L“ .1 '1 413 3-.- 4% L z) r’ ‘f M- *w 35‘9“ _ y 3?va n. a- . waif; "a. iffiggfifé’ffi .J. a “”719 {34: a“: » T“ nrwfifé» [afi‘xfihfiz { , a f"? .5»; - ”3..“ i ‘J 55'4“} 51;.z"‘é¢;«,: 29‘6” * VVf .1 firm . mm; .. .. . finw'fwc.» '3? ‘ 53““ « ”3.: a- {1'1}???- $123.”. 3 "r vv‘. -— “r: , pi. I ‘ .234: :Fr 3» f"? 5" y-yc-«vy ..‘ ‘ “v :- n ," , < cww‘t-fisyéffifig‘m‘fia‘v 1. .. «aux-"aerw UNIVERSITY LIBRAR RIES IIIIIIIIIIIIIIIIIIII I IIIIIIIIIIIIIIII 3 1293 010204 This is to certify that the dissertation entitled Characterization of the Role of the SpoIIID Switch Protein During Bacillus Subtilis Sporulation presented by Richard Brott Halberg has been accepted towards fulfillment of the requirements for Ph . D. degree in Biochemistry Ufa. X.ona\ Major professor Date 6'1’”71‘ MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY M”Moan State University PLACE IN RETURN BOX to romovothb chookout from your record. TO AVOID FINES Mum on or odor. duo duo. DATE DUE DATE DUE DATE DUE MSU!9An.“" “ “' ' CHARACTERIZATION OF THE ROLE OF THE SPOIIID SWITCH PROTEIN DURING BACILLUS SUBTILIS SPORULATION BY Richard Brott Halberg A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1994 ABSTRACT CHARACTERIZATION OF THE ROLE OF THE SPOIIID SWITCH PROTEIN DURING BACILLUS SUBTILIS SPORULATION BY Richard Brott Halberg A fundamental problem in biology is understanding how gene expression is regulated temporally and spatially during the development of living organisms. A particularly tractable system to address this problem is the gram-positive bacterium Bacillus subtilis, because it is amenable to both genetic and biochemical approaches. Under conditions of nutrient deprivation, B. subtilis undergoes a series of morphological changes which culminate in the formation of a spore. The first easily-observed morphological structure is an asymetrically—positioned septum, which divides the bacterium into two compartments, the mother cell and the forespore. Both of these compartments receive a copy of the genome, but they realize alternative developmental fates because gene expression is regulated temporally and spatially. Three key regulators of transcription in the mother cell are SpoIIID, OE, and 6K. The research presented in this dissertation is focused on characterizing the role of SpoIIID during Sporulation. The ability of SpoIIID to affect the transcription of 03- and OK-dependent genes was tested in vitro. SpoIIID activates and represses transcription by both forms of polymerase. SpoIIID appears to have these effects by binding to specific sequences in the -35 region of OE— and OK-dependent genes, as evidenced by DNase I footprinting. A comparison of the sequences in the protected regions revealed a putative consensus sequence for binding of SpoIIID. The fate of SpoIIID during Sporulation was determined by Western blot analysis. The level of this protein decreases sharply during the late stages. The decrease appears to be controlled by two independent mechanisms: one acts at the level of SpoIIID mRNA synthesis and/or stability, while the other involves the conversion of SpoIIID to a less stable 9 kDa form. The conversion appears to involve the removal of 7 amino acids from the C-terminus of SpoIIID, based on N- terminal sequencing and mass analysis. The proper level of SpoIIID is crucial for normal Sporulation. Cells engineered to produce an elevated level of SpoIIID throughout Sporulation produced fewer heat- resistant spores. This effect appears to result from a reduction in oKproduction and ofi-dependent gene expression. ‘ This dissertation is dedicated with love and respect to Sara Tim Emily Mom Dad Grandma Joe Lois Dave Carol Bob Kathy Kevin Theresa Bob Margie Dean Jane Steve Karen Doober Jane and Friends ACKNOWLEDGEMENTS When considering graduate school, I never believed that a person could be a Ph.D. candidate and have a family at the same time. Now, I can not imagine one without the other. I would like to thank my wonderful wife, Sara, for her love and support, and two children, Tim and Emily, for their warm smiles and innocent ways. I would also like to thank my parents for always encouraging me to pursue my dreams. I have been fortunate to have three outstanding mentors. Dr. Edward Buchanan Jr. sparked my research interest when I was only in my second year of college. Dr. Gary Small kept my interest in research alive and golf skills honed. Dr. Lee Kroos has taught me many things about the “art of science". He is an excellent researcher whom I will always respect and admire. Lee and his wife, Mary, have been very supportive during life's challenges, especially, the adoption of Tim and the birth of Emily. Lee and Mary are tremedous friends that I will greatly miss. I have also had the opportunity to work with great friends and collegues throughout my research career. I would like to thank: Ted for taking me under his wing and sharing L numerous dinners with Sara and me, Kevin for constantly arguing with me and making sure I saw George Bush in 1988, Dave for arguing with Kevin and showing me how golf should be played, Mark Sutton for his mutual interest in Mike Dikta and the Chicago Bears, Mark Sinton for computer advice and sharing risky experiences, Eugene for his expertise with MALDI mass spectroscopy and his fatherly advice, Sara for numerous meals and sharing Brian with me, Brian for sharing the joys of fatherhood and fighting numerous battles, Jamie for just being Jamie, Sijie for making scientific meetings enjoyable and my first experiences with cell death, Monica for her great energy and banana spilt bread, Makda for her methodically ways and all encompassing knowledge of camels, Hiroshi for doing all the experiments that I should of done and his multitude of hairstyles, Bin for doing more of the experiments I should have done and his ability to acquire free samples, and Janine for housing tips and humerous cat stories. I would like to thank the members of my committee: Jon Kaguni for allowing me to rotate in his lab and his cautious ways, Tom Deits for sharing his intriguing spore coat ideas, Arnold Revzin for his ability to reduce the most complex to simple terms, and Larry Synder for helping me see the "big" picture. I would like to thank Joe, Melanie, and Colleen for sequencing the 9 kDa protein, preparing primers, and advice vi A and Diane, Julie, Mary, Vickie, and Carol for getting me out of numerous dilemas. I would also like to thank George and Nick for their friendship. vii TABLE OF CONTENTS LIST OF TABLES ........................................... xiii LIST OF FIGURES ......................................... xiv LIST OF ABBREVIATIONS ................................. xviii INTRODUCTION ............................................. 1 CHAPTER I Overview ........................................... 5 Initiation of Sporulation ........................... 8 Role of Sigma Factors During Sporulation ........... 13 Forespore—specific sigma factors ............... 13 Mother-cell-specific sigma factors ............. 16 Intercompartmental Coupling of Gene Expression in Both Compartments ............................... 19 0F and GE ..................................... 19 0E and CG ..................................... 22 OS and 6K ..................................... 23 Role of DNA-binding Proteins During Sporulation ..... 25 SpoOA ......................................... 25 SpoIIID ....................................... 26 GerE ........................................... 31 viii A Signficance ......................................... 31 CHAPTER II Abstract ........................................... 34 Introduction ....................................... 35 Materials and Methods ............................... 38 Results ............................................. 41 SpoIIID binds to specific sequences in spoIVCA, sigK, bofA, and cotD ..................... 41 SpoIIID activates spoIVCA and sigK transcription, but represses bofA transcription by GE RNA polymerase in vitro ....................... 50 SpoIIID binding in the -35 region is sufficient to activate sigK transcription and repress cotD transcription by GK RNA polymerase in vitro ..................................... 55 Discussion ............................................... 59 CHAPTER III Abstract ........................................... 73 Introduction ....................................... 73 Materials and Methods ............................... 74 Results ............................................. 74 The level of SpoIIID changes during Sporulation ............................... 74 The SpoIIID decrease coincides with increases in the level of OK and spore coat gene expression ............................... 75 Mutants defective in OK production are defective in the SpoIIID decrease ................... 76 The SpoIIID decrease occurs earlier in cells that produce oKearlier ...................... . 77 Production of OKduring Sporulation leads to a decrease in the level of SpoIIID mRNA ..... 78 Discussion ......................................... 78 CHAPTER IV Abstract ........................................... 84 Introduction ....................................... 85 Materials and Methods ............................... 88 Results ............................................. 94 A 9 kDa protein that copurifies with SpoIIID appears to be a degradation product of SpoIIID ................................... 94 DNA—binding and transcriptional properties of the 9 kDa protein ......................... 100 SpoIIID is converted to the 9 kDa protein in a developmentally regulated fashion ......... 105 Conversion of SpoIIID to the 9 kDa protein is reduced in several Sporulation mutants ................................... 111 Discussion ......................................... 118 CHAPTER V Abstract ........................................... 127 Introduction ....................................... 129 Materials and Methods ............................... 132 Results ............................................. 137 A plasmid containing the SpoIIID gene permits continued production of SpoIIID late in Sporulation ............................... 137 Overproduction of SpoIIID reduces cotD, cotA, and gerE promoter activity ............... 138 Overproduction of SpoIIID reduces sigK promoter activity and the level of 0'K ..... 141 Overproduction of SpoIIID inhibits the production of heat-resistant spores ....... Production of SpoIIID from Pspac—spoIIID increases sigK expression and reduces cotD expression, but does not affect cotA and gerE expression, in cells engineered to produce 6“ during vegetative growth Additional evidence that cotD, but not cotA or gerE, is repressed by SpoIIID ............. Discussion ......................................... CHAPTER VI APPENDIX A Abstract ........................................... Introduction ....................................... Materials and Methods ............................... Results ............................................. Antibodies to pro-0'K detect pro-0'K and (3'K in sporulating B. subtilis ................... Levels of pro—(5'K and OKare developmentally regulated ................................. Mutations in many Sporulation genes block accumulation of CK ....................... Processing of pro-0'K to OK is required to produce an active 0 factor and is developmentally regulated ................. Discussion ......................................... APPENDIX B Abstract ........................................... Introduction ....................................... xi 145 149 153 178 179 179 179 180 Materials and Methods ............................... 185 Results ............................................. 188 Purification of GerE ........................... 188 Mapping the 5' terminus of cotC mRNA ........... 188 GerE binds to specific sequences ............... 190 GerE stimulates cotB and cotC transcription in vitro ..................................... 190 Effects of GerE on in vitro transcription of other mother—cell-expressed genes ......... 192 Discussion ......................................... 193 BIBLIOGRAPHY xii LIST OF TABLES CHAPTER I Table 1. Genes whose expression is affected by SpoIIID ... 28 CHAPTER V Table 1. Number of heat-resistant spores produced by Bacillus subtilis strains in the absence and presence of SpoIIID overproduction ............. 147 xiii LIST OF FIGURES CHAPTER I Figure 1. Schematic representation of the stages of Bacillus subtilis Sporulation ............. 7 Figure 2. Phoshorelay signal transduction pathway ....... 10 Figure 3. Criss—cross activation of sigma factors during Sporulation ....................... 21 Figure 4. Temporal switch in the mother-cell pattern of gene expression ........................... 30 CHAPTER II Figure 1. SpoIIID footprints on spoIVCA and sigK ......... 43 Figure 2. SpoIIID footprints on bofA and cotD ........... 48 Figure 3. Position of SpoIIID binding sites in spoIVCA, sigK, bofA and cotD ................... 52 Figure 4 Effects of SpoIIID on spoIVCA, sigK, and bofA transcription in vitro ............... 54 Figure 5. Effects of SpoIIID on the in vitro transcription of 319K and cotD templates containing different combinations of SpoIIID binding sites ..................... 57 Figure 6. Alignment of SpoIIID binding sites and the arrangement of matches to SpoIIID consensus sequence in binding sites ................. Figure 7. Regulatory effects of GE, SpoIIID, and (3'K on mother-cell gene expression ...... . ........ 69 xiv CHAPTER III Figure 1. Characterization of anti-SpoIIID antibodies by Western blot analysis ................... 75 Figure 2 Levels of SpoIIID, pro-OK, 0K, and cotD-directed B-galactosidase activity in sporulating B. subtilis. ................................. 75 Figure 3 Level of SpoIIID in Sporulation mutants with defects in OK production. ................. 76 Figure 4. Levels of SpoIIID, 0K, and cotD-directed B- galactosidase activity in mutants that produce OK earlier than normal. ........... 77 Figure 5 Level of SpoIIID mRNA in sporulating wild-type and 319K mutant cells ..................... 78 Figure 6 Model for the switch from sigK to cotD transcription in the mother cell during stage IV to stage V transition of Sporulation ............................... 79 CHAPTER IV Figure 1. A silver-stianed 18% polyacrylamide gel displaying Figure Figure Figure Figure Figure Figure proteins collected from a double-stranded DNA-cellulose column upon elution with a 0.5 M - 1.3 M potassium chloride gradient ................................. 96 Mass spectra of SpoIIID and SpoIIID digested with endoproteinase Asp-N. ..................... 98 Mass spectrum of the 9 kDa protein ........... 102 The SpoIIID and 9 kDa footprints on sigK and cotD. ..................................... 104 Effects of SpoIIID and the 9 kDa protein on $19K and cotD transcription in Vitro ........... 107 Mobility shift assay to monitor the levels of SpoIIID and the 9 kDa protein in wild-type and SpoIIID mutant cells .................. 110 7. Mobility shift assay to determine whether gel- purified SpoIIID is converted to the 9 kDa XV fi protein in extracts prepared from wild—type, spoIIG and spoIIID cells. ................. 113 Figure 8. Mobility shift assay to monitor the levels of SpoIIID and the 9 kDa protein in several Sporulation mutants ....................... 116 Figure 9. Model for the switch from sigK to cotD transcription in the mother cell during the stage IV to stage V transition of Sporulation ............................... 121 CHAPTER V Figure 1. Levels of SpoIIID and cotD-, cotA— , and gerE—directed B-galactosidase activity in spo+ cells containing either Pspac- spoIIID or the parental plasmid ........... 140 Figure 2. Levels of sigK‘directed B-galactosidase activity, pro-OK, and OK in spo+ cells containing either Pspac-spoIIID or the parental plasmid ................................... 144 Figure 3. Levels of SpoIIID, sigK—directed B-galactosidase activity, and OK in Pspac—PsigK-sigKAlQ cells containing either Pspac-spoIIID or the parental plasmid. ......................... 152 Figure 4. Levels of cotD-, cotA- and gerE-directed B- galactosidase activity in Pspac-PsigKe sigKAl9 cells containing either Pspac- spoIIID or the parental plasmid ........... 155 Figure 5. Effects of SpoIIID on cotD, sigK, cotA, and gerE transcription in vitro .................... 158 Figure 6. Levels of gerE and cotD mRNA in sprulating B. subtilis ................................. 160 APPENDIX A Figure 1. Production of pro-oK in E. coli ............... 179 Figure 2. Charactization of the anti-pro-O‘K antiserum by Western blot analyses .............. .... 179 xvi A Figure 3. Figure 4. Figure 5. APPENDIX Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Pro-OK and OK in sporulating B. subtilis ........ 180 Pro-OK and OK in B. subtilis Sporulation mutants harvested 6 hr after the end of exponential growth in D8 medium ........... 180 Effect of producing pro-0'K from a plasmid during B growth and Sporulation of B. subtilis ..... 181 The GerE binding sites in the 5'regions of cotB and cotC ............................. 186 Production of GerE in E. coli ................. 188 Mapping the S'terminus of cotC mRNA ........... 189 GerE footprints in cotB and cotC DNAs ......... 191 GerE stimulates cotB and cotC transcription in vitro. ................................. 192 Effects of GerE on cotD, sigK, gerE and cotA transcription in vitro ................... 192 Alignment of promoters transcribed by OKRNA polymerase. ............................... 194 Regulatory effects of SpoIIID and GerE during stages IV and V of Sporulation. ........... 195 Regulatory interactions controlling the levels of SpoIIID, OK and GerE govern the stage IV to stage V transition in the mother cell ..... 195 xvii LIST OF ABBREVIATIONS ADP adenosine-5’-diphosphate ATP adenosine-S’—triphosphate bp base pair BRL Bethesda Research Laboratories BSA bovine serum albumin CTP cytosine-5’-triphosphate Da dalton DNA deoxyribonucleic acid DPA dipicolinic acid DTT dithiothreitol EDF-A extracellular differentiation factor EDTA (ethylenedinitrilo)tetraacetic acid GDP guanosine-5’-diphosphate GTP guanosine—S’-triphosphate HCl hydrochloric acid IPTG isopropyl 8-D thiogalactopyranoside kb kilobases kDa kilodalton LB Luria—Bertani MALDI matrix-assisted laser desorption ionization xviii A MgCl; mRNA NaCl ng O.D. ONPG ORF PAGE pmole PMSF RNA SDS SM Tx TFA Mg Ill UTP V/v molar millimolar magnesium chloride messenger ribonucleic acid sodium chloride nanogram optical density o-nitrophenol-B—D-galactoside open reading frame polyacrylamide gel electrophoresis picomole phenylmethylsulfonyl fluoride ribonucleic acid sodium dodecylsulfate Sterlini-Mandelstam x hours after the onset of sporulation trifluoroacetic acid microgram microliter uracil-5’-triphosphate volume per volume weight per volume xix INTRODUCTION Under conditions of nutrient deprivation, Bacillus subtilis undergoes a series of morphological changes which culminate in the formation of a spore. The first easily observed morphological structure is an asymmetrically positioned septum, which divides the bacterium into two compartments, the mother cell and the forespore. Both of these compartments receive a copy of the genome, but they realize alternative developmental fates because gene expression is regulated temporally and spatially. Regulation in the mother cell involves two, small DNA—binding proteins, SpoIIID and GerE, as well as two sigma subunits of RNA polymerase, OE and OK. The experiments in Chapter II demonstrate that SpoIIID activates and represses transcription by both the OE and 0K forms of RNA polymerase. SpoIIID appears to have these effects by binding specific sequences in the promoter regions of 03- and OK-dependent genes. This work was submitted to the JOurnal of Molecular Biology. Based on the observation that SpoIIID can activate and A 2 repress transcription by OKRNA polymerase, it was proposed that inactivation and/or sequestering of SpoIIID would establish a switch in the pattern of gene expression. The experiments in Chapter III focus on testing this model. The level of SpoIIID does decrease sharply at the appropriate time to establish such a switch. This decrease is, in part, due to the production of active 6“, which leads to reduced synthesis and/or stability of spoIIID mRNA. This work was published in the Journal of Mblecular Biology. The decrease in the level of spoIIID mRNA is paralleled by a decrease in the level of SpoIIID, suggesting that a mechanism for degrading SpoIIID exists in sporulating B. subtilis. The experiments in Chapter IV provide evidence that SpoIIID is converted to a less stable 9 kDa form by removing 7 amino acids from its C-terminus. This conversion is developmentally regulated. The mass spectral data presented in this chapter was obtained through a collaboration with E. Zaluzec and D. Gage at Michigan State University. The proper level of SpoIIID is crucial for normal development. The experiments in Chapter V demonstrate that overproducing SpoIIID significantly reduces the production of heat-resistant spores. This effect appears to result from a reduction in OKproduction and Ox-dependent gene expression. The experiments in this chapter also provide evidence that A 3 SpoIIID represses cotD expression by OKRNA polymerase in vivo. V. Oke at Harvard University provided the strain which permits OKto be produced during growth. The experiments in appendix A demonstrate that the primary translation product of sigK is pro—6K, an inactive precursor, which is proteolytically-processed to the active form. This work was done by S. Lu, L. Kroos and myself. I was responsible for determining whether pro-0'K was capable of directing transcription in vitro. This work was published in the Proceedings of the National Academy of Science USA. The experiments in appendix B demonstrate that the switch in the mother-cell pattern of gene expression, which is established by the production of oxand subsequent decrease in the level of SpoIIID, is reinforced by GerE. This work was a collaboration between L. Zhang, S. Roels, R. Losick at Harvard University and H. Ichikawa, L. Kroos, and myself at Michigan State University. I was responsible for preparing some of the RNA for the primer extension experiments and the preparation of materials used in the in vitro transcription assays. This work was published in the JOurnal of Melecular Biology. CHAPTER I LITERATURE REVIEW "To be absolutely certain about something, one must know everything or nothing about it." Olin Miller Overview A fundamental problem in developmental biology is understanding how gene expression is regulated temporally and spatially during the development of living organisms. A particularly tractable system to address this problem is the gram-positive bacterium Bacillus subtilis, because it is amenable to both genetic and biochemical approaches. Under conditions of nutrient deprivation, B. subtilis sporulates. This process has been divided into several stages based on the morphological changes which occur (Figure 1; reviewed by Errington, 1993). The first easily observed change is the formation of an asymmetric septum, which divides the bacterium into two compartments, the larger mother cell and the smaller forespore (stage II). Migration of the septum results in the engulfment of the forespore as a free, double—membraned protoplast within the mother cell (stage III). A cell—wall-like material, called cortex, is deposited between the membranes of the forespore (stage IV) and then coat proteins, which are synthesized in the mother cell, are deposited on the outer surface of the forespore (stage V). The spore matures, gaining all of its resistance properties (stage VI), and is released by the lysis of the mother cell (stage VII). This process takes about six to ten hours at 37'C. A f “\ -eopum‘TI '- iI Stage a I I IIOII’AOhCOII . . vuI ' c::;I .— pjngfigéflpiualjamsdoa .1 910911 ranges M .t'lo.i isltnoqe “use“ , v.v“ss.ut a“: lieu-ISGJOM 5d: 03ni nulrejold '0! ed: 1 sdontq mu3qsz an: to noijsrpifl .(II epeJa) an: nidJiw fiesiquOIQ sea! 5 es ) Jpoo bne (VI spade) x9330: exoqe Siege“! he no: eroqe ad: 3991011 1511' lilo-V- Figure 1. Schematic representation of the stages of Bacillus subtilis sporulation. An asymmetric septum divides the bacterium into the mother-cell and forespore compartments (stage II). Migration of the septum pinches of the forespore as a free protoplast within the mother cell (stage III). The spore cortex (stage IV) and coat (stage V) are generated, which protect the spore from environmental insults. sssss m Stage II mother-cell forespore cccccc Stage IV germ cell wall Initiation of sporulation The key event in the initiation of sporulation appears to be the phosphorylation of SpoOA, which is the end—result of a phosphorelay system (Figure 2; Hoch, 1994). Two kinases have been identified. KinA is a cytoplasmic protein (Perego et al., 1989), while KinB is an integral membrane protein (Trach & Hoch, 1993). A kinA or kinB mutant sporulates (albeit at reduced efficiency), whereas a double mutant does not (Trach & Hoch, 1993). The target of KinA and KinB is the secondary messenger protein SpoOF (Peregoem al., 1989). The phosphoryl group of SpoOF—P is transferred to SpoOA via SpoOB (Burbulys et al., 1991). SpoOA—P activates and represses transcription of genes expressed during stationary phase and sporulation. Some of the phosphorelay proteins are similar to signal transduction proteins that are involved in numerous adaptive responses, including nitrogen utilization. KinA belongs to a family of histidine protein kinases, which receive signals from the extracellular environment and/or the cytoplasm (Stock et al., 1989). SpoOF and SpoOA belong to a family of response regulators, which alter the pattern of gene expression, allowing the bacterium to adapt to its environment (Stock et al., 1989). The phosphorylation of SpoOA appears to incorporate both A 1“: .vswdaeg Siv. ro\bae Ania/pcfiso 1: rain: IOOBISHI has relulieoeadxa a;x\ cod: :1 guorgg aJedqaoda gear? .300q8 easiyxodqeodq an aura ' gyfi¢ .DQQI .da mozi beiqo 0A .800qa aiv AOoqE o: berxeieaeis . w" {I ' 77. WC...