Production of and Response to the Cannibalism Peptide SDP in Bacillus subtilis

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Summer 2013 Production of and Response to the Cannibalism Peptide SDP in Bacillus subtilis Tiara G. Perez Morales University of Iowa Copyright 2013 Tiara G. Perez Morales This dissertation is available at Iowa Research Online: Recommended Citation Perez Morales, Tiara G.. "Production of and Response to the Cannibalism Peptide SDP in Bacillus subtilis." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Microbiology Commons

2 PRODUCTION OF AND RESPONSE TO THE CANNIBALISM PEPTIDE SDP IN BACILLUS SUBTILIS by Tiara G. Pérez Morales A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Microbiology in the Graduate College of The University of Iowa August 2013 Thesis Supervisor: Assistant Professor Craig Ellermeier

3 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Tiara G. Perez Morales has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Microbiology at the August 2013 graduation. Thesis Committee: Craig Ellermeier, Thesis Supervisor David Weiss John Kirby Timothy Yahr Jeffrey Banas

4 To my family who believed in me, to my dearest friends who kept me sane, and to my wonderful husband Joe; I am grateful for your love and support these years in Graduate School. ii

5 And, when you want something, all the universe conspires in helping you to achieve it. Paulo Coelho iii

6 ACKNOWLEDGEMENTS There are many things I have enjoyed about Iowa these years; the unpredictable weather, the numerous festivals, and the welcoming friendship of the people who live here. Most of all is the people who made my life in Graduate School great. I would like to thank Dr. Craig Ellermeier for his commitment to my work. I am grateful for all the helpful discussions, his minute details in seminar presentations and writing. These are all qualities that I treasure and hope I continue to develop. I would like to thank all the lab members for their advice during lab meetings and outside too. Thanks to T for her 3-D birthday cakes, especially the puffer fish cake. It has been very entertaining to see the lab grow in numbers and vertical in space. I would like to thank my thesis committee who guided me through my Master s and PhD training. I thank the Microbiology Department staff for always being so caring and obliging. Special thanks to Julie Nealson who has been a truthful friend and my godmother. To my friends who love all sorts of things that may be neglected in the non-science community. Dinosaur afternoons with Cindy, fantasy novels with Jed, and anime sagas with Beth; thanks for reaching out. To my family, I am thankful for their understanding and patience as I move in this career. Their support has never failed me. Lastly, to my husband Joe with whom I have managed to spend every Monday seminar in Grad School since I met him. I cannot thank you enough for your trust, comfort and love. It takes a village to raise a kid and it definitely takes all of you to make Grad School possible. iv

7 ABSTRACT The Gram positive soil dwelling bacteria Bacillus subtilis produces spores when encountered with a low nutrient environment. However, B. subtilis can delay spore production by a mechanism known as cannibalism. Cannibalism is a process by which B. subtilis delays commitment to sporulation by killing a subpopulation of its cells. This process involves production of two toxins, SDP and SKF. SDP is a 42 amino acid peptide with a disulfide bond derived from the internal cleavage of its precursor protein pro-sdpc. pro-sdpc is part of the sdpabc operon. Production of extracellular SDP induces expression of the sdpri operon. Encoded in this operon is the negative regulator SdpR and SdpI. SdpI is a dual function protein which acts both as a signal transduction protein and the immunity factor against SDP. The current model states that production of SDP is sensed via SdpI. SdpI will sequester SdpR to the membrane in response and allow for sdpri expression. The aims of this dissertation are to establish the requirements for SDP production and its response via SdpI/SdpR during cannibalism. Studies in Chapter II were carried out to determine the factors required for production of the antimicrobial peptide SDP. Site-directed mutagenesis of the leader signal peptide sequence in pro-sdpc demonstrated that proper signal peptide cleavage was required for SDP production. Additional site-directed mutants of the cysteine residues in pro-sdpc revealed that these are not required for SDP toxic activity. These studies also included deletions within the sdpabc operon and revealed that the two proteins of unknown function, SdpA and SdpB are required for SDP production. Using mass spectrometry analysis, we found that SdpA and SdpB together are required to produce the active 42 v

8 amino acid peptide SDP. Taken together we concluded that SDP production was a multi-step process which required proteins encoded within the operon and additional processing supplemented in the cell. In Chapter III we investigated the role of SdpI, specifically the residues required for the signaling and immunity functions. Our initial screen included site-directed mutagenesis of 20 highly conserved residues located between the 4 th and 5 th transmembrane domains of SdpI. Of those, only two SdpI mutants had defects in either signal transduction or SDP immunity. Additional localized mutagenesis was used to isolate two other mutants in SdpI which only affected signal transduction or SDP immunity. SdpI signaling - immunity + mutants presented a defect in SdpR membrane sequestration and sdpri induction. Our findings suggest these types of SdpI mutants may be important for the downstream effect of SdpR membrane sequestration. SdpI signaling + immunity - mutants revealed defects in SDP protection. Some of the residues mutated were conserved in other SdpI homologs. Site directed mutagenesis of these conserved residues in the SdpI ortholog YfhL showed these are also required for SDP resistance. For the first time, we were able to identify mutations which affected only SDP immunity and gained further insight into how SdpI signaling - immunity + mutants play a role during signal transduction. In Chapter IV we initiated studies to define what regions of the negative regulator SdpR are important for its function during cannibalism. We employed localized mutagenesis to identify SdpR mutants which decreased sdpri expression even in the presence of inducing signal. We isolated three such SdpR mutants, referred to as super repressors. We expect these SdpR super repressors are unable to be sequestered to the membrane in the presence of SDP. vi

