressin (VP1), and an adrenocorticotropic hormone fragment(1-10) (ACTH), were co-expressed with a pIL-derived plasmid carrying nisB (Table 1) in L. lactis NZ9000, in the absence nisC and nisT .

69 Chapter 5

Of all analyzed therapeutic peptides dehydrated variants (- 18 Da) could be detected in the supernatant by mass spectrometry (Table 2) (Figs. 2A, 5A). Dehydration of the peptides was confirmed by ethanethiol treatment (data not shown).

a TABLE 2. Masses of SP Usp45 -directed export of NisB-dehydrated therapeutic peptide variants

Peptide Sequence Dehydration Mass (M + H +) w/o Met1 extent Dehydration Observed Calculated

TPP II ITSISRASVA 4 3266 3267 3 3284 3285 2 3303 3303 1 3321 3321 0 3339 3339

LHRH1 QHWSYG CRPG 1 3508 3507 0 3527 3525

LHRH2 QHWSY SL RCG 2 3534 3535 1 3552 3553 0 3570 3571

EPO(1-13) YAS HFGPL GWVCK 1 3781 3781 0 3799 3799

VP1 AT FQ CA PRG 1 3268 3267 0 x 3285

ACTH(1-10) SYSME CFRWG 1 3581 3582 0 3598 3600 a The amino acids that are part of the potentially active part of the peptide are underlined. Amino acids in boldface correspond to mutations. b Observed masses are of peptides that contain the nisin leader without Met1. Peptide masses with or without dehydration are shown in Da. x, not observed.

Without expression of NisB, all peptides were secreted without dehydrations (data not shown). Observed and calculated masses of the modified peptides are summarized in Table 2. These data demonstrate that NisB is still capable of dehydrating the substrate peptides even when a Sec signal peptide (SP Usp45 ) is fused to the N-terminal part of the leader peptide, and that the versatility of the Sec pathway of L. lactis can be exploited to secrete a wide range of dehydrated peptides. In particular, the Sec pathway provides a good alternative for the peptides VP1, ACTH and EPO which are inefficiently exported by NisT (Fig. S1, compare respectively export of different peptides via Sec: lanes 2, 4, 6 with export via NisT: lanes 1, 3, 5).

70 Sec-mediated transport of dehydrated peptides

dehydration extent

A B

%Intensity

C

Mass (m/z)

Figure 2 . SP Usp45 and SP YwbN directed export via the Sec pathway of dehydrated variants of the peptide ITSISRASVA. Culture supernatant was analyzed by mass spectrometry as described in the methods section. (A) Supernatant of L. lactis NZ9000 pNG41tppii pILhpB (with inverted repeat). (B) Supernatant of L. lactis NZ9000 pNG41tppii pILB (without inverted repeat). (C) Supernatant of L. lactis NZ9000 pNG51tppii pILB. The variants of four, three, two, one time dehydrated and absence of dehydration are indicated by -4, -3, -2, -1 and 0, respectively. Intensity of mass spectrometry signals is in arbitrary units and left out in the mass spectra.

Increased peptide dehydration upon overexpression of NisB . We initially used a pIL-derived plasmid with an inverted repeat upstream of nisB, as it is encoded in the nisin gene cluster nisABTCIPRKFEG on the chromosome of L. lactis NZ9700. This inverted repeat results in a partial read-through by the RNA polymerase into genes downstream of the nisA gene, which is thought to yield an optimal stoichiometry of the NisA prepeptide, the modification enzymes, NisB and NisC, and the transporter NisT

(91, 142). After removal of the inverted repeat sequence between Pnis and nisB on pILhpB, a significant increase in the extent of dehydration of the peptides was observed as shown for the inhibitor of tripeptidylpeptidase II peptide, TPPII (Compare

71 Chapter 5

Figs. 2A and 2B). This is likely due to an increased activity caused by an elevated expression of NisB as verified by SDS-PAGE and Coomassie staining of cell lysates (Fig. 3). (See also Karakas Sen et al . 1999)

117 kDa

Figure 3. Increased NisB-level following removal of the inverted repeat in front of nisB . L. lactis cells were induced with nisin and grown overnight. Approximately 3.0 × 10 7 lysed cells were analyzed on Coomassie-stained SDS-PAGE for NisB detection. Lane 1 : NZ9000; Lane 2 : NZ9000 containing pNG41tppii and pILhpB (with inverted repeat); Lane 3 : NZ9000 pNG41tppii pILB (without inverted repeat).

Improvement of extent of dehydration by using the Tat-signal sequence. An intriguing question is whether modulation of the transport efficiency of the peptide influences the extent of dehydration of the exported peptides. Interestingly, we observed that the Tat signal sequence, SP YwbN can be used to secrete the prepeptide NisA by the Sec system into the supernatant (Fig. 1A lane 4). However, lower quantities of peptide were exported, which points at an inefficient targeting and/or recognition of SP YwbN by the Sec system. To investigate whether these slower kinetics increase the opportunity of the peptide substrate to interact with NisB, and thereby influence the extent of dehydration, SP YwbN was used to direct dehydrated peptides to the Sec pathway. From the induced cultures of L. lactis NZ9000 pNG51tppii pILB and NZ9000 pNG51lhrhr2 pILB, that thus lack NisT, the supernatant was analyzed by Western blotting and mass spectrometry. As expected, a lower level of secretion was observed when SP YwbN was used instead of SP Usp45 to direct the export of the TPPII – or LHRH2 peptide (Fig. 4). However, as can be seen on the Coomassie-stained gel, a reasonable amount of peptide was still secreted, even when SP YwbN was used to direct secretion (Fig. 4 lane 3 and 5). In a control experiment, no secreted peptides could be

72 Sec-mediated transport of dehydrated peptides

detected in the supernatant when the peptide was not preceded by a signal peptide together with absence of NisT (Fig. 4 lane 1), which result excludes lysis. The mass spectra of the more slowly secreted TPPII revealed that the peptide was nearly fully dehydrated (Compare Figs. 2B and 2C). Similarly, the extent of dehydration of

LHRH2 peptide was also increased when the peptide was secreted by means of SP YwbN (Compare Figs. 5A and 5B). Furthermore the main peaks detected by mass spectrometry corresponded with the peptides TPPII and LHRH2 nicely processed by the signal peptidase. The amount of the peptides in the supernatant and their complete processing are consistent with Sec-mediated secretion. These data demonstrate that a reduced secretion caused by a poor targeting signal is accompanied by an improved and more complete level of dehydration of the secreted peptides.

kDa 1 2 3 kDa 4 5

Figure 4 . Comparison of secretion levels of SP USP45 and SP YwbN preceded peptides . The equivalent of 2,5 ml supernatant was analyzed on a Coomassie-stained tricine gel. Lane 1 : NZ9000 pTPtppii (no signal sequence, no NisT); Lane 2 : NZ9000 pILB pNG41tppii (SP USP45 , no NisT); Lane 3 : NZ9000 pILB pNG51tppii (SP YwbN , no NisT); Lane 4 : NZ9000 pIL5B pNG41lhrh2; Lane 5 : NZ9000 pIL5B pNG51lhrh2. Cultures were induced at OD600 being 0.4 and supernatant was harvested at OD600 close to 1.0.

Spontaneous ring formation . Cysteinylation of an available cysteine in a peptide gives a mass shift of 119 Da that can be detected by mass spectrometry (Figs. 5A and 5B). Spontaneous ring formation by coupling of a dehydrated residue to a cysteine in the same peptide sequence prevents this cysteinylation. For instance, no cysteinylation of the two times dehydrated variant of LHRH2 (Figs. 5A and 5B) or of ACTH was observed in the mass spectra. On the other hand, cysteinylation of the non-dehydrated variant or the single dehydrated variant of LHRH2 was clearly detectable, suggesting spontaneous ring formation in the double dehydrated variant of these peptides. When a thioether bridge is formed in the LHRH2 peptide (QHWSYSLRCG), likely, the arginine becomes protected against cleavage by trypsin. Peptides precipitated from the supernatant of NZ9000 cells transformed with pNG41lhrh2 pILB or pNG51lhrh2 pILB, were treated with trypsin and analyzed by mass spectrometry (Fig. S2: A and B). The non-modified LHRH2 peptide from culture NZ9000 pNG41lhrh2 pILB was completely digested behind the arginine position. However, the single dehydrated LHRH2 variants of both cultures were only partially digested suggesting some degree

73 Chapter 5

of ring closure. Moreover, the double dehydrated variant of the LHRH2 peptide of both cultures was entirely trypsin resistant consistent with complete ring closure.

dehydration extent A B

Intensity

%

Mass (m/z)

Figure 5. Enhanced dehydration of LHRH2 when using the YwbN signal peptide compared to the Usp signal peptide. Culture supernatant was analyzed using Maldi-TOF as described in the methods section. (A) Supernatant of L. lactis NZ9000 pNG41lhrh2 pILB (SP USP ). (B) Supernatant of L. lactis NZ9000 pNG51lhrh2 pILB (SP YwbN ). The two, one time dehydrated and non-dehydrated peptides are indicated by -2, -1 and 0. Cysteinylation (+Cys) to an available cysteine results in a shift of 119 Da.

Discussion The general secretion (Sec-) pathway of L. lactis has been used successfully for the secretion of heterologous proteins. Here, we have investigated whether the Sec- pathway can also be used to secrete dehydrated therapeutic peptides. The study demonstrates that NisB-catalyzed dehydration of serine and threonine residues in peptide sequences can still take place even when the nisin leader is preceded by a Sec or Tat signal sequence. Since export is observed in a strain lacking the lantibiotic transporter NisT, it appears that the dehydrated peptide fusion proteins are secreted via the Sec pathway. Some peptides are even secreted to higher levels than observed with NisT-mediated transport. To enhance the extent of dehydration of Sec-transported peptides, two successful approaches were followed. First, the dehydration of the secreted peptide could be improved by raising the intracellular concentration of NisB. Thus, it is evident that the NisB to substrate ratio has a considerable effect on the extent of dehydration of the secreted peptide. Lowering of the intracellular peptide concentration, while raising the NisB concentration, will most likely result in an even higher extent of dehydration of the peptide. Second, less efficient targeting and

74 Sec-mediated transport of dehydrated peptides

transport also leads to a higher level of dehydration of the secreted peptide. Reduced translocation kinetics might increase the opportunity for NisB to dehydrate the fusion peptide. Indeed, the extent of dehydration improved when the Tat signal - instead of the Sec signal sequence was used. Since L. lactis does not contain a Tat translocase, the heterologous signal sequence of YwbN of B. subtilis most likely targets the peptide to the Sec system. It has been reported before that inefficient, low-level targeting and translocation of a Tat-directed substrate via the Sec pathway can occur (173). In conclusion, in vivo modulation of the enzyme:substrate ratio and/or the efficiency of transport, by modulating the signal sequence, can be valuable approaches to improve peptide dehydration, while maintaining an appreciable level of peptide export. Another important observation is that NisB is functional in the absence of NisT and NisC. Previously, we have demonstrated that NisT functions in the absence of NisB and NisC and is able to transport non-lantibiotic peptides (86). Recently, in vitro activity of NisC, in the absence of NisB and NisT, has been demonstrated (106). When combining these results, we now can conclude that the enzymes of the proposed membrane-associated multimeric lanthionine synthase complex (NisBTC) (75, 189) can function independently. Thus peptide modification and secretion are intrinsically uncoupled processes. Furthermore, NisC is still able to cyclize the dehydrated prepeptide NisA when preceded by the Sec signal sequence as evidenced by the antimicrobial assay. As a result of NisC-mediated modification, secretion of the fully modified prenisin via the Sec pathway appeared to be impossible. The (methyl)lanthionines in nisin give the peptide a bulky character which probably precludes its export via the translocation pore of the Sec transport system. The dimensions of the solvent accessible surface of a completely modified nisin is about 2.2 x 2.7 x 4.2 nm (14) whereas molecular dynamics simulations suggest that the monomeric SecY pore has a maximal pore diameter of about 1.6 nm (201). These data therefore suggest that the completely modified nisin molecule is too large to fit in the SecY pore. On the other hand, translocation of peptides with a single lanthionine via the Sec pathway might still be possible as our results with the therapeutic peptides did not establish if ring formation has occurred intra- or extracellulary. Possibly, fully modified nisin can be translocated via the Tat system. In order to establish such a phenomenon, it will be necessary to reconstitute the Tat pathway in L. lactis by introducing a system from another bacterial host, such as B. subtilis , or express the nisin system in an alternative host with an endogenous Tat system. In conclusion, here we show for the first time that at least one completely different transport machinery besides the dedicated lantibiotic transporter NisT can be used for the export of lantibiotic-enzyme-modified peptides. Our present findings imply that the Sec system with an established broad substrate range can be used for the export of peptides with dehydrated amino acids, although completely modified

75 Chapter 5

prenisin with its multiple thioether rings appears not to be tolerated. The use of the Sec-system provides a greater versatility to produce peptides with dehydroresidues followed by in vitro cyclization of the dehydroresidues to cysteines, either chemically or enzymatically. Taken together these findings underline the large potential to apply lantibiotic enzymes for the biotechnological production of modified peptides.

Acknowledgements We are grateful to Esther de Boef for technical assistance and to Jan Jongbloed and Helga Westers for supporting information.

Footnote Supplemental material for this article may be found at http://aem.asm.org/.

76 Translocation of a thioether-bridged azurin peptide fragment via the Sec pathway in Lactococcus lactis

Abstract

This study demonstrates for the first time that a thioether-containing peptide, an azurin fragment, can be translocated via the Sec pathway. This methyl-lanthionine was introduced by the nisin modification enzymes. The Sec pathway can therefore be a successful alternative for those cyclized peptides that are inefficiently transported via NisT.

Appl Environ Microbiol. 2009 ; 75 :3800-2. Chapter 6

Azurin, a cupredoxin produced by Pseudomonas aeruginosa , can selectively enter human cancer cells and induce apoptosis (229) via binding to the tumor- suppressor protein p53 (3). The azurin peptide fragment p28, containing the amino acids 50-77 (LSTAADMQGVVTDGMASGLDKDYLKPDD), still enters human cancer cells and inhibits tumor proliferation (197). Importantly, novel cancer treatments can be based on azurin peptide fragments and derivatives thereof (31). Although the pharmacokinetic property of therapeutic peptides is promising, l ack of biostability is the major hurdle for their successful application. Consequently, it is very relevant to explore the possibilities for enhancing biostability of these peptides. In our group we developed a technology to improve the stability of therapeutic peptides by exploiting the nisin synthetase enzymes NisB and NisC for the introduction of thioether bridges. We applied a two plasmid expression system (77, 78, 87), in which the NisBTC-encoding plasmid is compatible with the substrate-peptide- encoding plasmid. Lactococcus lactis containing this expression system can secrete non-lantibiotic peptides which are dehydrated or stabilized by a thioether ring (78, 160). NisB dehydrates serines and threonines in substrate peptides, NisC couples dehydrated residues stereo- and regioselectively to cysteines and NisT, the ABC transporter, translocates the modified peptides out of the cell (83, 86, 93, 153). The leader peptide is essential in targeting and modification of the propeptides (226). When transport via NisT is impaired or is less efficient, the Sec pathway of L. lactis is a successful alternative in translocation of dehydrated peptides. When the nisin leader is preceded by a Sec signal peptide or a Tat signal peptide of respectively 27 or 44 amino acids long, modification by NisB and NisC still occurs (87, 130). However, NisC-cyclized prenisin was not translocated via the Sec system (87). This is likely due to the dimensions of fully modified nisin (14) which is too large to fit in the SecY pore (87, 201). Here, we report for the first time that the Sec pathway of L. lactis can translocate a p28 azurin fragment analog with a thioether ring. We previously demonstrated that under culturing conditions the highly reactive dehydroalanines can spontaneously couple to cysteines, either intra- or extracellularly, whereas the less reactive dehydrobutyrines do not (160). To exclude spontaneous thioether ring formation by dehydroalanines we mutated serines in positions 51 and 67 to alanines, whereas a single dehydratable threonine was kept in position 52 of the 50- 77 azurin fragment. Position 56 was mutated to a cysteine to allow posttranslational introduction of a thioether bridge (Table 1, pNZ8048-derived plasmids).

78 Transport of a thioether bridged peptide via Sec

Table 1. Bacterial strains and plasmids Characteristics a Reference Strains NZ9000 MG1363 derivative; pepN::nisRK + (91) pIL253-derived (192) plasmids pIL3BTC nisBTC cloned behind the Pnis promoter, Cm-r (157) pILBC nisBC cloned behind the Pnis promoter, Em-r (87) pILB nisB cloned behind the Pnis promoter, Em-r (87) pNZ8048-derived (93) plasmids pNZazu Pnis + encoding nisin leader:: This study LATAAD CQGVVADGMA AGLDKDYLKPDD, Em-r pNG41azu Pnis + encoding SP Usp45 :: nisin leader:: This study LATAAD CQGVVADGMA AGLDKDYLKPDD, Cm-r pNG51azu Pnis + encoding SP YwbN :: nisin leader:: This study LATAAD CQGVVADGMA AGLDKDYLKPDD, Cm-r a Pnis is the nisin inducible nisA promoter; Em-r: erythromycin resistance; Cm-r: chloramphenicol resistance; Bold indicates mutations: S51A , M56C and S66A in Azurin (50-77) peptide fragment.