‘ Iii ’ “”03“” $190013 (3- 0364-— 33. f: n i CD? a, I ~r-. ‘u I Spooaoi ir" ". _ ; ’ ’ --=£§s at _ 5“ “ " ‘ L)?” V - ‘ L ‘ -.__‘r',_“;:.§~5 3. ~'- ._*xnu-v..oufl¢i|\fl Ext-.‘ei" .2 fi‘ . € 11;“ . . ' q \ .-.-. 1‘ - - . c;- ».J_”!..” __ ~_.'<:- ,- . . _. -“ 1‘10 ,_e.iv‘,‘ - ..- .- ‘ ;- I. .1. ‘~ Figure 2. Phosphorelay signal transduction pathway. Extracellular and intracellular signals cause KinA and/or KinB to phosphorylate SpoOF. The phosphate group is then transferred to SpoOA via SpoOB. Adapted from Hoch, 1994. KinB ATP <;PADP -———_SpoOK 2“». Y m SpoOF SpoOF—P >< ., SpoOB-P n Cycle GDP GTP SpoOA SpoOB ('— osc(—— Cell sp©©a=s 11 extracellular and intracellular signals. KinA activity may be regulated by an extracellular signal. Cells induced to sporulate at low cell density form heat—resistant spores about 104-fold less efficiently than cells induced at high cell density (Grossman & Losick, 1988). This effect is alleviated by the addition of filter-sterilized supernatant from cells sporulated at high cell density. The stimulatory factor(s) in the supernatant was called extracellular differentiation factor A (EDF—A) and appears to be, at least in part, an oligopeptide(s) (Grossman & Losick, 1988). It has been proposed that the entry of EDF—A into the cell via the spoOK—encoded transport system affects the activity of KinA (Antoniewskiet al., 1990). Consistent with this idea, kinA spoOK double mutants have a more severe sporulation defect than either single mutant and overproduction of KinA partially suppresses the sporulation defect in spoOK mutants (Rudner et al., 1991). Extracellular signals have been implicated in the initiation of development in other systems. Myxococcus xanthus cells induced to sporulate at low cell density formed spores less efficiently than cells induced at high cell density (Wireman & Dworkin, 1975). This effect was alleviated by the addition of A-factor, which is composed of amino acids, oligopeptides and proteases (Kuspaiet al., 1992; Plamann et al., 1992) . The role of the proteases appears to be to release amino acids and oligopeptides. The amino acids ‘ 12 and oligopeptides may enter the cell by a transport system similar to that encoded by spoOK, because two genes at the sasA (suppressor of A-signaling defect) locus resemble genes in the spoOK operon (H. Kaplan, unpublished data). The entry of amino acids and oligopeptides may affect a signal transduction pathway similar to the phosphorelay system, because asgA (a—signalling) encodes a protein that is similar to histidine kinases and response regulators (L. Plamann, unpublished data). The activity of other components of the phosphorelay system may be regulated by intracellular signals. It has been proposed that Obg links the activity of SpoOB to the cell cycle, because (1) obg is in the same operon as SpoOB (Trach & Hoch, 1989), (2) Obg is essential for growth (Trach & Hoch, 1989), (3) Obg contains a guanosine-5'—triphosphate (GTP) binding site, which is similar to that in Era, an E. coli Ras-like protein (Trach & Hoch, 1989), and (4) the phosphorylation of SpoOA is coupled to DNA synthesis (Ireton & Grossman, 1992b). In addition to cell density and cell cycle signals, nutritional signals also influence the initiation of sporulation. For example, starvation for sources of carbon, nitrogen, or phosphorous can induce sporulation. The mechanisms for sensing nutritional signals have not been elucidated. However, some experiments suggest GTP (and/or GDP) represents the key effector of the nutritional signals A 13 (Freeseaet al., 1985). Thus, Obg may incorporate nutritional signals into the phosphorelay system, if it really binds GTP and affects the activity of SpoOB. Role of Sigma rectors During Sporulation The temporal and spatial pattern of gene expression during sporulation is established, in part, by the sequential activation of compartment-specific sigma subunits of RNA polymerase. E _ 'E' . E spoIIA is a three-cistron operon, consisting of spoIIAA, spoIIAB, and spoIIAC (Fort & Piggot, 1984; Piggotet al., 1984; Errington et al., 21985; Stragier, 1986; Sun et al., 1989). Transcription of spoIIA is dependent upon SpoOA-P and 03 RNA polymerase (Zuber et al., 1989), the earliest acting sporulation-specific form of RNA polymerase (Errington, 1986; Savva & Mandelstam, 1986; Burbulys et al., 1991), and begins prior to septation (Gholamhoseinian & Piggot, 1989). However, the product of spoIIAC, OF, appears to be active only in the forespore, since B-galactosidase expressed from a OT-dependent lacZ fusion is localized to this compartment (Margolis et al., 1991). 0? activity is regulated by spoIIAA A 14 and SpoIIAB. Genetic studies demonstrated that GFactivity is antagonized by SpoIIAB and that SpoIIAB is antagonized by spoIIAA (Schmidt et al., 1990). Biochemical studies revealed that SpoIIAB is an anti-sigma factor that binds to 6F and blocks OF-directed transcription. SpoIIAB can also bind to SpoIIAA (Duncan & Losick, 1993). SpoIIAB binding to CF is stimulated by adenosine-5’-triphosphate (ATP) and its non-hydrolyzable analogs, while SpoIIAB binding to SpoIIAA is stimulated by adenosine-5’—diphosphate (ADP) (Alper et al., 1994). Based on these observations, it has been proposed that the concentration of ATP decreases and the concentration of ADP increases in the forespore, resulting in SpoIIAB binding SpoIIAA rather than GP and thereby OF becoming active specifically in the forespore. In contrast, the concentration of ATP remains high relative to ADP in the mother cell, resulting in SpoIIAB remaining bound to CF and thereby blocking OF—dependent gene expression. Consistent with this idea, the concentration of ATP relative to other adenosine nucleotides decreases dramatically in the forespore during sporulation of Bacillus megaterium (Singh et al., 1977). spoIIIG encodes 0G (Masuda et al., 1988; Karmazyn- Campelli et al., 1989; Sun et al., 1989), which is required 15 for the transcription of a family of small, acid—soluble proteins that protect the spore DNA from different types of environmental insults (Mascuiet al., 1988). spoIIIG appears to be transcribed by three different mechanisms. Prior to the formation of the asymmetric septum (stage II), spoIIIG is transcribed as the third member of the spoIIG operon (discussed below) by GA RNA polymerase, the vegetative form of RNA polymerase, and SpoOA-P (Masuda et al., 1988; Karmazyn- Campelli.et al., 1989). However, this probably does not lead to the production of 06, because there is a stem-loop structure located upstream of spoIIIG that is predicted to block its translation. After septation, spoIIIG is transcribed by GP RNA polymerase from a promoter located proximal to the spoIIIG ORF (Sun et al., 1989; Schmidt et al., 1990; Partridgeret al., 1991). However, this early transcription does not appear to lead immediately to active 06, because there is only a very low level of B—galactosidase expressed from fusions between OG-dependent promoters and lacz at this time (Mason et al., 1988). It has been proposed that like 0?, OG is held inactive by an anti-sigma factor. In fact, genetic and biochemical studies suggest that SpoIIAB inhibits CG activity (Rather & Moran, 1988; Kirchman et al., 1993). If SpoIIAB does repress both 6F and 06 activity, then A 16 its repression of these sigma factors must be relieved by different mechanisms, because GP and 05 are activated sequentially. After engulfment of the forespore by migration of the sporulation septum (stage III), CG becomes active and directs the transcription of its own gene (Karmazyn—Campelli et al., 1989) as well as other genes (Mascuiet al., 1988). Autoregulation provides a burst of spoIIIG expression exclusively in the forespore. llll _ J]_ 'E' . E I spoIIG locus contains two genes, spoIIGA and spoIIGB, involved in the production of active 63, which is required for the transcription of genes that are involved in engulfment, cortex formation, and coat formation (Trempy'et al., 1985a; Trempy'et al., 1985b; Jonas et al., 1988; Stragier et al., 1988). Transcription of spoIIG by (5'A RNA polymerase and SpoOA—P begins prior to formation of the asymmetric septum (Kenney'et al., 1988; Kenney'et al., 1989; Satola.et al., 1991; SatolaIet al., 1992). However, the product of spoIIGB, 63, appears to be active only in the mother cell, because B-galactosidase activity expressed from a cE-dependent lacz fusion is localized to the mother cell (Driks & Losick, 1991). GE is synthesized as an inactive precursor, pro-OE, A 17 with an additional 27—29 amino acids at its N-terminus (LaBell.et al., 1987). Proteolytic processing of pro-0’E to CE is dependent upon spoIIGA, which encodes a protein that contains some sequences conserved in aspartic proteases (Jonaset al., 1988; Stragieret al., 1988). Processing of pro-OE to OE appears to occur only after septation (stage II). sigK encodes 6K (Stragier et al., 1989), which is required for the transcription of several spore coat proteins that protect the spore from environmental insults (Donovanet al., 1987; Sandmanet al., 1988; Zheng & Losick, 1990; Cutting et al., 1991c). sigK is a composite gene that is generated by joining spoIVCB, encoding the N-terminal portion of OK, to spoIIIC, encoding the C—terminal portion of OK(Stragieret al., 1989). This event occurs exclusively in the mother cell, because it is dependent upon two genes, spoIVCA and spoIIID (Kunkelet al., 1990), which are transcribed by OERNA polymerase (Tatti.et al., 1991; Satr>et al., 1994). spoIVCA encodes the putative recombinase (Kunkel et al., 1990; Satc>et al., 1990), while spoIIID encodes a DNA—binding protein (Kroos et al., 1989; Kunkel et al., 1989; Stevens & Errington, 1990). spoIVCA transcription by 0S RNA polymerase is stimulated by SpoIIID (Satc>et al., 1994; Chapter II). In addition to affecting spoIVCA transcription, SpoIIID 18 stimulates sigK transcription by GE RNA polymerase (Kunkel.et al., 1988; Chapter II). However, this does not lead immediately to active OK, because like GE, OK is synthesized as an inactive precursor, in this case with an additional 20 amino acids at its N—terminus (Kroos et al., 1989; Stragier et al., 1989; Inlet al., 1990). Proteolytic processing is dependent upon several sporulation genes (discussed below) and does not occur until after the completion of forespore engulfment (stage III). 6K directs the transcription of its own gene (Kunkel.et al., 1988; Kroos et al., 1989) as well as other genes (Zheng & Losick, 1990; Cuttimget al., 1991c). OK-dependent transcription is affected by SpoIIID and by another small DNA-binding protein, GerE (discussed below). Specific proteolysis is a regulatory mechanism used to control gene expression in other organisms. For example, the p50 subunit of NF-kappa-B, a transcription factor that affects the expression of genes involved in immune function, inflammation, and cellular growth, is generated by removing the C-terminal portion of p105 through an ATP-dependent proteolytic pathway (Fan & Maniatas, 1991). The p50 subunit binds DNA, whereas p105 does not (Ghosh, 1990; Kieran, 1990), indicating that the C-terminus interferes with the interaction between the DNA-binding domain (located at the N- terminus) and DNA. Similarly, the additional 20 amino acids fl 19 at the N-terminus of pro—OK reduces the affinity of this protein for its cognate promoters (Dombroski et al., 1993). Interconpartmental Coupling of Gene Expression in Both Compartments Although transcription is driven by forespore- and mother-cell—specific sigma factors, gene expression in each compartment is coupled to events that occur in the other compartment, because CF is required for the activation of GE, GE is required for the activation of 05, and CG is required for the activation of OK(Figure 3). The mechanisms for intercompartmental coupling are still unclear. However, recent studies have provided some insight (reveiwed by Kroos and Cutting, 1994). 9*" mm: It has been proposed that OT-directed gene expression in the forespore leads to a modification of the sporulation septum that triggers pro-<3"E processing in the mother cell (Higgins & Piggot, 1992; Losick & Stragier, 1992). Consistent with this hypothesis are two observations. First, 6? activity appears to be required to remove a thin 20 Figure 3. Criss-cross activation of sigma factors during sporulation. The double line represents the membrane separating the forespore and mother-cell compartments. Adapted from Kroos and Cutting, 1994. 21 forespore a K mother cell 22 layer of peptidoglycan that initially forms in the sporulation septum (Illing & Errington, 1991a; Higgins & Piggot, 1992). Second, SpoIIGA (the putative processing enzyme or regulator of processing) is predicted to contain five membrane—spanning domains and appears to be an integral membrane protein (Peters & Haldenwang, 1991), so it may be in position to sense a morphological change in the sporulation septum. Thus, processing of pro—GEtx>(fi1may couple gene expression in the mother cell to gene expression in the forespore and formation of the sporulation septum. However, other experiments indicate that pro-0'E is processed to (5'E in both compartments (Carlson & Haldenwang, 1989; Kirchmanefi: al., 1993). The relative amounts of pro-0'E and CE in the two compartments are still being debated. Based on these observations, it was proposed that (3'F directs the transcription of two genes: one encoding a product required for processing and the other encoding a product that inhibits OE activity in the forespore (Stragier et al., 1994) . QEandQG 0E directs the transcription of spoIID and spoIIIA (Rong et al., 1986; Driks & Losick, 1991; Illing & Errington, 1991b), which are both required for the activation of CG. 23 spoIID encodes a 37 kDa protein, which resembles a modifier of a cell wall hydrolytic enzyme (Kuroda et al., 1992; Lazarevit:et al., 1992). It has been proposed that this protein may play a role in the release of the forespore in the mother cell (Kuroda et al., 1992; Lazarevic et al., 1992) . spoIIIA is a seven-cistron operon. All of these genes appear to encode proteins with membrane spanning domains (P. Stagier, unpublished data). It has been proposed that these proteins reside in the outer membrane of the forespore and affect OG-dependent transcription by affecting the stability of SpoIIAB (Kirchmanemzal., 1993). Consistent with this idea, the level of SpoIIAB remained high in both compartments during the development of spoIIIA mutant cells, but was greatly reduced in the forespore during the development of wild-type cells (Kirchman et al., 1993) . 96de CE in the forespore directs the transcription of spoIVB, which encodes a 43 kDa protein that is required for the processing of pro-(1’K to OK (Lu et al., 1990; Cutting et al., 1991a). SpoIVB may play a role in the synthesis of the germ cell wall and this may stimulate processing. Consistent with this hypothesis, alleviating the requirement for pro-GK processing does not rescue sporulation in SpoIVB mutant 24 cells, indicating that SpoIVB plays some other role in sporulation in addition to its role in processing (Cuttingefl: al., 1991a). Alternatively, SpoIVB may directly interact with proteins involved in processing. spoIVF is a two-cistron operon. Genetic studies indicate that one member of this operon, SpoIVFB, encodes either the protease responsible for the processing of pro—OK to OK or a regulator of the processing enzyme (Cutting et al., 1990; Cuttingtet al., 1991b). The N-terminal portion of SpoIVFB resembles zinc proteases (S. Lu & L. Kroos, unpublished data). SpoIVFB activity is regulated by proteins encoded by spoIVFA, the other member of the spoIVF operon, and bofA (bypass of forespore), since some mutations in these genes relieve the requirement for CC and spoIVB in the processing of pro-0K to 0K (Cutting et al., 1990) . SpoIVFA has a positive and negative effect (Cuttingem:al., 1991b). In its positive role, SpoIVFA appears to stabilize SpoIVFB activity, since a mutation in spoIVFA results in SpoIVFB becoming thermosensitive. In its negative role, SpoIVFA inhibits SpoIVFB until the 06-dependent signal is received from the forespore. BofA also has a negative effect until a signal is received from the forespore (Cutting et al., 1990) . spoIVF and bofA are transcribed by (3'E RNA polymerase and 25 their expression is thereby confined to the mother cell (Cutting et al., 1991b; Ireton & Grossman, 1992a; Ricca et al., 1992). The proteins encoded by these genes appear to have membrane spanning domains (Cutting et al., 1991b; Ricca et al., 1992). Based on these observations and the results from the genetic studies, it was proposed that these proteins form an oligomeric complex in the outer membrane of the forespore and govern pro-OK processing by sensing either a SpoIVB-dependent morphological change in the membrane and/or a spoIVB-dependent signal from the forespore (Cuttingemzal., 1991b; Ricca et al., 1992) . Role of DNA—binding Proteins During Sporulation The temporal pattern of gene expression appears to be established, in part, by DNA-binding proteins that act as both transcriptional activators and repressors. 52206 As already discussed, the key event in the initiation of sporulation appears to be the phosphorylation of SpoOA via the phosphorelay system. SpoOA represses aer transcription by 61* RNA polymerase (Perego et al., 1988) . aer encodes a transcriptional repressor, which blocks the transcription of several genes that are expressed during the 26 transition from exponential growth to stationary phase. SpoOA stimulates the transcription of spoIIA (encoding O?) and spoIIG (encoding OE) by (5'H and.0m RNA polymerase, respectively (Errington & Mandelstam, 1986; Savva & Mandelstam, 1986; Burbulys et al., 1991; Satola et al., 1991; Satola et al., 1992). The repression of aer occurs earlier than the stimulation of the spoII genes. Based on this observation, it was proposed that a low level of SpoOA results in the transcription of genes expressed during the transition from exponential growth to stationary phase and that the accumulation of SpoOA results in the transcription of spoII genes required for early stages of sporulation (Hoch, 1994). 59.911112 SpoIIID encodes a 10.8 kDa protein that contains a putative helix-turn-helix DNA-binding motif (Kroos«et al., 1989; Kunked.et al., 1989; Stevens & Errington, 1990). Genetic studies and in vitro transcription experiments indicate that this protein stimulates and inhibits the transcription of OH- and OK-dependent genes (Table 1; Kunkel et al., 1989; Stevens & Errington, 1990; Kroos«et al., 1989; Satc>et al., 1994; Ireton & Grossman, 1992a; Chapter II; J. Errington unpublished data; H. Ichikawa and L. Kroos unpublished data; B. Zhang and L. Kroos unpublished data). Since SpoIIID has both a positive and negative effect on 1 I6E Reculon bl .IIIZOQC yd bejretls q lgna g3 VIIfD Effect I05 Re I I 9 I a; p. ~‘ :‘n “ tn we 4‘ a}: “H :3 C- C) 0) E} 3 (J -ll. --4 PSIIS DOB 1-9'I0nw,:q 9d Giiloqa ssisv'bnr oq K—d I to a it mral .1 b Echi- Q }.iKHT;-.Iir¢q> I) b‘ O 0' O M o QHILQIQCIQ a; In In In La; .u 0! Inhibition i noizeeque suede ssnsfl .1 sidsT i o: obztd Ultisqa asssolbal egg: {3 nisze .oxalu a: not3qiaoene13 seebtsoJosIBo—a in aoieaaiqxo our .1051 0: sure”). so ”M --m‘“ we. I 27 Table 1. Genes whose expression is affected by SpoIIID. Bold type indicates SpoIIID binds to the promoter and affects transcription in vitro. Plain type indicates SpoIIID affects the expression of B-galactosidase from a fusion of the gene of interest to lacZ. 28 Nuou Q>omm noon soon . open 330% 83325 MSon mo>Homm Mono _ No.3 320% qoflnmassflm coaomom so coasmmm mo powwow. 333% 29 Figure 4. Temporal switch in the mother—cell pattern of gene expression. SpoIIID stimulates sigK, but represses cotD transcription by GK RNA polymerase. The inactivation and/or sequestering of SpoIIID (represented by a question mark) switches the mother-cell pattern of gene expression from sigK to cotD transcription. 30 (- 0K RNA polymerase \ sigK p cotD (stage IV) (stage V) SpoIIID j 31 transcription by (3'K RNA polymerase, it was postulated that the inactivation of SpoIIID establishes a switch in the mother-cell pattern of gene expression (Figure 4; Kroos & Losick, 1989). GELE gerE encodes an 8.5 kDa protein that contains a putative helix-turn-helix DNA-binding motif (Cutting & Mandelstam, 1986). This protein is similar to several response regulators (Kahn & Ditta, 1991). Genetic studies indicate that GerE stimulates or inhibits the expression of several spore coat genes by (‘5'K RNA polymerase. For example, in a gerE mutant background, cotB and cotC fail to be expressed, cotD is partially expressed, and cotA is overexpressed (Zhengwet al., 1992). Based on these observations, it was proposed that GerE may affect how coat proteins are deposited on the surface of the forespore and thereby affect the resistance properties of the mature spore (Zheng et al., 1992) . Consistent with this idea, gerE mutant spores are lysozyme- sensitive and germination defective (Feng & Aronson, 1986). Significance The regulatory features for controlling gene expression described above (i.e, sigma factors, anti-sigma factors, 32 proteolytic processing, sequence-specific DNA—binding proteins, and coupling to morphogenesis) are present in other systems. A excellent example is flagellum biosynthesis in Salmonella typhimurium. The flagellar genes are grouped into 13 operons (Kutsukakeem:al., 1988). These operons have been divided into three classes. Class I genes are required for the expression of class II genes and class II genes are required for the expression of class III genes (Kutsukakeefi: al., 1990). For example, fliA is a class II gene that encodes an alternative sigma factor, 0?, which is required for the transcription of class III genes (Ohnishi.et al., 1990). Interestingly, flgM is another class II gene that encodes an anti-sigma factor, which regulates OF activity (Hughes et al., 1993). The FlgM negative effect is relieved during flagellum biosynthesis by exporting it into the media (HugheSIet al., 1993). Export requires completion of part of the flagellum (Hughes et al., 1993) . This observation demonstrates the expression of class III genes is coupled to morphogenesis. Thus, studies on the model system of B. subtilis sporulation are likely to provide insight into the temporal and spatial regulation of gene expression in other organisms. CHAPTER II The SpoIIID Switch Protein Activates and Represses Transcription by Both Mother-Cell—Specific Forms of RNA Polymerase "The rose and the thorn, and sorrow and gladness are linked together." Saadi 33 34 Abstract Mother-cell-specific gene expression during sporulation of Bacillus subtilis is controlled by (3'E and 0“ RNA polymerases. GE is required for the expression of genes during stage III (engulfment of the forespore), while UK is required for the expression of genes during stage IV (formation of the spore cortex) and stage V (formation of the spore coat). Previous studies indicated that SpoIIID could influence transcription by (3'K RNA polymerase in vitro. We demonstrate here that SpoIIID is a DNA-binding protein that recognizes specific sequences in the promoter regions and open reading frames of both OE- and OK-dependent genes. We also show that SpoIIID binding can activate or repress transcription by both forms of RNA polymerase. These results support the idea that the appearance and subsequent disappearance of SpoIIID plays a major role in establishing the mother-cell pattern of gene expression during stages III to V of sporulation. 35 Introduction Under conditions of nutrient deprivation, the gram- positive bacterium Bacillus subtilis undergoes a series of morphological changes that culminate in the formation of an endospore (reviewed by Errington, 1993). The first easily observed morphological structure is an asymmetrically positioned septum that divides the bacterium into two compartments, the mother cell and the forespore. Both of these compartments receive a copy of the genome, but they realize alternative developmental fates because gene expression is regulated spatially. Spatial regulation is established by compartment-specific activation of sigma subunits of RNA polymerase (Losick & Stragier, 1992). Two mother-cell—specific sigma factors are 03 (Driks & Losick, 1991) and OK (Kroos et al., 1989; Stragier et al., 1989) . GE is required for the migration of the septum and engulfment of the forespore in a double membrane (stage III). OK is required for the deposition of cell-wall-like material called cortex between the membranes of the forespore (stage IV) (Cutting et al., 1991a) and the synthesis of spore coat proteins that assemble on the surface of the forespore (stage V) (Kroos et al., 1989; Zheng & Losick, 1990) . Gene expression in the mother cell is regulated temporally by the ordered appearance of GE, then 0K. Temporal 36 regulation in the mother cell also involves two transcription factors, SpoIIID and GerE, that affect gene expression driven by (3'E and/or (3'K RNA polymerase. Here we focus on transcriptional regulation by SpoIIID. A mutation in spoIIID, which is predicted to encode a 10.8 kDa protein with a putative helix-turn-helix DNA-binding motif (Kunkel.et al., 1989; Stevens & Errington, 1990), affects the expression of several OE-dependent genes. For example, in spoIIID mutant cells bofA (encoding a protein that appears to inhibit processing of pro-0K to UK) is overexpressed (Ireton & Grossman, 1992a), but spoIVCA (encoding a putative recombinase that generates the composite sigK gene (Kunkel et al., 1990; Sato et al., 1990; Popham & Stragier, 1992)) and sigK (encoding pro-0K) fail to be expressed (Kunkel et al., 1988; Sato et al., 1994). However, it was unknown whether the SpoIIID protein affects OE-dependent transcription of these genes directly or indirectly. SpoIIID was shown previously to stimulate sigK and inhibit cotD (encoding a spore coat protein) transcription by 0'K RNA polymerase in vitro (Kroos et al., 1989) . These effects appeared to result from SpoIIID binding to the DNA. Based on these observations, it was proposed that inactivation of SpoIIID establishes a switch in the 37 mother-cell pattern of gene expression from sigK transcription at stage IV to cotD transcription at stage V (Kroos & Losick, 1989). Consistent with this model, the level of SpoIIID decreases sharply at the proper time during development (Halberg & Kroos, 1992). Here we demonstrate that SpoIIID is a DNA-binding protein that recognizes specific sequences in the promoter regions and open reading frames (ORFs) of bofA, spoIVCA, sigK, and cotD. A consensus sequence for SpoIIID binding is proposed. We show that SpoIIID activates spoIVCA and sigK transcription, but represses bofA transcription by GE RNA polymerase in vitro. We also show that SpoIIID binding to the -35 region of sigK and cotD is sufficient to mediate the transcriptional effects that SpoIIID has on the transcription of these genes by (5‘K RNA polymerase in vitro. These results suggest that the appearance and subsequent disappearance of SpoIIID plays a central role in determining the temporal pattern of mother-cell gene expression during the transition from stages III through V of sporulation. 