9 TABLE OF CONTENTS LIST OF TABLES... x LIST OF FIGURES... xi LIST OF ABBREVIATIONS... xiii CHAPTER ONE: INTRODUCTION... 1 Bacillus subtilis... 2 The sporulation process in Bacillus subtilis... 2 Entry into sporulation... 4 Phosphatases and Aspartyl-phosphatases involved in suppresion of sporulation... 5 CodY inhibition of sporulation... 7 Spo0A regulation... 8 AbrB negative regulation... 9 Regulation of cannibalism The skf operon The sdpabc operon The sdpri operon The dual function protein SdpI Other SDP immunity factors Other examples of cannibalistic behavior in bacteria CHAPTER TWO: PRODUCTION OF THE CANNIBALISM TOXIN SDP IS A MULTI STEP PROCESS THAT REQUIRES SDPA AND SDPB Introduction Materials and Methods Bacterial strains and growth Construction of plasmids Site directed mutagenesis of SdpC β-galactosidase activity assay SDP mediated killing assay Subcellular fractionation of cells Immunoblot analysis of SdpC In situ assay to monitor SDP Mass spectrometry analysis of strains to detect SDP Results vii

10 Signal peptide cleavage is required for SDP activity SdpA and SdpB required for SDP activity Export and secretion of SdpC does not require SdpA and SdpB SdpAB are required for SDP production SDP disulfide bond is not essential for activity SdpC disulfide bond formation occurs independently of SdpAB SdpA is a cytoplasmic protein Discussion Production of SDP requires multiple steps Possible roles for SdpA and SdpB CHAPTER THREE: ANALYSIS OF SDPI RESIDUES REQUIRED FOR SIGNAL TRASNDUCTION AND IMMUNITY TOWARDS THE ANTIMICROBIAL PEPTIDE SDP Introduction Materials and Methods Media and Strains Plasmid construction Localized mutagenesis of SdpI SDP killing assay SdpI competition assay SdpR-GFP localization assay β-galactosidase Assay Immunoblot analysis of SdpI and YfhL Results Isolation of SdpI mutants with altered SDP immunity or signaling SdpI Signaling- Immunity + mutants fail to induce sdpri Expression SdpI Signaling- Immunity + mutants provide resistance to SDP SdpI Signaling- Immunity + mutants fail to localize SdpR-GFP to the membrane in response to SDP SdpI Signaling- Immunity + mutation is dominant to a constitutive SdpI mutation SdpI Signaling + Immunity - mutants fail to provide SDP resistance SdpI Signaling + Immunity - mutants retain signaling function SdpI Signaling + Immunity - mutants can sequester SdpR-GFP to the membrane viii

11 sdpri expression in SdpI Signaling + Immunity - mutants is SDP dependent SDP-protection in YfhL mutants SdpI C.difficile provides partial resistance to SDP Discussion CHAPTER FOUR: INITIAL CHARACTERIZATION OF SDPR SUPER REPRESSORS AND THEIR ACTIVITY DURING CANNIBALISM Introduction Materials and Methods Media and Strains Plasmid construction SdpR localized mutagenesis Results Isolation of SdpR super repressors Discussion Role of SdpR super repressors during signal response to SDP CHAPTER FIVE: DISCUSSION AND FUTURE DIRECTIONS SDP production overview Role of secretion and disulfide bond Proteins in SDP production Role SdpA and SdpB in generation of SDP SDP response via SdpI and SdpR overview SdpR membrane sequestration SDP mediated immunity by SdpI SdpI topology Characterization of SdpR super repressors Summary REFERENCES ix

12 LIST OF TABLES Table 1. Strains used in Chapter II Oligos used in Chapter II Plasmids used in Chapter II Strains used in Chapter III Localization of SdpI mutants and respective phenotypes Plasmids used in Chapter III Oligos used in Chapter III Competitive index of SdpI to SdpI mutants Strains used in Chapter IV Oligos used in Chapter IV Plasmids used in Chapter IV x

13 LIST OF FIGURES Figure 1. Bacillus subtilis sporulation process Regulatory pathways which lead to different cellular development programs in B.subtilis Cannibalism in B. subtilis The sdpabcir operons in B. subtilis SDP toxin production model Signal peptide cleavage is required for full secretion and activity of SDP SdpAB are required for induction of the sdpri operon and SDP toxicity Secretion of SdpC does not require SdpAB SdpAB are required for production of 42 amino acid peptide SDP SDP disulfide bond formation is not essential for activity and is independent of SdpAB SdpA is a cytosolic protein SdpI conserved residues in B. subtilis and SdpI homologs PsdpRI-lacZ induction and SDP-sensitivity screen Several highly conserved within the cytosolic loop of SdpI have no effect on sdpri expression SdpI signaling - immunity + mutants fail to function as a signal transduction protein SdpR-GFP cannot be sequestered by SdpI signaling - immunity + mutants xi

14 17. SdpI signaling - immunity + mutants can overcome an SdpI constitutive mutants SdpI F78IQ126L fails to sequester SdpR-GFP SdpI signaling + immunity - mutants are SDP-sensitive SdpI R137A and SdpI N123A effect on sdpri expression in the absence of SDP Amino acid sequence alignment of SdpI ortholog YfhL in B. subtilis, and the homologs in S. mutans UA195 and C. difficile SDP protection for SdpI ortholog YfhL and YfhL mutants SDP peptide protection for SdpI homologs SdpI S.mutans and SdpI C.difficile in B. subtilis SdpR topology and model for membrane sequestration via SdpI SdpR super repressors decrease sdpri expression in the presence or absence of SdpI xii

15 LIST OF ABBREVIATIONS AMP DSM ECF EMS GLA GLU IPTG IMS KO LB MALDI MS RIP SDS-PAGE SDP SKF VKD-γ carboxylase X-GAL ONPG antimicrobial peptide Difco sporulation media Extra-cytoplasmic function Ethyl methanosulfonate γ-carboxylated glutamic acid glutamic acid Isopropyl-β,D-thiogalactopyranoside imaging- mass spectrometry vitamin K 2,3, epoxide Luria Bertani matrix assisted laser desorption ionization mass spectrometry Regulated intramembrane proteolysis sodium dodecyl sulfate polyacrylamide gel electrophoresis Sporulation delaying protein Sporulation killing factor vitamin K dependent gamma carboxylase 5-bromo-4-chloro-3-indolyl-β,D-galactopyranoside Ortho-nitrophenyl-β-D-galactopyranoside xiii