When the azurin peptide fragment fused behind the nisin leader was co- expressed with the enzymes NisB, NisC and NisT in L. lactis , no secreted (un)modified peptides in the supernatant were detected. Hence, we made use of the Sec pathway of L. lactis for export of the azurin peptide fragment. The nisin leader of the substrate peptide was preceded by the Sec signal peptide of Usp45, SP Usp45 . This fusion peptide was co-expressed with NisB and NisC in L. lactis in the absence of NisT. The used strains and plasmids are listed in Table 1. The culture L. lactis NZ9000 (pNG41azurin)(pILBC) was grown in minimal medium. Peptides from 4 ml induced cultures were isolated and purified with bond elute C 18 cartridges from Varian. Dissolved peptides were analyzed by mass spectrometry directly or after TCEP (Tris(2-carboxyethyl)phosphine) - and subsequent CDAP (1-cyano-4-dimethylamino- pyridinium tetrafluoroborate) incubation (160). Mass spectra were recorded with a Voyager DE PRO MALDI-TOF mass spectrometer in the linear mode. We were able to detect Sec-secreted azurin peptides in the supernatant by mass spectrometry. Some of the dehydrated peptides in the supernatant contained disulfide bonded cysteine adducts meaning that not all the formed dehydrobutyrines were coupled to the peptide’s cysteines (data not shown). This observation of partial ring formation was confirmed by mass spectrometry. Peptides were first reduced with the phosphine TCEP, then alkylated with CDAP; formation of the thiocyanate results in a mass shift of +25 Da. (Fig. 1A). These data indicate that transport of an azurin peptide fragment with a thioether ring via the Sec pathway is possible, when preceded by the

SP Usp45 and the nisin leader.

79 Chapter 6

sity

n

%Inte TCEP ------TCEP + CDAP

CDAP addition 25 Da

Mass (m/z) Fig. 1ABC. Transport of cyclized and dehydrated azurin peptide fragments via the Sec pathway. Culture supernatant was analyzed by Maldi-TOF. Expected mass of the processed, protonated and fully dehydrated fusion peptide (STKDFNLDLVSVSKKDSGASPR:: LATAADCQGVVADGMAAGLDKDYLKPD) is 5141 Da. (A) Supernatant of L. lactis NZ9000 (pILBC)(pNG41azu), fusion peptide preceded by the SP Usp45 . (B) Supernatant of L. lactis NZ9000 (pILBC)(pNG51azu), fusion peptide preceded by the SP YwbN . (C) Supernatant of control L. lactis NZ9000 (pILB)(pNG51azu).

80 Transport of a thioether bridged peptide via Sec

As demonstrated before, replacement of the SP Usp45 with a Tat signal peptide of

YwbN from Bacillus subtilis , SP YwbN , resulted in reduced transport efficiency and simultaneously enhanced extent of dehydration of the substrate peptide (87). To examine the effect of reduced transport on the extent of NisC-mediated cyclization,

SP YwbN was fused N-terminally to the nisin leader and azurin peptide fragment (Table 1) and coexpressed with NisB and NisC. Peptides in the supernatant of induced cultures were analyzed by mass spectrometry as described above and analyzed on a tricine gel (20) by silver staining (Invitrogen). As a control the substrate peptide was also coexpressed with only NisB. In the case of coexpression of NisB and NisC the secreted peptides were almost fully dehydrated and no cysteinylations were seen (data not shown), suggesting that the peptides were fully ring-closed. TCEP treatment and CDAP incubation (Fig. 1B) of the purified isolated peptides confirmed this observation. As expected, in the supernatant of the control which had only coexpression of NisB (pILB), fully dehydrated peptides were seen. No cyclization took place and therefore all the free cysteines reacted with CDAP (Fig. 1C). This control experiment clearly demonstrated that no spontaneous cyclization had occurred in the absence of NisC. Hence, the thioether bridge formation observed in the experiments presented in Fig 1A and 1B should result from intracellular NisC- mediated cyclization. The data therefore convincingly prove that indeed this intrinsically stable thioether-bridged peptide is transported via the Sec system. Analyses on silver-stained gel showed that no transport at all via the transporter NisT had occurred when the azurin peptide fragment was preceded by only the nisin leader (Fig. 2, lane 1). In full contrast, transport of the modified azurin peptide fragment via the Sec pathway was successful. The transport of the peptide fragment was less efficient when the SP YwbN was used instead of the SP Usp45 (Fig. 2), which is in full agreement with previous data (87).

M 1 2 3 7.8 kDa 3.9 kDa

Figure 2. Amount of secreted peptides in the supernatant. Peptides from 2 ml supernatant of induced cultures were applied on a gel. Lanes: M, kaleiodoscopic marker (Biorad); 1, NZ9000 (pIL3BTC)(pNZazu); 2, NZ9000 (pIL3BC)(pNG41azu); 3, NZ9000 (pIL3BC)(pNG51azu).

These data demonstrate for the first time that a peptide with an intramolecular thioether bridge can be translocated in vivo via the Sec pathway of L. lactis . A more in detail studied Sec system is that of Escherichia coli. The Sec translocase in the membrane is composed of a highly conserved protein conducting channel, SecYEG (43). The Sec translocase transports unfolded proteins which is driven by the ATPase

81 Chapter 6

SecA. Homologues of SecYEG and SecA are also found in L. lactis (12, 82). In vitro studies demonstrated that the translocon SecYEG of E. coli can also translocate the polypeptide proOmpA with a disulfide-bridge, which can have a loop of 18 amino acids or smaller (206). Other in vitro data demonstrated that proOmpA labeled with bulky fluorescent probes, assessing up to 16 Å were also transported (33). These in vitro data with the Sec system of E. coli suggested that the SecY translocon is not that rigid and could be used for other purposes such as the in vivo translocation in L. lactis of peptides with thioether bridges. Interestingly, the efficiency of translocation in E. coli can be drastically enhanced by prlA (secY ) mutations (43). Likely, in the near future such mutations in the SecY translocon of L. lactis can contribute to an even more successful application of the Sec pathway for transport of therapeutic peptides with thioether bridges.

Footnote This project was cofinanced by the European Fund for Regional Development and the Dutch Ministry of Economic Affairs.

82

Angiotensin-(1-7) with thioether-bridge: an angiotensin- converting enzyme-resistant, potent Angiotensin-(1-7) analogue

Abstract The in vivo efficacy of many therapeutic peptides is hampered by their rapid proteolytic degradation. Cyclization of these therapeutic peptides is an excellent way to render them more resistance against breakdown. Here we describe the enzymatic introduction of a thioether ring in angiotensin (Ang-(1-7)), a heptapeptide that plays a pivotal role in the renin-angiotensin system and possesses important therapeutic activities. The lactic acid bacterium Lactococcus lactis equipped with the plasmid- based nisin modification machinery was used to produce thioether bridged Ang-(1-7). The resulting cyclized Ang-(1-7) is fully resistant against purified angiotensin- converting enzyme, has significantly increased stability in homogenates of different organs and in plasma derived from pig, and displays a strongly (34-fold) enhanced survival in Sprague Dawley (SD) rats, in vivo . With respect to functional activity, cyclized Ang-(1-7) induces relaxation of precontracted SD rat aorta rings, in vitro . The magnitude of this effect is two-fold larger than that obtained for natural Ang-(1-7). The Ang-(1-7) receptor antagonist D-Pro 7-Ang-(1-7) which completely inhibits the activity of natural Ang-(1-7) also abolishes the vasodilation by cyclized Ang-(1-7), providing evidence that also cyclized Ang-(1-7) interacts with the Ang-(1-7) receptor. Taken together, applying a highly innovative enzymatic peptide stabilization method we generated a stable Ang-(1-7) analogue with strongly enhanced therapeutic potential.

J Pharmacol Exp Ther. 2009 ; 328 :849-54 Chapter 7

Introduction Angiotensin-(1-7) is an endogenous heptapeptide that plays an important role in the renin-angiotensin system (RAS) (204). Its alleged biological inactivity was first refuted by Schiavone et al., who demonstrated a dose-related stimulation of arginine vasopressin release by Ang-(1-7) equipotent to angiotensin-II (AngII) (184). Since then, its significance in the complex RAS has been slowly unraveled. Whereas in some cases, Ang-(1-7) acts similar to AngII (164, 165, 184), Ang-(1-7) mainly exerts effects which oppose those of AngII, like vasoconstriction, fibrosis and proliferation (175, 185). Discoveries such as the counteractive effect of Ang-(1-7) on the cardiovascular actions of AngII (46, 167) and its ability to attenuate the development of heart failure (112) have subsequently supported this concept. More recently, it has been demonstrated that Ang-(1-7) also plays a protective role in hepatic diseases such as liver fibrosis (150), and can even inhibit the growth of human lung cancer cells in vivo (123). Furthermore, the discovery that the G protein-coupled receptor Mas is involved in Ang-(1-7) signaling has greatly contributed to the elucidation of the heptapeptide’s biological role (177). Its variety of beneficial actions against a range of diseases would make Ang-(1- 7) an ideal therapeutic agent. However, its rapid in vivo catabolism by angiotensin- converting enzyme (ACE) and other proteases has so far reduced its potential as a drug. Cyclization of therapeutically important peptides has been proven to be a successful method to create more stable peptide analogues with improved pharmacodynamic properties. Especially intramolecular thioether bridge formation is an effective way to protect peptides against proteolytic degradation. Introduction of thioether bridges in peptides has been effectively applied on peptides like somatostatin, enkephalin and an epitope of the herpes simplex virus glycoprotein D. For all these peptides an increased catabolic stability was observed compared to their linear counterparts (155, 205). Thioether bridges are more stable than peptide bonds and disulfide bridges (205). Cyclization imposes conformational constraints, which likely caused the significantly enhanced receptor interaction of thioether enkephalin (155). Multistep chemical synthesis of thioether-bridged peptides is costly and cumbersome. Recently, we described a fermentative procedure to stabilize therapeutic peptides by introducing thioether bridges (77). This procedure involves plasmid- encoded nisin-modification enzymes expressed by Lactococcus lactis: a dehydratase (NisB) enzymatically dehydrates a Ser (or Thr) into a dehydroalanine (or dehydrobutyrine) (86). The generated dehydroresidues either enzyme (NisC)-catalysed or spontaneously react with a cysteine thiol group, which yields an intramolecular thioether bridge in the peptide (77, 160).

84 ACE-resistant, potent Ang(1-7) analog

We here present a prime example of the successful biological production of a stable and active thioether-cyclized peptide hormone, cyclized Ang-(1-7) (cAng-(1- 7)), by the nisin biosynthesis machinery. We demonstrate via in vitro and in vivo (rat) assays that the stability of cAng-(1-7) is strongly enhanced compared to its natural counterpart. We furthermore demonstrate that cAng-(1-7) has enhanced relaxing activity on precontracted rat aorta rings via the Ang-(1-7) receptor indicating enhanced receptor interaction.

Methods Chemicals . Ang-(1-7) and other chemicals were obtained as described (Supplemental Materials and Methods). cAng-(1-7) was designed, produced and purified as described (Supplemental Materials and Methods). Animals . Specified pathogen-free male Sprague Dawley rats (SD) (Harlan, Zeist, The Netherlands), weighing 350–450 gram were used. Prior to the experiment, the animals were housed together with free access of tap water and solid chow (Harlan, Zeist, The Netherlands) in a temperature and humidity controlled room and a 12/12h light/dark cycle. All protocols described were approved by the University of Groningen Committee for Animal Experimentation. Ang-(1-7) stability in ACE solution and plasma . Natural and cyclized Ang-(1-7) (50 µM) was incubated with angiotensin-converting enzyme (ACE) at 37°C over a period of 7 hour in a buffer containing 100 mM Tris-HCl pH 8.3, 300 mM NaCl, 10

µM ZnCl 2. ACE was used in final concentrations of 0.03 units/mL. Incubation in plasma occurred at concentrations of 100 µM natural or cAng-(1-7), in 15-fold diluted plasma, 20 mM phosphate buffer, pH 7.4. Analysis and Ang-(1-7) detection in these in vitro studies were carried out as described (Supplemental Materials and Methods). Ang-(1-7) stability in pig organ homogenates . Natural and cyclized Ang-(1-7) were incubated with pig homogenate of kidney cortex or liver (0.2 and 0.1 µmol Ang- (1-7)/mg homogenate, respectively) at 37°C. Incubation was performed in 20 mM phosphate buffered at pH 7.4 (cytosolic) and pH 5 (lysosomal) over a period of 3 h. The reaction was stopped by heat incubation (100°C, 5 min). Preparation of organ homogenates, analysis and Ang-(1-7) in vitro detection were performed as described (Supplemental Materials and Methods). Ang-(1-7) infusion studies in Sprague-Dawley rats . The in vivo kinetics of cAng-(1-7) and natural Ang-(1-7) were determined by continuous intravenous infusion of the peptides in male Sprague-Dawley rats of approximately 400 gram. Animals were kept under O 2/Isoflurane (1.5-2%) anaesthesia throughout the study. After a 30 min stabilization period with 0.9 % NaCl 1 mL/h, 100 µM cAng-(1-7) and 100 µM Ang-(1-7) in 0.9 % NaCl was infused with a constant rate of 1 mL/hour. Blood (200- 300 µl) was sampled via the carotid artery every 30 min for 2 hours. Creatinine levels in plasma were measured according to Bartels and Böhmer, 1971; n = 6. Further

85 Chapter 7

analysis and Ang-(1-7) detection in the in vivo studies were performed as described (Supplemental Materials and Methods). To ensure that enzyme saturation was not the determinant of the peptide clearances, the study was repeated once using a similar infusion dose of cAng-(1-7) (100 µM) but a 10-fold higher dose of natural Ang-(1-7) (1 mM). Indeed, the clearances of natural and cAng-(1-7) were not depending on the used dose of natural Ang-(1-7) indicating that enzyme saturation did not play a role. Vasodilation on precontracted aorta-rings . Arterial rings for organ bath experiments were prepared as described in Supplemental Materials and Methods. Before testing the vasodilating effects of the peptides the rings were precontracted to 50% of their maximum contraction level with 30 nM phenylephrine (PE). Natural or cyclized Ang-(1-7) were added cumulatively in a range of 0.1 nM to 1 µM. If appropriate Ang-(1-7) receptor blockers A-779 (191) and D-Pro 7-Ang-(1-7) (176) were applied 10 min prior to addition of 30 nM PE at a concentration of 0.1 µM. Vasodilation data are represented as % relaxation of 30 nM PE contraction. The vasodilating effect of the peptides was washed-out, followed by a full concentration- response curve of PE (0.1 nM to 10 µM), wash-out (two times) and again by establishing tonus by 30 nM PE to test the endothelium responsiveness of the rings. All preparations were sensitive to acetylcholine (10 µM) reducing this tonus by 51 ± 3 %. ACE activity . The inhibitory effect of cAng-(1-7), natural Ang-(1-7), Ang-(1-5) and AngI (Ang-(1-10)) on the activity of ACE was spectrophotometrically examined in rat plasma using hippuryl-L-histidyl-leucine (HHL) as a substrate (Cushman and Cheung, 1971). As a control, the specific ACE inhibitor captopril was used. Statistical analysis and calculation . All data were presented as mean ± S.E.M.. Statistical comparison of individual data was done using unpaired Student’s t-statistics except for the in vivo infusion study in which paired analysis was used. Repeated measures ANOVA with Bonferoni for post-hoc test was used to compare the curves of proteolytic stability, concentration-response curves of aorta ring dilatation and ACE activity. Differences were considered significant at p ≤ 0.05. Asterixes indicate: *, p < 0.05; **, p, 0.01 and ***, p < 0.001. Plasma clearance of Ang-(1-7) was calculated by dividing the infusion rate by the plasma concentration of the peptide.

Results Introduction of a thioether bridge in Ang-(1-7) linking amino acids 4 to 7 .The heptapeptide Ang-(1-7) consists of the residues H-Asp 1-Arg 2-Val 3-Tyr 4-Ile 5-His 6-Pro 7- OH. As the introduction of a thioether bridge requires the presence of a dehydratable residue (e.g. Ser) and a thiol-containing Cys, both corresponding amino acids were introduced at position 4 and 7, respectively, by site-directed mutagenesis. Figure 1 shows the chemical structure of the generated thioether-cyclized peptide Ang-(1-7).

86 ACE-resistant, potent Ang(1-7) analog

The thioether bridge was introduced by a fermentative method using L. lactis as a host organism containing the nisin modification enzymes. Maldi-TOF mass spectrometry confirmed successful dehydration (-18 Da mass shift) and the presence of a thioether bridge (no mass shift of +25 Da after incubation with 1-cyano-4-dimethylamino- pyridinium tetrafluoroborate), as explained previously in detail (77, 160).

Figure 1 . Hypothetical chemical structure of cAng-(1-7). The thioether bridge from position 4 to 7 (N-terminus on the left) is depicted in the D-L configuration expected on the basis of the DL configuration of the thioether bridges in nisin and other lantibiotics.

Resistance of cAng-(1-7) against proteolysis by purified angiotensin-converting enzyme (ACE) and plasma. In the circulation, endogenous Ang-(1-7) is rapidly catabolized, primarily by ACE (228). In the present study, cAng-(1-7) and natural Ang-(1-7) were incubated with purified ACE or plasma and the amount of remaining intact peptide was determined. The study showed that cyclization of Ang-(1-7) clearly resulted in enhanced proteolytic resistance. Whereas natural Ang-(1-7) was completely degraded within 90 min in a purified ACE solution, equi-amounts of cAng-(1-7) were completely resistant against ACE degradation over the full period of 4 hours ( p < 0.001) (Fig. 2A). After 4 hours incubation in plasma, natural Ang-(1-7) was largely degraded (5% remaining), whereas most of cAng-(1-7) remained intact (75% remaining) ( p < 0.001) (Fig. 2A). Resistance of cAng-(1-7) to proteolysis in pig organ homogenates . Besides ACE, other proteases are able to cleave Ang-(1-7) into smaller fragments (175). The proteolytic stability of cAng-(1-7) was compared to natural Ang-(1-7) using homogenates of pig-kidney and -liver, organs rich of a wide range of proteases. In kidney homogenate at pH 7.4, natural Ang-(1-7) was fully degraded within 3 hours, whereas more than 80% of the initial amount of cAng-(1-7) remained intact ( p <

87 Chapter 7

0.001). Similar results were obtained when both peptides were incubated with liver homogenate ( p = 0.002) (Fig. 2B). ACE-inhibition by lisinopril did not affect the degradation rate of Ang-(1-7) in kidney and liver homogenate indicating that other proteases than ACE are responsible for the degradation in these organs. Also at lysosomal pH 5.0, the degradation of cAng-(1-7) was slower than of natural Ang-(1-7) (data not shown). These data consistently demonstrate strongly enhanced resistance of cAng-(1-7) against proteolytic degradation in protease-rich organ homogenates compared to natural Ang-(1-7).