38 Materials and Nethods DNaSLLfmtnrinting DNA fragments labelled at only one end were prepared as follows. For the analysis of spoIVCA, 10 ug of pUCllBIVCP (Satt>et al., 1990) was digested with EcoRI and labelled using either the fill-in reaction of the Klenow fragment of DNA polymerase I and (a-xm) dATP or by treating with alkaline phosphatase followed by phage T4 polynucleotide kinase and (7- 32P) ATP. In both cases, the labelled DNA was digested with HindIII and the 654 bp EcoRI-HindIII fragment was purified after electrophoresis in a non-denaturing polyacrylamide gel using the crush and soak method described previously (Sambrookmet al., 1989), except double-stranded poly (dI-dC) (10 pg) was added to serve both as a carrier during the ethanol precipitation and as a competitor during the footprinting experiments. Similar end-labelling and recovery methods were used to prepare DNA fragments for analysis of cotD. A 279 bp EcoRI-NarI fragment from pLRKlOO (Kroosefl: al., 1989) was labelled at the EcoRI site and a 443 bp EcoRI-HindIII fragment from pLRKlOO was labelled at the HindIII site. For the analysis of sigK, 10 ug of pBK16 (Kroos et al., 1989) was digested with XbaI and end-labelled as described above. The labelled DNA was digested with HindIII, which releases a 368 bp fragment containing a 39 portion of sigK for footprinting, and EcoRI, which releases a fragment from the vector DNA that is sufficiently small (27 bp) so as not to interfere with footprinting. In this case, the vector DNA serves as competitor during footprinting reactions. Similarly, for the analysis of bofA, an approximately 400 bp EcoRI-BamHI fragment from pIK132 (Ireton & Grossman, 1992a) was labelled at the EcoRI site for footprinting and XhoI was used to release a 33 bp labelled fragment from the vector DNA. DNase I footprinting experiments were performed according to method 2 described by Zhengwet al., 1992, except 0.5 pmole of end-labelled DNA was used. SpoIIID was gel— purified from fractions of partially purified (5'K RNA polymerase as described previously (Kroosem al., 1989). Weiss End-labelled DNA fragments (described above) were subjected to the chemical cleavage reactions of Maxim and Gilbert as described previously (Maniatisiet al., 1982). I 'l I . I' (fliand (5'K RNA polymerases were partially purified from sigK (BK410; Kunkel et al., 1989) and gerE (SC104; Kroos et 5.1., 1989) mutant cells, respectively, following the Procedure for the partial purification of OK RNA polymerase 40 described previously (Kroos et al., 1989) . The UK RNA polymerase was comparable in protein composition and in cotD- and sigK; transcribing activities to fraction 24 shown in Figure 2 of Kroosem:al., 1989. Transcription reactions (45 ul) were performed as described previously (Carter & Moran, 1986), except that RNA polymerase was allowed to bind to the DNA template for 10 minutes at 37‘C before the addition of nucleotides. The labelled nucleotide was (oz-321?) CTP. Heparin (6 pg) was added 2 minutes after the addition of nucleotides to prevent reinitiation. After the reactions were stopped, 20 ul of each reaction mixture was subjected to electrophoresis and transcripts were detected by autoradiography. The signal intensities were quantitated using a Visage 110 Imager Analyzer (BioImage). 41 Results SpoIIID binds to specific sequences in spoIVCA, sigK,.bo£A, and cotD. Gel-mobility shift assays indicated that SpoIIID binds to DNA fragments containing the spoIVCA, sigK, bofA or cotD promoter region (data not shown). To more precisely localize the binding of SpoIIID in these genes, DNase I protection experiments were performed. Radioactive DNA probes labelled separately on the non- transcribed or transcribed strand were incubated with SpoIIID and then mildly digested with DNase I. The resulting fragments were separated by gel electrophoresis and visualized by autoradiography. For the spoIVCA and sigK genes, sites in both the promoter region and ORF were protected by SpoIIID from DNase I digestion (Figure 1). In the case of spoIVCA, protection in the promoter region (site 1) spanned from -31 to -18 on the non-transcribed strand and from -37 to -12 on the transcribed strand (panel A), while protection in the ORF (site 2) spanned from +138 to +160 on the non-transcribed strand and from +137 to +154 on the transcribed strand (panel B). Extensive protection of site 1 was observed with 30 ng of SpoIIID, whereas only weak protection of site 2 was observed with 240 ng of SpoIIID, indicating that binding to site 1 is significantly stronger than binding to site 2. In the case of sigK, SpoIIID IProtected one site in the promoter region and two sites in 42 Figure 1. SpoIIID footprints on spoIVCA and sigK. Radioactive DNA probes separately end-labelled on the non- transcribed or transcribed strand were incubated in separate reactions with no protein (lane 1), 30 ng (lane 2), 60 ng (lane 3), 120 ng (lane 4), or 240 ng (lane 5) of gel-purified SpoIIID and then mildly digested with DNase I. The resulting DNA fragments were separated by electrophoresis on a 7% polyacrylamide gel containing 8 M urea alongside a sequencing ladder generated by chemical cleavage of one of the end- labelled DNA probes. (A) and (B) Footprints in the spoIVCA promoter region and ORF, respectively, identified using probes labelled at the EcoRI site located downstream of the transcriptional start site. (C), (D), and (E) Footprints in the sigK promoter region and ORF identified using probes labelled at the XbaI site located downstream of the transcriptional start site. Arrows indicate the boundaries of protection and numbers indicate positions relative to the transcriptional start site. 43 -37 all. I... I. l.‘. ..1. .. :l- II... transcribed non—transcribed 7 3 1 + 3 4 5 2 0 ...(+154 1 .0- 88:3 transcribed non-transcribed 44 G non-transcribed transcribed 5 ..v m .m e a 3 . . m 2T!!!“ ...!!I 3...... ...!!I 2.2.... 3's... 8.38.“... lo... . . 2.23m ital-... ...Im ... 3 ee 45 2 4 ! ‘1‘+106 . e 1 (+111 - I (+126 i‘HBO 3 O non-transcribed transcribed 0m I ”9 «09 Hi to 9‘01 e I. 0 "I" 1.00.!»- H ’ O ”Male... ”W’s“... 0o . 46 the ORF. Protection in the promoter region (site 1) spanned from -37 to -19 on the non-transcribed strand and from -39 to -19 on the transcribed strand (Figure 1C). Protected regions in the ORF (sites 2 and 3, respectively) spanned from +80 to +97 and +111 to +130 on the non-transcribed strand and from +75 to +91 and +106 to +126 on the transcribed strand (Figures 1D and 1E). Binding to site 2 was significantly stronger than binding to sites 1 and 3. Similarly, for the bofA and cotD genes, sites in both the promoter region and ORF were protected by SpoIIID from DNase I digestion (Figure 2). In the case of bofA, protection in the promoter region (site 1) spanned from -13 to +13 on the non-transcribed strand and from -17 to +10 on the transcribed strand (panel A), while protection in the ORF (site 2) spanned from +57 to +74 on the non-transcribed strand and +54 to +73 on the transcribed strand (panel B). Binding to site 1 was slightly stronger than binding to site 2. In the case of cotD, SpoIIID protected two sites in the promoter region and one site in the ORF. Protected regions in the promoter region (sites 1 and 2, respectively) spanned from -73 to -59 and -40 to -21 on the non-transcribed strand and from -75 to -61 and -37 to -23 on the transcribed strand (Figures 2C and 2D). Protection in the ORF (site 3) spanned from +61 to +79 on the non—transcribed strand and from +54 to +72 on the transcribed strand (Figure 2E). Binding to sites 2 and 3 is comparable in strength and much stronger than 47 Figure 2. SpoIIID footprints on bofA and cotD. Radioactive DNA probes separately end—labelled on the non-transcribed or transcribed strand were incubated in separate reactions with no protein (lane 1), 30 ng (lane 2), 60 ng (lane 3), 120 ng (lane 4), or 240 ng (lane 5) of gel-purified SpoIIID and then mildly digested with DNase I. The resulting DNA fragments were separated by electrophoresis on a 7% polyacrylamide gel containing 8 M urea alongside a sequencing ladder generated by chemical cleavage of one of the end-labelled DNA probes. (A) and (B) Footprints in the bofA promoter region and ORF, respectively, identified using probes labelled at the EcoRI site located downstream of the transcriptional start site. (C) and (D) Footprints in the cotD promoter region identified using probes labelled at the EcoRI site located upstream of the transcriptional start site. (E) Footprints in the cotD ORF identified using probes labelled at the HindIII site located downstream of the transcriptional start site. Arrows indicate the boundaries of protection and numbers indicate positions relative to the transcriptional start site. 48 ‘~ 4: 5 1 2 3 4 5 ""(-17 o e .- (+13 ‘ 3 (+10 ; - a non—transcribed transcribed (+54 (+57 (+73 non—transcribed rtranscribed _, 49 C .3 1.2 E 12 5 -..-... n "e-e .. ' (—61 n ”('59 .... ...; "e-e r— (45 I. ('73 ...--. w.- trans rited non—transcribed 2 non-transcribed D G G/ A .‘b (—21 5 3(40 ”IIII .. Ill. “- II II - M- - A .A I5-- 13‘.G cm 12 3 4 5 .---.-.. -..—«-— + \J N z. -— -u transcribed 50 binding to site 1. The results of the DNase I protection experiments are summarized in Figure 3. SpoIIID activates spoIVCA and sigK transcription, but rcprcsscs botA transcription by or RNA polymarasa in vitro. To determine how SpoIIID binding affects the transcription of spoIVCA, sigK, and bofA, linearized DNA templates were transcribed with partially purified OE RNA polymerase alone or in the presence of gel-purified SpoIIID (Figure 4). (5'E RNA polymerase produced run-off transcripts of the expected sizes from the spoIVCA, sigK, and bofA templates (panels A, B, and C, respectively). The presence of SpoIIID increased the spoIVCA (panel A, indicated by arrowheads, compare lanes 1 and 2 or lanes 3 and 4) and sigK (panel B, compare lanes 1 and 2) signals 5-fold and 6-fold, respectively, but it markedly reduced the bofA signal (panel C, compare lanes 1 and 2 or lanes 3 and 4). Thus, SpoIIID activates spoIVCA and sigK transcription, but it represses bofA transcription, by (5'E RNA polymerase in vitro. Activation of spoIVCA and sigK transcription and repression of bofA transcription by SpoIIID in vitro are consistent with the effects of a spoIIID mutation on the expression of lacz fusions to these promoters in vivo (Kunkelmet al., 1988; Ireton & Grossman, 1992a; Satt>et al., 1994). Surprisingly, (5'E RNA polymerase produced another run- 51 Figure 3. Position of SpoIIID binding sites in spoIVCA, sigK, bofA and cotD. Overlining and underlining indicate regions on the non-transcribed and transcribed strands, respectively, protected by SpoIIID from DNase I digestion (Figures 1 and 2). The dashed portion of a line indicates a region of uncertain protection due to a lack of DNase I I digestion in this region. Asterisks indicate positions of enhanced cleavage by DNase I upon SpoIIID binding. Numbers refer to positions relative to the transcriptional start site. Nucleotide sequences upstream of the transcriptional start sites of spoIVCA, sigK, and bofA are aligned with respect to conserved nucleotides in the -10 and —35 regions of promoters transcribed by of RNA polymerase, shown at the top of the diagram (Roelsem al., 1992). k means G or T and m means A or C. Nucleotide sequences upstream of the transcriptional start sites of sigK and cotD are aligned with respect to conserved nucleotides in the -10 and -35 regions of promoter transcribed by (3'K RNA polymerase, shown above the sigK and cotD sequences (Zhengwet al., 1992). Matches to the consensus sequences are shown as bold, capital letters. 52 ouou uuuuuuuuuoUUumuuuuUUOUUOuooua... onuuum¢auou49¢0uuuuuauuuouocuuua0«uoaouuoouuu ...uuouoauuuuonuuoouoamuaauuuu — _ - IOOIIO _ _ 8+ q 3+ 3 a S... u a no- I now» dunnouuuoo4340uouooooouououoaocuduooooanoon ~+ «a 538 u an S u on on: an- IQ ¢ (won uuuuouuouuuuuouuoouUuooouuuuuuuao ... aounmauououoooomuuoayouauou¢fi¢0uouyouooouoouonudaouoouooouoo _ . u 3+ 2+ T. a C rawn opoouuooouuuuuuoomuoooocnuaOumoouuououoouuuuuouOuuoooouoouuo ...mucususu¢049<0nuunaccouoooooco‘u¢0uooooumoo . _ _ ~2+ ... A 2+ 2 N C I i Homn ummoouoooouuouaoooauooooooouoooououuooao ...cocoauncooda40cuuouoooouaoomoaaua4uaoououom - - — .... oma+ m~a+ H+ a 35.40.. an 3-2 -9532... can an- 53 Figure 4. Effects of SpoIIID on spoIVCA, sigK, and bofA transcription in vitro. Linearized plasmid DNA (1 ug) was transcribed with partially purified OE RNA polymerase (200 ng) alone or in the presence of gel-purified SpoIIID (120 ng). Run-off transcripts were electrophoresed in a 5% polyacrylamide gel containing 8 M urea. Arrowheads denote the positions of run-off transcripts of the expected sizes, while asterisks denote the positions of run-off transcripts of unexpected sizes. The sizes of run-off transcripts were estimated from the migration of end-labelled fragments of MspI-digested pBR322. (A) spoIVCA transcription from pUC118IVCP digested with KpnI (lanes 1 and 2, 193-base transcript) or EcoRI (lanes 3 and 4, 201-base transcript) with GE RNA polymerase alone (lanes 1 and 3) or in the presence of SpoIIID (lanes 2 and 4). (B) sigK transcription from pBK16 digested with XbaI (l70-base transcript) with GE RNA polymerase alone (lane 1) or in the presence of SpoIIID (lane 2). (C) bofA transcription from pIK132 digested with EcoRI (lanes 1 and 2, 134-base transcript) or XbaI (lanes 3 and 4, 164-base transcript) with GE RNA polymerase alone (lanes 1 and 3) or in the presence of SpoIIID (lanes 2 and 4). 54 55 off transcript which is approximately 40 bases longer than the expected product from the spoIVCA template (Figure 4A, indicated by asterisks). The longer transcript appears to originate from a site upstream of the spoIVCA transcriptional start site since its size varies in the same manner as the size of the spoIVCA transcript when templates are cleaved at different downstream restriction sites. The presence of SpoIIID markedly reduced the longer transcript signal (Figure 4A, compare lanes 1 and 2 or lanes 3 and 4). A possible explanation for these results is presented in the Discussion. SpoIIID binding in the -35 region is sufficient to activate sigK transcription and repress cotD transcription by on RNA polymerase in vitro. SpoIIID was shown previously to stimulate sigK transcription and inhibit cotD transcription by 0’“ RNA polymerase in vitro (Kroos et al., 1989) . To determine which SpoIIID binding sites are required for these effects, sigK and cotD templates containing different combinations of SpoIIID binding sites were transcribed with partially purified (3'K RNA polymerase alone or in the presence of gel—purified SpoIIID (Figure 5). The presence of SpoIIID increased sigK transcription 3-fold from a HindIII-Xbal fragment containing sites 1-3, and 3-fold from this template after it was digested with HhaI (panel A). The presence of SpoIIID also increased transcription to a 56 Figure 5. Effects of SpoIIID on the in vitro transcription of sigK and cotD templates containing different combinations of SpoIIID binding sites. Isolated fragments containing a portion of sigK or cotD served as templates directly, or were digested with restriction enzymes prior to serving as templates. The fragments (300 ng total DNA in each case) were transcribed separately with partially purified 0K RNA polymerase (200 ng) alone or in the presence of gel-purified SpoIIID (120 ng). Run-off transcripts were electrophoresed in a 5% polyacrylamide gel containing 8 M urea. Arrowheads denote the positions of run—off transcripts of the expected sizes in each panel, as judged from the migration of end- labelled fragments of MspI-digested pBR322. (A) For sigK, a HindIII-Xbal fragment isolated from pBK16, and this fragment digested with HhaI, were transcribed with OK RNA polymerase alone (lane 1) or in the presence of SpoIIID (lane 2). The expected sizes of the run-off transcripts from the HindIII- XbaI fragment and the HhaI subfragment are 170 and 60 bases, respectively. (B) For cotD, an EcoRI-HindIII fragment isolated from pLRKlOO, and this fragment digested with DraI, NarI, or both DraI and NarI, were transcribed with OK RNA polymerase alone (lane 1) or in the presence of SpoIIID (lane 2). The expected size of the run-off transcript from the EcoRI-HindIII fragment or this fragment digested with DraI is 225 bases. The expected size of the run-off transcript from the EcoRI-HindIII fragment digested with NarI or with DraI and NarI is 54 bases. HindIII +1 XbaI HhaI HhaI HindIII—XbaI HhaI digested l 2 l 2 >~ _.< 57 EcoRI +1 HindIII DraI Natl ECORI-HindIII DraI digested 1 2 1 2 i >' . 4 NarI digested DraI/NarI digested 1 2 f 1 .i— ii 58 similar degree from the isolated HhaI fragment (data not shown). Thus, SpoIIID binding in the -35 region of sigK is sufficient to activate transcription by 0'K RNA polymerase in vitro. The presence of SpoIIID markedly reduced cotD transcription from an EcoRI-HindIII template containing sites 1-3, and from this template after it was digested with DraI, NarI, or both DraI and NarI (panel B). The presence of SpoIIID also markedly reduced transcription from the isolated DraI-HindIII and EcoRI—NarI fragments (data not shown). Thus, SpoIIID binding in the -35 region of cotD is sufficient to repress transcription by (3'K RNA polymerase in vitro. 59 Discussion We have shown that SpoIIID is a DNA-binding protein that positively and negatively affects transcription by both mother-cell-specific forms of RNA polymerase. SpoIIID activates spoIVCA and sigK transcription, but represses bofA transcription, by (5'E RNA polymerase in vitro. SpoIIID also activates sigK transcription, but represses cotD transcription, by (3'K RNA polymerase in vitro. These results are consistent with the effects of a spoIIID mutation on gene expression in vivo (Kunked.et al., 1988; Zheng & Losick, 1990; Ireton & Grossman, 1992a; Satc>et al., 1994). Taken together, these observations suggest that SpoIIID plays a direct role in establishing the temporal pattern of mother-cell gene expression by both activating and repressing transcription directed by two forms of RNA polymerase that appear sequentially during sporulation. The -10 and -35 regions of the promoters used in this study are aligned with the proposed consensus sequences for GE and OK-dependent promoters in Figure 3. The spoIVCA, sigK, and bofA promoters all show considerable similarity to the consensus for 03-dependent promoters. It is not obvious from this sequence comparison why, in the absence of SpoIIID, the bofA promoter is more highly transcribed by GE RNA polymerase 60 in vitro than the spoIVCA and sigK promoters (Figure 4). Similarly, both the sigK and cotD promoters match the consensus sequence for Om-dependent promoters quite well, yet the cotD promoter is more highly transcribed by (3'K RNA polymerase in vitro unless SpoIIID is added (Figure 5). Comparison of sequences in the SpoIIID binding sites in the promoter regions and open reading frames of spoIVCA, sigK, bofA, and cotD reveals an apparent consensus for SpoIIID binding, WWRRACAR-Y (W=A or T, R=purine, and Y=pyrimidine; Figure 6A). Interestingly, the ACA sequence in this consensus is similar to the ATA and AC sequences found in the -35 region of strong 03- and OK-dependent promoters, respectively (Figure 3). This may reflect similarity between the putative helix-turn-helix DNA-binding motif of SpoIIID (Kunkelwet al., 1989; Stevens & Errington, 1990) and regions 4.2 of (TE and OK, which are predicted to adopt helix-turn-helix structures that interact with the -35 region of cognate promoters (Helmann & Chamberlin, 1988; Lonettc>et al., 1992). For example, all three proteins have a serine residue at the same position in their putative recognition helix. This serine may hydrogen bond to adenine in the consensus sequences. Some proteins containing a helix-turn-helix DNA-binding motif are dimeric (e.g., catabolite activator protein and tryptophan repressor) and bind to sequences that exhibit dyad 61 Figure 6. Alignment of sequences within SpoIIID binding sites and the arrangement of additional matches to the proposed consensus sequence for SpoIIID binding. (A) Nucleotide sequences protected from DNase I by SpoIIID are aligned with a proposed consensus sequence, which is shown at the bottom. Nucleotides that match this consensus are shown as bold, capital letters. R means purine, Y means pyrimidine, and W means A or T. Note: The sequences shown for sigK site 2 and cotD site 3 are from the opposite DNA strand of that shown in Figure 3. (B and C) Arrangement of matches to the SpoIIID consensus in strong and weak SpoIIID binding sites, respectively. Arrows point 5' to 3' and indicate sequences matching the SpoIIID consensus, with tick marks indicating mismatches. Note: There is a third 7 out of 9 nucleotide match to the SpoIIID consensus to cotD site 1, but it is not shown because it contains two mismatches within the highly conserved ACA sequence. -31 TTGGACAAaC -22 +143 AgGAACAAgC +152 -32 ACAGACAGCC -23 +90 AAAGACAAgC +31 +106 AAAAACAAtg +115 -1 TAGAACAAgC +9 +60 TAGGACtGgT +69 -79 AAAGACAGCT -70 ~37 cAGAACAth -28 +73 ATGGACAAtT +64 -31 +74 +60 -36 +106 -79 WWRRACAR-Y .JL______; * TTGGACAAACAGCTGTTACA v—--TT- _____L___a TTAAAGAGCTTGTCTTT F * GCTTGTACTAGAACAAGC v---r-- ______i__a TAGGACTGGTTAT t-———1———- .1________a _L_L_____A AGACACAGACAGCC (52 spoIVCA site 1 spoIVCA site 2 sigK sigK sigK bofA bofA cotD cotD cotD site site site site site site site site Consensus 12 +90 +72 _________L ..LL_____i AAAAACAATGCCTTTCCACAACC _________A ___J___J_; AAAGACAGCTTAATTGCACACTT -23 +128 -57 WNHNHwNH spoIVCA site 1 sigK site 2 both site 1 bofA site 2 sigK site 1 sigK site 3 cotD site 1 63 symmetry (Otwinowski et al., 1988; Schultz et al., 1991) . Each sequence of the dyad is bound by one of the helix-turn-helix motifs present in the dimer. The consensus sequence we propose for SpoIIID binding does not exhibit dyad symmetry. Consistent with this observation, SpoIIID purified from sporulating B. subtilis is primarily in a monomeric state (B. Zhang and L. Kroos, unpublished data). SpoIIID probably binds to the proposed consensus as a monomer. However, some of the sites bound most strongly by SpoIIID exhibit a second good match (i.e., at least 7 out of 9) to the consensus in inverted orientation relative to the best match (Figure 6B), suggesting possible cooperative interactions between monomers. Some of the weakly bound sites show a second good match in direct orientation (Figure 6C). Mutational studies, and biochemical experiments to determine the number of SpoIIID monomers bound at particular sites, will be required to determine the significance of various arrangements of sequences matching the proposed consensus. Why does SpoIIID bind to both the promoter region and open reading frame of all four genes tested? On the basis of chance alone, a perfect match to the proposed consensus should occur once in about 4 kb of random sequence. In the case of sigK, three SpoIIID binding sites were found within a 170 bp region. However, binding of SpoIIID to site 1 in the promoter region is sufficient to stimulate transcription, as evidenced by our in vitro result (Figure 5A) and the spoIIID 64 dependence in vivo of a fusion between the sigK promoter region (~106 to +4) and the E. coli lacZ gene (Kunked.et al., 1988). These results do not rule out the possibility that SpoIIID binding sites 2 and 3 in the sigK'ORF may help activate transcription in vivo. AlgRl, a response regulator protein from Pseudomonas aeruginosa, stimulates algC promoter activity by binding to three sites located at ~94 to -81, +161 to +174 (in the leader sequence), and +389 to +403 (in the structural gene) (Fujiwaramet al., 1993). However, binding to the upstream site alone is sufficient to mediate 14% of the stimulatory effect in vivo (Fujiwaramet al., 1993). SpoIIID binds to three sites within a 160 bp region encompassing the cotD transcriptional start site, but binding to site 2 in the promoter region is sufficient to repress transcription in vitro (Figure SB). It remains possible that SpoIIID binding to sites 1 and/or 3 facilitates repression. This could involve interactions between bound SpoIIID molecules and looping of intervening DNA, as appears to be important for repression of the gal and ara operons in E. coli (Martin et al., 1986; Mandal et al., 1990). SpoIIID binding to DNA enhanced cleavage by DNase I at one or both ends of each protected region (Figures 1 to 3), perhaps indicating that SpoIIID bends the DNA upon binding. Why does SpoIIID binding in the -35 region of sigK and cotD have an opposite effect on transcription in vitro? In the consensus sequence for SpoIIID binding, there is an 65 absolutely conserved ACA sequence (Figure 6A). The C of this trinucleotide is at position —27 in the sigK promoter and at position -32 in the cotD promoter (Figure 