16 CHAPTER I:INTRODUCTION 1

17 2 Bacillus subtilis The Gram-positive spore forming bacteria Bacillus subtilis has been used in laboratory settings for many years as a model organism to understand basic and complex cellular mechanisms in other bacteria. The majority of B. subtilis strains used in the laboratory are derivatives of an isolate from Marburg strain (Burkholder and others 1947). The undomesticated wild isolate NCIB 3610 is widely known for studies involving multi-cellular functions, including fruiting body formation, biofilm formation and swarming motility. The B. subtilis strain used in our studies is PY79. PY79 is a laboratory domesticated derivative strain of B. subtilis 168, which is unable to swarm due to having several non-functional genes (Youngman and others 1984). Several studies have shown there are differences in the genome sequence between PY79 and 168, which include deletion of different sections within the PY79 genome and addition of genes which are not present in 168 (Zeigler and others 2008). Strains 168 and PY79 are naturally more competent than 3610 and thus the ability to genetically manipulate these strains has made them an attractive model. The sporulation process in Bacillus subtilis B. subtilis begins the process of sporulation when vegetative cells have exhausted any available nutrients (Trach 1991). Entry into sporulation is

18 3 governed by Spo0A (discussed below). In its first stages, cells will divide into two compartments, the mother cell and the forespore, via asymmetric cell division (Parker and others 1996). This asymmetric septation is driven by Spo0A and σ H regulated genes (Levin and others 1994). Proper spore formation requires inter-compartmental regulation between the mother cell and the forespore (Li and others 2001). During this time, which lasts approximately six hours (Errington 1993), the mother cell will nurture the forespore into a mature spore (Figure 1). This regulation is an intricate set of events where both mother cell and forespore activate cell-specific sigma factors important for spore development. The sigma factor σ F is the first member of this cascade and is produced in an inactivated form prior to asymmetric cell division (Margolis and others 1991). After asymmetric cell division has ended, σ F will be activated. σ F is negatively regulated by the anti-sigma factor, SpoIIAB. The interaction between σ F and SpoIIAB can be disrupted via the anti-anti sigma factor SpoIIAA. SpoIIAA interacts with SpoIIAB to release σ F (Errington 2001). σ F will then induce genes important for activation of σ E within the mother cell (Gholamhoseinian and others 1989; Narula and others 2012) (Figure 1). σ E activation occurs via protein cleavage of pro-σ E into σ E by a membrane bound protease, SpoIIGA (LaBell and others 1987; Stragier and others 1988). Both σ F, in the forespore, and σ E, in the mother cell, induce genes important for later stages in spore development. This in turn allows for forespore

19 4 engulfment and initiation of spore maturation via activation of the sigma factor σ G via σ F and later itself (Sun and others 1991). σ G is inhibited by SpoIIAB (Kellner and other 1996) and is activated only after forespore engulfment (Errington and others 1992). Later stages involve activation of σ K in the mother cell via regulation of σ G in the forespore (Oke and others 1997; Cutting and others 1991). σ K is made after forespore engulfment and is produced as an inactive form (Kunkel and others 1990). A signal made in the forespore by SpoIVB (Wakeley and others 2000) will trigger the next steps for σ K activation. σ K will then be activated via the protease SpoIVFB (Resnekov 1999; Ricca and others 1992; Lu and others 1995); active σ K will then govern the later stages of spore development. The proteinaceous spore coat, which allows the spore to be resistant to many environmental stresses, will then be assembled. Once the spore has matured, the cells will release the spore in the environment (Smith and others 1995; Nugroho and others 1999) (Figure 1). Entry into sporulation Nutrient limiting conditions such as amino acid starvation and low phosphate contribute to initiation of sporulation (Heinze and others 1978; Lopez and others 1979) (Figure 2). Detection of these signals is made via the protein kinases KinA and KinB (LeDeaux and others 1995; Trach and others 1991; Antoniewski and others 1990) (Figure 2). KinA and KinB sense and

20 5 respond to changes in ATP levels, and a decrease in cell transcription (Eswaramoorthy and others 2010; Tojo and others 2013). Other kinases such as KinC, and KinD are important for sporulation but have additional roles. KinC and KinD are activated in response to unknown signals and are important for biofilm formation in the wild type strain (LeDeaux and others 1995; Kobayashi and others 1995; Castilla Llorente and others 2008; Shemesh and others 2010) while KinE does not play a role for sporulation (Fujita and others 2005). Upon autophosphorylation, KinA and KinB indirectly phosphorylate Spo0A, the master regulator of sporulation, via a phosphorelay. KinA and KinB transfer a phosphate group to Spo0F (Burbulys and others 1991). Spo0F-P then transfers the phosphate group to the phosphotransferase Spo0B. Finally, Spo0B will transfer the phosphate to Spo0A, resulting in activated Spo0A-P (Burbulys and others 1991; Hoch 1993). Spo0A-P then initiates the process of sporulation by controlling expression of genes involved in asymmetric cell division (Parker and others 1996; Levin and others 1996). Phosphatases and Aspartyl-phosphatases phosphatases involved in suppression of sporulation As has been described, commitment to sporulation is regulated very tightly in B. subtilis. There are several pathways that inhibit cells from phosphorylating Spo0A (Figure 3). The phosphatase Spo0E is expressed from