Figure 2AB . Enhanced proteolytic resistance of cAng-(1-7). Proteolytic resistance was measured of natural ( ∆, ○, □, ∇) and cyclized Ang-(1-7), ( ▲, ●, ■,▼), against ACE (A: ∆, ▲), plasma (A: ○, ●), liver (B: □, ■) and kidney homogenate (B: ∇, ▼) at pH 7.4. Each point represents the mean ± S.E.M., generated from at least 3 separate experiments.

Long lasting in vivo survival of cAng-(1-7). The in vivo survival of cAng-(1-7) relative to natural Ang-(1-7) was determined by analyzing the drug concentration in plasma of anesthetized Sprague-Dawley (SD) rats during a continuous intravenous infusion of both peptides. Cyclization resulted in a strong, 34 ± 2 -fold, reduction of plasma clearance (P = 0.01) (Fig. 3). Hence the high resistance of cAng-(1-7) against proteolytic degradation, as shown in vitro, apparently results in a large increase in survival rate of the drug in vivo . This makes cAng-(1-7) much more attractive for therapeutic application than the natural Ang-(1-7). Creatinine level in the plasma was 49.82 ± 1.33 µM at 0 minutes of infusion, and 50.08 ± 1.38 µM at 120 min. Hence renal functioning was not affected by the infusion.

88 ACE -resistant, potent Ang(1 -7) analog

Figure 3. Long lasting in vivo survival of cAng -(1 -7). Plasma clearance (ml/min) of natural (ope n bar) and cAng -(1 -7) (filled bar) in Sprague Dawley rats under anesthesia during a constant intravenous infusion of both peptides. Each bar represents the mean ± S.E.M. (n = 5).

Potent dilating effect of cAng -(1 -7) on precontracted aorta rings of SD rats . Thoracic aortic rings with intact endothelium were precontracted with phenylephrine (PE) and subsequently relaxation of the established muscle tone was measured with cumulative additions of cAng -(1 -7) or natural Ang -(1 -7) (Fig. 4). Both natural and cAng -(1 -7) displayed a significant vasodilating effect, indicating that the introduction of a thioether bridge in Ang -(1 -7) does not impede its vasodilative properties. Moreover, the observed vasodilating effects by cAng-(1 -7) occurred at a lower concentration than by natural Ang -(1 -7).

Figure 4. Enhanced relaxation activity of cAng -(1 -7). Relaxation of SD rat aorta rings during cumulative addition of natural ( ○) or cyclized Ang -(1 -7) ( ●) or in control incubation (×), after precontraction with 0.3 µM phenylephrine, related to a vehicle experiment. Each point represents the mean ± S.E.M. generated from at least seven (natural) and eleven (cyclized) separate experiments.

89 Chapter 7

A 10 -fold lower amount of cAng -(1 -7) was needed to achieve 10% dilatation (0.3 nM cAng -(1 -7) versus 3 nM natural Ang -(1 -7)) while even 100 -fold less cAng -(1 -7) was required to dilate the precontracted aortic rings for 40% (0.03 µM cAng -(1 -7) and 3 µM natural Ang -(1 -7)). Additionally, the maximal degree of vasodilation that was obtain ed was almost doubled with cAng -(1 -7) compared to natural Ang -(1 -7). At the highest concentration tested, 1 µM, cAng -(1 -7) caused relaxation of 63 ± 5% whereas 1 µM natural Ang -(1 -7) resulted in 33 ± 9% relaxation, only (P < 0.01). After the cumulative ad dition of cAng -(1 -7) and natural Ang -(1 -7), aortic rings were put back into the PE solution only, to determine the wash -out rate of the induced dilation by the Ang -(1 -7). With both peptides, the contraction was rapidly back to pre - challenge level with a wa sh -out half -life of 2.1 ± 0.3 min and 1.7 ± 0.3 min of cAng - (1 -7) and natural Ang -(1 -7) respectively (n=6, not significantly different). This is a strong indication that the vasodilating effect of cAng -(1 -7) is a result of ligand/receptor interaction. The involvement of the Ang -(1 -7) receptor on the vasodilating properties of cAng -(1 -7) was examined using the Ang -(1 -7) receptor antagonists D -Ala 7-Ang -(1 -7) (A -779) and D -Pro 7-Ang -(1 -7) at 0.1 µM (Fig. 5). In line with recent findings on aorta -rings of Spragu e-Dawley rats ( 191 ), natural Ang -(1 -7) dilation was abolished by D-Pro 7-Ang -(1 -7) and insensitive for A -779 antagonism (Fig. 5A). cAng -(1 -7) - induced dilation was fully blocked by D -Pro 7-Ang -(1 -7), providing evidence for the involvement of the Ang -(1 -7) rec eptor in the vasorelaxant properties of cAng -(1 -7). Interestingly, at low concentrations of cAng -(1 -7), its activity was also antagonized by A-779 (Fig. 5B).

Figure 5. Effects of Ang -(1 -7)-receptor -antagonists .Vasodilating effect of natural (A ) and cyclized (B) Ang -(1 -7) in the presence of Ang -(1 -7) -receptor -antagonists A -779 or D -Pro 7-Ang -(1 -7) (0.1 µM) compared to control (control curves are similar to the ones in figure 3). A: interrupted line:Ang -(1 -7); ∆: Ang -(1 -7) + A -779; ∇: Ang -(1 -7) + D -Pro 7-Ang -(1 -7). B: interrupted line: cAng -(1 -7); ▲: cAng - (1 -7) + A -779; ▼: cAng -(1 -7) + D -Pr o7-Ang -(1 -7). Each point in the inhibition experiment represents the mean ± S.E.M. generated from at least six (D -Pro 7-Ang -(1 -7)) and eight (A -779) separate experiments.

90 ACE -resistant, potent Ang(1 -7) analog

Experiments on the in vivo vasodilating effect of cAng -(1 -7) and its in vivo receptor specificity are under way. No effect of cAng -(1 -7) on ACE activity . Besides being a substrate of the N - terminal part of ACE for cleavage into Ang -(1 -5), at high concentrations natural Ang - (1 -7) inhibits ACE activity ( 32 ) by direct binding to the C -terminal domain of ACE. The inhibitory effect of cAng -(1 -7) on the activity of ACE was analyzed using hippuryl -histidyl -leucine as a substrate. Figure 6 shows that natural Ang -(1 -7) inhibits the activity of ACE at concentrations in the micromolar range, with an IC 50 around 10 µM. The breakdown product of natural Ang -(1 -7), Ang -(1 -5), at concentrations up to 0.1 mM did not induce ACE inhibition whereas the ACE -inhibitor captopril was very effective (EC 50 = 1 nM). In case of cAng -(1 -7), ACE activity was not inhibited . Even at a concentration 10 -fold higher than the IC 50 for natural Ang -(1 -7), ACE activity was unaffected. These data demonstrate that cAng -(1 -7) has largely or completely lost its capacity to act as an ACE inhibitor due to the cyclization.

Figure 6 . cAng -(1 -7) has no or little ACE inhibiting activity. Inhibitory effect of natural Ang -(1 -7) (○), Ang -(1 -5) ( ◊), cAng -(1 -7) ( ●), and captopril (×) on the activity of ACE. Each point represents the mean ± S.E.M. generated from 3 separate experime nts. The curve of cAng -(1 -7) is significantly different from that of natural Ang -(1 -7) (p < 0.05).

Discussion Ang -(1 -7) has a variety of interesting properties for therapeutic purposes. However, the very rapid breakdown in the circulation ( 228 ) makes the natural peptide unsuitable as a drug. Therefore, we aimed at introducing a thioether bridge that stabilizes the peptide without causing loss of activity. The role of the Ang -(1 -7) residues in receptor binding and activity has not been elucidated. Except f or Norleucine3 ( 165 ) and two Ang -(1 -7) antagonists with a D -Ala (A -779) or D -Pro at

91 Chapter 7

position 7 (174, 176) the effect that exchanging residues has on the peptide’s functionality is not known. We here clearly demonstrated that Ang-(1-7) with a thioether bridge between position four and seven has strongly increased stability and significantly increased activity. In vivo , circulating Ang-(1-7) is rapidly inactivated by lung ACE, which cleaves off two amino acids at the C-terminus, resulting in Ang-(1-5). By introducing a thioether bridge between position four and seven of the peptide, this major catabolic pathway was blocked for the thioether Ang-(1-7) analogue. Ring introduction also reduced peptide catabolism by other enzymes than ACE as indicated by the enhanced stability in liver and kidney homogenates. Since the extent of catabolism largely determines the in vivo clearance and thus the exposure time of the body to the peptide a major improvement was expected by cyclization of the peptide. Indeed, the body clearance of cAng-(1-7) was more than 30-fold lower than of natural Ang-(1-7). Several studies are known that aim at enhancement of the endogenous levels of Ang-(1-7) for therapeutic benefit. Since it was established that Ang-(1-7) is mainly degraded by ACE, approaches to increase Ang-(1-7) levels have been focused on inhibiting this protease, often in combination with blockers for the AngII receptor AT 1. However, ACE inhibition also results in lower AngII levels and, subsequently, less available substrate for generation of Ang-(1-7). With the discovery of ACE2, a second important protease in the RAS that generates Ang-(1-7) from AngI via Ang-(1-9) and from AngII directly (218), treatments have lately been focusing on increasing the ACE2 concentration. Techniques such as vector-mediated overexpression of ACE2 in SHR rat models demonstrated a protective effect of ACE2 on high blood pressure and cardiac pathophysiology induced by hypertension (41). Until now, administration of exogenous ACE2 is hampered by suboptimal solubility and activity. Very recently, promising effects in SHR rats have been shown using small-molecule ACE2 activators (58). Unfortunately, the contribution of ACE2 to cardiovascular physiology and disease also lies in depletion of the AngII pool rather than the synthesis of Ang-(1-7) (54). Exogenous Ang-(1-7)-induced attractive effects in vivo (85, 6, 112, 95, 123, 164, 165) while no signs of toxicity were observed in phase I/II clinical studies using Ang- (1-7) for bone marrow protection during anticancer therapy (165). However, the rapid degradation of the natural peptide makes it necessary to use high doses. A small but significant bioavailability of oral Ang-(1-7) seems possible using liposomes or cyclodextrin for encapsulation (179). Ang-(1-7) fused to a furin-cleavable protein in transgenic rats (178) provides a good analytical alternative, but will be more complicated when patient administration is desired. Only one Ang-(1-7) agonist, the nonpeptide AVE 0991, has been described. However, as reviewed (180), the value of this compound for therapeutic application is yet unknown, since only few of the benefits of Ang-(1-7) have been established for AVE 0991 too and toxicity data are not available. In general, all therapeutic Ang-(1-7) shortcomings listed above indicate

92 ACE-resistant, potent Ang(1-7) analog

that a stable and functional analogue of Ang-(1-7) itself is the ideal compound for – cardiovascular- therapy. The proteolytic resistant cAng-(1-7) presented in the present study proved to be a highly active vasodilator on the isolated aorta ring of the SD rat. Both the affinity and maximal effect of cAng-(1-7) were shown to be higher than that of its natural counterpart, despite the fact that the introduction of a thioether bridge entails a drastic change of the C-terminal part of Ang-(1-7). Given the fact that the introduction of a thioether bridge from position four to seven enhances the activity and since known antagonists of Ang-(1-7) all contain a mutation at position seven, it is evident that not only the amino acids at those positions, but also the established conformation plays a crucial role in the activity. It is known that peptide cyclization may result in a more constrained entity with reduced conformational freedom, which may confer a higher receptor affinity and/or stronger activity. The enhanced effect of cAng-(1-7) points into this direction. As reviewed (64), multiple Ang-(1-7) receptors seem to exist. Several studies have shown that Ang-(1-7) mainly acts via the Mas receptor and that its activity can be fully prevented by A-779 and D-Pro 7-Ang-(1-7). This suggests that both compounds are full Mas receptor antagonists (176b). However, in cerebral arteries of the canine brain, endothelium-dependent relaxation by Ang-(1-7) was not blocked by A-779 (47). Also in another study Ang-(1-7) activity was not prevented by A-779 (191). In precontracted aorta rings of SD rats, the Ang-(1-7) induced vasodilation appeared mediated via an A-779-insensitive but D-Pro 7-Ang-(1-7)-sensitive receptor (191). In the present study, we compared the effect of the two antagonists on the vasodilation of both natural and cAng-(1-7) in precontracted aorta-rings of SD rats. The Ang-(1-7) receptor antagonist A-779 did not prevent the vasodilation of natural Ang-(1-7) and only partially the vasodilation of cAng-(1-7). On the other hand, the antagonist D-Pro 7-Ang-(1-7) completely prevented the vasodilation by both peptides. These results indicate that after cyclization, the peptide maintained the receptor profile of the natural Ang-(1-7). Like the dilatation curve of natural Ang-(1-7) in SD-rat aorta rings.shown by Silva and co-workers, the curves of natural.and cyclic Ang-(1-7) in the present study are not the result of interaction with a single receptor or a single binding state, indicated by the lack of steepness of the curve. The observed cAng-(1-7)-induced relaxation spans a concentration range of at least four orders of magnitude, which is much broader than the two orders expected for basic agonist-receptor interaction. Together with the observed antagonism by A-779 at low cAng-(1-7) concentrations these results suggest interaction with two receptor binding sites or receptor subtypes, one inducible at low concentrations of cAng-(1-7), sensitive to both A-779 and D- Pro 7-Ang-(1-7) and another one inducible with natural Ang-(1-7) and high

93 Chapter 7

concentrations of cAng-(1-7), sensitive to D-Pro 7-Ang-(1-7) but not to A-779. Further investigations are needed to warrant a final conclusion. Natural Ang-(1-7) serves as a cleavable substrate for the N-terminal-domain of ACE and as an inhibitor of cleavage by the C-terminal domain of ACE, which cleaves Angiotensin-I (32). We demonstrated that in contrast to natural Ang-(1-7), cAng-(1-7) in the micromolar range has no inhibitory effect on activity of the C-terminal part of ACE. Many C-domain ACE inhibitors contain a C-terminal proline, which is thought to be an important residue in the interaction of ACE inhibitors with the enzyme (145). In cAng-(1-7), this residue was replaced by a thioether bridged amino acid, thereby apparently reducing its affinity for the binding site. This lack of ACE-inhibitory activity of cAng-(1-7) indicates that a more specific therapeutic agent is obtained by the ring introduction. Although already two decades ago suggested as a tool to incorporate unnatural amino acids in therapeutic peptides (183), the use of the lantibiotic enzymes has now, for the first time, resulted in the production of an active and stable therapeutic agent. Recent developments show that multiple thioether bridges can be incorporated in designed nonlantibiotic peptides (160). Therefore, the ever growing knowledge on the lantibiotic modification enzymes will only expand the possibilities to use them in stabilizing peptides with interesting biological properties. The cyclized Ang-(1-7) described in the present study combines strongly enhanced proteolytic resistance with improved activity apparently due to better receptor interaction compared to its natural counterpart. This makes cAng-(1-7) highly interesting for therapeutic purposes.

Acknowledgements. We thank Boukje Mulder-Bosman and Christina E. Reitzema-Klein for purifying cyclized Ang-(1-7) and additional animal studies, respectively. Annie van Dam is acknowledged for the LC-MS analyses and Anton J. M. Roks for fruitful discussions.

Footnotes Abbreviations used : Ang, angiotensin; RAS, renin-angiotensin system; ACE, angiotensin-converting enzyme; cAng-(1–7), 4–7 thioether-bridgedangiotensin-(1–7); SD, Sprague-Dawley; PE, phenylephrine; A-779, D-Ala7-Ang-(1–7); AVE 0991, 5- formyl-4-methoxy-2-phenyl1-[[4-[2-ethylaminocarbonylsulfonamido)-5-isobutyl-3- thienyl]-phenyl]-methyl]-imidazole . Supplemental material for this article may be found at http://jpet.aspetjournals.org

94 Summary, general discussion and perspectives

This thesis describes the successful biological production and secretion of thioether- stabilized non-lantibiotic peptides. The lantibiotic modification- and transport enzymes NisBTC and LtnM2T involved in the synthesis of the lantibiotics nisin and lacticin 3147, respectively, were exploited for the introduction of thioether bridges in nonlantibiotic peptides. Exploiting the nisin modification enzymes NisB and NisC, we were able to demonstrate for the first time the posttranslational introduction of a thioether bridge in a therapeutic peptide, an analog of angiotensin-(1-7). This therapeutic peptide variant has a significantly improved stability and the effectivity of its interaction with the angiotensin-(1-7) receptor is even enhanced (Chapter 7). Angiotensin-(1-7) plays an important role in counterbalancing many actions of angiotensin II, which is a regulator of fluid and sodium balance, haemodynamics, cellular growth and cardiovascular remodelling (50). Hence, stabilized angiotensin-(1- 7) can be a potential therapeutic agent in the control of a range of diseases. Also other methods are used for stabilization of therapeutic peptides and thereby prolonging the half-life of these peptides in the blood. Examples are coupling of peptide drugs to polyethyleneglycol, by modifications such as glycosylation, N-terminal acetylation or C-terminal amidation or by making peptide analogs comprising nonproteinogenic amino acids. While there are hundreds of medically highly important therapeutic peptides, the pharmaceutical market of already a single therapeutic peptide can have a size of over a billion dollar. Consequently, stabilization of already FDA-approved therapeutic peptide hormones and development of new effective stabilized peptides has a tremendous potential. Additionally, via development of a genetically designed peptide library a huge number of biologically produced peptide variants can easily be screened. Utilization of lantibiotic enzymes in vivo for introduction of thioether bridges in peptides is therefore a promising technology.