3). Thus, SpoIIID may bind to opposite faces of the DNA in the -35 region of the sigK and cotD promoters. If so, the interactions between SpoIIID and OK RNA polymerase would be expected to be different, perhaps accounting for the opposite effects on sigK and cotD promoter activity. In particular, we propose that repression of cotD transcription may result from direct competition between SpoIIID and 6K RNA polymerase for contacts with the AC sequence at positions -33 and —32 (Figure 3). Activation of sigK transcription may result from interactions between SpoIIID and either OK or the a.subunits of RNA polymerase, based on analogy with emerging evidence for several prokaryotic transcriptional activators (Ishihama, 1993). SpoIIID activates spoIVCA transcription, but represses bofA transcription by GE RNA polymerase (Figure 4). In the case of spoIVCA, a perfect match to the consensus sequence for SpoIIID binding is present at nearly the identical position (one nucleotide closer to the transcriptional start site) as in the sigK promoter. This observation is consistent with the idea that SpoIIID activates transcription by binding to a particular face of the DNA helix. Moreover, it suggests that SpoIIID activates transcription by GE and OK 66 RNA polymerase via a similar mechanism. However, the possible role of SpoIIID binding to site 2 in the spoIVCA ORF has not been explored since site 2 was present in all of our in vitro experiments, as well as in the reported in vivo experiments with a spoIVCA-lacz fusion (Satc>et al., 1994). In the case of bofA, SpoIIID binding site 1 is centered at the transcriptional start site. This region is exclusively occupied by repressors when considering E. coli transcriptional regulators, because a protein bound here probably either prevents RNA polymerase from recognizing the promoter or prevents a later step in transcriptional initiation (Collado-Videsem:al., 1991). Interestingly, other proteins that function as both an activator and repressor (e.g., fumarate and nitrate regulatory protein) activate transcription by binding in the -35 region of a promoter and repress transcription by binding near the transcriptional start site (Eiglmeier et al., 1989) . Binding of SpoIIID to site 2 in the bofA ORF may be required for repression. Both sites 1 and 2 were present in our in vitro experiments and in a bofA-lacz fusion that showed seven-fold higher expression in spoIIID mutant cells than in wild-type cells. (W RNA polymerase produced an unexpected transcript from the spoIVCA template in vitro (Figure 4A). This transcript appears to originate approximately 40 bp upstream of the in vivo spoIVCA transcriptional start site. Inspection of this region reveals a 7 bp sequence (TCATTGA) at positions -78 to 67 -72 and an 8 bp sequence (TATAGTTA) at positions -57 to —50, which resemble the -35 and -10 regions of OE-dependent promoters, respectively. This promoter does not appear to be active in vivo, based on both 51 nuclease protection and primer extension experiments (Sato et al., 1990; Sato et al., 1994). Interestingly, transcription that appears to be initiated from this promoter is repressed by SpoIIID in vitro (Figure 4A). The observed repression may result from SpoIIID binding to site 1, which lies just downstream of the approximate transcriptional start site of the in vitro promoter. SpoIIID plays a critical role in establishing the mother-cell pattern of gene expression by affecting transcription by GE and 0K RNA polymerase (Figure 7) . OE RNA polymerase transcribes spoIIID (Tatti.et al., 1991). As SpoIIID accumulates, it affects the transcription of other 03- dependent genes. For example, SpoIIID represses bofA transcription (Figure 4; (Ireton & Grossman, 1992a)), but activates the transcription of its own gene (Kunked.et al., 1990; Stevens & Errington, 1990) as well as transcription of spoIVCA and sigK (Figure 4; (Kunkel et: al., 1988; Sato et al., 1994)). Other 03-dependent genes repressed by SpoIIID may include spoIIIA and spoVD, since (1) spoIIIA— and spoVD-directed B-galactosidase activity are significantly 68 Figure 7. Regulatory effects of GE, SpoIIID, and (5'K on mother- cell gene expression. Arrows represent activation, while bars represent repression. 69 Sou .thm Qwou Mb All. Mme» «5&on $8 Bream Scam 70 higher in spoIIID mutant cells than in wild-type cells (Illing & Errington, 1991b; Danielem:al., 1994), and (2) SpoIIID represses spoVD transcription by (5’E RNA polymerase in vitro (B. Zhang and L. Kroos, unpublished data). Other 03-dependent genes activated by SpoIIID may include cotE and spoVK, since cotE— and spoVKedirected B-galactosidase activity are markedly reduced in spoIIID mutant cells (Zheng et al., 1988; Erringtcw1et al., 1989). Thus, an increase in the level of SpoIIID switches the mother-cell pattern of gene expression, repressing the transcription of some 03-dependent genes and activating transcription of others. The primary translation product of sigK is pro-OK, a transcriptionally inactive precursor protein (Kroos‘et al., 1989; Stragierwet al., 1989). pro-0'K is proteolytically processed to active OKhw'removing amino acids from its N-terminus (Imlet al., 1990). As <3“K becomes available, it transcribes its own gene (Kunkel et al., 1988; Kroos et al., 1989) as well as OK-dependent genes whose expression is unaffected by SpoIIID (e.g., cotA and gerE; R. Halberg and L. Kroos, unpublished data). Accumulation of <3“K causes a decrease in the level of SpoIIID, allowing cotD transcription to begin (Halberg & Kroos, 1992). Thus, a decrease in the 71 level of SpoIIID appears to switch the mother-cell pattern of gene expression from OK-dependent genes like sigK, whose expression is activated by SpoIIID, to OK-dependent genes like cotD, whose expression is repressed by SpoIIID. Transcription of two other OK-dependent genes, cotC and cotX, is repressed by SpoIIID in vitro (H. Ichikawa and L. Kroos, unpublished data). GerE is a DNA—binding protein that reinforces the switch brought about by the disappearance of SpoIIID, since its level rises as UK accumulates and its effects on sigK, cotD, cotC, and cotX transcription are opposite the effects of SpoIIID (Zheng et al., 1992; Zhang et al., 1994). In addition, GerE activates the transcription of several other cot genes and represses cotA transcription (Zheng et al., 1992; Zhang et al., 1994) . GerE appears to be the principal regulator of Ox-dependent gene expression, while SpoIIID may "fine-tune" the expression of particular cot genes by affecting the time and/or level of expression, possibly enhancing spore coat assembly. The results presented here demonstrate that SpoIIID is also a direct regulator of 03-dependent gene transcription, including transcription of spoIVCA and sigK, which is essential for production of OK. CHAPTER I I I Fate of the SpoIIID Switch Protein During Bacillus subtilis Sporulation Depends on the Mother-Cell Sigma Factor,om "To be or not to be..." William Shakespeare 72 J. Mol. Iiiol. (1992) 228. 340-349 73 Fate of the SpoIIID Switch Protein during Bacillus subtilis Sporulation Depends on the Mother-cell Sigma Factor, cK Richard Halberg and Lee Kroos Department of Biochemistry Michigan State University East Lansing. All 48824. U.S.A. (Received 14 April 1992.- accepted 20 July 1992) Sporulation of Bacillus subtilis involves the dilferehtiation of two cell types. the mother cell and the forespore. Two key regulators of mother-cell gene expression are SpoIIID. a DNA-binding protein that activates or represses transcription of many different genes. and d‘. a subunit of RNA polymerase that directs the enzyme to transcribe genes encoding proteins that form the spore coat. Previous studies showed that SpoIIID is_needed to produce 0‘. but suggested that SpoIIID represses c‘-directed transcription of genes encoding spore coat proteins. Here we show that a feedback loop connects'the levels of d" and SpoIIID. such that production of 0‘ leads to a decrease in the level of SpoIIID. The existence of the feedback loop was demonstrated by using antibodies prepared against SpoIIID to measure the level of SpoIIID during sporulation of wild-type cells. mutants defective in 0" production. and a mutant engineered to produce 0‘ earlier than normal. The feedback loop operates at the level of synthesis and/or stability of spoIIID mRNA. as demonstrated by measuring the level of apolIlB mRNA during sporulation of wild-type cells and mutants defective in 6" production. Our results suggest that a rise in the level of a" during the stage (IV) of spore cortex formation causes a decrease in the level of SpoIIID. which. at least in part. establishes the switch to the stage V (spore coat formation) pattern of mother-cell gene expression. Keywords: transcription factor; Bacillus subtilis; sporulation; a factor; feedback loop 9‘ I. Introduction In response to starvation. the Gram~positive bacterium Bacillus subtilis a series of morphological changes that result in the formation cfan endospore (Smith at al., 1989). The first easily observed morphological structure that is specific to the sporulation process is an asymmetrically posi- tioned septum. which divides the bacterium into mother-cell and forespore compartments. Both of these compartments receive a copy of the genome. but difl'erential gene expmmion in the two compart- ments drives further morphological change including migration of the septum and engulfment of the forespore in a double membrane (stage III). disposition of cell-wall-like material called cortex between the membranes surrounding the forespore (stage IV). and synthesis in the mother cell ofspore coatproteinathataasemblecntheaurfacaet'the {m («page V). The developmental process culminates with the lysis of the mother cell to release the endospore. Gene expression in the mother-cell compartment during stages III to V ofsporulation is controlled. in WWW “DON 840 part. by two regulatory proteins. SpoIIID and c‘. The gene encoding SpoIIID (spoIIID) is tran- scribed predominantly. if not exclusively. in the mother cell by RNA polymerase containing 0‘ (Kunkel d al., 1989; Stevens to Errington. [990; Tatti ct cl.. 1991). which appears to be active only in the mother cell (Driks & Losick. 1991). SpoIIID la a 108 kDa DNA.binding protein that activities or represses transcription of many difl'erent (Kroos et al., 1989; our unpublished results). of is a sigma subunit of RNA polymerase that directs the enzyme to transcribe genes whose products are neededforformationofthesporecortexandcoat during morphological stages IV and V. respectively (Kroos d (IL. 1989; Zheng et al., 1992). SpoIIID regulates production of c‘ by at least two mechanisms. First. SpoIIID is required for the chromosomal DNA rearrangement that generates the composite gene (sigK) encoding c‘ (Stragier 4 al., 1989; Kunkel d al., 1990). Second. SpoIIID is required for the transcription of sigK (Kunkel 6 al., 1988). which appears to be driven initially by 0" RNA polymerase (our unpublished results). then by c‘ RNA polymerase (Kroos at al., 1989). Another GIMAcademicPre-Uraited 74 ltepulalory Switch in Bacillus subtilis 84l mechanism controlling a" production is a proteo- lytic processing event that generates active 0" from the primary translation product of sigK. called prO-d". and several lines of evidence support the view that pro-o" processing couples mother-cell gene expression to events occurring in the forespore (Cutting st 01.. l990. l99la.b; Lu et al.. l990). In addition to its role in activating transcription of sigK. SpoIIID markedly represses transcription of call) (encoding a spare coat protein; Donovan et ah. l987) by a“ RNA polymerase in vitro (Kroos el al., l989). Here we demonstrate that the level of SpoIIID decreases at the appropriate time during sporulation to produce a switch from sigK to cotD expression. We also present evidence that the SpoIIID decrease is controlled by the production of active 0‘ via a negative feedback loop acting on the synthesis and/or stability of spoIIID mRNA. These results suggest that a rise in the level of c" and a concomitant decrease in the level of SpoIIID govern thetransitionfromthestageIVtothestageV pattern of mother-cell gene expression. 2. Materials and Methods (a) Bedenol' drains B. subtilis strains were provided by R. Losick except for 88.33. which was generated by transforming comMnt 523.2 (germ. trpc2; Errington l. Mandelstam. I986) as described previously (Dubnau t Davidolf-Abelaon. 19'”) with chromosomal DNA isolated from V048 (spol VOBAI9; Cutting at al.. 1990) and selecting chlorarnphenicol-resistant colonies. All strains are hogenic with the Spo‘ strain PY79 (Youngman d al., [984) except for 522.: and BRH3. which are iscgenic with the Spo‘ strain 8038 (Errington & Mandelstam, I980).Use of the specialized transducing phage SszzcctD-loez (obtained from L. Zhang and R. Losick) has been described (Zheng & Losick. I990). (Naturalist-(sporulation Sporulationwaainducsdbyrssupand'mggrowingalll 'm 8! medium as dssaibed (Starlini & Wlmt'l‘lrsonsstcfsporulatlon(1‘.)ia definedaathetimsofresuspenfion. (e) Preparation a] antibodies SpoIIID was pardslly purified from sporuladng B.ssdtilis as described previously (Kroos 4 IL. I989). Proteins in DNA-cellulose column fractions were precipi- tated in 10% trichloroacetic add. and boiled for 5 min in sample butler (0125 Ir-Trir HG. pH 0'8. 2% lwlv) SIB 5% (v/v) tmeroaptosthanol. 10% (vlv) glycerol. (H 95 (em bromophenol tr...» «.4 “bigoted to BMAGE (18% polyacrylamide gel; Thomas & Kornberg. I978). SpoIIID was embed from the gel. elec- troaluted. acetone precipitated. and dissolved in phos- phate-bummed saline (Harlow t. Lans.l988).10ng was .emuhiflsd with heund's complete adjuvant (BRL) and mpctsd intocrnearthepoplitealglandofaNewZealand White rabbit. After a period of 3 weeks. the rabbit received at the same location a booster injection containing 6 pg of SpoIIID emulsified in heund's incom- plete adjuvant (BRL). One week later. the rabbit was bladandssrumwaaprepared(Harlow&Lane. l988). (d) Western blot analysis livery hour after the onset of sporulation. cells were harvested by centrifugation (H.000 g for 5 min) and whole-cell extracts were prepared as described previously (floaty er al.. l99l) except the lysis buffer did not contain DNase I. The amount of protein present in the extracts was quantified by the Bradford method (Bradford. I970). After the addition of 05 vol. 3x sample buffer. [noteina were separated by SDS/PAGE “8% polyacryl- amide gel. Thomas a Kornberg. l978) and electroblotted to a poly(vinylidene dilluoride) membrane (Matsudaira. I987). The membrane was incubated in blocking bufl'er (20 mN-Tris'llCl. pH 7'5. 05 sr-NaO. 2% nonfat dry milk) for 2 h at room temperature with shaking in order to block non-specific interactions between the primary antibodies and the membrane. The membrane was then probed for at least 2 h with shaking at room temperature with either polyclonal antiserum prepared against SpoIIID (this study) or polyclonal antiserum prepared against. ( ' previoudy; Lu 4 al.. I990) diluted l:l000 in antibody buffer (:0 mar-TrirHCl. pH 7-5. 0-5 rr-NaG. 2% nonfat dry milk. 005% Tween 20). Immunodetcction utilising goat anti-rabbit alkaline phosphatase conjugate was performed following the manufacturer's instructions (Bio-Rad). Signals were quantified using a Visage Digital Imager. All signals that werequantiliedwereinthelinearresponaerangeofthe lmager as determined by scanning the signals produced by different amounts of gel-purified SpoIIID in a Western blot experiment utilising the anti-SpoIIID antibodies (data not shown). (a) fi-Galecloaidaas assays fi—Galactoaidass activity was determined using the substrate o-nitrophenol-fi-o-galsctoside (ONPG) as described previously (Killer. I972). One unit of Griyme hydrolyzes 1 pmol of substratelmin per 4”, of initial cell density. (flNortAenblclaaolyais AthourlyintervalsbetweenSand'lhafter-theonsetof tion. cells were harvested by csrrtrifrgation (ll,950gfor 10 min) and RNA waaprepared aadeaaibed ° & Losick. 1980) except the RNA was edinlOOplcfwaterthathadbeentreatedwlth 01% (v/v) ' ylpyrocarbonate. The RNA was treated with DNaasI to remove contaminating chromosomal DNA.thentheRNAwasfractionatedbyelectrophocssis on a 14% (w/v) agaroas gel containing [-11% (v/v) formaldehyde. transferred to nitrocellulose. and hybrid- bsd at “'0 to nick-translated p81!” (Kunkel d al., l989)asdascribsd(Aasubsldal.. l989).Tlraaignalswere visualisedbyaotcradicgraphy. 3. Results (a) The level qfspoIIID changes during sporulation lAshowathelevel ofSpoIIIDinwild-type cells harvested at hourly intervals during sporula- tion. We measured the level of SpoIIID in whole- cell extracts by using antibodies generated against SpoIIID in Western blot analysis. The anti- SpoIIID antibodies detected three polypeptides in extracts from ° . wild-type cells (Fig. 1A). The SpoIIID polypeptide comigrated with gel-puri- 75 842 R. Halberg and L. Kroos A Tr T: T: Tr Ts Ts T1 Ts Ts :-' 1g” ”A“? :1flu‘ifan‘ic‘3w‘bmm‘ndlui Tag: q... . < . . 7 Hfiéu’nu ‘Ix-‘WE _.l._'§.|3'-3’_ f .- army) mm. :I :r (114::an . ._'.: SpoIIID ; Mr? qr £21.an I .L ..rivn ,1: H4 .9 "fnrf‘ 12345678910 “In-j,“ 37,7; Jail! 21“”: “'8‘” 1'. 1', T, T. T; T. 1', 'r. 1 234567810 Figure l. Characterization of anti-SpoIIID antibodies by Western blot analyn. Wholeeell extracts (5 a) were prepared from csTIs collected at hourly intervals afler the utilising anti-SpoIIID antibodies. Gel-purified SpoIIID (l0 ng) served as a positive control (lane I0). Panel A show! the wild-type Spo' strain PY'N and panel B shows the spoIIID mutant strain BKMI (spolllDAem). which has an erythromyein n-resistance canette replacing codona I5 to 37 ofspoIIID (Kunkel d al., I989). tied SpoIIID (Fig. IA. lane l0) and was not de in extracts pre repared from sporulating. spoIIID mutant cells (Fig. IB). SpoIIID was first three hours alter the onset of sporulation (73). its level increased until 7". and its level sharply by 1“. This pattern is consistent with the pattern of fi-galactosidase expression observed from a spell lD-lacZ fusion during sporulation. except that the level of fl- -galactosidase tant afler 'l', (Kunkel et al. I989). The other two polypeptides detected by the anti- Spollll) antibodies were larger than SpoIIID and were present in extracts prepared from both wild- type and spoIIID mutant cells. Apparently. the antibodies cross-react with two polypeptides that are not precursor or modified forms of SpoIIID since these forms would be expected to be absent in the spoIIID mutant. These polypeptides are. T2 T; T4 T5 Ts T1 Ts era ' ... E100- 0 -l I580- 3 E :60“ s 340- G u 220- C 0 0 so n.12345678 Hours After Resuspenslon Figure 2. levels of SpoIIID. pro-cl s‘ and urb- fl BRHI cellscollectcdathourly intervalsaltsrtheonaetcl sporulation in SI! medium and were subjected to Western blotan analyses utilising anti-SpoIIID antibodies (upper inset) or anti- pro-cl antibodies (lower inset). The signals obtained were quantified using a Visage DigI tal Imager. fi-Gal actusidaseactivity reached a maximum level of300 Miller units. The shows the level of SpoIIID (C1). pro-0" (A). 0‘ (Al. and cotD-directed 5 activity (0). each ass percentage of the maxi- mum lsvel achieved during sporulation. however. sporulation-specific. since these poly- peptides were not detected in wing cells or in sporulating. sigE (encoding a sporulation- spccific sigma factor that functions earlier than ; reviewed by Moran. l989) mutant cells (data not shown). (b) The SpoIIID decrease coincides with Increases nthe level of a" and spore coal gene expression SpoIIID markedly represses transcription of the cotD promoter by 4:" RNA polymerase in vitro (roosstol1989).'l‘hcdccreaseinthclevclof SpoIIID after '1', seemed to ofl'er a possible explana- tion of how a repressive effect of SpoIIID on expression of 00:0 in rice might be removed. Expression of 0010 was shown previously to begin unldory Switch in Bacillus subtilis between r,.nd1‘.ofsporulation.asdetermined by measuringthelevelofcoleRNAuF' ' dase activity from . coco-rad fusion (Zhens & M, 1990). Also. the level of a‘ was shown previously to increase after 7". although sporulation was induced differently than in the experiment shown in Figure I (Lu ct al., l990). Figure 2 shows a direct comparison of the relationship between the SpoIIID decrease and the levels of cotD expression and a“. We collected samples of the wild-type strain carrying a cotD-lacz fusion at hourly intervals during sporulation and measured the levels of SpoIIID. pro-o". o“ and cothirected fi-galactcsi- dase activity. The levels of pro-6‘ and a" were monitored by Western blot analysis using anti-prov" antibodies (Lu ct al., l990). As the level ofSpolllD decreased fourfold betwwn 7', and T1. cothirected fi-galectosidase activity increased l2- fold. During the same interval the level of prov" decreased threefold and the level of 6‘ increased twofold. Thus. there is a reciprocal relationship between the levels of 6‘ and SpolllD in sporulating cells; as the level of 0" increases. the level of SpoIIID decreases. Both of these efl’ects may play a releinthesharpriseincotD-laclexpression between 1', and 1',. If SpolllD does repress cotD transcription in rice. as predicted from in vitro studies (Kroos cl al., l989). these results suggest that a mechanism(s) exists to decrease the level of Spell") at the appropriate time during sporulation. The results shown in Figure 2 are consistent with the idea that SpoIIID activates sigK transcription (Kunkel et al., l988; Kroos t! al., ”89). if Spollll) activates sigK transcription. expression of sigK (as A B d 8. - n A e A Percentage ot Ilsxlmum Level 1230801. flemAtterlesuspenslen 76 843 reflectedbythetotelamountofpro-e‘pluse‘) wouldnotbeexpectedtoincreaseafter‘l‘,.beeause thelevelofSpolllDdecreasessharply afterT.As shown in Figure 2.thetotal amount ofpro and it" remained aboutthesameafter T,.The level of pre-a‘declinedastlielevelofe‘increased.presum- ablyduetotlieprccessingofpre-e‘tee‘(budal.. 1990). (c) Mutants defective in 6‘ production are defective in the SpoIIID decrease The results shown in Figure 3 demonstrate that theproductionofaetivee‘playsaroleinthe Spollll) decrease. The level of SpoillD was moni- tored in sporulation mutants shown previously to be defective in the production of 6‘ (Lu et al., l9”). Mutants that fail to produce pro-6‘ and 0‘. due either to a mutation in sigK (spa! V08 and spoIIIG encode the N-terminal and C-terminal portions of 6". respectively; Stragier at al., l989) or a mutation in spa! VOA (encoding a recombinase that is e-en- tial for the chromosomal rearrangement which forms the composite sigK gene; Kunkel at al., ”90; Soto d 01.. l990) accumulated SpolllD throughout sporulation (Fig. 3A). Mutants that produce pro-e“, but do not produce a detectable amount of 0‘. exhibited a partial (Fig. 3B. spelllE. spoIIIG and spa! VB) or delayed (Fig. 3C. epolllA and spa! VF) decrease in the level of SpolllD during sporulation. These mutants may produce a low level of a“ that cannot be detected by Western blot analysis using anti-pro-o" antibodies (Lu at al., I990). but iiulfi- cienttoelicitapartialordelayeddecreaseinthe heureMterlIeauspenelon Figure 3. level of SpoIIID in sporulation mutants with defects in e‘ production. Whole-cell extracts (5 n) were prepared from cells collected at hourly intervals after the onset of sporulation in 8" medium and were subjected to Western blot analyses utilising anti-SpoIIID antibodies. For each mutant. the experiment was repeated at lead. once and in all cases changes in the level of SpolllD during sporulation were qualitatively reproducible. A representative blot for each mutant was chosen and the SpoIIID signal was quantified using a Visage Digital integer. Visible signals below the sen-'tivity of the imager were auigned a value of 5%. For each mutant. the SpoIIID level at each time point 5 plotted as a percentage of the maximum level achieved during sporulation. The maximum lech of SpoIIID achieved was dmilar in all mutants except for spell [094. in which Spollll) reached a 3-fold higher level. For comparison. the level of Spell") in the wild.type Spo’ strain PY79 (Cl) is shown in each panel. A. SpoIIID remained at a high level in drab BKrlo (spoIIIC94; I). BKsss (spol V0.4 13.3; A) and K88"! (ml Vcszz'r-orrfluzls; O) 8. SpoIIID dean-d partially in strains 8cm (mu/GA: an; A). 60022 (vellum: I) and BK750 (cm mm; 0). C. The 890"") do... was delayed in strain- KSI: (spoIIlA ::'l‘n917fle13; I). seem (spol VIE)”; Al and RBI“ (‘70! VA gwliflglu; 0). These mutant strains have been described previously (Sandman d el.. "87; Kunkel d el.. I”; Cutting a en. "00. lflle: Du et el.. IOOO). 77 844 R. Ham.“ L. Kroos level of SpolllD. Consistent with this idea is the observation that a ape! VA mutant. which produces a low level of c" that is barely detectable by Western blot analysis (Lu et al., “90). exhibited a eligth delayed decrease in the level of SpolllD (Fig. 3C) Also. the spelllA. epolllG. epol VB and F! VF mutants do retain the capacity to produce . because when pro-o" is overexpressed from a multicopy plasmid in these mutants. a low level of of (detectable by Western blot analysis) is produced and the e‘-transcribed cotD gene is expressed (S. Lu & L. Kroos. unpublished results). Alternatively or in addition. the pro-6‘ produced in these mutants may afi'ect the Spollll) level. Clearly. all mutants tested that were shown previously to be defective in 0‘ production were shown here to be defective in the Spollll) decrease. Taken together. these results strongly support the hypothesis that in wild-type cells the production ofactive 0" causes the Spollll) decrease. (d) The SpoIIID decrease occurs earlier in cells that produce e‘ earlier lftheS "10 decrease is caused by the appear- anoe of of? then earlier production of 6" should result in an earlier decrease in the level of SpoIIID. FiguretAshowsthatthelevelofSpolllDbeganto decreeseaboutonehourearlierthannormalina A T: T3 T4 T. T. T7 T. Sporrro— ’ "'"”-"' ......” fins}? 