21 6 a SigB dependent promoter under stress conditions (Reder and others 2012). Studies have shown that spo0e expression is negatively regulated by the transcriptional regulators AbrB and Rok (Reder and others 2012; Strauch and others 1989a; Strauch and others 1989b). Spo0E levels are also regulated post-translationally by the metalloprotease FtsH (Le and others 2009). Spo0E was found to maintain a balance via de-phosphorylation of Spo0A-P (Perego and others 1987; Perego and others 1991; Ohlsen and others 1994). Additionally, residues have been identified in both Spo0E and Spo0A which affect the ability of Spo0E to de-phosphorylate Spo0A (Stephenson and others 2002; Diaz and others 2008 and 2012). The aspartyl-phosphatases RapA, RapB and RapE are also known to inhibit the Spo0A phosphorelay pathway (Antoniewski and others 1990; Perego and others 1996; Tzeng and others 1998; Veening and others 2005). RapA, RapB, and RapE dephosphorylate Spo0F-P to allow for cell developmental pathway, such as competence, to occur. The aspartylphosphatase RapB is produced constitutively in the cell (Tzeng and others 1998; Perego and others 1994) while RapA is activated via the two-component system ComPA (Mueller and others 1992). ComPA is in turn activated via the pheromone ComX at high cell densities (Magnuson and others 1994). RapA can be inhibited via the pheromone PhrA so that sporulation occurs (Perego and others 1996). Residues important for the catalytic interaction of RapA and RapB with Spo0F have been determined (Stephenson and others 2002;

22 7 Diaz and others 2012). Lastly, RapE also de-phosphorylates Spo0F-P similar to RapA and is inhibited by PhrE (Figure 2) (Perego and others 1994; Jiang and others 2000). CodY inhibition of sporulation Spo0A activation can also be inhibited via CodY. CodY can sense changes in cellular GTP and branched chain amino acid levels (Ratnayake Lecamwasam and others 2001; Shivers and others 2004; Slack and others 1995; Guedon and others 2001). CodY functions mostly as a negative regulator (Molle and others 2003b). In exponentially growing cells, CodY inhibits expression of genes involved in cell metabolism as well as competence (Ratnayake Lecamwasam and others 2001; Molle and others 2003b; Lazazzera and others 1999). This is mediated by CodY interaction with GTP and the branched chain amino acid signals (Handke and others 2008; Villapakkam and others 2009) Levels of CodY repression are dictated by CodY binding sites located at these promoter regions (Figure 2) (Belitsky and others 2013; Belitsky and others 2011). A drop in GTP or amino acid levels will relieve CodY repression of genes important for amino acid transport, the sporulation kinase KinB, and the phosphatase/pheromone inhibitors RapA/PhrA and RapE/PhrE (Molle and others 2003b). These last are important for high levels of Spo0F phosphorylation during entry into sporulation.

23 8 Spo0A regulation Spo0A is the master regulator of sporulation in B. subtilis. The spo0a promoter is divided into Pv and Ps; which indicate vegetative and sporulation respectively (Chibazakura and others 1991). Pv is a σ A -dependent promoter while Ps is a σ H dependent promoter (Fujita and others 1998; Predich and others 1992). Within the promoter sequence of Spo0A, there are four elements or Spo0A boxes but only three are activated (O1, O2, and O3) which play a role in Spo0A levels in exponential and stationary phase (Strauch and others 1992; Chastanet and others 2011). Box O2 was shown to have a role in inhibiting Ps during exponential growth. Boxes O1 and O3 are important for induction of spo0a during entry to sporulation via the Ps promoter (O3) and inhibiting induction from the Pv promoter (O1) (Chastanet and others 2011). This regulation was suggested to be via Spo0A-P (Strauch and others 1992). As mentioned before, once Spo0A is phosphorylated, Spo0A-P can activate or repress more than 200 genes (Chibazakura and others 1991; Molle and others 2003a). Spo0A-P positively regulates itself, spo0f and sigh (Fujita and others 2005). Previous work has determined that differences in Spo0A-P levels govern which genes will be activated during stationary phase (Fujita and others 2005; Mirouze and others 2011). High levels of Spo0A-P will initiate sporulation and expression of SinI (sporulation inhibitor) (Shafikhani and others 2004). SinI is a repressor of the negative regulator SinR (Bai and others 1993). SinI repression of SinR relieves rok expression (Kearns and

24 9 others 2005). Production of Rok (repressor of ComK) will inhibit the competence master regulator protein ComK, thus blocking competence and pushing the cells towards sporulation (Figure 2) (Hoa and others 2002). Rok can also inhibit expression of genes involved in antimicrobial peptide synthesis (Albano and others 2005). Lastly, other genes involved in later stages of sporulation development are induced by high levels of Spo0A-P (Fujita and others 2005). Additional induction of the Rap inhibitors is mediated by σ H, which we have mentioned is induced via Spo0A-P. Low concentrations of Spo0A-P can regulate other genes in the cell. It is here that Spo0A-P creates a positive feedback loop by activating expression of kina, spo0a and spo0f (Fujita and others 2005). Low levels of Spo0A-P turn on genes important for production, export and immunity to toxins directly or indirectly via repression of AbrB; this includes the skf and sdp operons (discussed below). The difference in activation from these different promoters is due to the binding affinity of Spo0A-P to DNA promoters (Fujita and others 2005). AbrB negative regulation One of the most important regulatory effects of Spo0A-P is repression of the negative regulator AbrB (Perego and others 1988). AbrB is produced at high levels during exponential phase (Strauch and others 1989a). At this stage, AbrB represses expression of genes required for sporulation such as