Mechanistic aspects of the enzymes involved in biosynthesis of lantibiotics . The nisin biosynthesis machinery NisBTC proved to be highly versatile and can be used for the introduction of thioethers in a broad spectrum of nonlantibiotic peptides (Chapter 3, 157, 158). A number of designed nonlantibiotic peptides were Chapter 8

efficiently dehydrated by NisB but failed to be cyclized by NisC, indicating that the substrate specificity of NisB might be more relaxed than that of NisC. Threonines were more easily dehydrated than serines. Furthermore successful dehydration of threonines/serines seems to be influenced by the flanking residues. Hydrophobic flanking residues on one or both sides may favour dehydration, whereas the simultaneous presence of hydrophilic flanking residues on both sides seems to disfavour dehydration (158). Dehydroalanines are highly reactive and can spontaneously form a lanthionine when reacting with a cysteine. On the other hand, methyllanthionines are not formed spontaneously. A methyllanthionine in a microbially produced peptide by utilization of NisBTC is therefore definitely formed by its dehydration by NisB and the subsequent cyclization by NisC. When NisC fails to cyclize a peptide, thioether formation between the dehydrobutyrine and the cysteine can be induced by incubation of the dehydrated peptide at high pH e.g. pH 10. However, non-enzymatic ring closure can result in a mixture of stereoisomers (19). Thioether linkages can also be introduced by base-assisted sulphur extraction from disulfide bridged peptides (51). Under mild alkaline conditions such chemical introduction of a thioether bridge can result in more than one stereo-isomer (LD, DL, LL and DD). Different isomers can have different retention times when analyzed by HPLC. When biologically produced thioether-stabilized peptides were purified by HPLC, in all studied cases only one peak could be detected. However, when thioether rings were introduced chemically, often more peaks were seen. This suggests that by biological introduction of a thioether bridge, only one isomer, probably the DL variant, is formed. Proof of the presence of only a DL isomer, the only isomer found for the few lantibiotics studied in this respect, can be obtained by Edman degradation of peptides that are desulfurized (52). For lantibiotic-enzyme-cyclized non-lantibiotic peptides, identifying the formed isomer is still one of the future experiments. Not only small thioether-bridged peptides can be synthesized by utilization of NisBTC, also a more complicated substrate peptide with the sequence ITPGCKATVECKITGPCKATVECK can be successfully modified to a peptide with four thioether linkages. Also introduction of intertwined thioether rings is possible thanks to the stereo- and regiospecificity of NisC (160). Moreover, a substrate peptide with the length of 28 amino acids and with the sequence DSRWARVALIDSQKAAVDKAITDIAEKL was produced, with selective dehydration of the threonine at position 22 (157). Especially for longer peptides and peptides with more thioether rings, biological production may have an advantage in reducing the costs and time of synthesis. Also the lacticin 3147 enzymes LtnT and LtnM2 were successfully used in vivo for the introduction of thioether bridges in nonnatural peptide substrates. However, this system is more complex to examine. Secretion of the studied peptides appears to be the bottleneck. Before translocation of the nonnatural substrate by LtnT, the peptide

96 Summary, general discussion and perspectives

first has to be processed. The LtnA2 leader is intracellularly cleaved off by the same LtnT enzyme. It is not clear whether the processing or the translocation itself is the restriction in this process (Chapter 4). When introduction of thioether bridges in peptides is performed with lantibiotic enzymes in vitro , this processing/ translocation drawback can be avoided. In 2004 a paper in Science was published that describes the reconstitution of lacticin M in vitro for the synthesis of lacticin 481 (226). The enzyme LctM and the substrate LctA were both produced in E. coli. When the His-tagged substrate peptide LctA and the modifying enzyme LctM are incubated in vitro in the presence of adenosine triphosphate (ATP) and Mg 2+ , dehydration and cyclization of the substrate occurs effectively. Though LctM doesn’t display an evident ATP-binding domain, ATP is necessary for functionality of LctM (22). Also in vitro reconstitution of the two-component haloduracin has been successful. After incubation of the prepeptides HalA1 and HalA2 with the modifying enzymes HalM1 and HalM2, respectively, a bioactive haloduracin is formed (121). Concerning the modifying enzymes NisB and NisC, only NisC has successfully been reconstituted in vitro . Incubation of dehydrated prenisin with NisC in vitro , results in bioactive nisin after removal of the nisin leader (106). With the powerful in vitro thioether modification system, the substrate specificity and mechanistic aspects of the LctM 481 enzyme were further explored. Just like NisB, LctM has a high substrate promiscuity. LctM can dehydrate a range of nonlantibiotic peptides when attached to the leader peptide (23, 103, 147). While NisB disfavors hydrophilic amino acid residues, especially negatively charged ones, surrounding its dehydratable residues, this is not the case for LctM (23). Furthermore, semi-synthetic LctA derivates with nonproteinogenic amino acids like β-amino acids, D-amino acids and N-alkyl amino acids, derived by expressed protein ligation, are successfully modified by LctM (104, 105, 234). In the case of in vivo utilization of lantibiotic enzymes this incorporation of nonproteinogenic amino acids before posttranslational (enzymatic) modification will be very complicated if not impossible. For efficient modification of the structural peptides the precence of the leader peptide is essential. However, it has been demonstrated that presenting the leader in trans, not attached to the substrate, still leads to modification of the structural peptide (102, 147). Moreover, other interesting mechanistic aspects of LctM were revealed. LctM phosphorylates the serines and threonines of its substrate prior to the dehydration of these residues (22, 231). An LctM T405A mutant was not affected in phosphorylation of serines and threonines of its substrate but was hampered in the phosphate- elimination step and therefore lost the ability of dehydration of its substrate. This mutant shows to be a highly efficient kinase for a broad range of peptide substrates with serines when preceded by the lacticin leader (232). Recently, LctM-mediated dehydration was demonstrated to occur via a distributive mechanism (96).

97 Chapter 8

Furthermore a tendency to an apparently directional behavior of the LctM enzyme was revealed (96). LctM has a high, though not strict, propensity for an apparent N-to-C directionality. When the leader peptide is present in trans, dehydrations were nondirectional. Also intermediates were found which were not completely dehydrated but already contained rings within their N-terminal region. This latter finding suggests that the dehydration - and the cyclization activities of LctM can alternate. This alternating feature was also observed for NisB- and NisC activity, outlined in this thesis in Chapter 4. The latter data were supported by data obtained by Lubelski et al., who also suggested that NisB and NisC are acting in an N- to C- direction (114). Chapters 4 and 5 of this thesis demonstrate that the nisin synthetase complex and the lacticin 3147 synthetase complex can be dissected and that enzymes still can be functional without the presence of other lantibiotic enzymes. For the nisin biosynthetic enzymes it was demonstrated that NisT can transport peptides when preceded by the nisin leader in the absence of all other nisin enzymes. The same is true for the dehydratase activity of the nisin dehydratase NisB. In the absence of other nisin enzymes, NisB-dehydrated peptides were exported via the Sec pathway when the nisin leader was preceded by a Sec signal sequence. When the same peptide substrate was preceded by a Tat signal sequence and exported via the Sec pathway, an identical level of dehydration of the peptide was obtained compared to the case of dehydration and export via NisBT. Moreover prenisin without preceding Sec or Tat signal was intracellularly fully modified by NisB and NisC in the absence of NisT, which precludes the necessity of either NisT or Sec action. In view of the above it is difficult to comprehend the proposed enzyme- complex-dependent NisT-driven-modification working model for nisin biosynthesis by Lubelski et al. 2009 (114). In this model it was speculated that NisT, that uses ATP for the transport, is also involved in pulling the substrate through the active sites of the modification enzymes NisB and NisC.

Production and transport of posttranslationally modified peptides. The ABC transporter NisT is capable of translocating a broad range of substrates. The substrates that have been examined up to now, varied in size from 6 amino acids up to 47 amino acids, were linear, dehydrated and/or had various (intertwined) thioether rings. Also the leader peptide itself, without any propeptide was translocated (158). Although the NisT transporter is very successful in translocation of nonnatural substrate peptides, not all tested peptides are transported equally well via NisT. Whether these differences in transport efficiency are due to size, polarity, hydrophobicity or conformation of the peptides is not established yet. Furthermore, it was noticed that fusion of peptides of interest behind the first two rings of nisin improved the production level significantly (161). This improvement can be caused by

98 Summary, general discussion and perspectives

the autoinducing effects of rings A and B of nisin (91) or may be based on the positive effect of these rings on translocation via NisT. Another alternative route for transport of modified peptides in L. lactis is the Sec pathway, which is successfully demonstrated in Chapters 5 and 6 of this thesis. For transport of peptides or proteins via the well studied Sec pathway substrates have to be preceded by a Sec signal peptide. It was demonstrated that an N-terminal fusion to the nisin leader of a signal peptide up to 44 amino acids long didn’t prevent dehydration by NisB and cyclization by NisC. This feature has a high potential. For instance, N-terminal extensions to the leader peptide can be a promising tool in the development of combining the thioether modification technology with phage display. With these united techniques huge libraries of thioether-constrained peptides can be made and screened for dedicated purposes. Furthermore, the Twin arginine signal sequence of YwbN still targets the peptide to the Sec pathway in L. lactis . This targeting signal sequence had a positive effect on the dehydration efficiency possibly thanks to a reduced translocation efficiency of the peptide. By altering the signal sequence a more optimal system may be created with improved modification and transport efficiency. Although the Sec pathway can successfully transport linear dehydrated peptides and an azurin peptide fragment with a small methyllanthionine, the route was not fuctional for producing prenisin. It was estimated that the size of the SecY pore in L. lactis is too small for translocation of prenisin. In E. coli it has been demonstrated that the translocon SecYEG could also transport the polypeptide proOmpA with disulfides or labeled with a bulky fluorescent probe (33, 206). Possibly in the future improved transport via the Sec pathway in L. lactis can be obtained by making SecY mutants with a a more relaxed fitting that are able to transport also more bulky peptides. Studies on lantibiotic biosynthesis systems for nisin, subtilin and nukacin ISK1 revealed that the modification enzymes and transporters are arranged in multimeric membrane-associated enzyme complexes (75, 133, 189). A nisin synthetase complex composed of two NisT molecules, two NisC molecules and one NisB molecule in the WT strain has been suggested. Whether this stoichiometry of enzymes is also present in the two plasmid production strain that we used for the production of modified peptides has not yet been determined. The optimal stoichiometry of substrate, modification enzymes and transporter probably results from a fine-tuned balance in expression. Raising the intracellular concentration of NisB and thereby changing the ratio substrate: NisB, led to an improved dehydration level of the secreted peptides (72, Chapter 5). Additionally, in a study of van den Berg van Saparoea et al. a channeling of prenisin of NisB to NisT was suggested to obtain efficient transport of prenisin. In the presence of the enzymes NisB and NisC a higher production level of prenisin was found when compared with the production level of the dehydrated prepeptide NisA in the absence of NisC. Moreover, in the absence of NisB and NisC

99 Chapter 8

the production level of unmodified precursor NisA was even lower (211). These data underline the presence of a multimeric enzyme complex and the importance thereof for optimal production of prenisin. However, in the absence of the cyclase NisC thioether rings are absent. The presence of thioether rings might impose a structure which is transported more efficiently by NisT. The latter possibility is revealed by data outlined in Chapter 4 of this thesis. Furthermore, rings A and B are involved in induction of the Pnis promoter. Even a low extent of removal of the leader by peptidases can have an impact on production of prenisin by autoinduction. The technology described in this thesis is unique, especially with respect to the regio-, stereo- and chemospecific synthesis. After the production of the successful thioether angiotensin-(1-7), challenging applied perspectives are now the selection of new relevant candidate substrate peptides for thioether stabilization. Improvement of production levels of peptides of interest may be obtained by several adjustments of the biological production system. Examples for reaching the latter goal are: using another production host / host engineering, raising the amount of the intracellular substrate and modification enzymes by raising the copy numbers of the encoding genes, adjusting the nisin leader or the NisT transporter for better transport or improving growth conditions of the producing culture. From the more fundamental point of view, the lantibiotic enzymes constitute a fascinatingly rich and diverse field of research. Important new mechanistic insights may follow from the crystallization of lantibiotic enzymes other than NisC. In vitro reconstitution of NisB activity, which has not been attained despite many efforts in several laboratories, will hopefully be realized one day and facilitate further mechanistic studies.

100 References

1. Allgaier, H., G. Jung, R. G. Werner, U. Schneider, and H. Zahner . 1986. Epidermin: sequencing of a heterodetic tetracyclic 21-peptide amide antibiotic. Eur. J. Biochem. 160 :9-22.

2. Arnau, J., E. Hjerl-Hansen, and H. Israelsen. 1997. Heterologous gene expression of bovine plasmin in Lactococcus lactis . Appl. Microbiol. Biotechnol. 48 :331-338.

3. Apiyo, D. and P. Wittung-Stafshede. 2005. Unique complex between bacterial azurin and tumor-suppressor protein p53. Biochem. Biophys. Res. Commun. 332 :965-968.

4. Appleyard, A. N., S. Choi, D. M. Read, A. Lightfoot, S. Boakes, A. Hoffmann, I. Chopra, G. Bierbaum, B. A. Rudd, M. J. Dawson, and J. Cortes. 2009. Dissecting structural and functional diversity of the lantibiotic mersacidin. Chem. Biol . 16 :490-498.

5. Bartels, H., and M. Böhmer . 1971. Micro-determination of creatinine. Clin Chim Acta 32: 81-85.

6. Benter, I. F. , C. M. Ferrario, M. Morris, and D. I. Diz . 1995. Antihypertensive actions of angiotensin-(1-7) in spontaneously hypertensive rats. Am J Physiol 269: H313-H319.

7. Bierbaum, G. and H. G. Sahl . 1985. Induction of autolysis of staphylococci by the basic peptide antibiotics Pep 5 and nisin and their influence on the activity of autolytic enzymes. Arch. Microbiol. 141: 249-254.

8. Bierbaum, G., C. Szekat, M. Josten, C. Heidrich, C. Kempter, G. Jung, and H. G. Sahl . 1996. Engineering of a novel thioether bridge and role of modified residues in the lantibiotic Pep5. Appl. Environ. Microbiol. 62: 385-392.

9. Bierbaum, G . 1999. Genetics of lantibiotic biosynthesis, in Amino acids, peptides, porphyrins, alkaloids (Kelly, J. W., Barton, D., Nakanishi, K., and Meth-Cohn, O., Eds.) pp 275-301, Elsevier Science Publishers B.V., Amsterdam.

10. Bierbaum, G. and H. G. Sahl . 2009. Lantibiotics: mode of action, biosynthesis and bioengineering. Curr. Pharm. Biotechnol. 10: 2-18.

11. Bhunia, A. K., M. C. Johnson, and B. Ray. 1987. Direct detection of an antimicrobial peptide of Pediococcus acidilactici in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. J. Ind. Microbiol. 2:319–322.

Chapter 9

12. Bolotin, A., P. Wincker, S. Mauger, O. Jaillon, K. Malarme, J. Weissenbach, S. D. Ehrlich, and A. Sorokin. 2001. The complete genome sequence of the lactic acid bacterium Lactococcus lactis ssp. lactis IL1403. Genome Res. 11 :731-753.

13. Breukink, E., I. Wiedemann, C. van Kraaij, O. P. Kuipers, H. Sahl, and B. de Kruijff. 1999. Use of the cell wall precursor lipid II by a pore-forming peptide antibiotic. Science 286:2361-2364.

14. Breukink, E., van Heusden, H. E., Vollmerhaus, P. J., Swiezewska, E., Brunner, L., Walker, S., Heck, A. J., and de Kruijff, B. 2003. Lipid II is an intrinsic component of the pore induced by nisin in bacterial membranes. J. Biol. Chem. 278 : 19898-19903.

15. Breukink, E. 2006a. A lesson in efficient killing from two-component lantibiotics. Mol. Microbiol. 61:271-273.

16. Breukink, E. and B. de Kruijff. 2006b. Lipid II as a target for antibiotics. Nat. Rev. Drug Discov. 5:321-332.

17. Brötz, H., M. Josten, I. Wiedemann, U. Schneider, F. Götz, G. Bierbaum, and H. G.Sahl . 1998. Role of lipid-bound peptidoglycan precursors in the formation of pores by nisin, epidermin and other lantibiotics. Mol. Microbiol. 30: 317-327.

18. Buist, G., J. Kok, K. J. Leenhouts, M. Dabrowska, G. Venema, and A. J. Haandrikman . 1995. Molecular cloning and nucleotide sequence of the gene encoding the major peptidoglycan hydrolase of Lactococcus lactis , a muramidase needed for cell separation. J. Bacteriol. 177: 1554-1563.

19. Burrage, S., T. Raynham, G. Williams, J. W. Essex, C. Allen, M. Carno, V. Swali, and M. Bradley . 2000. Biomimetic synthesis of lantibiotics. Chem. Eur. J. 6:1455-1466

20. Chakicherla, A., and J. N. Hansen. 1995. Role of the leader and structural regions of prelantibiotic peptides as assessed by expressing nisin-subtilin chimeras in Bacillus subtilis 168, and characterization of their physical, chemical, and antimicrobial properties. J. Biol. Chem. 270 :23533-23539.

21. Chatterjee, C., M. Paul, L. Xie, and W. A. van der Donk. 2005a. Biosynthesis and mode of action of lantibiotics. Chem. Rev. 105:633-684.

22. Chatterjee, C., L. M. Miller, Y. L. Leung, L. Xie, M. Yi, N. L., Kelleher., and W. A. van der Donk. 2005b. Lacticin 481 synthetase phosphorylates its substrate during lantibiotic production. J. Am. Chem. Soc. 127: 15332-15333 .

23. Chatterjee, C., G. C. Patton, L. Cooper, M. Paul, and W. A. van der Donk. 2006. Engineering dehydro amino acids and thioethers into peptides using lacticin 481 synthetase. Chem. Biol. 13:1109-1117.

24. Chen, P., J. Novak, M. Kirk, S. Barnes, F. Qi,.and P. W. Caufield . 1998. Structure-activity study of the lantibiotic mutacin II from Streptococcus mutans T8 by a gene replacement strategy. Appl Environ Microbiol 64: 2335-2340.

25. Chen, P., F. X. Qi, J. Novak, R. E. Krull, and P. W Caufield . 2001. Effect of amino acid substitutions in conserved residues in the leader peptide on biosynthesis of the lantibiotic mutacin II. FEMS Microbiol. Lett. 195: 139-144

102 References

26. Coghlan, P. A., and C. J. Easton . 1999. b-nitro-a-amino acids as latent a,b-dehydro-a-amino acid residues in peptides. Tetrahedron Lett. 40 :4745-4748.