3 . g 1” J E so‘ 3 E E ”‘ 3 40‘ C D 8 20‘ c! r 2 s r s s 7 r Hours Atter Resuspenslon Figure 4. Levels ofSpolllD. a‘ and cotD-directed 3 mutant engineered to express 0‘ earlier than normal (compare Fig. 4A to Fig. 2). We monitored the level of SpoIIID in a mutant (epol VCBAI9; Cutting et at.. I990) expected to produce or" earlier during lation than the wild-type strain. The spa! VCBAI9 mutation deletes codons 2 to so of spa! VCB. which encodes the N terminus of pro-or" (Stragier et al., I989). After the chromosomal rearrangement generates the composite riglr' gene in this strain. the primary translation product is active 0‘ rather than pro-o". Samples from the spa! VCBAI9 mutant carrying a cotD-lacz fusion were collected every hour after the onset of sporula- tion and mayed for Spollll). 0" and cotD-directed p-galactosidase activity. 0“ was produced about one hour earlier than normal in the spa! VCBAI9 mutant; 0" was first detected at 1', and reached a maximum level at 1" (Fig. 4A). as was observed for pro-o" in the wild-type strain (Fig. 2). The Spollll) level decreased about one hour earlier in the spa! VCBAIQ mutant than in wild-type cells. As documented in Figure 4A. and as was observed in three additional experiments (data not shown). the level of Spollll) reproducibly decreased about twofold between 1" and 1', in the spa! VCBAI9 mutant. ln wild-type cells. the level of SpoIIID reproducibly decreased from a high level at 1', to ls.thanhalfthemaximallevelat1'..asdocu- mentedin FigureLandaswasoMrvedinthree Ts Ts Ts Ts Te 1'1 Te " Percentage of Maxlmurn Level 8. r 2 s 3'; i 717 HoursAtterllesuspenalon ' activity in mutantsthatprodueee‘earlbrthan -galactnmdase normal. The B. subtilis strains V048 (spol VCBAIQ; Cutting st el.. l9”) and BRH3 (spol VCBAI9. germ: see Materials and Methods) were lysogenised with phage BPfizmotD-tacz to construct strains BR“! and Bill“. respectively. Whole—cell extracts (5 pg) were prepared from cells collected at hourly intervals alter the onset ofsporulation in 8” medium and were subjected to Western blot analyses utilising anti-SpoIIID antibodies (upper inset) or anti-pro-e‘ antibodies (lower inset). The signals obtained were quantified using a Visage Digital Imager. Visible signals below the unsitivity of the imager were aligned a value of 5%. fl-Galactceidase activity reached maximum levels of 550 and 320 Miller units in strains BRH2 (A) and BRHt (B). respectively. The graphs show the levels ofSpolllD (U). 6‘ (A) and cotD-directed fl-galaetosidase activity (0). each plotted as a percentage of the maximum level adrieved during sporulation. 78 Regulatory Switch in Bacillus subtilis 845 additional expriments (data not shown . In addi- tion. the earlier production of in the spoll’CBAI9 mutant appears to limit SpolllD production. since the maximum level of SpoIIID was lower to the mutant (Fig. 4A) than in wild-type cells (Fig. 2). These results support the idea that the production of active 6‘ causes the SpoIIID decrease. Figure 4A also shows that the time ofcotD-lacZ induction was about one hour earlier in the spa! VCBAI9 mutant than in wild-type cells (comm Fig.4Ato Fig.2) In bothcases. an increase in cotD-lad expression coincided with a decrease in the level of SpoIIID However. cotD« directed fi-galactosidase activity lagged consider- ably behind o“ production in the spoIVCliAI9 mutant (Fig. 4A). whereas only a slight lag was observed in wild-type cells (Fig. 2). SpoIIID-mediated repression of cotD transcription might explain these observations. but we also con- sidered the possibility that the lag represents time required for o‘-directed synthesis of GerE. a DNA binding protein that in wild- -type cells greatly stimulates cotD expression (Cutting & Mandelstam, I986; Cutting ct al.. I989; Zheng & Losick. l000; Zheng et al., 1092). Frgure‘ 4B shows that GerEynthesis does not account for the lag between production and cotD-lacZ expression in the spa! VCBA19 mutant. We constructed a spa! VCBAI9. 90536 double mutant and monitored the levels of SpolllD. o" and cotD-directed fl- -galactosidase activity during sporulation. The gerE mutation did not all'ect the time of cotD-lacz induction. which still lagged con- siderably behind 0" production (compare Fig. 48 to Fig. 4A). However. the major increase in cotlHacZ expression did coincide with the SpoIIID decrease in the spa! VCBAI9. gerEJG double mutant (Fig. 4B). which is comistent with the idea that the SpoIIID decrease derepresses cotD transcription in ace. (e) Pralrrctiorr of a" during sporulation leads to a decrease in the level of spoIIID mRNA The decrease in the level of SpoIIID during the later stages of sporulation could. at least in part. be due to a decrease in the level of spoIIID mRNA available for translation. Figure 5 shows that the level of spa! I ID mRNA decreased sharply between five and six hours after the onset of sporulation in wild-type cells. but not in sigK mutant cells. We measured the level of spoIIID mRNA in sporu- lating wild-type and sick mutant cells using Northern blot analysis. The probe rn the experiment was pBK39. which contains a l- l kbl’ fragment of B. subtilis DNA centered about the spoIIID coding sequence (Kunkel et al., l989). Similar results were obtained when a 306 bp ApaLI—Xrnrrl DNA frag- ? Abbreviations used: lrb. l0’ base-pairs; bp. base-pair“). wlle-type sigK spemo r—-———r r———r H r, r. r, 1,?! .t' r. r, r. 1,01, r. l 2 3 O I. 7 0010"“ Figure 5. level of spoIIID mRNA in sporulating wild- .type and sigK mutant cells. RNA (20 ug) was prepared from cellscollected atthe indicated timesaflertheonset of sporulation in 8H medium and was analysed by Northern blot analysis. B. subtilis strain PY79 was the wild-type 8po‘ strain (lanes l to 5). BK4|0 (sputum; Kunkel ct al.. lm) was the sigK mutant (lanes ii to l0) and 8K5“ (spoIllDAsr-m; Kunkel d al.. “89) was the spoIIID mutant (lanes ll and I2). The poeitionsof RNA standards (Old to l'T'l lrb RNA ladder. BRL) are indicated. ment extending from the first codon of spoIIID to 25 bp beyond the translational stop codon of spoIIID served as the probe. An RNA of approxi- mately I340 bases was detected in RNA isolated from sporulating. wild-type cells (lanes l to 5) and this RNA was not detected in RNA isolate from a spoIIID mutant bearing an insertion of a drug resistance gene near the 5' end of spoIIID (lanes II and I2). Therefore. we believe the I340 base RNWis . the mRNA transcript derived from spell ID. Based on the transcript sire and the position of the tran- scriptional start site (Tatti et al., l99l). the tran- script is predicted to extend about 830 bases beyond the end of the spa! l l D coding sequence. suggesting the possibility of an additional downstream gene(s) at the spoIIID locus. In sporulating. wild-type cells the level of spoIIID mRNA decreased sharply btween T, and 7" (Fig. 5. lanes 3 and 4). Thus. the level ofspolllD mRNA is markedly reduced late in sporulation and this could. at least in part. account for the decrease in the level of SpoIIID late in sporulation (Fig. 2). Figure 5 also shows that in sporulating sigK mutant cells. the level of spoIIID mRNA remained high throughout sporulation (lanes 6 to IO). The level of SpoIIID protein also remained high throughout sporulation in sigK mutant cells (Fi 3).These results suggest that in wild-type cells a! negatively regulates the synthesis and/or stability of spoIIID mRNAandthatthisleadstoadecrease in thelevel ofSpolllD late in sporulation. 4. Discussion We have demonstrated that the level of SpolllD in sporulating B.srrbtilir decreases at the appro- priatctimetopreduceaswitchinthepatternof mother-cell gene expression from transcription of sigK at morphological stage IV to transcription of 846 cotD at stage V. The following results strongly suggest that the decrease in the level of SpoIIID. and hence the switch. is controlled by the produc- tion of e‘: (l) in wild-type cells the SpoIIID decrease coincides with an increase in the level of c‘ (Fig. 2); (2) mutants defective in 6‘ production are defective in the SpoIIID decrease (Fig. 3); (3) in cells engineered to express 6“ earlier than normal. the SpoIIID decrease occurs earlier than normal and the maximum level of SpoIIID produced is lower than normal (Fig. 4); and (4) the level of spoIIID mRNA in sigK mutant cells remains high late in sporulation. whereas it decreases in wild-type cells (Fig. 5). Thus. production of active 6‘ appears to initiate a feedback loop that leads to a decrease in the level of SpoIIID. How does the negative feedback loop connecting a" production to the SpoIIID decrease operate! The production’of e‘ normally leads to a decrease in the level of l I I D mRNA late in sporulation (Fig. 5). Thus. aflppears to influence the rate of spoIIID transcription and/or the stability of the transcript. The observation that a spa! l I Dth fusion is over- expressed in mutants (spa/HA. spoIIIC, rmIIIG. spol VGA and spa! VCB; Kunkel d IL. l989) that fail to make a detectable amount of e“ (Lu et al., l990) suggests that the appearance of e“ normally decreases the rate of spoIIID transcription. Several lines of evidence indicate that rpol I ID is transcribed by 0" RNA polymerase (Kunkel er al., l989; Stevens a Errington. l990; Tatti er al., l99l). 0' is a sporulation-specific sigma factor that func- tions earlier than c‘ (for a review. see Moran. l989). 0‘ appears to be a highly labile protein that is stabilised by binding to core RNA polymerase (Jonas er al.. l990). Perhaps it" competes with e‘ for binding to core RNA polymerase. thereby destab- lising o" and decreasing transcription of spoIIID. Alternatively or in addition. 0‘ RNA polymerase may direct transcription of a gene whose product represses spoIIID transcription directly or influences spolll D transcription indirectly by aliecting a! production. The decrease in the level of spoIIID mRNA between 7', and 1" (Fig. 5) is paral- leled by a decrease in the level of SpoIIID protein (Fig. 2). suggesting that during this period of sporulation the mother cell is capable of rapidly degrading SpoIIID. . Figure 6 illustrates our current model for the switch from sigK to cotD transcription in the mother cell during the transition from morpho- logical stage IV (cortex formation) of sporulation to stage V (coat formation). We propose that the 6‘ RNA polymerase produced initially during stage IV is stimulated by SpoIIID to transcribe sigK and is prevented by SpoIIID from transcribing cotD (Fig. 6A). Also. during this period gerE would be transcribed by a“ RNA polymerase (Cutting el al., l989; Zheng et al., l992). As the level ofc" rises. it would cause a decrease in the level of SpoIIID end an increase in the level of 06f”. eventually "immw-cen gene expression to the stage V pattern ( g. 68). Because SpoIIID and GerE exert 79 R. Halberg and L. Kroos A "a I I u v 0" 0‘03 i some“ ".46 s..- ahead—em sec-rs e'—e.us-ee~r see. so. Flgure‘. Model fortheswitch from sigh'torutl) traruxaription in the mother cell during the stage IV to stage V transition of sporulation. A. During stage IV (rrrrtex formation). Spollll) stimulates sigK transcription b 6‘ RNA polymerase and reprearss coll) transcription. RNA polymerase also transcribes gerb‘. leading to synthesis of GerE. B. The accumulation of e‘. resulting fromtheproosmingofpro-e‘toe‘.causssadecreasein the level ofSpoIIlD. producing a switch to the stage \' (coat formation) pattern of gene expel-ion (Le. sigK transcription is no longer stimulated and call) transcrip- 'tion by e‘ RNA polymerase is no longer repressed). Continued production of GerE reinforcm the switch since Geri-I represses sigK transcription and stimulates coll) transcription (Zheng et al., "92). opposite efi'ects on e‘directed transcription of sigK and cotD (Kroos er al.. l989; Zheng at al., "92). a declining level of SpoIIID and a rising level ofGerE would produce a reinforced switch from sigK to cotD transcription (Fig. 6). The model proposes that IV of sporulation is a period during which the RNA polymerase produced initially is stimulated by SpoIIID to tran- scribe sigK and is prevented by SpoIIID from transcribing cotD. This period could extend from between 1', and T. to between 1" and 1', in wild- type cells. since rr'I is first detected at 1'. and cotD- directed pogalactosidase activity is first detected at 1', (Fig. 2). SpoIIID reaches its maximum level during this period (Fig. 2). so it is available to affect transcription. Evidence that Spol II D stimulates sigK transcrip- tionbye'" RNA polymerasecomesfrombothirrsiuo and in vitro studies. Expression of a sigK-incl fusion is reduced in sigK mutant cells. demon- strating that rr‘ does autoregulate its own expression. and sigK-lecz expremion is undetec- table in spoIIID mutant cells. demonstrating that SpoIIID is essential for sigK transcription in viva (Kunkel et el.. 1988). In vitro. SpoIIID greatly stimulates sigK transcription by 0‘ RNA poly- merase (Kroos er al., l989). The suggestion that SpoIIID prevents cotD tran» scription by 0‘ RNA polymerase for a short period during stage IV of sporulation is based primarily on the finding that SpoIIID can completely repress transcription of cotD by a‘ RNA polymerase in vitro (Kroos er al., l989). Consistent with this idea is the finding that an increasing level of cotD-directed Waco activity coincided’ with a decreasrng' level of SpoIIID in both wild-type cells (Fig. 2) and cells engineered to express it" earlier than normal (Fig. 4A). However. as illustrated in Figure 68. 0010 expression is stimulated by GerE (Zheng t 80 Regulatory Switch in Bscillus subtilis 847 Losick. I990; Zheng er al., I992). To eliminete the efleet of GerE on cotD-Incl expression. e per]? mutsnt wes employed (Fig. 43). In this strein. en increesing level of cotD-directed fl.gelsctosidsse ectivity still coincided with e decreesing level of SpoIIID. even though the level of 6" was decreesing during the seme period. This finding is consistent with the idee thet SpoIIID represses roll) trenscrip- tion in ru'uo. In vitro trenscription studies elso show thet SpoIIID cen repress trenscription of co“! (encoding s spore cost protein; Donovsn er al., I987) by 0‘ RNA polymer-see plus GerE (R. Helherg. H. Ichikews & L. Kroos. unpublished results). Thus. SpoIIID mey repress s set of spore cost genes for e short period during stege IV of sporulstion. SpoIIID does not inhibit trenscription of the gerb‘ promoter by a" RNA polymerese is eilro (L. Kroos & R. Losick. unpublished results). This result thet perk: expression would not be repressed by b‘polll!) during stege IV (Fig. 6A). We find thet gerlr‘ is expressed ebout one hour eerlier during sporulstion thet coil). es determined by meesuring fi-geleetnsidese ectivit_v from lacZ fusions or by Northern blot enelysis of the mRNAs (our unpublished results). One gene encoding e spore cost protein. ecu. eppeers to be expressed with similer timing es perE. es determined by rneesuring fi-gelectosidese ectivity from s locZ fusion (our unpublished results). Although S ioIIID cen inhibit eoUl trenscription in vitro by RNA yrnersse (I.. Kroos & R. Losick. unpublished results). this efi'ect requires e higher molsr retio of SpoIIID to DNA then do the effects of Spoil") on sigK. cotD end cotC trenscription (our unpublished results). One or more genes involved in cortex forrnetion mey elso be trenscribed by a" RNA polymerese end not subject to repression by SpoIIID. since siglr' mutents ere defective in cortex formetion (Cutting er al., I99lo). Our results suggest thet es the level of c‘ rises during stege IV. it ceuses s decreese in the level of SpoIIID. As the emount of SpoIIID eveileble to stimulete trenscription of sick (end perheps other stege IV genes) by 6‘ RNA polymerese end to repress trenscription of cotD (end perheps other stege V genes) diminishes. the pettern of mother-cell gene expression would switch to the stege V pattern (Fig. 68). The results of Zheng el al. (I992) suggest thet en increese in the level of GerE during stege IV would reinforce this switch in the pettern of mother- ”ll gene expression. been.” the efl'ects 0f GerE on sigK end coil) trenscription in vitro by 6" RNA polymerese ere just the opposite of the effects exerted by SpoIIID (Fig. 6). GerE csn completely inhibit sigK trenscription by 6" RNA polymerese in vitro (Zheng et al., I992). coM is elso repmsed by GerE (Sendmen et al., I988; Cutting cl al., I989; Zheng er al., I992). On the other hend. both in viva end in vitro studies support the idee thet GerE not only stimuletes trenscription of cotD by 0" RNA polymerese (es shown in Fig. 68). but it slso stimu- letes e‘directed trenscription of rolB end 6010 during stege V (Zheng & Losick. I990: Zheng et al., I992). Thus. both SpoIIID end GerE een positively or negetively efl’ect d"-directed trenscription of severe! different mother-cell-specilic genes end the levels of SpoIIID. 6" end GerE eppeer to be controlled by the reguletory intersetions illustreted in Figure 6 so es to produce e molecqu switch governing the stege IV to stege V trensition of sporuletion. An interesting question is whether this switch is coupled to morphogenesis. A priori. it seems likely thet mechenisms must exist to co-ondinete sporule- tion gene expression with rnorphogenic progress during development. Two such coupling points heve been proposed v‘previously. The proteolytic processing of pro to 0‘ mey couple formetion ol' the sporuletion septum to the production of 0‘ end the ensuing new pettern of gene expression (Leliell er al., I987: Strsgier et al., I988). Similerly. severe! lines of evidence support the idee thet proteolytic proccuing of pro-o" to a" is e reguletory device thet couples forespore morphogenesis to the production ofo‘ in the mother cell (Cutting e1 al., I990. I99Ib; Lu d al., I990). Is the proposed switch governing the stege IV to stege V trensition somehow coupled to the completion of cortex formetion. or do the retes of synthesis end degrsdstion of SpoIIID. 0‘ end GerE constitute e developments! timer thet determines the length of stege IV? These poss- ibilities ere not mutuelly exclusive. For exemple. the rete(s) of synthesis end/or degrsdetion of SpoIIID. rt" end/or GerE my be responsive to e signs! genereted upon completion of the spore cortex. This would be enelogous to the mechenism controlling the decision of becteriophsge 1 to lyse-or lysogenise its E. coli host. where reguletory inter- ections between severe! reguletory proteins produce e molecqu switch (Pteshne. I987) end degredetion of the all protein is effected by virel. host end environments! fsetors (Herslrowits & Regen. I980; Hoyt er al., I982). WethenlrLZhengendR.hrsiekforproviding beeterielstreinsendphsges.Wethenk8.Ierfor providing snti-pro-e‘ entibodies. We thenlr R. M. A. Omen end A. Bosme for helpful edvice on the menu- script. This reseereh wss supported by the Hichigen Agriculture! Experiment Stetion end by grent (”“3585 from the Netionel Institutes of Heelth. References Ausubel. R. Brent. R.. Kingston. IL. Moore. 0.. Seidmen. J.. Smith. J. & Struhl. K. (I989). In (fume! Protocols in Molecular Biology. John Wiley t Sons. New York. Bredl'ord. M. (I976). A repid end sensitive method for the quentitetion of microgrem quentitins of protein utilizing the principle of protein-dye binding. Anal. Bicelrers. 72. 248-254. Cutting. S. & Mendelstem. J. (INC). The nucleotide sequence end the trenscription during sporuletion of thepengeneofBoeillussuttilis.J.0¢e. Hicrobiel. I32. 30l3-m24. Cutting. 8.. Psnser. 8. & Losick. R. (I989). Reguletory 848 mudiesontbepromoterforsgenegoverningsynthe. n'sandamemblyoftbesporecoatin Bacillussublilis. J. Md. Bid. 207. 393—404. Cutting. 8.. Oke. V.. Driks. A.. Losick. R.. Lu. 8. a Kroos. L. (I990). A forespore checkpoint for mother-cell gene expremion during development In Bacillus subli- lis. Cell. 62. 239-250. Cutting.8 .Driks. A. Schmidt. R.. Kunkel. 8. & Losick. R. (I99Ia). Pompom-specific transcription ofa gene in the signal transduction pathway that governs S. 456-468. Cutting.8.. Roels. 8. tLoaick. R. (I99Ib). Sporulation operon spol VP and the dreraeteriution of mutations that uncouple mother-cell from forespore gensexpresaioninBacillussubtilis. J. Md. Biol. 22!. I237 I256. Donovan. W.. Zheng. L.. Sandman. K. a Losick. R. (res-n. Genes encodirg spore coat polypeptides from Bacillus subtilis. J. Md. Biol. I96. I-IO. Driks.A.&Iosick...R(I99I)(‘r‘ "" expressionofageneunderthecontrolofsporulation transcription factoro'I rn Bacillussublilis. Proc. Nat. Acad. Sci.. U.S..A 88. 9934-9938. Dubnau. D. & Davidofl-Abelson. R. (I97I). Fate of trans- forrning DNA following uptake by competent B. subtilis. J. Md. Bid. 56. mil-22!. Errington. J. 8. Mandelstam, J. (I986). Use ofslacl gene fusion to determine the dependence pattern of sporulation operon spoIIA in spa mutants of Bacillus subtilis. J. Gen. Microbid. I32. 2987-2976. Harlow. 8. 8 Lane. D. (I988). In Antibodies. Cold Spring Harbor Press. Cold Spring Harbor. NY. Huly. J. Weir. J.. Smith. I. I. Losick. R. (I99I). Post-transcriptional control of a sporulation regula- torygeneencoding transcriptionfaetorolin Bacillus subtilis. Md. Microbid. 5. 477-487. Her-downs. I. & Hagen. D. (Im). The Iysia~lysogeny decision of phage l: explicit programming and responsiveness. Annu. Rev. Genet. I4. 399—445. Hoyt. H. A.. Knight. D. It. Des. A.. Miller. H. I. a We. H. (Im). Control ofphage 1. development- by uability and synthesis of cII protein: role of the viral an and host AM. AimA and AirsD genes. Cell. 31. 866-673. Igo. H. M. 8. Losick. R. (IM). Regulation ofa promoter region that is utilesd by minor forms ofRNA poly- merase holoerrsyrns hr Bacillus subtilis. J. Md. Bid. I91. 6I5-824. Jonas. R. It. Peters. H. K.. III 8 Haldenwang, W. 0. (I990). Phenotypes of Bacillus subtilis mutants altered in the precursor-specific region of 0". J. Bacterial. I72. 4I78-4I86. Kroos. L.. Kunkel. 8. I. Losick. R. (IU9). Switch protein alhrs specificity of RNA polymerase containing a compartment—specific sigma factor. Science. 243. 526-529. Kunkel. 8.. Sandman. K.. Panssr.8.. Youngman. P. 8. Kodak. B. (I988). Thepromoterforasporulation geneinthespoli’ClocusofBacillussntlilisand its ussinstudiesoftem andspatielcontrolofgene expremion. J. Bacterial. I70. 35I3—3522. Kunkel. 8.. Kroos. I... Poth. H.. Youngman. P. 8. Losick. R. (I989). Temporal and spatial control of the mother-cell regulatory gene spoIIID of Bacillus subtilis. GencsDevelop. 3. I735-I744. Kunkel. 8.. Losick. R. tStragier. P. (I990). TheBacillus subtilis for the developmental transcription factor isgerreratd by excision of a dispensable 81 R. Halberg and L. Kroos DNA elernent containing a sporulation recombinase genchncs Develop. 4. W535 LsBell. T. L. Trempy. J. E a Haldenwang, W. G. (I987). Sporulation-specific sigma factor. a” . of Bacillus subtilis rs synthesized from a precursor pro- tein. P". Proc. Nat. Acad. Sci" USA. 84. I784- I788. Lu. 8.. Halberg. R. a Kroos. L. (I990). Processing ofthe mother-cell a factor. a‘. may depend on events occurring in the forespore during Bacillus subtilis development. Proc. Nat. Acad. Sci.. U.S.A. 87. 9722- 9726. Hatsudaira. P. (I987). Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J. Biol. Chem. 262. “”35- I0038. Hiller. J. H. (I972). Experiments in Mdecular Genetics. Cold Spring Harbor Laboratory Prom. Cold Spring Harbor. NY. Horan. C. P.. Jr(I989). Sigmefsctorsandtheregulstion of transcription. In Regulation of “energetic (Smith. I. Slepccky. R. A. & Setlow. P.. eds) PP l67- I84. AmericanSocietyol’Microbiology. Washington. DC. Ptashne. ll. (I987). In A Gendic Stretch. Cell Prom. Cambridge. IIA. Sandman. K.. Losick. R. tYoungrnan. P. (I987). Genetic analysis of Bacillus subtilis spa mutations gensreted by Tn917 mediated insertions! mutagenesis. Gendics II7. 603-6I7. Sandman. K.. Kroos. L.. Cutting. 8.. Youngman. P 8. Losick. R. (I988). Identification ofthe promoter fora sporecoatproteingenein Bacillussulrlilisand studiesontheregulationofitsinductionatalete stage ofsporulstion. J. Md. Bid. 200. 46I—472. Sate. T.. Samori. Y. a Kobayashi. Y. (I990). The cisA cistron of Bacillus subtilis sporulation gene WC encodes a protein homologous to a site-specific recombinase. J. Boater-id. I72. I092-I098. Smrth I Slepecky. R. A. lSetlow. P. (I989). liq-lotion o] Procarydic Development. American Society for Microbiology. Washington. DC Sterlini. J. It. a Mandelstam, J. (I969). Commitmsntto sporulation in Bacillus subtilis and its relationship to development of actinomyein resistance. Biochem. J. II3. 29-37. Stevens. C. H. a Errington. J. (I990). Difl’erentia! gene expression during sporulation in Bacillus subtilis: structure and regulation of the spoIIID gene. Md. Microbid. 4. 843-552. Stragier. P.. Bonarny. C. a Karrnasyn-Campelli. C. (I988). Processing of a sporulation sigma factor in Bacillus subtilis: how morphological structure could control gene expression. Cdl. 52. 697-704. Stragier. P.. Kunkel. 8.. Kroos. L. 8. Losick. R. H“). Chromosome! rearrangement generating a composite gene for a developmental transcription factor. Science. 243. 607-5I2. Tatti. K. K.. Jones. C. H. a C. P. Moran. J. "99”. Genetic evidence for interaction of o‘ with the spoIIID promoter in Bacillus subtilis. J. Baalcrid. I73. 7828-7833. Thomas. J. D. a Kornberg. R. D. (I978). The study of histone-histane associations by chemical cross. linking. Methods Cell Bid. I8. 429-440. Youngman. P.. Perkins. J. 8. & Losick. R. (I984). (bnstructionofscloningsitenearoneendofthe Tn917 into which foreign DNA may be inserted without aflecting transposition in Bacillus subtilis or 82 Regulatory Sroildr in Bacillus subtilis 849 expression of the transposonoborne errn gene. Zlnng. I... Halberg. R.. Reels. 8.. Ichikawa. H.. Kroos. L. Plasmid. I1. I-9. & Losick. R. (I992). Sporulation regulatory protein Zheng.L.&Lom'ck.R.(I990).CsscadereguIationofapore GerEl'rom Bacillussuuilisbindetoandcanactivste coat genesxpremion in Bacillus subtilis. J. Mal. Bid. or repress transcription from promoters for mother- 2I2. 