25 10 spo0a (Strauch and others 1989a; Robertson and others 1989). In addition, AbrB represses expression of several genes encoding antibiotics during vegetative growth either directly or by repression of sigh (Figure 2) (Albano and others 2005; Stein 2005). Some of these regulated antimicrobial operons include: sdp (SDP), skf (SKF), sbo-alb (Subtilosin), spa (Subtilin), tasa, bcl (Bacilysin), and srf (Surfactin) (Stein 2005; Ellermeier and others 2006a; Gonzalez-Pastor and others 2003; Zheng 2000; Stöver and others 1999; Karataş and others 2003). When cells reached nutrient-limiting conditions, Spo0A-P levels begin to rise. Spo0A-P induces expression of abba (Banse and others 2008) and can directly inhibit AbrB. In turn, AbrB levels decrease via direct repression of abrb by Spo0A-P (Greene and others 1996; Strauch and others 1990a; Strauch and others 1990b). With respect to cannibalism we are specifically interested in the operons which are induced by Spo0A via repression of AbrB (Fujita and others 2005); the sdpabc, sdpri, and the skfabcedfgh operon (Ellermeier and others 2006a; Gonzalez-Pastor and others 2003). Regulation of cannibalism The skf and sdp operons are induced indirectly by low levels of Spo0A- P (Fujita and others 2005; Chen and others 2006) via repression of AbrB (Strauch and others 1990b) (Figure 2). In addition, activation of skf expression requires Spo0A-P to directly bind to an extended -35 region of the

26 11 skf promoter at low Spo0A-P levels. At high Spo0A-P levels, Spo0A-P can repress expression of sdp transcription by binding to a low affinity binding site in the sdpa promoter (Fujita and others 2005). These differences in levels of Spo0A-P create a bi-stable switch; thus, activation of the sdp and skf operons occurs stochastically as it depends on the levels of Spo0A-P in each cell in the population (Chastanet and others 2010; Chung and others 1994; Dubnau and others 2006). This will result in a sub-population of cells which have activated Spo0A-P (Spo0A-ON) and one that has not reached this threshold (Spo0A-OFF) (Figure 3). Expression of these operons allows cells to delay the process of sporulation while possibly increasing the nutrient availability (Ellermeier and others 2006a; Gonzalez-Pastor and others 2003). This delay is possible due to the expression of toxins and immunity proteins encoded in the operons (Figure 3) (discussion below). Cells that do not activate sdp or skf are unable to produce toxin or immunity, rendering them sensitive to their siblings (Figure 3) (Ellermeier and others 2006a; Gonzalez- Pastor and others 2003). Nonetheless, once cells have again depleted all available nutrients, cell will cycle back to sporulation and Spo0A-P will negatively regulate expression of these genes, allowing for sporulation to initiate (Fujita and others 2005).

27 12 The skf operon The skf operon encodes for SkfABCEFGH. skfa encodes the toxic peptide SKF. SKF is one of two killing factors involved in B. subtilis cannibalism (Gonzalez-Pastor and others 2003). SKF was first described as a toxin which could kill the plant pathogen Xanthomonas oryzae (Lin and others 2001). Later work demonstrated that SKF is a 26 amino acid peptide that is post-translationally modified. These modifications included cyclization and the generation of a disulfide and a thioether bond (Liu and others 2010). The other genes in the operon are predicted to be involved in the posttranslational modification of SkfA to SKF and resistance to SKF. From these, only SkfB has been described to be vital for the creation of the thioether bond in SKF (Flühe and others 2013). SkfC is predicted to be responsible for the cyclization step of SKF while SkfH would create the disulfide bond via its thioredoxin-oxidoreductase domain (Liu and others 2010; Dorenbos and others 2002). Lastly, the skf operon also encodes an ABC transporter (SkfEF) which is hypothesized to provide export or immunity towards SKF (Gonzalez-Pastor and others 2003; Liu and others 2010). The sdpabc operon The sdpabc and sdpri operons were first described to be important for cannibalism in B. subtilis (Figure 4) (Gonzalez-Pastor and others 2003) and it was suggested that sdpabc encodes a signaling molecule SdpC of 63

28 13 amino acids. Later, it was determined that the signaling molecule was in fact a toxin factor (Ellermeier and others 2006a). SDP was shown to have toxic activity against several bacteria and its mechanism of action involves disruption of the proton motive force in sensitive cells (Liu and others 2010; Lamsa and others 2012). More recently, it was described that SdpC is an internal 42 amino acid peptide which contains a disulfide bond at residues C141-C147 and is derived from the precursor protein pro-sdpc ((Liu and others 2010), Chapter II). It has been shown that this di-sulfide bond is not necessary for SDP toxic activity (Chapter II). pro-sdpc is secreted via the Sec system in B. subtilis (Linde and others 2003a) and the signal peptidases SipS and SipT are the most important for signal peptide cleavage of the pro-sdpc to SdpC The chaperone CsaA is known to bind pro-sdpc and is thought to aid in secretion of SdpC (Linde and others 2003b). I have established that signal peptide cleavage of pro-sdpc is required for efficient production of the final product SDP (Chapter II). The first part of the sdpabc operon also encodes two proteins, SdpA and SdpB, required for the production of SDP toxin (See Chapter II). SdpA is a cytosolic protein that can localize to the membrane in an SDP-independent manner (Chapter II). SdpB is a predicted six transmembrane protein with distant homology to vitamin K dependent carboxylases (See Discussion Chapter II). Other SdpB orthologs have carboxylase activity but their function is unknown (Rishavy and others 2005). These proteins were

29 14 hypothesized to be involved in the final step of SDP production where SdpC is cleaved between positions 140 and 141 and 181 and 182 to release the final toxic SDP peptide. The resulting peptide will induce expression of the sdpri operon. The sdpri operon The second part of the operon includes the sdpri genes which encode for a regulator, SdpR, and an immunity/signal transduction protein, SdpI (discussed in next section). The sdpri operon is subject to regulation via AbrB in the absence of SDP (Fujita and others 2005). This AbrB-dependent regulation occurs when cells have not activated Spo0A-P. Negative autoregulation is provided also via SdpR (Ellermeier and others 2006a). SdpR is a 90 amino acid protein that belongs to the ArsR family of dimeric regulators (Busenlehner and others 2003). SdpR contains a helix-turn-helix DNA binding motif and a dimerization domain at its N-terminal sequence and binds its own promoter sequence at the transcriptional start site thereby preventing transcription (Ellermeier and others 2006a; Marchler Bauer and others 2013). At the C-terminal domain, metal binding sites in Firmicutes include a CxC/G motif, which is utilized to sense and bind diverse metals (Busenlehner and others 2003). Protein binding to the metals destabilizes the protein-dna binding interaction. Sequence comparison of SdpR with other family members shows that the metal binding sites are absent. Nonetheless,