27. Cotter, P. D., C. Hill, and P. P. Ross. 2005a. Bacterial lantibiotics: strategies to improve therapeutic potential. Curr. Protein Pept. Sci. 6: 61-75.

28. Cotter, P. D., P. M. O'Connor, L. A. Draper, E. M. Lawton, L. H. Deegan, C. Hill, and R. P. Ross. 2005b. Posttranslational conversion of L-serines to D-alanines is vital for optimal production and activity of the lantibiotic lacticin 3147. Proc. Natl. Acad. Sci. U. S. A. 102:18584-18589.

29. Cotter, P. D., L. A. Draper, E. M. Lawton, O. McAuliffe, C. Hill, and R. P. Ross. 2006a. Overproduction of wild-type and bioengineered derivatives of the lantibiotic lacticin 3147. Appl. Environ. Microbiol. 72:4492-4496.

30. Cotter, P. D., L. H. Deegan, E. M. Lawton, L. A. Draper, P. M. O'Connor, C. Hill, and R. P. Ross. 2006b. Complete alanine scanning of the two-component lantibiotic lacticin 3147: generating a blueprint for rational drug design. Mol. Microbiol. 62:735-747.

31. Das Gupta, T., and A. Chakrabarty. 2008. Modifications of cupredoxin derived peptides and methods of use thereof. Patent application WO2008033820.

32. Deddish, P. A., B. Marcic, H. L. Jackman, H. Z. Wang, R. A. Skidgel, and E. G. Erdos . 1998. N-domain-specific substrate and C-domain inhibitors of angiotensin-converting enzyme: angiotensin-(1-7) and keto-ACE. Hypertension. 31: 912-917.

33. De Keyzer, J., C. van der Does, and A. J. Driessen . 2002. Kinetic analysis of the translocation of fluorescent precursor proteins into Escherichia coli membrane vesicles. J. Biol. Chem. 277 :46059-46065.

34. De Lisa, M. P., D. Tullman, and G. Georgiou . 2003. Folding quality control in the export of proteins by the bacterial twin-arginine translocation pathway. Proc. Natl. Acad. Sci. U. S. A 100 :6115-6120.

35. Delves-Broughton, J., P. Blackburn, R. E. Evans, and J. Hugenholz . 1996. Antonie van Leeuwenhoek 69: 193-202.

36. Delves-Broughton, J. 2005. Nisin as a food preservative, Food Australia 57: 525-527.

37. De Vos, W.M., G. Jung, and H. G. Sahl . 1991. In: Nisin and Novel Lantibiotics: Proceedings of the First International Workshopon Lantibiotics , Jung, G. and Sahl, H. G (eds.), Leiden: Escom Publishers, 457-464.

38. De Vos, W. M., J. W. Mulders, R. J. Siezen, J. Hugenholtz, and O. P. Kuipers . 1993. Properties of nisin Z and distribution of its gene, nisZ, in Lactococcus lactis . Appl. Environ. Microbiol . 59 :213-218.

39. De Vos, W. M., O. P. Kuipers, d. M. van, Jr., and R. J. Siezen. 1995. Maturation pathway of nisin and other lantibiotics: post-translationally modified antimicrobial peptides exported by gram-positive bacteria. Mol. Microbiol. 17 :427-437.

40. De Ruyter, P. G., O. P. Kuipers, and W. M. de Vos. 1996 . Controlled gene expression systems for Lactococcus lactis with the food-grade inducer nisin. Appl. Environ. Microbiol. 62 :3662-3667.

103 Chapter 9

41. Diez-Freire, C., J. Vazquez,M. F. Correa de Adjounian, M. F. Ferrari, L. Yuan, X. Silver, R. Torres, and M. K. Raizada. 2006. ACE2 gene transfer attenuates hypertension- linked pathophysiological changes in the SHR. Physiol Genomics. 27: 12-19.

42. Driessen, A. J. M., H. W. van den Hooven, W. Kuiper, M. van de Kamp, H. G. Sahl, R. N. Konings, and W. N. Konings . 1995. Mechanistic studies of lantibiotic-induced permeabilization of phospholipid vesicles. Biochemistry 34: 1606-1614.

43. Driessen, A. J. and N. Nouwen. 2008. Protein translocation across the bacterial cytoplasmic membrane. Annu. Rev. Biochem. 77:643-667.

44. Eichenbaum, Z., M. J. Federle, D. Marra, W. M. de Vos, O. P. Kuipers, M. Kleerebezem, and J. Scott . 1998. Use of the lactococcal nisA promoter to regulate gene expression in gram- positive bacteria: comparison of induction level and promoter strength. Appl. Environ. Microbiol. 64 :2763-2769.

45. Fekkes, P., and A. J. Driessen. 1999. Protein targeting to the bacterial cytoplasmic membrane. Microbiol. Mol. Biol. Rev. 63 :161-173.

46. Ferrario, C. M . 2002. Angiotensin I, angiotensin II and their biologically active peptides. J Hypertens. 20: 805-807.

47. Feterik, K., L. Smith, and Z. S. Katusic . 2000. Angiotensin-(1-7) causes endothelium- dependent relaxation in canine middle cerebral artery. Brain Res. 873: 75-82.

48. Field, D., B. Collins, P. D. Cotter, C. Hill, and R. P. Ross. 2007. A system for the random mutagenesis of the two-peptide lantibiotic lacticin 3147: analysis of mutants producing reduced antibacterial activities. J. Mol. Microbiol. Biotechnol. 13:226-234.

49. Field, D., P. M. Connor, P. D. Cotter, C. Hill, and R. P. Ross . 2008. The generation of nisin variants with enhanced activity against specific gram-positive pathogens. Mol. Microbiol. 69: 218-230.

50. Fyhrquist, F., and O. Saijonmaa . 2008. Renin-angiotensin system revisited. J. Intern. Med. 264: 224-236.

51. Galande, A. K., J. O. Trent, and A. F. Spatola . 2003. Understanding base-assisted desulfurization using a variety of disulfide-bridged peptides. Biopolymers 71: 534-551 .

52. Gross, E. and J. L. Morell . 1971. The structure of nisin. J. Am. Chem. Soc. 93: 4634-4635.

53. Gross, E., H. H. Kiltz, and E. Nebelin . 1973. Subtilin, VI: the structure of subtilin. Hoppe Seylers. Z. Physiol Chem. 354 :810-812.

54. Gurley, S. B. and T, M. Coffman . 2008. Angiotensin-converting enzyme 2 gene targeting studies in mice: mixed messages. Exp Physiol. 93: 538-542.

55. Haas, M., L. D. Kluskens, A. Kuipers, R. Rink, S. A. Nelemans, G. N. Moll. 2008. Cyclic angiotensin analogues. Patent Application WO 2008/018792A2

56. Hasper, H. E., N. E. Kramer, J. L. Smith, J. D. Hillman, C. Zachariah, O. P. Kuipers, B. de Kruijff, and E. Breukink. 2006. An alternative bactericidal mechanism of action for lantibiotic peptides that target lipid II. Science 313:1636-1637.

104 References

57. Håvarstein, L. S., D. B. Diep, and I. F. Nes. 1995. A family of bacteriocin ABC transporters carry out proteolytic processing of their substrates concomitant with export. Mol. Microbiol. 16:229-240.

58. Hernández Prada, J. A., A. J. Ferreira, M. J Katovich, V. Shenoy, Y. Qi, R. A. Santos, R. K. Castellano, A. J. Lampkins, V. Gubala, D. A. Ostrov, and M. K. Raizada . 2008. Structure-based identification of small-molecule angiotensin-converting enzyme 2 activators as novel antihypertensive agents. Hypertension . 51: 1312-1317.

59. Herranz, C., and A. J. Driessen. 2005. Sec-mediated secretion of bacteriocin enterocin P by Lactococcus lactis . Appl. Environ. Microbiol. 71 :1959-1963.

60. Hindré, T., J. P. Le Pennec, D. Haras, and A. Dufour . 2004. Regulation of lantibiotic lacticin 481 production at the transcriptional level by acid pH. FEMS Microbiol. Lett . 231 :291-298.

61. Holo, H. and I. F. Nes. 1995. Transformation of Lactococcus by electroporation. Methods Mol. Biol. 47:195-199.

62. Horn, M. J., D. B. Jones, and S. J. Ringel . 1941. Isolation of a New Sulfur-Containing Amino Acid (Lanthionine) from Sodium Carbonate-Treated Wool. J Biol. Chem. 138 :141- 149.

63. Islam, M. R., K. Shioya, J. Nagao, M. Nishie, H. Jikuya, T. Zendo, J. Nakayama, and K. Sonomoto . 2009. Evaluation of essential and variable residues of nukacin ISK-1 by NNK scanning, Mol. Microbiol. 72: 1438-47.

64. Iusuf, D., R. H. Henning, W. H. van Gilst, and A. J. M. Roks . 2008. Angiotensin-(1-7): pharmacological properties and pharmacotherapeutic perspectives. European J Pharmacol. 585: 303-312.

65. Izaguirre, G., and J. N. Hansen . 1997. Use of alkaline phosphatase as a reporter polypeptide to study the role of the subtilin leader segment and the SpaT transporter in the posttranslational modifications and secretion of subtilin in Bacillus subtilis 168. Appl. Environ. Microbiol. 63 :3965-3971.

66. Jensen, P. R. and K. Hammer. 1993. Minimal requirements for exponential growth of Lactococcus lactis . Appl. Environ. Microbiol. 59:4363-4366.

67. Johnson, D. L., F. X. Farrell, F. P. Barbone, F. J. McMahon, J. Tullai, K. Hoey, O. Livnah, N. C. Wrighton, S. A. Middleton, D. A. Loughney, E. A. Stura, W. J. Dower, L. S. Mulcahy, I. A. Wilson, and L. K. Jolliffe . 1998. Identification of a 13 amino acid peptide mimetic of erythropoietin and description of amino acids critical for the mimetic activity of EMP1. Biochemistry 37 :3699-3710.

68. Jongbloed, J. D., U. Grieger, H. Antelmann, M. Hecker, R. Nijland, S. Bron, and J. M. van Dijl . 2004. Two minimal Tat translocases in Bacillus. Mol. Microbiol. 54 :1319-1325.

69. Jošt, K . 1987. Modifications in the disulfide bridge (Carba-analogs), in Handbook of neurohypophyseal hormone analogs (Jošt, K., Lebl, M., and Brtnik, F., Eds.) pp 144-156, CRC Press, Boca Raton.

105 Chapter 9

70. Jung, G . 1991. In: Nisin and Novel Lantibiotics: Proceedings of the First International Workshopon Lantibiotics , Jung, G. and Sahl, H. G (eds.), Leiden: Escom Publishers, 457-464.

71. Kaletta, C., K. D. Entian, R. Kellner, G. Jung, M. Reis, and H. G. Sahl . 1989. Pep5, a new lantibiotic: structural gene isolation and prepeptide sequence. Arch. Microbiol. 152 :16-19.

72. Karakas Sen, A., A. Narbad, N. Horn, H. M. Dodd, A. J. Parr, I. Colquhoun, and M. J. Gasson . 1999. Post-translational modification of nisin. The involvement of NisB in the dehydration process. Eur. J. Biochem. 261 :524-532.

73. Kelleman, A., R. H. Mattern, M. D. Pierschbacher, and M. Goodman. 2003. Incorporation of thioether building blocks into an a vb3-specific RGD peptide: synthesis and biological activity. Biopolymers 71 :686-695.

74. Kido, Y., T. Hamakado, Y. Yoshida, M. Anno, Y. Motoki, T. Wakamiya, and T. Shiba . 1983. Isolation and characterization of ancovenin, a new inhibitor of angiotensin I converting enzyme, produced by actinomycetes . J. Antibiot. 36: 1295-1299.

75. Kiesau, P., U. Eikmanns, Z. Gutowski-Eckel, S. Weber, M. Hammelmann, and K. D. Entian. 1997. Evidence for a multimeric subtilin synthetase complex. J. Bacteriol. 179:1475- 1481.

76. Kleerebezem, M., M. M. Beerthuyzen, E. E. Vaughan, W. M. de Vos, and O. P. Kuipers . 1997. Controlled gene expression systems for lactic acid bacteria: transferable nisin-inducible expression cassettes for Lactococcus, Leuconostoc, and Lactobacillus spp. Appl. Environ. Microbiol . 63 :4581-4584.

77. Kluskens, L. D., A. Kuipers, R. Rink, E. de Boef, S. Fekken, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2005. Post-translational modification of therapeutic peptides by NisB, the dehydratase of the lantibiotic nisin. Biochemistry 44:12827-12834.

78. Kluskens, L. D., A. Nelemans, R. Rink, L. de Vries, A. Meter-Arkema, Y. Wang, T. Walther, A. Kuipers, G. N. Moll, M. Haas. 2009. Angiotensin-(1-7) with thioether-bridge: an ACE-resistant, potent Ang-(1-7) analogue. J. Pharmacol. Exp. Ther. 328: 849-854.

79. Kodani, S., M. E. Hudson, M. C. Durrant, M. J. Buttner, J. R. Nodwell, and J. M. Willey . 2004. The SapB morphogen is a lantibiotic-like peptide derived from the product of the developmental gene ramS in Streptomyces coelicolor. Proc. Natl. Acad. Sci. U. S. A 101: 11448-11453.

80. Kodani, S., M. A. Lodato, M. C. Durrant, F. Picart, and J. M. Willey . 2005. SapT, a lanthionine-containing peptide involved in aerial hyphae formation in the streptomycetes. Mol. Microbiol. 58: 1368-1380.

81. Kohli, R. M., J. W. Trauger, D. Schwarzer, M. A. Marahiel, M. A., and C. T. Walsh . 2001. Generality of peptide cyclization catalyzed by isolated thioesterase domains of nonribosomal peptide synthetases. Biochemistry 40 :7099-7108.

82. Koivula, T., I. Palva, and H. Hemila . 1991. Nucleotide sequence of the secY gene from Lactococcus lactis and identification of conserved regions by comparison of four SecY proteins. FEBS Lett. 288 :114-118.

106 References

83. Koponen, O., M. Tolonen, M. Qiao, G. Wahlstrom, J. Helin, and P. E. Saris. 2002. NisB is required for the dehydration and NisC for the lanthionine formation in the post-translational modification of nisin. Microbiology 148:3561-3568.

84. Kordel, M., F. Schüller, and H. G. Sahl . 1989. Interaction of the pore forming-peptide antibiotics Pep 5, nisin and subtilin with non-energized liposomes. FEBS Lett. 244: 99-102.

85. Kucharewicz, I., R. Pawlak, T. Matys, D. Pawlak, and W. Buczko . 2002. Antithrombotic effect of captopril and losartan is mediated by angiotensin-(1-7). Hypertension. 40: 774-779.

86. Kuipers, A., E. de Boef, R. Rink, S. Fekken, L. D. Kluskens, A. J. Driessen, K. Leenhouts, O. P. Kuipers, and G. N. Moll. 2004. NisT, the transporter of the lantibiotic nisin, can transport fully modified, dehydrated, and unmodified prenisin and fusions of the leader peptide with nonlantibiotic peptides. J. Biol. Chem. 279:22176-22182.

87. Kuipers, A., J. Wierenga, R. Rink, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2006. Sec-mediated transport of posttranslationally dehydrated peptides in Lactococcus lactis . Appl. Environ. Microbiol. 72:7626-7633.

88. Kuipers, O. P., H. S. Rollema, W. M. Yap, H. J. Boot, R. J. Siezen, and W. M.de Vos . 1992. Engineering dehydrated amino acid residues in the antimicrobial peptide nisin. J. Biol. Chem. 267 :24340-24346.

89. Kuipers, O. P., M. M. Beerthuyzen, R. J. Siezen, and W. M. de Vos. 1993a. Characterization of the nisin gene cluster nisABTCIPR of Lactococcus lactis . Requirement of expression of the nisA and nisI genes for development of immunity. Eur. J. Biochem. 216:281-291.

90. Kuipers, O. P., H. S. Rollema, W. M. de Vos, and R. J. Siezen . 1993b. Biosynthesis and secretion of a precursor of nisin Z by Lactococcus lactis , directed by the leader peptide of the homologous lantibiotic subtilin from Bacillus subtilis . FEBS Lett. 330:23-27.

91. Kuipers, O. P., M. M. Beerthuyzen, P. G. de Ruyter, E. J. Luesink, and W. M. de Vos . 1995. Autoregulation of nisin biosynthesis in Lactococcus lactis by signal transduction. J. Biol. Chem. 270 :27299-27304.

92. Kuipers, O. P., G. Bierbaum, B. Öttenwalder, H. M. Dodd, N. Horn, J. Metzger, T. Kupke, V. Gnau, R. Bongers, B. P. van den Bogaard. P., H. Kosters, H. S. Rollema, W. M. de Vos, R. J. Siezen, G. Jung, F. Gotz, H. G. Sahl, and M. J Gasson . 1996. Protein engineering of lantibiotics. Antonie Van Leeuwenhoek 69 :161-169.

93. Kuipers, O. P., P. G. de Ruyter, M. Kleerebezem, and W. M. de Vos. 1997. Controlled overproduction of proteins by lactic acid bacteria. Trends Biotechnol. 15:135-140.

94. Kunst, F., N. Ogasawara, I. Moszer, A. M. Albertini, G. Alloni, V. Azevedo, M. G. Bertero, P. Bessieres, A. Bolotin, S. Borchert, R. Borriss, L. Boursier, A. Brans, M. Braun, S. C. Brignell, S. Bron, S. Brouillet, C. V. Bruschi, B. Caldwell, V. Capuano, N. M. Carter, S. K. Choi, J. J. Codani, I. F. Connerton, A. Danchin, and. 1997. The complete genome sequence of the gram-positive bacterium Bacillus subtilis . Nature 390 :249-256.

95. Langeveld, B., W. H. van Gilst, R. A. Tio, F. Zijlstra, and A. J. M. Roks . 2005. Angiotensin-(1-7) attenuates neointimal formation after stent implantation in the rat. Hypertension 45: 138-141.