646-660. cell-specific genes. J. Md. Bid. 226. I037—I050. Edited by M. Gdtesrnan Note added in proof. The level ofrpolllD mRNA remains high late in sporulation (i.e. between 1', and 7' ) in spa! VOA mutant cells (strain 8K6“ (spol VOA l”); Kunkel d al.. I989). which fail to produce . supporting the idea that e‘ production negatively regulates the synthesis and/or stability of spell I D mRh' A. CHAPTER IV The SpoIIID Switch Protein Is Converted to a 9 kDa Form During Bacillus subtilis Sporulation "There is nothing permanent except change." Heraclitus 83 84 Abstract Mother-cell-specific gene expression during sporulation of Bacillus subtilis is, in part, controlled by (3'K RNA polymerase. UK is required for the formation of the spore cortex, stage IV, and the formation of the spore coat, stage V. It has been demonstrated that SpoIIID stimulates the transcription of a stage IV gene, but represses the transcription of several stage V genes by 0* RNA polymerase in vitro. Based on these observations, it was proposed that the inactivation of SpoIIID establishes a switch from the stage IV to stage V pattern of gene expression in the mother cell. Consistent with this idea, the level of SpoIIID decreases sharply at the appropriate time during sporulation. Here we provide evidence that this decrease involves the conversion of SpoIIID to a less stable 9 kDa form. We demonstrate that this conversion is developmentally regulated. The conversion of SpoIIID to the 9 kDa protein may be important to relieve the repressive effect that SpoIIID has on the expression of several stage V genes. 85 Introduction Under conditions of nutrient deprivation the gram- positive bacterium Bacillus subtilis undergoes a series of morphological changes that culminate in the formation of a mature spore (reviewed by (Errington, 1993)). The first easily observed morphological change is an asymmetrically positioned septum, which divides the bacterium into two compartments, the mother cell and the forespore. Each of these compartments receives a copy of the genome, but they realize alternative developmental fates because gene expression is regulated temporally and spatially. Spatial regulation is established by the compartment-specific activation of sigma subunits of RNA polymerase. Two mother- cell-specific sigma factors are 03(Driks & Losick, 1991) and OK (Kroos et al., 1989; Stragier et al., 1989) . GE is required for the migration of the septum and engulfment of the forespore in a double membrane (stage III). 0K is required for the deposition of cell-wall—like material called cortex between the membranes of the forespore (stage IV) (Cuttingefl: al., 1991a) and the synthesis of spore coat proteins that assemble on the surface of the forespore (stage V) (Kroosen: al., 1989; Zheng & Losick, 1990). OK is initially synthesized as an inactive precursor, 86 pro-OK, which has an additional 20 amino acids at its N- terminus (Kroos et al., 1989; Stragier et al., 1989; Lu et al., 1990). Proteolytic processing of pro-(5'K to UK is dependent upon forespore—specific gene expression (Cuttingwet al., 1990; Inlet al., 1990). Thus, the processing event appears to couple gene expression in both compartments. OK-dependent transcription is affected by SpoIIID, a 10.8 kDa DNA-binding protein [Kunkel et al., 1989; Stevens & Errington, 1990; Chapter II]. For example, SpoIIID stimulates sigK (encoding pro-OK) transcription, but represses cotC, cotD, and cotX (encoding spore coat proteins) transcription by GK RNA polymerase in vitro (H. Ichikawa, R. Halberg, L. Kroos, unpublished data; H. Ichikawa and L. Kroos, unpublished data; Kroosemzal., 1989). Previously, it was proposed that the inactivation and/or sequestering of SpoIIID establishes a switch in the mother-cell pattern of gene expression from the transcription of sigK at stage IV to the transcription of spore coat genes at stage V (Kroos & Losick, 1989). Consistent with this idea, the level of SpoIIID decreases at the appropriate time during sporulation (Halberg & Kroos, 1992). This decrease is dependent upon the processing of pro-(SK to 0K (Halberg & Kroos, 1992) . The accumulation of (5'K reduces the level of spoIIID mRNA with 87 similar timing during sporulation as the reduction in the level of the SpoIIID protein (Halberg & Kroos, 1992). The parallel decrease in the levels of spoIIID mRNA and SpoIIID suggests that a mechanism exists for degrading SpoIIID in sporulating cells. Here we provide evidence that SpoIIID is converted to a less stable 9 kDa form by removing 7 amino acids from its C-terminus. We demonstrate that this conversion is developmentally regulated. The conversion of SpoIIID to the 9 kDa protein may be necessary to relieve the repressive effect that SpoIIID has on the expression of spore coat genes during sporulation. 88 Matsrials and Methods B I . 1 SI . B. subtilis strains were provided by R. Losick at Harvard University. All strains were isogenic with PY79 (Youngman et al., 1984) except for 91 (spoVB91; Piggot and Coote, 1976), SCSO (spoVC134; Coote, 1972), and SC53 (spoVF224; Coote, 1972). Wan Sporulation was induced by resuspending growing cells in SM medium as described previously (Sterlini & Mandelstam, 1969). The onset of sporulation (To) is defined as the time of resuspension. S | | . E J . SpoIIID and the 9 kDa protein were partially purified from sporulating B. subtilis following the procedure for partially purifying SpoIIID and (3'K RNA polymerase described previously (Kroosem al., 1989). Proteins in DNA-cellulose column fractions were precipitated in 10% trichloroacetic acid, resuspended and boiled 5 minutes in sample buffer (0.125 M Tris-HCl, pH 6.8, 2% (w/v) SDS, 5% (v/v) 2-mercaptoethanol, 10% (v/v) glycerol, 0.1% (w/v) bromophenol blue), and subjected to SDS-PAGE (18% polyacrylamide; (Thomas 89 & Kornberg, 1978)). SpoIIID and the 9 kDa protein (approximately 10 pg each) were excised from the gel, eluted following the procedure described previously (Hager & Burgess, 1980) except the elution buffer did not contain BSA, acetone precipitated, and dissolved in 1 pl of 6 M guanidine-HCl and 19 pl of 0.1% trifluoroacetic acid (TFA). The samples (2 pl) were applied to membranes (Zetabind; 0.45 pm pore size and 50 pm thickness) attached to a stainless steel probe tip. The sample was allowed to air dry and then the membrane was washed by immersion in deionized water for 15-20 seconds at room temperature. The SpoIIID digest with endoproteinase Asp—N was performed as follows. SpoIIID was applied to membrane attached to a probe tip as described above. The probe tip was placed in a glass vial containing a small amount of 100 mM ammonium bicarbonate, pH 8.0. Endoproteinase Asp-N (2 pl of 0.04 pg/pl) was applied to the membrane and the glass vial was sealed. The digest was incubated 14 hours at 25°C. To analyze the samples, a saturated solution of a-cyano-4-hydroxycinnamic acid in 1:1 acetonitrile/O.1% TFA (2 pl) was applied to the membrane and air dried. Mass spectra were obtained utilizing a VT-200 linear time of flight mass spectrometer equipped with a nitrogen laser (337 90 nm, 3 ns pulse) as described previously (Zaluzecemzal., 1994). EH I E l . I' DNA probes labelled at only one end were prepared as follows. For the analysis of sigK, 20 pg of pBK16 (Kroosefl: al., 1989) was digested with XbaI and labelled by treating with alkaline phosphatase followed by phage T4 polynucleotide kinase and (YéfiP) ATP. The labelled DNA was digested with PstI and the 352 bp XbaI-PstI fragment was purified after electrophoresis in a non-denaturing polyacrylamide gel using the crush and soak method described previously (Sambrookefl: al., 1989), except 20 pg of double-stranded poly (dI-dC) (Pharmacia) was added to serve both as a carrier during ethanol precipitation and as a competitor during footprinting experiments. Similar end-labelling and recovery methods were used to prepare a DNA fragment for the analysis of cotD. A 324 bp EcoRI-Tan fragment from pLRKlOO (Kroos et al., 1989) was labelled at the EcoRI site. DNase I footprinting experiments were performed according to method 2 described (Zheng et al., 1992), except 0.5 pmole of end-labelled DNA was used. SpoIIID and the 9 kDa protein were gel-purified following the procedure described for SpoIIID previously (Kroos et al., 1989) . 91 I 'l I . I' (fiiRNA polymerase was a gift from Kathleen Tatti and Charlie Moran at Emory University. 0'K RNA polymerase was partially purified from gerE mutant cells (SC104; gerE36; Kroosem al., 1989), following the procedure described previously (Kroos et al., 1989) . The GK RNA polymerase was comparable in protein composition and in cotD— and sigK- transcribing activities to fraction 24 shown in Figure 2 of (Kroos et al., 1989) . Transcription reactions (45 pl) were performed as described previously (Carter & Moran, 1986), except that RNA polymerase was allowed to bind to the DNA template for 10 minutes at 37‘C before the addition of nucleotides. The labelled nucleotide was (a-32P) CTP. Heparin (6 pg) was added 2 minutes after the addition of nucleotides to prevent reinitiation. After the reactions were stopped, 20 pl of each reaction mixture was subjected to electrophoresis and transcripts were detected by autoradiography. M lil'l SI'EI E The DNA probe was either a synthetic oligonucleotide or a 106 bp Eco47III-XbaI DNA fragment prepared from pBK16 (Kroos et al., 1989), containing the SpoIIID binding sites in the sigK open reading frame (Chapter II). The former was prepared as follows. Two complementary oligonucleotides ‘92 (AGATACTAAAAAGACAAGCTCTTT and GTTAAAGAGCTTGTCTTTTTAGTA) were synthesized (MSU Macromolecular Structure Facility) and combined in annealing buffer (67 mM Tris-HCl, pH 7.6, 13 mM MgC12, 6.7 mM DTT, 1.3 mM spermidine, and 1.3 mM EDTA). The mixture was incubated at 88°C for 2 minutes, 65°C for 10 minutes, 37°C for 10 minutes, and room temperature for 5 minutes. The double-stranded oligonucleotide was labelled by treating with T4 polynucleotide kinase and (YénP) ATP. The 106 bp Eco47III-XbaI fragment was prepared as follows. pBK16 (Kroosem:al., 1989) was digested with HindIII and XbaI. The DNA fragments were labelled by treating with alkaline phosphatase followed by T4 polynucleotide kinase and (yéfiP) ATP. The labelled DNA was digested with Eco47III. B. subtilis cells (1 ml) were harvested at hourly intervals after the onset of sporulation by centrifugation (14,0009 for 5 minutes at room temperature). The supernatant was removed. Cell pellets were quickly frozen in a dry ice/ethanol bath and stored at -80°C. Extracts were prepared by resuspending cell pellets in 200 pl of Buffer I (Shorenstein & Losick, 1973) supplemented with 5% (v/v) phenylmethylsulfonyl fluoride (PMSF) (6 mg/ml in 95% ethanol), sonicating the cells, and centrifuging the lysate (14,0009 for 2 minutes at 4°C). The protein concentration of the extracts was determined by Bradford analysis (Bradford, 1976). 93 The DNA probes and a 10-fold excess of double—stranded poly (dI-dC) (Pharmacia) were combined with extracts, gel-purified SpoIIID, or gel-purified 9 kDa protein in binding buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM DTT, 1 mM EDTA and 5% (v/v) glycerol). The mixtures were incubated for 20 minutes at 37°C and electrophoresed on non-denaturing polyacrylamide gels. The unbound DNA and DNA-protein complexes were visualized by autoradiography. 94 Results A 9 kDa protein that copurifies with SpoIIID appears to be a degradation product of SpoIIID. A decrease in the level of spoIIID mRNA is paralleled by a decrease in the level of SpoIIID, suggesting a mechanism for degrading SpoIIID exist in sporulating B. subtilis (Halberg & Kroos, 1992). A clue to how this mechanism may work was obtained during the purification of SpoIIID. A 9 kDa protein co-eluted with SpoIIID upon salt-gradient elution of a double-stranded DNA-cellulose column (Figure 1). The first 20 amino acids of this protein were identical to those of SpoIIID, as determined by amino-terminal amino acid sequencing using sequential Edman degradation in an automated gas-phase sequentor (MSU Macromolecular Structure Facility; data not shown). This observation suggested that the 9 kDa protein may arise by removing amino acids from the C-terminus of SpoIIID. Consistent with this idea were the results of the mass spectral analysis of SpoIIID and the 9 kDa protein. Six independent experiments revealed that gel-purified SpoIIID consists of a single polypeptide with an average mass of 10813 i 13 Da (Figure 2A; a representative spectrum), which is in good agreement with the predicted mass of SpoIIID (10803 Da; Kunkel et al. 1989). Digestion of gel—purified SpoIIID with endoproteinase Asp-N yielded seven peptides, which had masses in good agreement with the predicted masses 95 Figure 1. A silver-stained 18% polyacrylamide gel displaying proteins collected from a double-stranded DNA—cellulose column upon elution with a 0.5 M - 1.3 M potassium chloride gradient. The positions of SpoIIID and the 9 kDa protein are indicated. 96 1.3 M KCl 0.5 M KCl oIIID —' 9 kDa— SP 97 Figure 2. Mass spectra of SpoIIID and SpoIIID digested with endoproteinase Asp-N. Spectra of (A) SpoIIID and (B) SpoIIID digested with endoproteinase Asp-N were obtained by matrix- assisted laser desorption (MALDI) mass spectroscopy as described in the Materials and Methods. The intensity is in arbitrary units. 98 ooovfi BE: 30: Comm? OOON P momoa _ --_ 1 . : ;.++e; ,. .-.: ;+d;Il .g 23.2.... ....si oomv: ZLQOI. 0009 tom .00— on— Kggsuazul 99 mac: mmo: OpMm coho comm Gown come oowe 00mm opwn comm ooow . . .. . “a c. 79 .o . C. 9 V .7 C 9 7o m .. ...H. m 7. .00“ Honov p.mmav H.~mae . an; 982. 932 v I... . . w z 3% H 22 m 32 .9 w 8-3 manna can?” 9 9 Hanan ~.nmmm m.vmmm 3-8 933 82.8 n. .004 8-3 p.28 83$ m Insuown 100 of SpoIIID peptides expected to result from such a digest (Figure 28). These fragments spanned the entire SpoIIID sequence (Figure 28). Four independent experiments revealed that gel-purified 9 kDa protein consisted of two polypeptides with average masses of 10078 i 5 Da and 9896 t 15 Da (Figure 3; a representative spectrum). The relative amounts of these polypeptides varied from sample to sample. A mass of 10078 Da is in good agreement with the predicted mass (10073 Da) of a SpoIIID peptide spanning from amino acids 1 through 86 (i.e., missing the C-terminal 7 amino acids). A mass of 9896 Da does not match that of any predicted SpoIIID peptide, suggesting that it may be a contaminating protein. Taken together, N-terminal amino acid sequencing and mass spectral data suggest that SpoIIID is converted to a 9 kDa form by removing 7 amino acids from its C-terminus during sporulation. Additional experiments are required to determine the C-terminus of the 9 kDa protein with certainty. DNA-binding and transcriptional properties of the 9 kDa protein. The putative helix-turn-helix DNA-binding motif in SpoIIID spans from amino acid 23 to amino acid 42 (Kunkefl.et al., 1989; Stevens & Errington, 1990). Based on this observation, the 9 kDa protein may, like SpoIIID (Chapter II), bind to sigK and cotD. To test this idea, DNase I footprinting experiments were performed (Figure 4). The 9 kDa protein protects the same regions of sigK and cotD 101 Figure 3. Mass spectrum of the 9 kDa protein. The spectrum was obtained by MALDI mass spectroscopy as described in the Materials and Methods. The observed mass at 11464 is (2M+H)+ species of insulin (internal standard). The intensity is in arbitrary units. 102 BE: 80: ooomp oooFF oooop ooom. ooow _ b n - 00F . i .o . 6 l. 0 nvuL mw.9 c. .oom ( - . .oon u _ ... ... c.82— .? _ w 2.2: 2.2: 4-, i -I.. i . 11: 4 K318119311] 103 Figure 4. SpoIIID and 9 kDa footprints on sigK and cotD. Radioactive DNA probes separately end-labelled were incubated in separate reactions with no protein, 120 ng gel—purified 9 kDa protein, or 120 ng gel—purified SpoIIID, and then digested with DNase I. The resulting DNA fragments were separated by electrophoresis on a 6% polyacrylamide gel containing 8 M urea. (A) Footprints in the sigK promoter region and open reading frame (ORF) identified using a probe labelled at the XbaI site located downstream of the transcriptional start site. (B) Footprints in the cotD promoter region and ORF identified using a probe labelled at the EcoRI site located upstream of the transcriptional start site. 104 co tD sigK I SpoIIID - - + [ 9 kDa '0 O I... 'e... ......D...’ o. '....O. ,Ocoa ... o.....o...'... 0 es. . .0. 0:. ’3. ...e’eeo...’ .. ea: . 105 from DNase I digestion as SpoIIID. SpoIIID binding stimulates sigK transcription and represses cotD transcription by (J?K RNA polymerase in vitro (Kroos et al., 1989) . To test whether the 9 kDa protein affects the transcription of these genes, linearized DNA templates were transcribed by partially purified 0'K RNA polymerase alone or in the presence of gel-purified 9 kDa protein. The presence of the 9 kDa protein greatly activated sigK transcription, but failed to repress cotD transcription (Figure 5). This observation suggests that the C-terminus of SpoIIID is required for cotD repression. However, the gel-purified 9 kDa protein used in these experiments may have been contaminated with the 9896 Da polypeptide identified by mass spectral analysis. These experiments need to be repeated with more highly purified 9 kDa protein in order to determine unequivocally its DNA-binding and transcriptional properties. SpoIIID is converted to the 9 kDa protein in a developmentally regulated fashion. The 9 kDa protein was not detected in extracts of sporulating cells by Western blot analysis with polyclonal anti-SpoIIID antibodies (Halberg & Kroos, 1992); however, these antibodies recognized gel-purified 9 kDa protein weakly. Because the experiments presented above suggested that the 9 kDa protein would bind to specific sites in DNA with similar affinity as SpoIIID, we 106 Figure 5. Effects of SpoIIID and the 9 kDa protein on sigK and cotD transcription in vitro. DNA templates (1pg of pLRKlOO, which contains the cotD promoter, digested with HindIII and 1pg of pBK16, which contains the sigK promoter, digested with XbaI; Kroos et al., 1989) were transcribed with partially purified OK RNA polymerase (200 ng) alone (lane 1) or in the presence of either 120 ng of gel-purified 9 kDa protein (lane 2) or 120 ng of gel-purified SpoIIID (lane 3). Run-off transcripts were electrophoresed in a 5% polyacrylamide gel containing 8 M urea. Sizes of transcripts were estimated from the migration of end-labelled fragments of MSpI-digested pBR322. 107 cotD M 225 ——> sigK +1 — 178 —> SpoIIID - - + 9kDa - + - 225 178 108 used a mobility shift assay to detect the 9 kDa protein in extracts of sporulating cells. Extracts prepared from cells harvested at hourly intervals between T1 and T8 were incubated with a radiolabelled DNA probe and then electrophoresed on a non-denaturing polyacrylamide gel. Two shifted complexes were observed with extracts prepared from wild—type cells (Figure 6A). The upper complex comigrated with a complex produced by gel-purified SpoIIID and DNA. This complex was first detected at T3, its level increased until T5, and its level decreased beginning at T6. This pattern is consistent with the level of SpoIIID detected by Western blotting during sporulation (Halberg & Kroos, 1992). The lower complex comigrated with a complex produced by gel-purified 9 kDa protein and DNA. This complex was first detected at T4 and its level increased until T6. The upper and lower complexes were not observed in extracts prepared from spoIIID mutant cells (Figure 6A). Taken together, these results indicate that the upper and lower complexes correspond to binding of SpoIIID and the 9 kDa protein to the DNA probe, respectively. In addition, the absence of a 9 kDa protein complex in extracts of spoIIID mutant cells is consistent with the idea that the 9 kDa protein is derived from SpoIIID. If this is the case, the results shown in Figure 6A suggest that cells acquire the ability to convert SpoIIID to the 9 kDa protein beginning at T4 of sporulation. The low abundance of the 9 109 Figure 6. Mobility shift assay to monitor the levels of SpoIIID and the 9 kDa protein in wild-type and spoIIID mutant cells. A labelled oligonucleotide (50 ng) designed to correspond to a SpoIIID/9 kDa protein binding site in the sigK ORF (+73 to +96) was incubated with extracts (5 pg total protein) prepared from wild-type and spoIIID mutant cells that were collected at the indicated times after the onset of sporulation, gel-purified SpoIIID (30 ng) and gel—purified 9 kDa protein (5 ng), or gel-purified 9 kDa protein (30 ng). The mixtures were electrophoresed on a 20% polyacrylamide gel. (A) Extracts were prepared immediately before incubation with labelled DNA probe. (B) Extracts were incubated overnight at 4'C prior to incubation with the labelled DNA probe. 110 9 ~49 4. °) SP0 spoIIID 0' x 7- '7' m I» , . 17 l T T T T T T T T «I Q I l 3 4 5 6 7 8. 1 3 5 O ‘4- _ I fiH 9Q D) tn spo + £1 T2 T3 T4 T5 T6 T7 T8. I a 111 kDa protein complex suggests that the 9 kDa protein never accumulates to a very high level in cells. Surprisingly, the relative amounts of the SpoIIID and 9 kDa protein complexes changed when the extracts were incubated overnight at 4’C (Figure 6B). The SpoIIID complex was only detected between T3 and T5. The 9 kDa protein complex was still detected between T4 and T3, but its level was significantly higher in this experiment as compared to the experiment performed with freshly prepared extracts (Figure 6A). A simple explanation for these results is that SpoIIID may be converted to the 9 kDa protein during the overnight incubation at 4°C. The presence of phenylmethylsulfonyl fluoride (PMSF, a serine protease inhibitor) in the extract buffer may stabilize the 9 kDa protein in vitro, allowing it to accumulate to a higher level than in vivo, because the levels of both SpoIIID and the 9 kDa protein were significantly reduced when PMSF was omitted from the extract buffer (data not shown). Conversion of SpoIIID to the 9 kDa protein is reduced in several sporulation mutants. To test whether mutants that failed to produce SpoIIID were capable of converting SpoIIID to the 9 kDa protein, gel-purified SpoIIID was added to crude extracts prepared from these mutants, the mixture was incubated overnight at 4’C, and then the levels of SpoIIID and the 9 kDa protein were determined using the mobility shift assay (Figure 7). As a control, gel-purified SpoIIID was added to extracts prepared from wild-type cells. 112 Figure 7. Mobility shift assay to determine whether gel— purified SpoIIID is converted to the 9 kDa protein in extracts prepared from wild-type, spoIIG and spoIIID cells. Extracts (5 pg total protein) prepared from wild-type, spoIIG and spoIIID cells collected at the indicated times were incubated overnight at 4°C alone, or with gel-purified SpoIIID (15 ng). The extracts were then incubated with a labelled 106 bp Eco47III-HindIII fragment (100 ng) described in Materials and Methods. Gel-purified 9 kDa protein (15 ng; this particular preparation had low DNA-binding activity) and gel-purified SpoIIID (15 ng) were also incubated with the labelled DNA fragment. The mixtures were electrophoresed on an 8% polyacrylamide gel. 9 O 08' 0" fig 99 99 T1 T6 T6 T6 0 H H H oé‘rx 00 SpoIIID - + - + — + _ ... 9‘2 c, .I..- ., "u‘ 114 Gel-purified SpoIIID failed to be converted to the 9 kDa protein in extracts prepared from wild-type cells harvested at T1, but was converted in extracts prepared from wild-type cells harvested at T6. Extracts prepared from spoIIG (encoding (fiiwhich is required for spoIIID transcription) and spoIIID mutant cells harvested at Tefailed to convert gel-purified SpoIIID to the 9 kDa protein. These results demonstrate that the ability to convert SpoIIID to the 9 kDa form is a developmentally-regulated event that depends upon (3'E and SpoIIID. To determine whether other sporulation genes affected the conversion, the levels of SpoIIID and the 9 kDa protein in sporulation mutants were monitored by mobility shift assays (Figure 8). Mutants that fail to produce pro-(3K or (SR (Lu et al., 1990), due to a mutation in sigK (Stragier et al., 1989) or a mutation in spoIVCA (encoding the putative recombinase that generates the composite sigK gene (Kunkelefi: al., 1990; Satt>et al., 1990; Popham & Stragier, 1992)), produced the 9 kDa protein. Similarly, mutants that produce pro-<5K but fail to produce a detectable amount of OK, due to a mutation in spoIVB or spoIVF (Lu et al., 1990), produced the 9 kDa protein. These results demonstrate that the conversion does not depend on OK. In fact, the level of the 9 kDa 115 Figure 8. Mobility shift assay to monitor the levels of SpoIIID and the 9 kDa protein in sporulation mutants. A labelled 106 bp Eco47III-HindIII fragment (100 ng) described in Materials and Methods was incubated with extracts (2 pg) prepared from wild—type or mutant cells that were collected at T6, gel-purified 9 kDa protein (30 ng; lane 11), or a combination (lane 12) of gel-purified SpoIIID (30 ng) and 9 kDa protein (30 ng). These mixtures were electrophoresed on an 8% polyacrylamide. Note: Extracts were incubated overnight at 4°C prior to incubation with the labelled DNA fragment. Lanes: 1, PY79 (spo+), 2, BK183 (spoIVA), 3, (spoIVB), 4, BK558 (spoIVCA), 5, BK103 (spoIVCB), BSL51 (spoIVF), 91 (spoVB), SCSO (spoVC), SC53 (spoVF), and BZ216 (cotE). 123456789101112 --- DNA-SpoIIID DNA-9 kDa s’ 117 protein is significantly higher in fresh extracts prepared from mutants that fail to produce 0“, as compared to wild-type cells (data not shown), suggesting that active OK may interfere with the conversion and/or cause destabilization of the 9 kDa protein. Mutants that produce abnormal spores with altered cortex and/or coat layers (Piggotem:al., 1981) either fail to convert SpoIIID to the 9 kDa protein (spoIVA) or are impaired in the conversion (spoVB, spoVC and spoVF). These observations suggest that the conversion of SpoIIID to the 9 kDa protein is coupled to spore morphogenesis. 