30 15 ArsR-like regulators with no metal binding sites have been described to possess different functions. For example, PyeR in Pseudomonas aeruginosa can modulate biofilm formation (Mac Aogain and others 2012). In the case of B. subtilis, SdpR is sequestered to the membrane via SdpI in the presence of the cannibalism peptide SDP (Ellermeier and others 2006a). The dual d function protein SdpI Cells that have activated Spo0A-P (Spo0A-ON) produce the toxin SDP. The presence of SDP in turn will induce of expression of sdpri (Ellermeier and others 2006a). Encoded in sdpri is the immunity and signal transduction protein SdpI. SdpI can sense SDP and, in response, it sequesters SdpR to the membrane. De-repression of the sdpri operon allows for production of higher quantities of SdpI. SdpI is a six-transmembrane protein which forms part of a family of proteins with predicted immunity towards anti-microbial peptides (Povolotsky and others 2010). Proteins in this family range in size from three to twelve transmembrane domains. Except for the paralog YfhL (discussed below), no other orthologs of SdpI have been studied. SdpI constitutive mutations within the first three transmembrane domains of SdpI can sequester SdpR to the membrane in an SDPindependent fashion (Ellermeier and others 2006a). This allowed for a model were SdpI senses via these transmembrane domains. However, it was not

31 16 known if this region was also important for SdpR membrane sequestration and subsequent induction of sdpri. A careful sequence examination of several SdpI homologs shows only very small sections of conservation (Bailey and others 1994; Sievers and others 2011) (Chapter III). The most conserved region of SdpI homologs is between the predicted 4th and 5th transmembrane domains. Previously, two residues have been shown to be important for induction of sdpri and SdpR membrane sequestration (Ellermeier and others 2006a). In Chapter III, these residues will be discussed in detail along with other isolated SdpI mutants. SdpI can also provide immunity towards SDP by an unknown mechanism (Ellermeier and others 2006a). In Chapter III we will describe several residues found within a predicted cytosolic loop in SdpI which affect immunity to SDP only. These SdpI immunity- mutants can induce sdpri and sequester SdpR to the membrane. In addition, two orthologs of SdpI in Clostridium difficile and Streptococcus mutans are examined for their ability to provide resistance to SDP (Chapter III). Other SDP immunity factors B. subtilis can provide immunity towards SDP via other pathways. In the absence of SdpI, it was shown that the extra-cytoplasmic function (ECF) sigma factor σ W can induce yfhl expression (Butcher and others 2006). Interestingly, YfhL is a three transmembrane domain paralog of SdpI in B.

32 17 subtilis. YfhL provides resistance to SDP also in an unknown manner (Butcher and others 2006). In Chapter III, YfhL immunity will be examined based on work done in SdpI. A third possible mechanism for resistance lies within the yknwxyz operon (Yamada and others 2012). This operon encodes for a transporter that is expressed constitutively in the cells regardless of SDP presence. It was determined that YknWXYZ was able to provide resistance to SDP in the absence of SdpI (Yamada and others 2012). Other examples of cannibalistic behavior in bacteria Microbial cannibalism is an area which has only recently been studied. There are, however, other examples of cannibalistic behavior in bacteria. In the Gram-positive spore-former Paenibacillus dendritiformis, the production of a killing factor occurs in response to the presence of sibling bacteria during swarming (Be'er and others 2009a; Be'er and others 2009b). In this example, P. dendritiformis swarms via secretion of a surfactant (subtilisin) at low concentrations (Be'er and others 2010). However, cells that grow too close to each other will generate a toxic molecule that will kill the nearby cells. It was concluded that this toxin-mediated killing occurred even in the presence of nutrients (Be'er and others 2009a). Slf (sibling lethal factor) is generated via cleavage of subtilisin when cells are growing at high concentrations and in

33 18 close proximity (Be'er and others 2010). The toxin Slf alters cell morphology and motility in P. dentritiformis (Be'er and others 2011). Another example is found in Streptococcus pneumoniae, where the cannibalistic behavior was termed fratricide (Guiral and others 2005). In this example, cells induce fratricide during natural competence. Competent cells produce the toxin or fratricin CbpD, which will cleave the cell wall of non-competent cells. Cell wall stress will in turn induce the lytic factors LytA and LytC thus killing the non-competent cells (Wei and others 2012). CbpD can activate LytC while a different protein CbpF inhibits LytC (Perez Dorado and others 2010; Molina and others 2009). In combination with CbpD, cells also produce the bacteriocins CibAB to kill their neighboring non-competent cells (Guiral and others 2005). CibAB activity can be inhibited by the immunity protein CibC (Guiral and others 2005), similar to the B. subtilis mechanism where surviving cells make both toxin and immunity. Killing of non-competent cells in Streptococcus was suggested to provide genetic material and an additional release of intracellular virulence proteins (Guiral and others 2005). It has also been shown that fratricide aids the cells to acquire DNA when in biofilms (Wei and others 2012). However, B. subtilis remains the only bacteria to cannibalize in order to delay its cell fate to begin sporulation.

34 19 Figure 1. Bacillus subtilis sporulation process. Vegetative cells which encounter nutrient limiting conditions and have exhausted all other cellular pathways will undergo sporulation. The cell divides into the mother cell and the forespore via asymmetric cell division mediated by the master regulator Spo0A. Intercompartmental regulation between the mother cell and the forespore enables forespore engulfment via SigF and SigE. Spore maturation is regulated via SigG and SigK. The matured spore is then released by cell death into the environment.