107 Chapter 9

96. Lee, M. V., L. A. Ihnken, Y. O. You, A. L. McClerren, W. A. van der Donk, and N. L. Kelleher. 2009. Distributive and directional behavior of lantibiotic synthetases revealed by high-resolution tandem mass spectrometry. J. Am. Chem. Soc. 131: 12258-12264

97. Lee, P. A., D. Tullman-Ercek, and G. Georgiou . 2006. The bacterial twin-arginine translocation pathway. Annu. Rev. Microbiol. 60: 373-395.

98. Leenhouts, K., G. Buist, A. Bolhuis, A. ten Berge, J. Kiel, I Miereau, M. Dabrowski, G. Venema, and J. Kok . 1996. A general system for generating unlabelled gene replacements in bacterial chromosomes . Mol. Gen. Genet. 253 :217-224.

99. Leenhouts, K., A. Bolhuis, G. Venema, and J. Kok. 1998. Construction of a food-grade multiple-copy integration system for Lactococcus lactis . Appl. Microbiol. Biotechnol. 49 :417- 423.

100. Le Loir, Y., S. Nouaille, J. Commissaire, L. Bretigny, A. Gruss, and P. Lan gella. 2001. Signal peptide and propeptide optimization for heterologous protein secretion in Lactococcus lactis . Appl. Environ. Microbiol. 67 :4119-4127.

101. Le Loir, Y., V. Azevedo, S. C. Oliveira, D. A. Freitas, A. Miyoshi, L. G. Bermudez- Humaran, S. Nouaille, L. A. Ribeiro, S. Leclercq, J. E. Gabriel, V. D. Guimaraes, M. N. Oliveira, C. Charlier, M. Gautier, and P. Langella. 2005. Protein secretion in Lactococcus lactis : an efficient way to increase the overall heterologous protein production.Microb. Cell Fact. 4:2.

102. Levengood, M. R., G. C. Patton, and W. A. van der Donk . 2007. The leader peptide is not required for post-translational modification by lacticin 481 synthetase. J. Am. Chem. Soc. 129 :10314-10315 .

103. Levengood, M. R., and W. A. van der Donk. 2008. Use of lantibiotic synthetases for the preparation of bioactive constrained peptides. Bioorg. Med. Chem. Lett. 18: 3025-3028.

104. Levengood, M. R., C. C. Kerwood, C. Chatterjee, and W. A. van der Donk . 2009a Investigation of the substrate specificity of lacticin 481 synthetase by using nonproteinogenic amino acids. Chembiochem. 10 :911-919 .

105. Levengood, M. R., P. J. Knerr, T. J. Oman, and W. A. van der Donk . 2009b. In vitro mutasynthesis of lantibiotic analogues containing nonproteinogenic amino acids. J. Am. Chem. Soc. 131 :12024-12025.

106. Li, B., J. P. Yu, J. S. Brunzelle, G. N. Moll, W. A. van der Donk, and S. K. Nair. 2006. Structure and mechanism of the lantibiotic cyclase involved in nisin biosynthesis. Science 311:1464-1467.

107. Li, B. and W. A. van der Donk. 2007. Identification of essential catalytic residues of the cyclase NisC involved in the biosynthesis of nisin. J. Biol. Chem. 282:21169-21175.

108. Li, P., and P. P.Roller . 2002. Cyclization strategies in peptide derived drug design. Curr. Topics Med. Chem. 2:325-341.

109. Lian, L. Y., W. C. Chan, S. D. Morley, G. C. Roberts, B. W. Bycroft, and D. Jackson . 1992. Solution structures of nisin A and its two major degradation products determined by n.m.r. Biochem. J. 283 ( Pt 2): 413-420.

108 References

110. Liu, W. and J. N.Hansen . 1992. Enhancement of the chemical and antimicrobial properties of subtilin by site-directed mutagenesis. J. Biol. Chem. 267 :25078-25085.

111. Lombardi, A., B. D'Agostino, F. Nastri, L. D. D'Andrea, A. Filippelli, M. Falciani, F. Rossi, and V. Pavone. 1998. A novel super-potent neurokinin A receptor antagonist containing dehydroalanine. Bioorg. Med. Chem. Lett. 8:1153-1156.

112. Loot, A. E., A. J. M. Roks, R. H. Henning, R. A. Tio, A. J. H. Suurmeijer, F. Boomsma, and W. H. van Gilst. 2002. Angiotensin-(1-7) attenuates the development of heart failure after myocardial infarction in ats. Circulation 105: 1548-1550.

113. Lubelski, J., R. Rink, R. Khusainov, G. N Moll, and O. P. Kuipers . 2008. Biosynthesis, immunity, regulation, mode of action and engineering of the model lantibiotic nisin. Cell Mol. Life Sci. 65 :455-476.

114. Lubelski, J., R. Khusainov, and O. P. Kuipers. 2009. Directionality and coordination of dehydration and ring formation during biosynthesis of the lantibiotic nisin. J. Biol. Chem. 284 :25962-25972.

115. Marki, F., E. Hanni, A. Fredenhagen, and J. van Oostrum . 1991. Mode of action of the lanthionine-containing peptide antibiotics duramycin, duramycin B and C, and cinnamycin as indirect inhibitors of phospholipase A2. Biochem. Pharmacol. 42 :2027-2035.

116. Martin, N. I., T. Sprules, M. R. Carpenter, P. D. Cotter, C. Hill, R. P. Ross, and J. C. Vederas. 2004. Structural characterization of lacticin 3147, a two-peptide lantibiotic with synergistic activity. Biochemistry 43:3049-3056.

117. Martinez-Cuesta, M. C., G. Buist, J. Kok, H. H. Hauge, J. Nissen-Meyer, C. Pelaez, and T. Requena. 2000. Biological and molecular characterization of a two-peptide lantibiotic produced by Lactococcus lactis IFPL105. J. Appl. Microbiol. 89:249-260.

118. McAuliffe, O., M. P. Ryan, R. P. Ross, C. Hill, P. Breeuwer, and T. Abee. 1998. Lacticin 3147, a broad-spectrum bacteriocin which selectively dissipates the membrane potential. Appl. Environ. Microbiol. 64:439-445.

119. McAuliffe, O., C. Hill, and R. P. Ross. 2000. Each peptide of the two-component lantibiotic lacticin 3147 requires a separate modification enzyme for activity. Microbiology 146:2147- 2154.

120. McAuliffe, O., R. P. Ross, and C. Hill. 2001. Lantibiotics: structure, biosynthesis and mode of action. FEMS Microbiol. Rev. 25:285-308.

121. McClerren, A. L., Cooper, L. E., C. Quan, P.M. Thomas, N. L. Kelleher, and W. A. van der Donk . 2006. Discovery and in vitro biosynthesis of haloduracin, a two-component lantibiotic. Proc. Natl. Acad. Sci. U. S. A 103 :17243-17248.

122. Meens, J., E. Frings, M. Klose, and R. Freudl. 1993. An outer membrane protein (OmpA) of Escherichia coli can be translocated across the cytoplasmic membrane of Bacillus subtilis . Mol. Microbiol. 9:847-855.

123. Menon, J., D. R. Soto-Pantoja, M. F. Callahan, J. M. Cline, C. M. Ferrario, E. A. Tallant, and P. E. Gallagher. 2007. Angiotensin-(1-7) inhibits growth of human lung adenocarcinoma xenografts in nude mice through a reduction in cyclooxygenase-2. Cancer Res. 67: 2809-2815.

109 Chapter 9

124. Meyer, H. E., M. Heber, B. Eisermann, H. Korte, J. W. Metzger, and G. Jung . 1994. Sequence analysis of lantibiotics: chemical derivatization procedures allow a fast access to complete Edman degradation. Anal. Biochem. 223 :185-190.

125. Meyer, C., G. Bierbaum, C. Heidrich, M. Reis, J. Suling, J. Iglesias, and I. Wind . 1995. Nucleotide sequence of the lantibiotic Pep5 biosynthetic gene cluster and functional analysis of PepP and PepC. Evidence for a role of PepC in thioether formation . Eur. J. Biochem. 232: 478-489.

126. Miller, L. M., C. Chatterjee, W. A. van der Donk, and N. L. Kelleher. 2006. The dehydratase activity of lacticin 481 synthetase is highly processive. J. Am. Chem. Soc. 128:1420-1421.

127. Moll, G. N., G. C. K. Roberts, W. N. Konings, and A. J. M Driessen. A. J. M . 1996. Mechanism of lantibiotic-induced pore-formation . Antonie van Leeuwenhoek 69 :185-191.

128. Moll, G. N., J. Clark, W. C. Chan, B. W. Bycroft, G. C. K. Roberts, W. N. Konings and A. M. J. Driessen. 1997. Role of transmembrane pH gradient and membrane binding in nisin pore formation. J. Bacteriol. 179 :135-140.

129. Moll, G. N., W. N. Konings, and A. J. M. Driessen . 1998. The lantibiotic nisin induces transmembrane movement of a fluorescent phospholipid J. Bacteriol. 180 :6565-6570.

130. Moll, G. N., A. Kuipers, R. Rink, A. J. M. Driessen, O. P. Kuipers. 2006. Methods for the production and secretion of modified peptides. Patent application WO 2006062398.

131. Morris, S. L., R. C. Walsh, and J. N. Hansen. 1984. Identification and characterization of some bacterial membrane sulfhydryl groups which are targets of bacteriostatic and antibiotic action. J. Biol. Chem. 259:13590-13594.

132. Murkin, A. S., and M. E. Tanner. 2002. Dehydroalanine-based inhibition of a peptide epimerase from spider venom. J. Org. Chem. 67 :8389-8394.

133. Nagao, J., Y. Aso, T. Sashihara, K. Shioya, A. Adachi, J. Nakayama, and K. Sonomoto . 2005. Localization and interaction of the biosynthetic proteins for the lantibiotic, Nukacin ISK-1. Biosci. Biotechnol. Biochem. 69 :1341-1347.

134. Nagao, J., Y. Aso, K. Shioya, J. Nakayama, and K. Sonomoto. 2007. Lantibiotic engineering: molecular characterization and exploitation of lantibiotic-synthesizing enzymes for peptide engineering. J. Mol. Microbiol. Biotechnol. 13:235-242.

135. Neis, S., G. Bierbaum, M. Josten, U. Pag, C. Kempter, G. Jung, and H. G. Sahl . 1997. Effect of leader peptide mutations on biosynthesis of the lantibiotic Pep5. FEMS Microbiol. Lett. 149 :249-255.

136. Nissen-Meyer, J., H. Holo, L. S. Håverstein, K. Sletten, and I. F. Nes . A novel lactococcal bacteriocin whose activity depends on the complementary action of two peptides . 1992. J. Bacteriol. 174 :5686-5692.

137. Novak, J., M. Kirk, P. W. Caufield, S. Barnes, K. Morrison, and J. Baker. 1996. Detection of modified amino acids in lantibiotic peptide mutacin II by chemical derivation and electrospray ionization-mass spectroscopic analysis. Anal. Biochem. 236 :358-360.

110 References

138. Ösapay, G., L. Prokai, H. S. Kim, K. F. Medzihradszky, D. H. Coy, G. Liapakis, T. Reisine, G. Melacini, Q. Zhu, S. H. Wang, R. H. Mattern, and M. Goodman . 1997. Lanthionine-somatostatin analogs: synthesis, characterization, biological activity, and enzymatic stability studies. J. Med. Chem. 40 :2241-2251.

139. Okeley, N. M., Y. Zhu, and W. A. van Der Donk . 2000. Facile chemoselective synthesis of dehydroalanine-containing peptides. Org. Lett. 2: 3603-3606.

140. Oliver, D. B., R. J. Cabelli, K. M. Dolan, and G. P. Jarosik . 1990. Azide-resistant mutants of Escherichia coli alter the SecA protein, an azide-sensitive component of the protein export machinery. Proc. Natl. Acad. Sci. U. S. A. 87 :8227-8231.

141. Öttenwalder, B., T. Kupke, S. Brecht, V. Gnau, J. Metzger, G. Jung, and F. Götz . 1995. Isolation and characterization of genetically engineered gallidermin and epidermin analogs. Appl. Environ. Microbiol. 61: 3894-3903.

142. Pag, U., C. Heidrich, G. Bierbaum, and H. G. Sahl . 1999. Molecular analysis of expression of the lantibiotic pep5 immunity phenotype. Appl. Environ. Microbiol. 65 :591-598.

143. Pag, U. and H. G. Sahl . 2002. Multiple activities in lantibiotics--models for the design of novel antibiotics?. Curr. Pharm. Des 8: 815-833.

144. Palmer, D. E., Pattaroni, C., Nunami, K., Chadha, R. K., Goodman, M., Wakamiya, T., Fukase, K., Horimoto, S., Kitazawa, M., Fujita, H., Kubo, A., and Shiba, T. 1992. Effects of dehydroalanine on peptide conformations. J. Am. Chem. Soc . 114 :5634-5642.

145. Patchett, A. A., and E. H. Cordes . 1985 The design and properties of N- carboxyalkyldipeptide inhibitors of angiotensin-converting enzyme. Adv Enzymol Relat Areas Mol Biol. 57: 1-84.

146. Patton, G. C., and W. A. van der Donk . 2005. New developments in lantibiotic biosynthesis and mode of action. Curr. Opin. Microbiol. 8:543-551.

147. Patton, G. C., M. Paul, L. E. Cooper, C. Chatterjee, and W. A. van der Donk . 2008. The importance of the leader sequence for directing lanthionine formation in lacticin 481. Biochemistry 47 :7342-7351 .

148. Paul, L. K., G. Izaguire, and J. N. Hansen . 1999. Studies of the subtilin leader peptide as a translocation signal in Escherichia coli K12 . FEMS Microbiol. Lett. 176 :45-50.

149. Paul, M., G. C. Patton, and W. A. van der Donk. 2007. Mutants of the zinc ligands of lacticin 481 synthetase retain dehydration activity but have impaired cyclization activity. Biochemistry 46 :6268-6276.

150. Pereira, R. M., R. A. Dos Santos, M. M. Teixeira, V. H Leite, L. P. Costa, F. L. da Costa Dias, L. S. Barcelos, G. B. Collares, and A. C. Simoes e Silva . 2007. The renin-angiotensin system in a rat model of hepatic fibrosis: evidence for a protective role of Angiotensin-(1-7). J Hepatol. 46: 674-681.

151. Pohlschröder, M., M. I. Gimenez, and K. F. Jarrell . 2005. Protein transport in Archaea: Sec and twin arginine translocation pathways. Curr. Opin. Microbiol. 8:713-719.

111 Chapter 9

152. Poquet, I., V. Saint, E. Seznec, N. Simoes, A. Bolotin, and A. Gruss . 2000. HtrA is the unique surface housekeeping protease in Lactococcus lactis and is required for natural protein processing. Mol. Microbiol. 35 :1042-1051.

153. Qiao, M. and P. E. Saris. 1996. Evidence for a role of NisT in transport of the lantibiotic nisin produced by Lactococcus lactis N8. FEMS Microbiol. Lett. 144:89-93.

154. Ra, R., M. A. Beerthuyzen, W. M. de Vos, P. E. J. Saris, and O. P. Kuipers . 1999. Effects of gene disruptions in the nisin gene cluster of Lactococcus lactis on nisin production and producer immunity. Microbiology 145: 1227-1233.

155. Rew, Y., S. Malkmus, C. Svensson, T. L. Yaksh, N. N. Chung, P. W. Schiller, J. A. Cassel, R. N. DeHaven, J. P. Taulane, and M. Goodman. 2002. Synthesis and biological activities of cyclic lanthionine enkephalin analogues: delta-opioid receptor selective ligands. J. Med. Chem. 45 :3746-3754.

156. Rince, A., A. Dufour, P. Uguen, J. P. Le Pennec, and D. Haras. 1997. Characterization of the lacticin 481 operon: the Lactococcus lactis genes lctF , lctE , and lctG encode a putative ABC transporter involved in bacteriocin immunity. Appl. Environ. Microbiol. 63:4252-4260.

157. Rink, R., A. Kuipers, E. de Boef., K. J. Leenhouts, A. J. Driessen, G. N. Moll, and O. P. Kuipers. 2005. Lantibiotic structures as guidelines for the design of peptides that can be modified by lantibiotic enzymes. Biochemistry 44:8873-8882.

158. Rink, R., J. Wierenga, A. Kuipers, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007a. Production of dehydroamino acid-containing peptides by Lactococcus lactis . Appl. Environ. Microbiol. 73: 1792–1796.

159. Rink, R., J. Wierenga, A. Kuipers, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007b. Dissection and modulation of the four distinct activities of nisin by mutagenesis of rings A and B and by C-terminal truncation. Appl. Environ. Microbiol. 73:5809-5816.

160. Rink, R., L. D. Kluskens, A. Kuipers, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007c. NisC, the cyclase of the lantibiotic nisin, can catalyze cyclization of designed nonlantibiotic peptides. Biochemistry 46:13179-13189.

161. Rink, R., A. Arkema-Meter, I. Baudoin, E. Post, A. Kuipers, S. A. Nelemans, M. Haas Jimoh Akanbi, and G. N. Moll. 2010. To protect peptide pharmaceuticals against peptidases. J Pharmacol. Toxicol. Methods, submitted.

162. Rivier, J. E., J. Porter, L. A. Cervini, S. L. Lahrichi, D. A. Kirby, R. S. Struthers, S. C. Koerber, and C. L. Rivier. 2000a. Design of monocyclic (1-3) and dicyclic (1-3/4-10) gonadotropin releasing hormone (GnRH) antagonists. J. Med. Chem. 43 :797-806.

163. Rivier, J. E., G. Jiang, R. S. Struthers, S. C. Koerber, J. Porter, L. A. Cervini, D. A. Kirby, A. G. Craig, and C. L. Rivier. 2000b. Design of potent dicyclic (1-5/4-10) gonadotropin releasing hormone (GnRH) antagonists. J. Med. Chem. 43: 807-818.

164. Rodgers, K. E., S. Xiong, and G. S. diZerega . 2002. Accelerated recovery from irradiation injury by angiotensin peptides. Cancer Chemother Pharmacol. 49: 403-411.