118 Discussion The results presented here provide evidence that SpoIIID is converted to a 9 kDa form by removing 7 amino acids from its C-terminus. While this conversion did not alter the ability of the protein to bind to specific sites in DNA (Figure 4), it did appear to alter the ability of the protein to affect transcription (Figure 5). The conversion is developmentally regulated, since (1) SpoIIID present at T3 was not converted to the 9 kDa protein during sporulation (Figure 6) and (2) the conversion is defective in several sporulation mutants (Figures 7 and 8). The 9 kDa protein does not appear to accumulate to a very high level in cells, suggesting that it is unstable. Hence, conversion of SpoIIID to the 9 kDa form may explain the parallel decrease in the levels of spoIIID mRNA and SpoIIID protein during sporulation (Halberg & Kroos, 1992). Because SpoIIID influences the transcription of many mother-cell genes (Kroos et al., 1989; Zheng et al., 1992), the change in its transcriptional effects and cellular stability brought about by conversion to the 9 kDa form presumably has a profound effect on the pattern of mother- cell gene expression. How might the SpoIIID protein be converted to the 9 kDa form? The total amount of SpoIIID and the 9 kDa protein is comaparable in fresh and overnight extracts, but the relative amounts of SpoIIID and the 9 kDa protein are different in 119 these extracts (i.e., the amount of SpoIIID is lower and the amount of the 9 kDa protein is higher in overnight extracts as compared to fresh extracts; Figure 6). This observation suggests that SpoIIID is converted to a 9 kDa form. A simple explanation for the conversion is proteolytic processing involving either a single endoproteolytic cleavage or several exoproteolytic cleavages. In these cases, the convertase (i.e., the factor that catalyzes the conversion) would be a protein. Another possible explanation is that SpoIIID may autoproteolyze. In this case, the convertase may or may not be a protein. Alternatively, SpoIIID may not be converted to a 9 kDa form. The 9 kDa protein may be generated by translational frameshifting. If this is the case, translation would have to occur in vitro, since the amount of the 9 kDa protein in the extracts increases during the overnight incubation (Figure 6). The frameshift would have to occur downstream of the valine at position 20, because the first 20 amino acids of the 9 kDa protein is identical to those of SpoIIID. In addition, it would most likely occur downstream of the threonine at position 42, because the 9 kDa protein and SpoIIID have identical binding sites in two genes (Figure 5) and the putative helix turn helix DNA-binding motif of SpoIIID is located between positions 23-42 (Kunkel.et al., 1989). Translational frameshifting often requires two sequence elements: a slippery site, which is composed of 120 Figure 9. Model for the switch from sigK to cotD transcription in the mother cell during the stage IV to stage V transition of sporulation. (A) During stage IV, SpoIIID stimulates sigK and represses cotD transcription by 0* RNA polymerase. (3'K RNA polymerase also transcribes gerE, leading to the synthesis of GerE. (B) The accumulation of OK, resulting from the processing of pro-(3'K to 0K, causes a decrease in the level of spoIIID mRNA. This decrease is paralleled by a decrease in the level of SpoIIID, which appears to be due to the conversion of SpoIIID to a less stable 9 kDa form. The SpoIIID decrease causes a switch to the stage V pattern of gene expression (i.e., sigK transcription is no longer stimulated and cotD transcription is no longer repressed). Continued production of GerE reinforces the switch since GerE represses sigK transcription and stimulates cotD transcription. 121 \I'Q+oo on“ m urootlmtouAllxoli @233 + v.05... go; > 095 m egos] mcoollmteulllxo + x992. + x3» >_ 3.5 9:26 122 stretch of adenosines, and a stem-loop structure. A potential slippery site exists in the C-terminus of the spoIIID gene and a -1 shift would result in a premature stop. However, there is no stem-loop structure and the predicted mass of the polypeptide produced by such a frameshift is smaller than the observed mass of the 9 kDa protein. What role could the conversion of SpoIIID to the 9 kDa form play in establishing the temporal pattern of mother-cell-specific gene expression? SpoIIID activates sigK transcription and represses cotC, cotD, and cotX transcription by GK RNA polymerase in vitro (H. Ichikawa, R. Halberg, L. Kroos, unpublished data; H. Ichikawa and L. Kroos, unpublished data; Kroose¢:al., 1989). Previously, it was proposed that the inactivation and/or sequestering of SpoIIID switches the mother-cell pattern of gene expression from the transcription of sigK during stage IV to the transcription of spore coat genes during stage V (Kroos & Losick, 1989). Consistent with this idea, the accumulation of OK, resulting from the processing of pro-(5’K to OK, reduces the levels of spoIIID mRNA and SpoIIID (Halberg & Kroos, 1992). The reduction in the level of spoIIID mRNA would not be concomitant with a reduction in the level of SpoIIID protein unless there was a mechanism to inactivate the SpoIIID protein. Thus, the conversion of SpoIIID to the 9 kDa protein may be required to relieve the repressive effect that SpoIIID 123 has on the expression of spore coat proteins during sporulation (Figure 9). Protein stability plays a role in controlling the pattern of gene expression in other biological systems. For example, the stability of the cII protein is a key element in the lysogeny-lysis decision of bacteriophage l (i.e., the presence of cII favors lysogeny (Echols & Green, 1971; Reichardt & Kaiser, 1971; Shimatake & Rosenberg, 1981; Ho et al., 1983; McMacken et al., 1970), whereas the absence of cII, resulting from its degradation by host factors (Hoyt et al., 1982; Rattray et al., 1984; Banuett et al., 1986; Chengwet al., 1988), favors lysis). How is the production of the convertase regulated during sporulation? The production of convertase depends on (3'E and SpoIIID (Figure 7) . Perhaps OE RNA polymerase transcribes a gene required to produce convertase and this transcription may be stimulated by SpoIIID. SpoIIID does activate and repress the transcription of several 03-dependent genes (Errington, 1993; Chapter II). Errington (1993) proposed that the transcriptional effects of SpoIIID may divide the OE regulon into three temporal classes. The first class is composed of genes whose transcription is unaffected by SpoIIID. The second class is composed of genes whose expression is repressed by SpoIIID. Expression of class I and II genes would begin as soon as 0? activity appears. The 124 third class is composed of genes whose expression is stimulated by SpoIIID. Expression of class III genes would be delayed relatively to classes 1 and 2, because transcription of the spoIIID gene by 03 RNA polymerase is itself partially dependent on SpoIIID (Kunkel.et al., 1989). Thus, a gene required to produce convertase may belong to temporal class III of the (5'E regulon. Alternatively or in addition, perhaps OE and SpoIIID are required indirectly to permit synthesis and/or activation of the convertase. No matter whether SpoIIID affects convertase directly or indirectly, a lag between the appearance of SpoIIID and appearance of the 9 kDa protein would be expected, and a lag of about one hour is observed (Figure 6). Conversion is inhibited by mutations in two genes that are known to be (3'E dependent, spoIVA (Roels et al., 1992) and spoVB (Popham & Stragier, 1991). Interestingly, mutations in two of these genes, spoIVA and spoVB, impair formation of the spore cortex (Roels et al., 1992; Popham & Stragier, 1991). Thus, cortex formation may be required for full convertase activity. An appealing feature of this model is that the relief of the SpoIIID repressive effect on the transcription of spore coat genes is coupled to the morphological event preceding the formation of the spore coat. However, mutations that block OK production also inhibit cortex 125 formation, yet SpoIIID is converted to the 9 kDa protein. One possible explanation for this observation is that the convertase does not gain full activity, but convertase accumulates due to the sporulation block caused by the absence of OK. Consistent with the former is the finding that a mutation in spoVF, a locus whose transcription depends on OK RNA polymerase, inhibits the conversion (Figure 8). Consistent with the idea that convertase accumulates due to the absence of OKis the observation that the level of SpoIIID is higher in mutants that fail to produce (3'K than in wild-type cells (Halberg & Kroos, 1992), which demonstrates that (3'K can affect the accumulation of a protein whose expression depends on GB. The spoVF operon is composed of two genes, dpaA and dpaB, which encode subunits of dipicolinic acid (DPA) synthetase (Daniel & Errington, 1993). DPA synthetase is required for the formation of DPA from an intermediate in the lysine biosynthetic pathway (Bach & Gilvarg, 1966). DPA is a small, polar molecule that appears to play a role in spore heat-resistance (Church & Halvorson, 1959). Hence, conversion of SpoIIID to the 9 kDa protein may couple release of spore coat repression to acquisition of heat-resistance as well as formation of the spore cortex. CHAPTER V Overproduction of SpoIIID Negatively Regulates O’K Accumulation During Bacillus subtilis Sporulation and Reduces cotD Expression in Cells Producing OK During Growth "In everything the middle course is the best; all things in excess bring trouble." Plautus 126 127 Abstract SpoIIID is a DNA-binding protein that activates or represses transcription of many different genes in the mother-cell compartment of sporulating Bacillus subtilis. Previous studies showed that SpoIIID represses cotD (encoding a spore coat protein) transcription by (I?K RNA polymerase in vitro and that a decrease in the level of SpoIIID coincides with an increase in cotD expression during sporulation. Hence, SpoIIID was proposed to repress cotD transcription in vivo until the level of SpoIIID falls. We attempted to test this hypothesis by engineering continued production of SpoIIID late in sporulation. We found that elevating the level of SpoIIID reduced the expression of all OK-transcribed genes tested, including some not expected to be repressed by SpoIIID. The production of heat-resistant spores was also reduced. These effects appear to result from a reduction in the level of CK. Thus, maintaining the proper levels of SpoIIID and OR is crucial for normal sporulation. To circumvent the problem of SpoIIID overproduction reducing the level of (3'K during sporulation, we employed a strain engineered to produce (5'K during growth. Expressing SpoIIID in this strain during growth reduced cotD expression, but not the expression of other OK-transcribed genes tested. These 128 results suggest that SpoIIID is capable of repressing cotD expression in vivo. 129 Introduction During sporulation of Bacillus subtilis, an asymmetrically positioned septum divides the bacterium into two compartments, the mother cell and the forespore (Smithefi: al., 1989). Both of these compartments receive a copy of the genome, but realize alternative developmental fates because gene expression is regulated spatially. Two key regulators of gene expression in the mother-cell compartment are SpoIIID, a small DNA-binding protein that functions as a transcriptional activator and repressor (Chapter II; Kroosefi: al., 1989; Kunked.et al., 1989), and UK, a sigma subunit of RNA polymerase (Kroos et al., 1989; Stragier et al., 1989; Zheng & Losick, 1990; Zheng et al., 1992) . Production of UK is regulated positively by SpoIIID at several levels. First, SpoIIID is required for a chromosomal rearrangement that generates the composite sigK gene encoding 6K (Stragier et al., 1989; Kunkel et al., 1990) . Second, SpoIIID stimulates sigK transcription (Kunkefl.et al., 1988; Kroos et al., 1989) . Third, SpoIIID is needed to permit proteolytic processing of the sigK'primary translation product, pro-OK, to produce active 0K (S. Lu & L. Kroos, unpublished data). Several lines of evidence indicate that processing of pro-(3'K couples events occurring in the mother 130 cell to events occurring in the forespore (Cuttingwet al., 1990; Lu et al., 1990; Cutting et al., 1991a; Cutting et al., 1991b). In addition to positively regulating Ofijproduction, SpoIIID represses cotD (encoding a spore coat protein (Donovan et al., 1987)) transcription by GK RNA polymerase in vitro (Kroos et al., 1989) . Hence, it was proposed that inactivation of SpoIIID switches the pattern of mother-cell gene expression from sigK transcription at stage IV (cortex formation) of sporulation to cotD transcription at stage V (coat formation) (Kroosem al., 1989). Previously, we demonstrated that the level of SpoIIID does decrease during sporulation and that the decrease is controlled, at least in part, by the production of 0K(Halberg & Kroos, 1992). We also noted that the SpoIIID decrease coincides with an increase in cotD expression during sporulation in wild-type cells and cells engineered to produce (3'K earlier than normal, suggesting that SpoIIID represses cotD expression in vivo (Halberg & Kroos, 1992). Here we attempted to test the idea that SpoIIID represses cotD expression in vivo by engineering continued production of SpoIIID late in sporulation. Unexpectedly, we found that elevating the level of SpoIIID not only reduced cotD expression, but the expression of all OK-transcribed 131 genes tested and the production of heat—resistant spores. We show that overproducing SpoIIID reduces (3'K accumulation during sporulation. To circumvent this problem, we employed a strain that was engineered to permit the production of OK during vegetative growth. We show that expression of SpoIIID in this strain reduces cotD promoter activity, but not the promoter activity of other OK-transcribed genes, suggesting that SpoIIID is capable of repressing cotD expression in vivo. 132 Materials and Methods B I . 1 SI . E. coli strain AG115 (araDl39A(ara,leu)7697, Alacx74, gaer, galKe, hsr-, hsm+, strA, (F', proAB, lachZ::Tn5)) was obtained from A. Grossman (Massachusetts Institute of Technology) and served as the host for construction and maintenance of plasmids. B. subtilis strains PY79 (spo+), BK395 (spoIIID83), BK541 (spoIIIDAerm), and V0536 (Pspac- PsigKesigKA19) have been described (Youngmanwet al., 1984; Kunked.et al., 1989; Oke & Losick, 1993). Pspac-spoIIID was derived from pBK39 (Kunkel.et al., 1989) which contains spoIIID, and pDGl48 (Stragierwet al., 1988) which contains the isopropyl B-D-thiogalactopyranoside (IPTG)-inducible promoter, spac, and is stably maintained in E. coli or B. subtilis. pDGl48 was digested with HindIII. The ends were rendered blunt by the fill-in reaction of Klenow enzyme and then dephosphorylated with calf intestinal phosphatase. The linearized plasmid was ligated to a 536 bp XMnI fragment from pBK39 containing the spoIIID promoter region and open reading frame (ORF). Ampicillin-resistant E. coli transformants were obtained and the structure of Pspac- spoIIID, a plasmid containing the insert in the proper orientation to fuse spoIIID transcription to spac, was verified by restriction mapping. 133 Pspac-spoIIIDA is identical to Pspac-spoIIID except it lacks the spoIIID ORF. It was generated following the procedure outlined above except the insert was a 234 bp XMnI- ApaLI fragment from pBK39 containing only the spoIIID promoter region. The ApaLI end was rendered blunt using the fill-in reaction of Klenow enzyme prior to ligation. Pspac-spoIIID-cat is an integrational version of Pspac- spoIIID. It was constructed by replacing the EcoRI fragment encoding kanamycin resistance and the origin of replication that functions in B. subtilis with a chloramphenicol resistance-encoding EcoRI fragment from pMIllOl (Igo & Losick, 1986). Competent B. subtilis cells were prepared and transformed as described (Dubnau & Davidoff-Abelson, 1971). Transformants containing Pspac-spoIIID and Pspac-spoIIIDA were selected on LB agar containing kanamycin sulfate (5 mg/ml). Transformants containing Pspac-spoIIID-cat integrated into the chromosome were selected on LB agar containing chloramphenicol (5 pg/ml). Use of specialized transducing phages SPB::cotA-lacz, SPB::cotD—lacz, SPB::gerE-lacz, and SP8::sigK-lacz has been described (Kunkel et al., 1988; Cutting et al., 1989; Cutting et al., 1990; Zheng & Losick, 1990). 134 Wall Sporulation was induced by resuspending growing cells in SM medium as described (Sterlini & Mandelstam, 1969). The onset of sporulation (To) is defined as the time of resuspension. At three hours after sporulation, IPTG was added to a final concentration of 1 mM. Production of heat-resistant spores was assayed as described (Cutting & Horn, 1990). Pspac-PsigK—sigKA19 cells were grown in 2 x YT medium (Maniatislet al., 1982). When cells reached the mid-log phase (O.D.an of 0.3-0.5), IPTG was added to a final concentration of 1 mM. Hesternmlntjnalxsis Samples (1 ml) were harvested by centrifugation (14,000 g for 5 minutes) at the indicated times during growth and sporulation. Whole-cell extracts were prepared as described (Halberg & Kroos, 1992) and the amount of protein present was quantified by the Bradford method (Bradford, 1976). Polypeptides (5 pg) were separated by SDS-PAGE (18% polyacrylamide; Thomas & Kornberg, 1978) and electroblotted to poly(viny1idene difluoride) membrane (Matsudaira, 1987). Immunoblot analyses using polyclonal anti-SpoIIID and anti- pro-(3’K antibodies were performed as described (Halberg & Kroos, 1992). 135 833W B-galactosidase activity was determined using the substrate O-nitrophenol-B-D-galactoside (ONPG) as described (Miller, 1972). One unit of enzyme hydrolyzes 1 pmol of substrate per min per O.D.am of initial cell density. Wears In vitro transcription assays were performed utilizing (fl RNA polymerase which was partially purified from gerE- cells and SpoIIID which was gel-purified as described (Kroos et al., 1989). MW At hourly intervals between T4 and T3, cells were harvested by centrifugation (11,950g for 10 minutes) and RNA was prepared as described (Halberg & Kroos, 1992). The RNA was treated with DNase I to remove contaminating chromosomal DNA. The RNA (20 #9) was fractionated by electrophoresis on a 1.2% (w/v) agarose gel containing 1.11% (v/v) formaldehyde, transferred to nitrocellulose, and hybridized at 55°C to nick-translated pSC146 (Zhengwet al., 1992) or pLRKlOO (Kroos et al., 1989). The signals were visualized by autoradiography and quantitated using a Visage Digital Imager. The size of the mRNAs was estimated by comparing the positions of the signals to the positions of RNA standards 136 (0.16 kb to 1.77 kb RNA ladder from BRL). 137 Results A plasmid containing the spoIIID gene permits continued production of SpoIIID late in sporulation. To produce SpoIIID, a multicopy plasmid (Pspac-spoIIID) was constructed containing the IPTG-inducible spac promoter fused to the spoIIID promoter region and ORF (see Materials and Methods). Pspac-spoIIID and its parental plasmid pDGl48 (Stragierwet al., 1988) were transformed into spoIIIDB3 cells which fail to produce SpoIIID (R. Halberg and L. Kroos, unpublished data). Sporulation was induced in the resulting strains by the resuspension method and IPTG was added at T3 (i.e., three hours after sporulation, which is the time when SpoIIID is first detected in wild-type cells (Halberg & Kroos, 1992)). The level of SpoIIID was monitored by Western blot analysis utilizing polyclonal antiserum against SpoIIID. spoIIID- cells containing Pspac-spoIIID accumulated more SpoIIID than wild-type (spo+) cells, whereas spoIIID— cells containing the parental plasmid failed to produce SpoIIID (data not shown). Pspac-spoIIID not only elevated the amount of SpoIIID present between T3 and T5, but permitted a significant level of SpoIIID to be maintained between T6 and T3, when the level of SpoIIID is normally very low (Halberg & Kroos, 1992). Interestingly, the level of SpoIIID was also elevated in Pspac-spoIIID—containing cells that were not treated with IPTG (data not shown). This result suggests 138 that overproduction of SpoIIID is due to the presence of multiple copies of the spoIIID gene (including the promoter) and/or incomplete repression of the spac promoter. Nonetheless, the goal of elevating the level of SpoIIID late in sporulation was achieved. Overproduction of SpoIIID reduces cotD, cotA, and gar! promoter activity. If SpoIIID does repress cotD transcription in vivo, then an elevated level of SpoIIID should reduce cotD expression. To test this prediction, a cotD-lacz fusion was introduced into wild-type cells containing either Pspac-spoIIID or the parental plasmid. Sporulation was induced in the resulting strains by the resuspension method and IPTG was added at T3. The levels of SpoIIID and cotD-directed B-galactosidase activity were monitored. Cells containing Pspac-spoIIID exhibited an elevated level of SpoIIID between T4 and T3 relative to that observed in cells containing the parental plasmid (Figure 1A). Overproduction of SpoIIID reduced cotD expression two- fold (Figure 1B). These results are consistent with the hypothesis that SpoIIID represses cotD expression in vivo. However, overproduction of SpoIIID also reduced the expression of two other Ox-transcribed genes, including one not expected to be repressed by SpoIIID. Preliminary studies indicated that SpoIIID represses cotA (encoding a spore coat protein (Donovan et al., 1987)) transcription by GK RNA 139 Figure 1. Levels of SpoIIID and cotD-, cotA- , and gerE- directed B-galactosidase activity in spo+ cells containing either Pspac-spoIIID or the parental plasmid. The level of SpoIIID (panel A) in spo+ cells containing Pspac-spoIIID or the parental plasmid (pDGl48), collected between T4 and T8 was determined by Western blot analysis as described in the Materials and Methods. cotD-, cotA—, and gerE-directed B- galactosidase activity (panels B, C, and D, respectively) detected in spo+ cells containing either Pspac-spoIIID (I) or the parental plasmid, pDGl48 (Ch. Points are the average of three determinations for cotD-directed B-galactosidase activity and the average of two determinations for cotA— and gerE—directed B-galactosidase activity. Error bars indicate one standard deviation of the data. 140 co_mcoamsmom .22 also: a n o m e n u p P F D P W- .om 6.23233: .83 952.. a h o n v n N w (suun Jamw) Aunuov asepgsozoemfi-g .ov I P b .om D co_mcoomsmo.m .33 230: m h w m e a N _. (mun mum) Awmov essmsoxosusB-g (suun Jaluw) Awmov asepgsowelefi-g 9:2.ml I. 1114...... 32.22. nEoueéeQea +°Qfi +0“. 0 141 polymerase in vitro, but has no effect on gerE (encoding a regulator of spore coat synthesis (Cutting & Mandelstam, 1986)) transcription by (5'K RNA polymerase in vitro (L. Kroos & R. Losick, unpublished data). Based on these observations, overproduction of SpoIIID was expected to reduce cotA promoter activity, but have no effect on gerE promoter activity. To test this prediction, lacz fusions to cotA and gerE were introduced into wild-type cells containing either Pspac-spoIIID or the parental plasmid. Sporulation was induced in the resulting strains by the resuspension method and IPTG was added at T3. cotA-directed B-galactosidase activity was reduced four-fold (Figure 1C) and gerE—directed B-galactosidase activity was reduced three-fold (Figure 1D) in cells containing Pspac-spoIIID as compared to cells containing the parental plasmid. The unexpected reduction in gerE-lacz expression suggests either that SpoIIID does repress gerE expression in vivo or that the overproduction of SpoIIID indirectly affects the expression of all OK-dependent genes by negatively regulating the level of (57K or its activity. Overproduction of SpoIIID reduces sigK promoter activity and the level of 0!. Since SpoIIID greatly activates sigK transcription by GE RNA polymerase (Chapter II) and 0K RNA polymerase (Kunkel et al., 1988; Kroos et al., 142 1989), overproduction of SpoIIID was expected to have a positive effect on sigK promoter activity. To test this prediction, wild-type cells containing either Pspac-spoIIID or the parental plasmid were transduced with phage carrying a sigK—lacz fusion. Sporulation was induced in the resulting strains by the resuspension method and IPTG was added at T3. sigKedirected B-galactosidase activity was elevated slightly during the early stages of sporulation and reduced significantly during intermediate to late stages of sporulation in cells containing Pspac-spoIIID as compared to cells containing the parental plasmid (Figure 2A). Elevated B-galactosidase activity at T3 and T4 may reflect SpoIIID stimulating sigK transcription by (3’E RNA polymerase. The reduced B—galactosidase activity at T5 through Th may reflect a reduction in sigK transcription by (3'K RNA polymerase. To determine whether the level of (3'K was lower in cells containing Pspac-spoIIID, we used anti-pro-(S’K antibodies (Lu et al., 1990) to detect pro-(SK and 6K in Western blot analyses (Figure 2B). The level of pro-(3'K was similar in strains containing Pspac-spoIIID or the parental plasmid, suggesting that overproduction of SpoIIID does not hinder the chromosomal rearrangement that generates the composite sigK gene or its initial expression. However, the level of (3'K was 143 Figure 2. Levels of sigKedirected B-galactosidase activity, pro—OK, and 0K in spo+ cells containing either Pspac-spoIIID or the parental plasmid. (Panel A) sigKedirected B- galactosidase activity detected in spo+1cells containing either Pspac-spoIIID (I) or the parental plasmid, pDGl48 “3). Points are the average of three determinations. Error bars indicate one standard deviation of the data. (Panel B) The levels of pro—(3'K and 6K in spo+ cells containing either Pspac-spoIIID or the parental plasmid (pDGl48), collected between T3 and T3 were determined by Western blot analyses utilizing anti-pro-(S'K antibodies as described in the Materials and Methods. «9312,22,22333121123-- w W W i. a r I. I 4.14.3.13). III\ . . 1...! I. . , .I . . .1ing LIV, , 1 IN . . ,r’ ...I‘ .‘klhluefl‘dffl‘asti . f1.llilk.u.n‘.lvg'i’,.‘1 Q} 55.0: I'.