35 20

36 21 Figure 2. Regulatory pathways which lead to different cellular development programs in B. subtilis. Nutrient stress will activate a signal response via the protein kinases KinA and KinB. This will generate a phosphorelay from the kinases to Spo0F, then to the phosphotransferase Spo0B, and finally to the master regulator Spo0A. Spo0A-P can then inhibit other cellular events such as competence and mediate indirect positive regulation of antimicrobial synthesis via inhibition of AbrB. Several signals can inhibit sporulation; DNA damage inhibits KinA via Sda, Nutrient availability induces the repressor CodY thus inhibiting sporulation. Competence can negatively impact sporulation via activation of the aspartyl-phosphatases Rap. Commitment to sporulation will inhibit Rap via induction of the cognate pheromones Phr.

37 Figure adapted from (Schultz and others 2009).

38 23 Figure 3. Cannibalism in B. subtilis. Cells which have reached a Spo0A-P threshold will activate expression of toxins via repression of the negative regulator AbrB (Spo0A-ON cells). In contrast, cells which have yet reached a threshold do not activate toxin production (Spo0A-OFF cells). An advantage of the Spo0A-ON cells is that toxin production can induce expression of the immunity factor via sdpri. Spo0A-OFF cells are repressed by AbrB and therefore cannot provide immunity. The toxins will then only kill Spo0A-OFF cells, providing nutrients and delaying sporulation in Spo0A-ON cells.

39 24

40 25 Figure 4. The sdpabcir operons in B. subtilis. Two operons are involved in the production and response to one of the cannibalism peptide; SDP. The genes involved in SDP production include two proteins of unknown function SdpA, SdpB, and SdpC which encodes the precursor toxin pro-sdpc. Response and immunity to SDP requires the sdpri operon. This encodes for the regulator SdpR and the immunity and signal transduction protein SdpI.

41 26

42 27 CHAPTER II. PRODUCTION OF THE CANNIBALISM TOXIN SDP IS A MULTI-STEP PROCESS THAT REQUIRES SDPA AND SDPB 2 2 I performed all the experiments in this chapter except for Figure 5 which was made by Wei-ting Liu at the University of California. The following results have been published in Perez Morales et al (Journal of Bacteriology 2013)

43 28 Introduction In the environment, microorganisms face constant competition for nutrients. In times of severe nutrient limitation the Gram-positive soil bacterium Bacillus subtilis initiates sporulation. Sporulation is an energetically costly process which becomes irreversible after the asymmetric septum is formed (Parker and others 1996). B. subtilis can delay the commitment to sporulation by inducing cannibalism, a process by which the sporulating cells in the population kill the non-sporulating cells (Ellermeier and others 2006a; Gonzalez-Pastor and others 2003). There are two toxins responsible for cannibalism, SDP and SKF (Ellermeier and others 2006a; Gonzalez-Pastor and others 2003). These toxins have antimicrobial activity against other bacteria including Xanthomonas oryzae, Listeria monocytogenes, and Staphylococcus aureus (Lin and others 2001; Liu and others 2010; Palmer and others 2009). SKF is produced by the skfabcdefgh operon while SDP is produced by the sdpabc operon. Expression of both operons is controlled by the master regulator of sporulation Spo0A, which when phosphorylated can repress expression of abrb, a negative regulator of skfabcdefgh and sdpabc (Fujita and others 2005; Chen and others 2006). Since AbrB negatively regulates expression of sdp and skf, the expression of both toxin encoding operons increases during early stationary phase upon entry into sporulation (Gonzalez-Pastor and others 2003). However these toxins are only produced

44 29 by a subset of B. subtilis as activation of Spo0A is subject to a bistable regulatory mechanism (Chung and others 1994). While the mechanism of SKF killing is unknown, the SDP toxin appears to kill sensitive cells by disrupting the proton motive force (Lamsa and others 2012). Antimicrobial peptides (AMPs) can be ribosomal or non-ribosomally synthesized. Non-ribosomally synthesized AMPs are generated from protein complexes that build, modify, and release an active peptide. The AMP Mycosubtilin produced by B. subtilis is a non-ribosomally synthesized β- amino fatty acid-linked cyclic heptapeptide which is produced by the products of the fenf-mycabc operon (Peypoux and others 1986; Peypoux and others 1976; Walton and others 1949). Ribosomally-synthesized AMPs often require post-translational modification to produce an active form of the toxin. For example production of Subtilosin A requires the genes alba and albf for modifcation of Subtilosin A (Zheng 2000; Zheng and others 1999). SKF is a ribosomally-synthesized 26 amino acid peptide encoded by skfa (Gonzalez-Pastor and others 2003; Liu and others 2010). SKF is a posttranslationally modified cyclic peptide with disulfide and thioether bonds (Liu and others 2010). Several genes in the skf operon have been proposed to be involved in the post-translational modifications of SKF (Liu and others 2010). It was recently demonstrated that SkfB is a 4Fe-4S cluster-containing radical SAM enzyme which is required for formation of a thioether bond in SKF (Flühe and others 2013).

45 30 SDP is a 42 amino acid, ribosomally synthesized AMP which contains a disulfide bond between two cysteine residues located at the N-terminus (Liu and others 2010). The active form of SDP is derived from an internal fragment of full length pro-sdpc (Liu and others 2010) (Figure 5A). Although the mature form of SDP has been determined, little is known about the factors required to process pro-sdpc into the active SDP peptide. The pro- SdpC form is a 203 amino acid protein secreted via the general secretory pathway (Linde and others 2003a). Signal peptidases SipS and/or SipT can cleave pro-sdpc to SdpC when expressed in E.coli (Linde and others 2003a). However, it was not known what role signal peptide cleavage plays in SDP production. The sdpab genes are located in an operon with sdpc but it is not known if they are required for the production of the toxin SDP (Gonzalez-Pastor and others 2003). Here we provide evidence that SDP production requires multiple steps including; signal peptide cleavage of pro-sdpc which creates SdpC , formation of disulfide bonds in SdpC , and processing of SdpC into mature SDP (Figure 5B). We also provide evidence that SdpAB are essential for the production of active SDP toxin and are presumably required for processing SdpC into mature SDP.