112 References

165. Rodgers, K. E. J. Oliver, and G. S. diZerega . 2006. Phase I/II dose escalation study of angiotensin 1-7 [A(1-7)] administered before and after chemotherapy in patients with newly diagnosed breast cancer. Cancer Chemother Pharmacol 57: 559-568.

166. Rogers, L. A., and E. O. Whittier . 1928. Limiting factors in lactic fermentation. Journal of Bacteriology 16 :211–9.

167. Roks, A. J., P. P. van Geel, Y. M. Pinto, H. Buikema, R. H. Henning, D. de Zeeuw, and W. H. van Gilst. 1999. Angiotensin-(1-7) is a modulator of the human renin angiotensin system. Hypertension . 34: 296-301.

168. Ryan, M. P., M. C. Rea, C. Hill, and R. P. Ross. 1996. An application in cheddar cheese manufacture for a strain of Lactococcus lactis producing a novel broad-spectrum bacteriocin, lacticin 3147. Appl. Environ. Microbiol. 62:612-619.

169. Ryan, M. P., R. W. Jack, M. Josten, H. G. Sahl, G. Jung, R. P. Ross, and C. Hill. 1999. Extensive post-translational modification, including serine to D-alanine conversion, in the two-component lantibiotic, lacticin 3147. J. Biol. Chem. 274:37544-37550.

170. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

171. Sahl, H. G., R. W. Jack, and G. Bierbaum. 1995. Biosynthesis and biological activities of lantibiotics with unique post-translational modifications. Eur. J. Biochem. 230:827-853.

172. Sahl, H.-G., and G. Bierbaum. 1998. Lantibiotics: biosynthesis and biological activities of uniquely modified peptides from gram-positive bacteria. Annu. Rev. Microbiol. 52: 41-79.

173. Sanders, C., N. Wethkamp, and H. Lill . 2001. Transport of cytochrome c derivatives by the bacterial Tat protein translocation system. Mol. Microbiol. 41 :241-246.

174. Santos, R. A., M. J. Campagnole-Santos, N. C. Baracho, M. A. Fontes, L. C. Silva, L. A. Neves, D. R. Oliveira, W. M Caligiorne, A. R Rodrigues, C. Gropen Júnior, and et al. 1994. Characterization of a new angiotensin antagonist selective for angiotensin-(1-7): evidence that the actions of angiotensin-(1-7) are mediated by specific angiotensin receptors. Brain Res Bull. 35: 293-298.

175. Santos, R. A., M. J. Campagnole-Santos, and S. P. Andrade . 2000. Angiotensin-(1-7): an update. Regul Pept. 91: 45-62.

176. Santos, R. A., A. S. Haibara, M. J. Campagnole-Santos, A. C. Simoes e Silva, R. D. Paula, S. V. Pinheiro, M. F. Leite, V. S. Lemos, D. M. Silva, M. T. Guerra, and M. C. Khosla . 2003a. Characterization of a new selective antagonist for angiotensin-(1-7), D-pro7- angiotensin-(1-7). Hypertension. 41: 737-743.

177. Santos, R. A., A. C. Simoes e Silva, C. Maric, D. M. Silva, R. P. Machado, I. de Buhr, S. Heringer-Walther, S. V. Pinheiro, M. T. Lopes, M. Bader, E. P. Mendes, V. S. Lemos, M. J. Campagnole-Santos, H. P. Schultheiss, R. Speth, and T. Walther . 2003b. Angiotensin- (1-7) is an endogenous ligand for the G protein-coupled receptor Mas. Proc Natl Acad Sci U S A. 100: 8258-8263.

113 Chapter 9

178. Santos, R. A., A. J. Ferreira, A. P. Nadu, A. N. Braga, A. P. de Almeida, M. J. Campagnole-Santos, O. Baltatu, R. Iliescu, T. L. Reudelhuber, and M. Bader . 2004. Expression of an angiotensin-(1-7)-producing fusion protein produces cardioprotective effects in rats. Physiol Genomics. 17: 292-299.

179. Santos, R. A., F. Frezard, and A. J. Ferreira . 2005. Angiotensin-(1-7): blood, heart, and blood vessels. Curr Med Chem Cardiovasc Hematol Agents . 3: 383-391.

180. Santos, R. A., and A. J. Ferreira . 2006. Pharmacological effects of AVE 0991, a nonpeptide angiotensin-(1-7) receptor agonist. Cardiovasc Drug Rev . 24: 239-46.

181. Schägger, H., G. von Jagow. 1987. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 166:368-379.

182. Schmidt, U., A. Lieberknecht, and J. Wild . 1988. Didehydroamino acids (DDAA) and didehydropeptides (DDP). Synthesis 1988:159-172.

183. Schnell, N., K. D. Entian, U. Schneider, F. Götz, H. Zahner, R. Kellner, and G. Jung . 1988. Prepeptide sequence of epidermin, a ribosomally synthesized antibiotic with four sulphide-rings. Nature 333: 276-278.

184. Schiavone, M. T., R. A. Santos, K. B. Brosnihan, M. C. Khosla, and M. C. Ferrario . 1988. Release of vasopressin from the rat hypothalamo-neurohypophysial system by angiotensin-(1-7) heptapeptide. Proc Natl Acad Sci U S A. 85: 4095-4098.

185. Schindler, C., P. Bramlage, W. Kirch, and C. M. Ferrario . 2007. Role of the vasodilator peptide angiotensin-(1-7) in cardiovascular drug therapy. Vasc Health Risk Manag. 3: 125- 137.

186. Schuster, B., and J. Rétey . 1995 . The mechanism of action of phenylalanine ammonia-lyase: the role of prosthetic dehydroalanine . PNAS. 92 :8433-8437.

187. Sealfon, S. C., H. Weinstein, and R. P. Millar . 1997. Molecular mechanisms of ligand interaction with the gonadotropin-releasing hormone receptor. Endocr. Rev. 18 :180-205.

188. Siddiqui, M. I., S. Kataria, V. Ahuja, and G. S. Rao . 2001. A peptide inhibitor of HIV-1 protease using alpha, beta- dehydro residues: a structure based computer model. Indian J. Biochem. Biophys. 38 :90-95.

189. Siegers, K., S. Heinzmann, and K. D. Entian. 1996. Biosynthesis of lantibiotic nisin. Posttranslational modification of its prepeptide occurs at a multimeric membrane-associated lanthionine synthetase complex. J. Biol. Chem. 271:12294-12301.

190. Siezen, R. J., O. P. Kuipers, and W. M. de Vos. 1996. Comparison of lantibiotic gene clusters and encoded proteins. Antonie Van Leeuwenhoek 69: 171-184.

191. Silva, D. M., H. R. Vianna, S. F. Cortes, M. J. Campagnole-Santos, R. A. Santos, and V. S. Lemos . 2007. Evidence for a new angiotensin-(1-7) receptor subtype in the aorta of Sprague-Dawley rats. Peptides. 28: 702-707.

192. Simon, D. and A. Chopin. 1988. Construction of a vector plasmid family and its use for molecular cloning in Streptococcus lactis . Biochimie 70:559-566.

114 References

193. Stein, T., and K. D. Entian . 2002. Maturation of the lantibiotic subtilin: matrix-assisted laser desorption/ionization time-of-flight mass spectrometry to monitor precursors and their proteolytic processing in crude bacterial cultures. Rapid Commun. Mass Spectrom. 16 :103- 110.

194. Stohlmeyer, M. M., H. Tanaka, and T. J. Wandless . 1999. A stereospecific elimination to form dehydroamino acids: synthesis of the phomopsin tripeptide side chain. J. Am. Chem. Soc. 121: 6100-6101.

195. Svensson, C. I., Y. Rew, S. Malkmus, P. W. Schiller, J. P. Taulane, M. Goodman, and T. L. Yaksh . 2003. Systemic and spinal analgesic activity of a delta-opioid-selective lanthionine enkephalin analog. J. Pharmacol. Exp. Ther. 304 :827-832.

196. Szekat, C., R. W. Jack, D Skutlarek, H. Farber, and G. Bierbaum . 2003. Construction of an expression system for site-directed mutagenesis of the lantibiotic mersacidin. Appl. Environ. Microbiol . 69 :3777-3783.

197. Taylor, B. N., R. R. Mehta, T. Yamada, F. Lekmine, K. Christov, A. M. Chakrabarty, A. Green, L. Bratescu, A. Shilkaitis, C. W. Beattie, and T. K. Das Gupta. 2009. Noncationic peptides obtained from azurin preferentially enter cancer cells. Cancer Res. 69 :537-546.

198. Terzaghi, B. E. and W. E. Sandine. 1975. Improved medium for Lactic streptococci and their bacteriophages. Appl. Microbiol. 29:807-813.

199. Thomas, L. V., M. R. Clarkson, and J. Delves-Broughton . 2000. Nisin. In A. S. Naidu (ed.), Natural food antimicrobial systems. p. 463-524.

200. Thomas, L. V., R. E. Ingram, H. E. Bevis, E. A. Davies, C. F. Milne, and J. Delves- Broughton . 2002. Effective use of nisin to control Bacillus and Clostridium spoilage of a pasteurized mashed potato product. J. Food Prot . 65 :1580-1585.

201. Tian, P., and I. Andricioaei. 2006 Size, motion, and function of the SecY translocon revealed by molecular dynamics simulations with virtual probes. Biophys. J. 90 :2718-2730.

202. Tjalsma, H., A. Bolhuis, J. D. Jongbloed, S. Bron, and J. M. van Dijl. 2000. Signal peptide-dependent protein transport in Bacillus subtilis : a genome-based survey of the secretome. Microbiol. Mol. Biol. Rev. 64 :515-547.

203. Tomkinson, B., L. Grehn, B. Fransson, and O. Zetterqvist . 1994. Use of a dehydroalanine- containing peptide as an efficient inhibitor of tripeptidyl peptidase II. Arch. Biochem. Biophys . 314: 276-279.

204. Trask. A. J., and C. M. Ferrario . 2007. Angiotensin-(1-7): pharmacology and new perspectives in cardiovascular treatments. Cardiovasc Drug Rev. 25: 162-174.

205. Tugyi, R., G. Mezo, E. Fellinger, D. Andreu, and F. Hudecz . 2005. The effect of cyclization on the enzymatic degradation of herpes simplex virus glycoprotein D derived epitope peptide. J. Pept. Sci . 11 :642-649.

206. Uchida, K., H. Mori, and S. Mizushima. 1995. Stepwise movement of preproteins in the process of translocation across the cytoplasmic membrane of Escherichia coli . J. Biol. Chem. 270 :30862-30868.

115 Chapter 9

207. Ueda, K., K. Oinuma, G. Ikeda, K. Hosono, Y. Ohnishi, S. Horinouchi, and T. Beppu. 2002. AmfS, an extracellular peptidic morphogen in Streptomyces griseus . J. Bacteriol . 184 :1488-1492.

208. Uguen, P., T. Hindre, S. Didelot, C. Marty, D. Haras, J. P. Le Pennec, K. Vallee-Rehel, and A. Dufour. 2005. Maturation by LctT is required for biosynthesis of full-length lantibiotic lacticin 481. Appl. Environ. Microbiol. 71:562-565.

209. Van Asseldonk M., W. M. de Vos, and G. Simons. 1993. Functional analysis of the Lactococcus lactis usp45 secretion signal in the secretion of a homologous proteinase and a heterologous alpha-amylase. Mol. Gen. Genet. 240 :428-434.

210. Van den Hooven, H. W., F. Fogolari, H. S. Rollema, R. N. Konings, C. W. Hilbers, and F. J. M. Van de Ven, F. J. M. 1993 NMR and circular dichroism studies of the lantibiotic nisin in non-aqueous environments. FEBS Lett . 319: 189-194.

211. Van den Berg van Saparoea H. B., P. J. Bakkes, G. N. Moll, and A. J. Driessen . 2008. Distinct contributions of the nisin biosynthesis enzymes NisB and NisC and transporter NisT to prenisin production by Lactococcus lactis . Appl. Environ. Microbiol. 74 :5541-5548.

212. Van der Meer, J. R., J. Polman, M. M. Beerthuyzen, R. J. Siezen, O. P. Kuipers, and W. M. de Vos. 1993. Characterization of the Lactococcus lactis nisin A operon genes nisP , encoding a subtilisin-like serine protease involved in precursor processing, and nisR , encoding a regulatory protein involved in nisin biosynthesis. J. Bacteriol. 175:2578-2588.

213. Van der Meer, J. R., H. S. Rollema, R. Siezen, M. M. Beerthuizen, O. P. Kuipers and W. M. de Vos. M . 1994. Influence of amino acid substitutions in the nisin leader peptide on biosynthesis and secretion of nisin by Lactococcus lactis . J. Biol. Chem. 269: 3555-3562.

214. Van de Ven, F. J. M., H. W. Van den Hooven, R. N. Konings, and C. W. Hilbers. 1991. NMR studies of lantibiotics. The structure of nisin in aqueous solution. Eur. J. Biochem . 202 :1181-1188.

215. Van Kraaij, C., W. M. de Vos, R. J. Siezen, and O. P. Kuipers . 1999. Lantibiotics: biosynthesis, mode of action and applications. Nat. Prod. Rep. 16 :575-587.

216. Van Kraaij, C., E. Breukink, H. S. Rollema, R. S. Bongers, H. A. Kosters, B. de Kruijff, and O. P. Kuipers. 2000 . Engineering a disulfide bond and free thiols in the lantibiotic nisin Z. Eur. J. Biochem. 267: 901-909.

217. Van Wely, K. H., J. Swaving, R. Freudl, and A. J. Driessen . 2001. Translocation of proteins across the cell envelope of Gram-positive bacteria. FEMS Microbiol. Rev. 25 : 437- 454.

218. Vickers, C., P. Hales, V. Kaushik, L. Dick, J. Gavin, J. Tang, K. Godbout, T. Parsons, E. Baronas, F. Hsieh, S. Acton, M. Patane, A. Nichols, and P. Tummino . 2002. Hydrolysis of biological peptides by human angiotensin-converting enzyme-related carboxypeptidase. J Biol Chem. 277: 14838-14843

219. Vollbrecht, L., H. Steinmetz, G. Hofle, L. Oberer, G. Rihs, G. Bovermann, and P. Matt . 2002. Argyrins, immunosuppressive cyclic peptides from myxobacteria. II. Structure elucidation and stereochemistry. J. Antibiot. 55 :715-721.

220. Von Heijne, G. 1985. Signal sequences. The limits of variation. J. Mol. Biol. 184 :99-105.

116 References

221. Widdick, D. A., H. M Dodd, P. Barraille, J. White, T. H. Stein, K. F. Chater, M. J. Gasson, and M. J.Bibb. 2003. Cloning and engineering of the cinnamycin biosynthetic gene cluster from Streptomyces cinnamoneus cinnamoneus DSM 40005. Proc. Natl. Acad. Sci. U. S. A. 100 :4316-4321.

222. Wiedemann, I., E. Breukink, C. van Kraaij, O. P. Kuipers, G. Bierbaum, B. de Kruijff, and H. G. Sahl. 2001. Specific binding of nisin to the peptidoglycan precursor lipid II combines pore formation and inhibition of cell wall biosynthesis for potent antibiotic activity. J. Biol. Chem. 276:1772-1779.

223. Wiedemann, I., T. Bottiger, R. R. Bonelli, A. Wiese, S. O. Hagge, T. Gutsmann, U. Seydel, L. Deegan, C. Hill, P. Ross, and H. G. Sahl. 2006. The mode of action of the lantibiotic lacticin 3147: a complex mechanism involving specific interaction of two peptides and the cell wall precursor lipid II. Mol. Microbiol. 61:285-296.

224. Willey, J. M. and W. A van der Donk . 2007 Lantibiotics: peptides of diverse structure and function. Annu. Rev. Microbiol. 61: 477-501.

225. Wu, J., and J. T. Watson . 1997 A novel methodology for assignment of disulfide bond pairings in proteins. Protein Sci . 6:391-398.

226. Xie, L., L. M. Miller, C. Chatterjee, O. Averin, N. L. Kelleher, and W. A. van der Donk. 2004a. Lacticin 481: in vitro reconstitution of lantibiotic synthetase activity. Science 303:679- 681.

227. Xie, L., and W. A. van der Donk . 2004b Post-translational modifications during lantibiotic biosynthesis. Curr. Opin. Chem. Biol . 8: 498-507.

228. Yamada, K., S. N. Iyer, M. C. Chapell, D. Ganten, and C. M. Ferrario . 1998. Converting enzyme determines plasma clearance of angiotensin-(1-7). Hypertension. 32: 496-502.

229. Yamada, T., M. Goto, V. Punj, O. Zaborina, M. L. Chen, K. Kimbara, D. Majumdar, E. Cunningham, T. K. Das Gupta, and A. M. Chakrabarty. 2002. Bacterial redox protein azurin, tumor suppressor protein p53, and regression of cancer. Proc. Natl. Acad. Sci. U.S.A. 99 :14098-14103.

230. Ye, S. Y., O. Koponen, M. Qiao, T. Immonen, and P. E. Saris. 1995. NisP is related to nisin precursor processing and possibly to immunity in Lactococcus lactis. J. Tongji Med. Univ 15:193-197.

231. You, Y. O. and W. A. van der Donk . 2007. Mechanistic investigations of the dehydration reaction of lacticin 481 synthetase using site-directed mutagenesis. Biochemistry 46 :5991- 6000.

232. You, Y. O., M. R. Levengood, L. A. Ihnken, A. K. Knowlton, and W. A. van der Donk . 2009. Lacticin 481 synthetase as a general serine/threonine kinase. ACS Chem. Biol. 4:379- 385.

233. Yuan, J., Z. Z. Zhang, X. Z . Chen, W. Yang, and L. D. Huan. 2004. Site-directed mutagenesis of the hinge region of nisinZ and properties of nisinZ mutants. Appl. Microbiol. Biotechnol. 64 :806-815.

117 Chapter 9

234. Zhang, X., and W. A. van der Donk . 2007a. On the substrate specificity of dehydration by lacticin 481 synthetase. J. Am. Chem. Soc. 129 :2212-2213.