46 31 Materials and Methods Bacterial Strains and Growth All strains used in the study are isogenic derivatives of PY79, a prototrophic derivative of B. subtilis strain 168, and are listed in Table 1 (Youngman and others 1984). Strains were routinely grown in Luria-Bertani (LB) and Difco sporulation medium (DSM) at 37 C except for overnight cultures, which were grown at 30 C (Harwood, C.R and Cutting, S.M., eds 1990). Antibiotics were used at the following concentrations: chloramphenicol (10 µg/ml), erythromycin plus lincomycin, (1 µg/ml and 25 µg/ml respectively); kanamycin (5 µg/ml), spectinomycin (100 µg/ml), tetracycline (10 µg/ml), and ampicillin (100 µg/ml). The β-galactosidase chromogenic indicator 5-bromo-4- chloro-3-indolyl β-d-galactopyranoside (X-Gal) was used at a concentration of 100 µg/ml. Isopropyl β-d-1-thiogalactopyranoside (IPTG) was used at a final concentration of 1 mm. Bacterial strains were constructed by transformation of relevant genomic or plasmid DNA into B. subtilis competent cells prepared by the one-step method previously described (Wilson and others 1968). Construction of Plasmids All DNA oligonucleotides and plasmids used in this study are listed in Tables 2 and 3. The IPTG-inducible Phs-sdpC integrated at amye was constructed by PCR amplifying sdpc from B. subtilis using oligos CDEP126

47 32 and CDEP127. The resulting PCR product was digested with HindIII and SphI, and cloned into pdr111 (Ben Yehuda and others 2003) digested with the same enzymes to create pce106. The IPTG-inducible Phs-sdpA, Phs-sdpB and Phs-sdpAB genes were constructed by PCR amplifying sdpa (CDEP124 and CDEP566), sdpb (CDEP567 and CDEP125) or sdpab (CDEP124 and CDEP125) from B. subtilis. The resulting PCR products were digested with HindIII and SphI and cloned into pdp150 (Kearns and others 2005) digested with the same enzymes to create pce216 (sdpa), pce315 (sdpb), and ptp092 (sdpab). The resulting plasmids were confirmed by sequencing (Iowa State University) and transformed into the wild type B. subtilis strain PY79. A Gateway destination vector was constructed to build N-terminal gfpsdpa fusions (Invitrogen). This was generated by cloning the RfA cassette (Invitrogen) into pce236 (pdr111-gfp) which had been digested with SphI and EcoRI and blunt ended with Klenow (NEB) to generate pjh183. N- terminal GFP tagged SdpA + (GFP-SdpA + ) was constructed by PCR amplifying sdpa from B. subtilis using oligos CDEP890 and CDEP566, and cloning into pentrd-topo, resulting in pdt001. To construct a plasmid producing GFP- SdpA +, sdpa + was moved from pdt001 onto pce291 using LR Clonase II (Invitrogen) resulting in plasmid pdt002.

48 33 Site-directed mutagenesis of SdpC Site-directed mutagenesis of pce106 (Phs-sdpC) was performed using the QuickChange site-directed mutagenesis kit (Agilent Technologies) according to the manufacturer s instructions. The SdpC signal peptide cleavage site mutant (sdpc T30H ) was constructed using primer pairs CDEP640 and CDEP641 (sdpc T30H ) to generate plasmid pce260. The SdpC disulfide bond single mutants were constructed using the following oligo pairs CDEP912 and CDEP913 (sdpc C141A ) and CDEP892 and CDEP893 (sdpc C147A ). The sdpc C141A C147A mutant was constructed by site directed mutagenesis of ptp085 with CDEP1247 and CDEP1248 to generate ptp091 (sdpc C141A C147A ). The resulting plasmids ptp085 (sdpc C141A ), ptp076 (sdpc C147A ), and ptp091 (sdpc C141A C147A ) were confirmed by sequencing and transformed into B. subtilis PY79. β-galactosidase activity assay Cultures were grown overnight in LB broth at 30 C and 40 µl was spotted onto LB agar supplemented with 1 mm IPTG. Plates were incubated at 37 C for 4 hours. Samples were harvested and resuspended in 1 ml of Z buffer (60 mm Na2HPO4, 40 mm NaH2PO4, 10 mm KCl, 1 mm MgSO4, 50 mm β-mercaptoethanol ph 7.0) and the OD600 was determined. Lysozyme (10 µg) was added to samples and incubated for 30 minutes at 37 C or until clear (Harwood, C.R and Cutting, S.M., eds 1990). 200 µl of cell lysates were added

49 34 to 96 well plates with 10 mg/ml ortho-nitrophenyl-β-galactoside (ONPG). The activity of β-galactosidase was measured every minute at OD405 for 40 minutes total. Data were analyzed as previously described (Slauch and others 1991). SDP-mediated killing assay Reporter cells which lack the ability to produce the SDP toxins, SDPsensitive ( sigw sdpabcir; CDE433) or SDP-resistant ( sigw sdpabcir amye::phs-sdpi; TPM758), were grown to an OD600 of 0.8 in LB broth with 1 mm IPTG. The reporter cells (10 6 ) were inoculated into LB agar (0.7%) + 1 mm IPTG. An overnight culture of each of the Sdp producing strains was subcultured 1:100 and grown in LB + 1 mm IPTG for 4 hours at 37 o C. 20 µl of Sdp producing cells were spotted onto plates containing either SDP-sensitive or SDP-resistant cultures. Plates were incubated overnight at 37 C and the zone of inhibition was determined. Subcellular fractionation f of cells Overnight cultures were sub-cultured 1:100 in liquid DS media supplemented with 1 mm IPTG and grown for 4 hours at 37 C. The cultures were separated into whole cell and supernatant fractions by centrifugation. The supernatants were concentrated by methanol-chloroform extraction (Wessel and others 1984). Briefly, 2 ml of supernatant was mixed with 2 ml of