235. Zhang, X., W. Ni, and W. A. van der Donk. 2007b. On the regioselectivity of thioether formation by lacticin 481 synthetase. Org. Lett. 9:3343-3346

118 Samenvatting

De toepassing van peptiden (kleine eiwitten) als medicijn is de laatste jaren enorm toegenomen en is economisch gezien een veelbelovende ontwikkeling in de farmaceutische sector. Er wordt dan ook veel geld geïnvesteerd in onderzoek en ontwikkeling van nieuwe peptiden die ziekten kunnen genezen of behandelen, voorkomen dan wel vroegtijdig opsporen. Wanneer peptiden de werking van eiwitten en antilichamen kunnen benaderen of evenaren, heeft de toepassing van peptiden meerdere voordelen. Omdat peptiden veel kleiner zijn kunnen ze makkelijker in het weefsel doordringen en worden ze minder snel herkend en weggevangen door het immuunsysteem. In vergelijking met kleine organische moleculen hebben peptiden vaak een grotere effectiviteit en specificiteit. Verder hebben peptiden in de meeste gevallen minder of geen toxische bijeffecten. Medicinale peptiden in mensen zijn vaak peptide hormonen die een rol spelen in signaalfuncties. Het stimuleren (agoneren) of blokkeren (antagoneren) van processen in het lichaam, vindt plaats door binding van deze peptiden aan receptoren. In tal van ziekten is de balans in stimulatie/remming van deze processen verstoord. Medicatie is dan nodig voor herstel. Peptiden kunnen ook een rol spelen in het afremmen of stimuleren van bepaalde enzymactiviteiten en kunnen een antivirale of een antibacteriële werking hebben. Een voorbeeld onder meer van een peptide met een antibacteriële werking is nisine. Nisine wordt geproduceerd door de melkzuurbacterie Lactococcus lactis en wordt al sinds 1969 gebruikt als conserveringsmiddel in voeding. Momenteel is de toename van multiresistente bacteriën die een bedreiging vormen voor de volksgezondheid een groot probleem. Het is dus van groot belang dat nieuwe antibiotica en antibacteriële peptiden ontwikkeld worden.

Een grote beperking voor de toedieningsroutes en de effectiviteit van peptiden als geneesmiddel is vooral de instabiliteit van deze peptiden in het menselijk lichaam door gevoeligheid voor maagzuur en / of voor proteases. Orale toediening behoort daardoor vaak niet tot de mogelijkheden en in plaats daarvan moeten peptiden worden geïnjecteerd. Voor een effectieve toepassing van het peptide als medicijn is het vaak nodig dat peptiden gestabiliseerd worden om afbraak tegen te gaan. Hiervoor bestaan verschillende methoden. De methode die de aanleiding vormt tot de studies in dit proefschrift is stabilisatie van peptiden door middel van cyclisatie in de vorm van thioether verbindingen. Thioether verbindingen zijn van nature aanwezig in lantibiotica. Dit zijn zeer stabiele peptiden met antimicrobiële werking en deze stabiliteit en antimicrobiële werking worden veroorzaakt door de aanwezigheid van de thioether verbindingen. Het antibacteriële peptide nisine is ook een lantibioticum en bevat vijf van deze thioether verbindingen. Thioether-gekoppelde aminozuren worden lanthionines en methyllanthionines genoemd (Fig. 1). Samenvatting

Figuur 1: Het antibacteriële peptide nisine bevat vijf thioether verbindingen, waarvan de eerste een lanthionine (Ala-S-Ala) is en de overige vier ringen methyllanthionines (Abu-S-Ala) zijn.

Deze (methyl)lanthionines worden in de peptiden geïntroduceerd met behulp van enzymen. Bij de synthese van nisine in de bacterie Lactococcus lactis zijn meerdere enzymen betrokken. De eerste modificatie stap is de dehydratatie van het merendeel aan serines en threonines in de propeptide keten, gekatalyseerd door het dehydratase NisB. Door dehydratie van de aminozuren serine en threonine worden dehydroalanine en dehydrobutyrine gevormd die vervolgens, regio- en stereo-specifiek, aan een cysteine gekoppeld worden door het cyclase NisC. Translocatie van het gemodificeerde peptide over de celmembraan vindt plaats via het transport eiwit NisT. Uiteindelijk wordt de nisine leader verwijderd door het protease NisP, zodat het actieve antibacteriële peptide nisine in het medium terecht komt. Het leader peptide is essentieel voor efficiënte modificatie en transport van het substraat peptide. Naast nisine zijn er zo’n zestig andere natuurlijke lantibiotica bekend.

In 1988 werd voor het eerst het idee geopperd lantibiotica enzymen te gebruiken om therapeutische peptiden te stabiliseren. Stabilisatie heeft als voordeel dat het therapeutische peptide in een lagere dosis of minder vaak toegediend hoeft te worden en dat orale toediening tot de mogelijkheden behoort. Ook kan door introductie van een thioether verbinding de interactie van het peptide met een receptor in positieve zin veranderen bijvoorbeeld door verhoogde specificiteit en / of effectiviteit. Zo is bekend dat een thioether-gestabiliseerde enkephaline variant die op chemisch wijze verkregen is, ongevoelig is voor proteolytische afbraak en veel effectiever werkt als medicijn dan de niet-gestabiliseerde lineaire variant. Dit proefschrift beschrijft de biologische productie van thioether-gestabiliseerde peptiden door gebruik te maken van lantibiotica synthese enzymen. De melkzuurbacterie Lactococcus lactis is gebruikt om deze peptiden te produceren. Er is een twee-plasmiden systeem geïntroduceerd om modificatie en productie te kunnen realiseren. Op het ene plasmide liggen de genen die coderen voor de modificatie enzymen en het transport eiwit. Op het andere plasmide ligt het gen dat codeert voor de fusie van leader peptide en substraat peptide. Het

120 Samenvatting

leader peptide zorgt ervoor dat het substraat peptide naar de modificatie enzymen en transport eiwit geleid wordt (Fig. 2).

Figuur 2: Biologische productie van thioether-gestabiliseerde peptiden.

Na isolatie van het gemodificeerde substraat peptide kan het leader peptide specifiek verwijderd worden met behulp van een protease, zodat het therapeutische peptide overblijft. Vervolgens kan het therapeutische peptide verder geanalyseerd worden. Is het stabieler? Bindt het nog aan de receptor en vindt er nog steeds signaaltransductie plaats? Samenvattend behandelen de hoofdstukken van dit proefschrift de volgende onderwerpen. Hoofdstuk 1 bevat de algemene inleiding. In Hoofdstuk 2 wordt uiteen gezet dat het nisine transport enzym NisT ook andere, niet-lantibiotica-gerelateerde, peptiden kan transporteren mits voorafgegaan door de nisine leader. Ook wordt in Hoofdstuk 2 aangetoond dat het enzymcomplex NisBTC gesplitst kan worden in onafhankelijk van elkaar functionerende enzymen / enzym subcomplexen. NisT is functioneel zonder de aanwezigheid van de modificatie enzymen NisB en NisC. Tevens is aangetoond dat NisB een peptide kan dehydrateren dat totaal verschilt van nisine en dat NisB functioneel is in afwezigheid van het cyclase NisC. Aansluitend wordt in Hoofdstuk 3 beschreven dat NisB een brede substraat specificiteit heeft en dat NisC ook niet-lantibiotica peptiden kan cycliseren. Hiermee is dus aangetoond dat het mogelijk is om de melkzuurbacterie Lactococcus lactis een thioether-gestabiliseerde

121 Samenvatting

peptide te laten produceren dat kandidaat geneesmiddel kan zijn. In Hoofdstuk 4 wordt de mogelijke toepassing van enzymen die betrokken zijn bij de synthese van het lantibioticum lacticine 3147 voor introductie van thioether verbindingen bestudeerd. Hoewel het modificatie enzym LtnM2 dat zowel de dehydratase activiteit als de cyclase activiteit bezit, ook zeer verschillende peptiden modificeert, is het niet duidelijk of dit enzym een net zo grote substraat specificiteit heeft als NisB en NisC. Door Lactococcus lactis met het lacticine 3147 modificatie en export systeem worden namelijk verschillende peptiden niet uitgescheiden in het medium. Dit kan komen doordat de leaderpeptidase activiteit (het verwijderen van de leader) en de transport activiteit beiden uitgevoerd worden door één enzym, namelijk LtnT, wat het enzym misschien substraat specifieker maakt. In Hoofdstuk 5 en 6 worden de mogelijkheden van het gebruik van de algemene ‘Sec’ transport route van Lactococcus lactis voor translocatie van gemodificeerde peptiden over de celmembraan besproken. In Hoofdstuk 5 wordt aangetoond dat wanneer de nisine leader voorafgegaan wordt door een ‘Sec’ transport signaal peptide van wel 44 aminozuren lang, er nog steeds modificaties plaats vinden gekatalyseerd door NisB en NisC. Gedehyrateerde peptiden worden succesvol getransporteerd via het ‘Sec’ transportsysteem. Nisine wordt echter niet getransporteerd via deze transport route. Door de meerdere thioether ringen heeft nisine een te brede diameter voor transport via ‘Sec’. Maar zoals Hoofdstuk 6 laat zien, kan een peptide met één enkele thioether ring wel succesvol via het ‘Sec’ transportsysteem worden getransloceerd. Vervolgens wordt de succesvolle toepassing van het therapeutische peptide angiotensine-(1-7), biologisch gestabiliseerd met een thioether verbinding, beschreven in Hoofdstuk 7. Toekomstperspectieven en resultaten worden besproken en samengevat in Hoofdstuk 8.

122

Nawoord

Dit proefschrift is natuurlijk mede tot stand gekomen met hulp en begeleiding van vele andere mensen. Een dankwoord is dan ook wel op zijn plaats en een mooie afsluiting van dit proefschrift.

Na een aantal jaren thuis te hebben gezeten om de zorg op me te nemen voor mijn twee geweldige zonen Jos en Luco ben ik weer aan het werk gegaan als analiste in Haren. Dankzij Harma Karsens, een vertrouwde studie – en kamergenote, ben ik bij de vakgroep Moleculaire Genetica terecht gekomen waar ik anderhalf jaar gewerkt heb. Daarna ben ik samen met Kees Leenhouts vanuit Haren naar Groningen vertrokken om bij de net opgestarte stichting Biomade te beginnen. Hoewel de projectgroep Peptide Antibiotics, onder leiding van Kees, na anderhalf jaar gestopt werd heb ik met veel plezier aan dit onderzoek gewerkt. Gelukkig voor Kees heeft het andere project, Protein Anchor, waar hij ook leiding aan gaf wel succes gehad en is sinds 2007 verder gegaan als het spin-off bedrijf Mucosis B.V.. In de elf jaren die ik bij Biomade heb gewerkt zijn veel projecten gestart en gesneuveld en heb ik veel collega’s zien komen en gaan. Zelf heb ik bij drie verschillende projecten binnen Biomade gewerkt. Ik heb met zeer veel verschillende collega’s gebabbeld, gelachen en gemopperd. Ik ga jullie niet allemaal bij naam noemen maar wil jullie alsnog bedanken voor de goede werksfeer.

Mijn laatste projectgroep waar ik sinds 2003 onderdeel van ben is de groep Therapeutic Peptides. Ook dit project is veelbelovend gebleken en staat op het punt verder te gaan als het spin-off bedrijf LanthioPep. Deze projectgroep is sinds het begin succesvol geleid door doctor Gert N. Moll en dankzij Gert heb ik me verder in de wetenschap weten te bekwamen. Doordat ik de ruimte kreeg om zelfstandig onderzoek te doen en te publiceren, werd het mogelijk om onder niet gebruikelijke omstandigheden te kunnen gaan promoveren. Gert je bent een geweldige leermeester geweest en ik wil je hiervoor dan ook hartelijk bedanken. Professor Oscar P. Kuipers van de vakgroep Moleculaire Genetica uit Haren wil ik bedanken voor de begeleiding die ik heb gehad bij het tot stand komen van dit proefschrift en voor het feit dat hij mijn promotor wil zijn. Verder wil ik professor George T. Robillard, oprichter van Biomade, nog bedanken. Zonder Biomade zou deze promotie niet mogelijk zijn. Collega’s die direct betrokken zijn geweest bij het project en sinds enige tijd vertrokken zijn, maar ook een grote bijdrage hebben geleverd zijn: Leon Kluskens, Esther de Boef, Susan Fekken, Jenny Wierenga, Karin Scholtmeijer, Marijke Haas en Anita Meter-Arkema. Allemaal heel erg bedankt en in het bijzonder Leon die een groot aandeel heeft gehad in het tot stand komen van dit proefschrift. Van

123 Nawoord

Hoofdstuk 3 zijn we allebei eerste auteur en van Hoofdstuk 7 dat de kroon op ons werk beschrijft, is hij ook eerste auteur. Natuurlijk wil ik ook nog even mijn huidige collega’s noemen en bedanken. Rick Rink die met zijn ideeën altijd vooruit holt en altijd weer nieuwe mogelijkheden ziet. Rick, jouw gedrevenheid en uitbundigheid werken zeer aanstekelijk en ik vind het geweldig om nu al acht jaar met jou onderzoek te kunnen doen. Tjibbe Bosma die ook al vanaf het begin bij Biomade werkzaam is, is de rots in de branding, de nuchtere Fries. Tjibbe, je hebt jouw expertise, ‘display-technologie’, geniaal weten te combineren met de biologische productie van thioether-gestabiliseerde peptiden en ik hoop voor je dat dit een groot succes gaat worden. Louwe de Vries, de sportjunk Fries, houdt van facilitair managen en is een zeer belangrijke schakel in onze huidige TP-groep na het wegvallen van de in vivo projectgroep. De laatste welkome aanwinst voor onze TP-groep en mijn lunchmaatje is Annechien Plat. Gert, Rick, Tjibbe, Louwe en Annechien, ik hoop nog een flink aantal jaren met jullie vooruit te kunnen in de hopelijk snel op te starten Lanthiopep B.V..

Verder wil ik mijn andere Biomade collega’s succes wensen: Erna Bulten, Menno de Jong en Gerald Metselaar met hun op te zetten spin-off bedrijf Nano F.M., Zedef Karakayali en natuurlijk Jarmila Šmisterová die het onderzoek gaat verlaten en de lerarenopleiding scheikunde gaat doen. En ‘last but not least’ wil ik mijn grote liefde Arie nog bedanken. Arie, jouw reflectie, steun en aanmoedigingen die ik al die tijd heb mogen ontvangen zijn van onschatbare waarde.

124

List of publications

Kuipers, A., E. de Boef, R. Rink, S. Fekken, L. D. Kluskens, A. J. Driessen, K. Leenhouts, O. P. Kuipers, and G. N. Moll. 2004. NisT, the transporter of the lantibiotic nisin, can transport fully modified, dehydrated, and unmodified prenisin and fusions of the leader peptide with nonlantibiotic peptides. J. Biol. Chem. 279:22176-22182.

Rink, R., A. Kuipers, E. de Boef., K. J. Leenhouts, A. J. Driessen, G. N. Moll, and O. P. Kuipers. 2005. Lantibiotic structures as guidelines for the design of peptides that can be modified by lantibiotic enzymes. Biochemistry 44:8873-8882

Kluskens#, L. D., A. Kuipers#, R. Rink, E. de Boef, S. Fekken, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2005. Post-translational modification of therapeutic peptides by NisB, the dehydratase of the lantibiotic nisin. Biochemistry 44:12827-12834.

Kuipers, A., J. Wierenga, R. Rink, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2006. Sec-mediated transport of posttranslationally dehydrated peptides in Lactococcus lactis . Appl. Environ. Microbiol. 72:7626-7633.

Rink, R., J. Wierenga, A. Kuipers, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007a. Production of dehydroamino acid-containing peptides by Lactococcus lactis . Appl. Environ. Microbiol. 73: 1792–1796.

Rink, R., J. Wierenga, A. Kuipers, L. D. Kluskens, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007b. Dissection and modulation of the four distinct activities of nisin by mutagenesis of rings A and B and by C-terminal truncation. Appl. Environ. Microbiol. 73:5809-5816.

Rink, R., L. D. Kluskens, A. Kuipers, A. J. Driessen, O. P. Kuipers, and G. N. Moll. 2007c. NisC, the cyclase of the lantibiotic nisin, can catalyze cyclization of designed nonlantibiotic peptides. Biochemistry 46:13179-13189.

Kuipers, A., J. Meijer-Wierenga, R. Rink, L. D. Kluskens, and G. N. Moll . 2008. Mechanistic dissection of the enzyme complexes involved in biosynthesis of lacticin 3147 and nisin. Appl. Environ. Microbiol . 74 : 6591-6597.

Kluskens, L. D., A. Nelemans, R. Rink, L. de Vries, A. Meter-Arkema, Y. Wang, T. Walther, A. Kuipers, G. N. Moll, and M. Haas. 2009. Angiotensin-(1-7) with thioether-bridge: an ACE-resistant, potent Ang-(1-7) analogue. J. Pharmacol. Exp. Ther. 328: 849-854.

Kuipers, A., R. Rink, and G. N. Moll . 2009. Translocation of a thioether-bridged azurin peptide fragment via the sec pathway in Lactococcus lactis , Appl. Environ. Microbiol . 75 : 3800-3802.

Rink, R., A. Arkema-Meter, I. Baudoin, E. Post, A. Kuipers, S. A. Nelemans, M. Haas Jimoh Akanbi, and G. N. Moll. 2010. To protect peptide pharmaceuticals against peptidases. J Pharmacol. Toxicol. Methods, submitted.

Moll, G. N., A. Kuipers, and R. Rink . 2010. Microbial engineering of dehydro-amino acids and lanthionines in non-lantibiotic peptides. Antonie Van Leeuwenhoek. 97 :319-333.

Kuipers, A., R. Rink,and G. N. Moll . 2010. Genetics, biosynthesis, structure and mode of action of lantibiotics. In: Prokaryotic Antimicrobial Peptides: from genes to applications. Editors: D. Drider and S. Rebuffat, in press.

#: contributed equally to this paper

125