-' *{a 144 B-gslectosidase Activity (Miller Unlts) Hours After Resuspenslon um" um" Pspsc-spolllD parental I l I J [—7 I "'374'751'61'7'r T3 T4 1'51}; T71's 145 lower at T4 through T8 in cells containing Pspac—spoIIID as compared to cells containing the parental plasmid. This shows that the overproduction of SpoIIID adversely affects the accumulation of OK. The lower level of OK may reduce expression of (SK-dependent genes. Because the level of OK, but not the level of pro-0K, was reduced by SpoIIID overproduction, we reasoned that processing of pro-OK to OK might be the step in OK production affected by elevating the level of SpoIIID. Since bof (bypass of the forespore) mutations bypass many of the normal requirements for pro-(5K processing (Cutting et al., 1990), we examined expression of cotD-, cotA-, gerE- and sigKelacZ fusions in a bofB8 mutant containing either Pspac-spoIIID or the parental plasmid. The results were similar to those observed in wild-type cells (data not shown). Hence, a bof mutation did not overcome the negative effect of SpoIIID overproduction on expression of OK-dependent genes. Overproduction of SpoIIID inhibits the production of heat-resistant spores. The presence of Pspac-spoIIID in spoIIID- cells only partially restored the production of heat-resistant spores (Table 1). Interestingly, the presence of Pspac-spoIIID in wild-type cells reduced the production of heat-resistant spores to a number that was comparable to the number of heat-resistant spores produced in spoIIID- cells 146 Table 1. Number of heat-resistant spores produced by Bacillus subtilis strains in the absence and presence of SpoIIID overproduction. The number of heat—resistant spores produced by cells containing Pspac-spoIIID or the parental plasmid (pDGl48) was determined as described in the Materials and Methods. The values represent the number of heat-resistant spores per ml at Tm3divided by the O.D.$5 at T1 and are the average of at least two independent determinations. The values in parentheses are the percentages of the number of heat-resistant spores produced in spo+ cells containing the parental plasmid. Strain spoIIID83, parental spoIIID83, Pspac-spollID spo+, parental 5190+, Pspac-spoIIID Spo+, Pspac-spoIIIDA 147 Heat-Resistant Sp_ore§. 1.3 x 103 (4.6 x 10—4) 1.9 x 107 (6.8) 2.8 x 108 (100) 3.2 x 107 (11.4) 4.4 x 108 (157) 148 containing Pspac-spoIIID (Table 1). One interpretation of these observations is that overproduction of SpoIIID and/or the resulting reduction in the level of OK inhibits sporulation. An alternative interpretation is that the presence of multiple copies of the spoIIID promoter region might titrate a limiting factor that is important for sporulation. To test this possibility, a multicopy plasmid, Pspac-spoIIIDA, was constructed (see Materials and Methods). This plasmid is identical to Pspac-spoIIID except it lacks the entire spoIIID open reading frame. A cotD-lacz fusion was introduced into wild-type cells containing Pspac-spoIIIDA. Sporulation was induced in the resulting strain by the resuspension method and IPTG was added at T3. Cells containing Pspac-spoIIIDA exhibited normal levels of cotD-directed B-galactosidase activity (data not shown) and heat-resistant spore production (Table 1). Thus, the presence of multiple copies of the spoIIID promoter region is not responsible for the effects of Pspac-spoIIID on cotD-directed B—galactosidase activity (Figure 1B) and heat-resistant spore production (Table 1). Another interpretation of the effect of Pspac-spoIIID on wild-type cells is that the copy of the spoIIID gene in Pspac-spoIIID bears a mutation and its product exerts a dominant negative effect on sporulation. To test this possibility, the spoIIID gene in Pspac-spoIIID was sequenced. No mutations were found (data not shown). Thus, 149 overproduction of SpoIIID and/or the resulting reduction in the level of (3'K appears to be responsible for reducing cotD expression (Figure 1B) and inhibiting heat-resistant spore production (Table 1). We attempted to alleviate the negative effect of the multicopy Pspac-spoIIID plasmid on (3'K accumulation and spore production by integrating the Pspac-spoIIID fusion into the bacterial chromosome at single copy. We constructed an integrational version of the plasmid, Pspac-spoIIID—cat (see Materials and Methods), and allowed it to integrate into the chromosome of wild-type and spoIIIDAerm cells by single- reciprocal recombination. The resulting strains produced normal numbers of heat-resistant spores when IPTG was added at T3 (data not shown). However, the level of SpoIIID as determined by Western blotting was not elevated late in sporulation (data not shown), but decreased as observed in wild-type cells ((Halberg & Kroos, 1992), Figure 1A). Therefore, integrating the Pspac-spoIIID fusion into the chromosome did alleviate its negative effect on spore production, presumably by reducing the copy number, but the resulting strain was not useful for altering the level of SpoIIID late in sporulation. Production of SpoIIID from Pspac-spoIIID increases sigK expression and reduces cotD expression, but does not affect cotA and gar! expression, in cells 150 engineered to produce 6! during vegetative growth. Since overproduction of SpoIIID reduced the level of OK accumulated in cells during sporulation, a strain (Pspac- PsigKesigKA19) which contains the spac promoter fused to the sigK'promoter and a truncated form of the sigK ORF (Oke & Losick, 1993) was employed to permit 0K production during vegetative growth. The primary translation product of the truncated ORF is OK (with an additional methionine on its N-terminus) rather than pro-OK. By supplementing transcription of sigK with Pspac and eliminating the need for processing to produce active a“, we hoped that overproduction of SpoIIID from Pspac-spoIIID would not reduce the level of OK. We reasoned that expressing SpoIIID in this strain might actually increase the level of OK because SpoIIID stimulates sigK transcription by GK RNA polymerase (Kunkel et al., 1988; Kroos et al., 1989) . To test this prediction, a sigK-lacz fusion was introduced into Pspac-PsigKesigKA19 cells containing either Pspac-spoIIID or the parental plasmid. The resulting strains were grown in 2xYT medium and IPTG was added during the mid-log phase. The levels of SpoIIID, sigKedirected B-galactosidase activity, and 0* were monitored. Cells containing Pspac-spoIIID produced SpoIIID, whereas 151 Figure 3. Levels of SpoIIID, sigK-directed B-galactosidase activity, and OK in Pspac-PsigK—sigKAlQ cells containing either Pspac—spoIIID or the parental plasmid. (Panel A) The level of SpoIIID in Pspac—PsigK-sigKAl9 cells containing Pspac-spoIIID or the parental plasmid (pDGl48), collected between 1 and 5 hours after IPTG induction was determined by Western blot analysis. (Panel B) sigKrdirected B- galactosidase activity detected in Pspac-PsigKesigKAl9 cells containing either Pspac-spoIIID (I) or the parental plasmid, pDGl48 (Cb. Points are the average of three determinations. The background level of B-galactosidase activity in cells not treated with IPTG was subtracted and did not exceed 2 units. Error bars indicate one standard deviation of the data. (Panel C) The level of (3'K in Pspac-PsigK-sigKA19 cells containing Pspac-spoIIID or the parental plasmid (pDGl48), collected between 1 and 5 hours after IPTG induction was determined by Western blot analysis. 152 A Pspac-spoIIID parental 1234512345 --~- -- -SpolllD B 2: 50- E 3.3-: 40‘ c 5: 82 so '33 2v 3 20‘ s 9' 0 1o . - . - . 2 4 6 Hours After Induction C Pspac-spoIIID parental I I 1234512345 153 cells containing the parental plasmid did not (Figure 3A). Producing SpoIIID in Pspac-PsigKesigKA19 cells increased sigKedirected B-galactosidase activity about two-fold between 4 and 6 hours after induction (Figure 3B), indicating that SpoIIID does stimulate sigK promoter activity in vivo. The level of 0x in cells containing Pspac-spoIIID is comparable to that observed in cells containing the parental plasmid (Figure 3C). Thus, these strains are suitable for testing the ability of SpoIIID to affect cotD, cotA, and gerE expression in vivo. cotD-, cotA- and gerE-lacz fusions were introduced into Pspac—PsigK-sigKA19 cells containing either Pspac-spoIIID or the parental plasmid. The resulting strains were grown in 2xYT medium and IPTG was added during the mid-log phase. The levels of cotD-, cotA- and gerE—directed B-galactosidase activity were monitored. Producing SpoIIID reduced cotD-directed B-galactosidase activity about two-fold between 4 and 6 hours after induction (Figure 4A), but had no effect on cotA- (Figure 4B) or gerE-directed B-galactosidase activity (Figure 4C). These results suggest that SpoIIID is capable of repressing cotD expression in vivo, but provide no evidence for an effect of SpoIIID on cotA or gerE expression. Additional evidence that cotD, but not cotA or gert, is repressed by SpoIIID. The prediction that SpoIIID would repress cotA, but not gerE, expression in vivo was based on preliminary in vitro transcription studies (L. Kroos & R. Losick, unpublished data). The results shown in 154 Figure 4. Levels of cotD-, cotA- and gerE-directed B- galactosidase activity in Pspac-PsigKesigKA19 cells containing either Pspac-spoIIID or the parental plasmid. cotD-, cotA— and gerE-directed B-galactosidase activity (panels A, B and C, respectively) detected in Pspac-PsigK— sigKAl9 cells containing either Pspac-spoIIID (I) or the parental plasmid, pDGl48 03). Points are the average of three determinations for cotD-directed B-galactosidase activity and the average of five determinations for cotA- and gerE-directed B-galactosidase activity. The background level of B-galactosidase activity observed in cells not treated with IPTG was subtracted and did not exceed 7, 5, or 25 units for cotD-, cotA— or gerE-lacz, respectively. Error bars indicate one standard deviation of the data. 155 A 4 1 H 3 1 1 ‘ ‘ mmmmmmmo 32211 3...: i=5: 5.384 omen—mosoaiué Hours Alter Induction t- .s.. . i . .o - o m m .... .. . 9...: .352. 3.2:; seep—eoaouiué Hours After Induction C .225 Base >=>=o< 2323823.: 4 6 Hours Alter Induction é 156 Figure 4B provided no evidence for a repressive effect of SpoIIID on cotA expression, so we repeated the in vitro transcription studies (Figure 5). SpoIIID markedly repressed cotD transcription and greatly stimulated sigK transcription (compare lanes 1 and 2), whereas it exhibited a slight repressive effect on cotA transcription (compare lanes 3 and 4) and gerE transcription (compare lanes 5 and 6). These results demonstrate that the ability of SpoIIID to repress cotA and gerE transcription is very weak relative to its ability to repress cotD transcription. A possible reason for the SpoIIID repression of cotA not appearing as strong in this study as compared to the previous one is that we used a different radioactively-labelled nucleotide (UTP), which provided a much stronger in vitro transcription signal. Expression of gerE (Figure 1D) and cotA (Figure 1C) consistently began one hour earlier during sporulation than expression of cotD (Figure 18) as measured by translational fusions to lacz. A similar result was observed when the levels of cotD and gerE mRNA were measured in sporulating wild-type cells by Northern blot analysis (quantitatively summarized in Figure 6). The probes were pSC146 (Zhengwet al., 1992), which contains a 263 bp AluI fragment of B. subtilis DNA spanning from 96 bp upstream of the gerE transcriptional start site to codon 35 of the gerE ORF, and pLRK100 (Kroos et al., 1989), which contains a 430 bp EcoRI-HincII fragment of B. subtilis DNA spanning from 225 bp 157 Figure 5. Effects of SpoIIID on cotD, sigK, cotA, and gerE transcription in vitro. Linearized plasmid DNA (1 pg) was transcribed with partially purified (5'K RNA polymerase alone (0.2 pg), or with SpoIIID (0.24 pg) added immediately after the addition of RNA polymerase. Run-off transcripts were electrophoresed in 5% polyacrylamide gels containing 8 M urea and were detected by autoradiography. cotD transcription from HindIII-digested pLRKlOO (225 base transcript) and sigK transcription from XbaI-digested pBK16 (170 base transcript) with 0'K RNA polymerase alone (lane 1) or with SpoIIID added (lane 2). cotA transcription from EcoRI-digested pKSZ3 (149 base transcript) with CiK RNA polymerase alone (lane 3) or with SpoIIID added (lane 4). gerE transcription from HindIII- digested pSC146 (204 base transcript) with (3'K RNA polymerase alone (lane 5) or with SpoIIID added (lane 6). . {cotD .7 .1 {sigK 158 159 Figure 6. Levels of gerE and cotD mRNA in sporulating B. subtilis. spo+ cells were sporulated by the resuspension method. RNA was prepared from cells collected between T4 and T8 and subjected to Northern blot analysis. The signals obtained for cotD (A) and gerE “3) mRNA were quantitated using a Visage Digital Imager and plotted as the percentage of the maximum level achieved during sporulation. Percentage of Maximum Level 1007 80" 60‘ 40‘ 20' 160 5.1 s 6 7 8 After Resuspension 161 upstream of the cotD transcriptional start site to codon 58 of the cotD ORF. A similar result to that obtained with the 430 bp cotD probe was also obtained utilizing a nick-translated 169 bp PstI fragment from pLRKlOO (Kroosem: al., 1989) which contains codons 4 through 59 of the cotD ORF. mRNAs of approximately 400 bases were detected with all three probes (data not shown). This observation is in good agreement with the size of gerE mRNA that was reported previously (Cutting & Mandelstam, 1986). The size of cotD mRNA suggests that it is monocistronic. gerE mRNA is first detected at T3 and its level increased significantly between T4 and T3 , whereas cotD mRNA is first detected at T5 and its level increased significantly between T5 and T6 (Figure 6). This difference in timing was observed in two additional experiments (data not shown). The delay in cotD expression may be caused by SpoIIID-mediated repression. 162 Discussion While attempting to determine whether SpoIIID represses cotD expression during sporulation, we found that overproduction of SpoIIID reduced the level of OK. This was surprising since SpoIIID plays a positive role in several aspects of am production, including the chromosomal rearrangement that generates the intact sigK gene (Stragierem al., 1989; Kunked.et al., 1990), the transcription of sigK initially by 0:15: RNA polymerase (Chapter II) and later by GK RNA polymerase (Kunkel et al., 1988; Kroos et al., 1989), and processing of pro-OK to OK (S. Lu & L. Kroos, unpublished data). Overproduction of SpoIIID does not appear to interfere with the initial transcription of sigK by 03 RNA polymerase, the chromosomal rearrangement that generates sigK, or translation of sigK mRNA to produce pro-OK. Transcription of sigK by (5'E RNA polymerase actually appears to be enhanced in cells overproducing SpoIIID since sigKedirected B-galactosidase activity is slightly elevated at T3 and T4 (i.e., the time interval when SpoIIID (Halberg & Kroos, 1992) and CE (Trempy et al., 1985b) are both present) as compared to wild-type cells (Figure 2A). This suggests 163 that SpoIIID is limiting for sigK transcription by GE RNA polymerase in wild—type cells. The chromosomal rearrangement and translation of sigK'mRNA appear to be unaffected by SpoIIID overproduction because the amount of pro-0'K present between T3 and T3 is comparable in wild-type cells and cells overproducing SpoIIID (Figure 23). Overproduction of SpoIIID does appear to interfere with the processing of pro-0’K to O“ and/or decrease the stability of OK in sporulating cells. The amount of (3'K present at T4 is lower in cells overproducing SpoIIID than in wild-type cells (Figure 2B). Deficient processing of pro-(3'K and/or decreased stability of 6'K during early times in sporulation would be expected to reduce the level of sigK’promoter activity during late times in sporulation (i.e., T5 to T8), as was observed (Figure 2A), because sigK is transcribed by (3'K RNA polymerase (Kunkel et al., 1988; Kroos et al., 1989) . Reduced sigK promoter activity during late times in sporulation would not necessarily result in lower pro-OKleNels in cells overproducing SpoIIID (e.g., a defect in processing would tend to elevate the level of pro-OK), but the total amount of pro-(5'K and 0'K present should be less than in wild-type cells, as was observed (Figure ZB). SpoIIID positively or 164 negatively regulates the expression of many genes, including bofA (Ireton & Grossman, 1992a; Riccaemzal., 1992), spoIIIA (Illing & Errington, 1991b), spoIIIG (Karmazyn-Campellien: al., 1989), spoIVB (Cutting et al., 1991a), and spoIVF (Cutting et al., 1991b), whose products influence pro—(SK processing (Cutting et al., 1990; Lu et al., 1990). The requirement for spoIIIA, spoIIIG and spoIVB, but not for spoIVF, in processing can be bypassed by bof mutations. However, a bof mutation did not overcome the negative affect of SpoIIID overproduction on (5'K accumulation (data not shown). Overproducing SpoIIID may reduce the level of 0’K by altering the expression of spoIVF, bofA, and/or other genes involved in pro-(IK processing and/or OK stability. Maintaining the proper levels of SpoIIID and (3'K during sporulation is critical because the overproduction of SpoIIID and/or the resulting reduction in the level of OK markedly reduced the formation of heat—resistant spores (Table 1). A lowered level of OK RNA polymerase is expected to reduce expression of genes involved in spore cortex (Cutting et al., 1991a) and coat formation (Zheng & Losick, 1990; Zheng et al., 1992). Thus, altered expression of many genes in the mother cell presumably contributes to the reduction in heat- resistant spore formation by cells overproducing SpoIIID. 165 Normally, SpoIIID production, like OK production, is subject to multiple levels of regulation. First, the synthesis of spoIIID mRNA appears to be positively autoregulated since spoIIID-directed B-galactosidase activity is reduced 3- to 6-fold in a spoIIID mutant background (Kunked.et al., 1989; Stevens & Errington, 1990). Second, the synthesis and/or stability of spoIIID mRNA is negatively regulated by the production of (3'K (Halberg & Kroos, 1992). Third, spoIIID mRNA is polycistronic, and translation through the first ORF appears to be required for translation of the spoIIID ORF since a mutation in the ribosomal binding site of the first ORF reduces the expression of SpoIIID-dependent genes (A. Bosma & R. Losick, unpublished data). Fourth, SpoIIID appears to be rapidly degraded in the mother cell during late times in sporulation since a decrease in the level of spoIIID mRNA is paralleled by a decrease in the level of SpoIIID (Halberg & Kroos, 1992). The degradation of SpoIIID may involve the conversion of SpoIIID to a less stable intermediate (Chapter IV). This conversion appears to be developmentally regulated and involves the removal of amino acids from the C-terminal portion of SpoIIID (Chapter IV). All of these levels of regulation provide potential mechanisms for connecting the level of SpoIIID, and hence the expression of many mother-cell genes, to genetic regulatory, physiological, and morphological cues during the sporulation process. 166 ‘We used cells engineered to make (5'K during growth to demonstrate that SpoIIID can stimulate sigK and inhibit cotD expression, but does not affect cotA or gerE expression in vivo (Figures 3 and 4). The effects of SpoIIID on the transcription of all four genes tested in this in vivo system are, at least qualitatively, consistent with the results of in vitro transcription assays (Figure 5). This system provides a convenient method for examining the effects in vivo of SpoIIID on expression of OK-dependent genes. The slopes of the curves are noteworthy because they may reflect accumulation of GerE in cells. GerE stimulates cotD transcription (Zheng & Losick, 1990; Zheng et al., 1992) and cotD-lacZ expression increased with time after induction of OK production (Figure 4A), whereas GerE inhibits cotA transcription (Sandman et al., 1988; Cutting et al., 1989; Zheng et al., 1992) and cotA-lacz expression decreased with time after induction (Figure 48). Expression of gerE-lacz (Figure 4C) and sigK-lacz (Figure 3B) changed little with time, consistent with the absence of a gerE effect on the expression of these fusions (Kunkel et al., 1988; Cutting et al., 1989). The sigK-lacz fusion lacks a site for GerE binding that mediates the strong repression of sigK transcription by GerE observed in vitro (H. Ichikawa and L. Kroos, unpublished data; Zheng et al., 1992) . While our results demonstrate that SpoIIID can repress 167 cotD transcription in cells engineered to produce (3'K during growth, the question of whether SpoIIID represses cotD transcription during sporulation, thus altering its time or level of expression, remains unanswered. The finding that SpoIIID can repress cotD transcription by (3'K RNA polymerase in vitro originally led to the proposal that SpoIIID represses cotD transcription for a period during sporulation (Kroos«et al., 1989). We subsequently showed that induction of a cotD- lacZ fusion coincides with a decrease in the level of SpoIIID during sporulation in wild-type cells (Halberg & Kroos, 1992). Furthermore, cotD-lacz induction coincides with a decrease in the level of SpoIIID but lags behind an increase in the level of (5'K in cells engineered to produce (3'K earlier than normal; this was shown not to be an effect of GerE (which stimulates cotD transcription), suggesting that cotD-lacz induction was delayed until the level of SpoIIID decreased (Halberg & Kroos, 1992). In contrast, the induction of cotA- and gerE-lacz fusions coincides with the 0K increase in these cells and precedes the SpoIIID decrease (R. Halberg and L. Kroos, unpublished data), suggesting that transcription of cotA and gerE is not subject to repression by SpoIIID. Consistent with this idea is the finding that SpoIIID's ability to repress cotA and gerE transcription by GK RNA polymerase in vitro is weak relative to its ability to 168 repress cotD transcription (Figure 5). Moreover, the major increase in cotA- and gerE-lacz expression during sporulation occurs one hour earlier than for cotD-lacz ((Oke & Losick, 1993); Figure 1) and the appearance of gerE mRNA also precedes the appearance of cotD mRNA by about one hour (Figure 6). All these results are consistent with the model that SpoIIID represses cotD transcription, but not cotA or gerE transcription, for about one hour during sporulation. If SpoIIID does repress cotD transcription, but not cotA or gerE transcription, during sporulation, one might have expected an elevated level of SpoIIID to exert a more pronounced negative effect on cotD-lacz expression than on cotA or gerE-lacz expression; however, this was not observed (Figure l). Apparently, even the elevated level of SpoIIID is insufficient to markedly reduce cotD transcription late in sporulation (i.e., at Tg'to T3). The timing of cotD-lacz expression was unaffected by SpoIIID overproduction, remaining about one hour later than cotA- and gerE-lacz expression (Figure 1). Thus, it remains possible that SpoIIID prevents premature transcription of cotD for a short period during sporulation while the levels of (3'K RNA polymerase and GerE are low (Halberg & Kroos, 1992). An alternative explanation for the later expression of cotD is that a higher threshold level of 0’“ may be required for cotD transcription than for transcription of cotA and gerE (Oke & 169 Losick, 1993). cotD is expressed about 1 to 2 hours later than gerE when 0K production is induced with IPTG during growth in cells containing Pspac-PsigKesigKAIQ (Oke & Losick, 1993). Since SpoIIID is absent in this case, it cannot account for the difference in timing. We have mapped a site for SpoIIID binding in the cotD promoter region (Chapter II). Perhaps mutations that eliminate SpoIIID binding in the cotD promoter region will enable us to determine whether SpoIIID affects cotD transcription during sporulation. SpoIIID is just one component of a molecular switch proposed to govern the change in the mother-cell pattern of gene expression during the transition from morphological stage IV (cortex formation) to stage V (coat formation) (Halberg & Kroos, 1992; Zhengwet al., 1992). Two other components are (5'K RNA polymerase and GerE. Processing of pro-(5'K to (3’K appears to activate the switch, resulting in a falling level of SpoIIID due (at least in part) to a falling level of spoIIID mRNA (Halberg & Kroos, 1992) and a rising level of GerE due to transcription of gerE by (3'K RNA polymerase (Zhengemzal., 1992). SpoIIID and GerE exert opposite effects on transcription of certain genes in vitro. For example, transcription of the earlier-expressed sigK gene is stimulated by SpoIIID (Kunkel et al., 1988; Kroos et al., 1989, Figure 5) and inhibited by GerE (Zheng et al., 1992), 170 whereas transcription of the later-expressed cotD gene is inhibited by SpoIIID (Kroos et al., 1989, Figure 5) and stimulated by GerE (Zheng et al., 1992) . Hence, both a falling level of SpoIIID and a rising level of GerE would change the pattern from sigK to cotD transcription by (JrK RNA polymerase. Why the need for this elaborate switch involving dual control by SpoIIID and GerE? One clue comes from the observation that other cot genes (encoding spore coat proteins) may also be subject to dual control by SpoIIID and GerE. Transcription of cotC and cotX'by 0'K RNA polymerase, like that of cotD, is inhibited by SpoIIID and stimulated by GerE in vitro (H. Ichikawa & L. Kroos, unpublished data). We speculate that differing arrangements and affinities of binding sites for SpoIIID and GerE in cot promoters may influence the time and level of production of different spore coat proteins to facilitate assembly of the multilayered spore coat. In summary, we have shown that overproduction of SpoIIID reduces the level of (5'K during sporulation and reduces cotD expression during vegetative growth. Overproduction of SpoIIID probably reduced the level of (3’K during sporulation by altering the expression of genes encoding proteins involved in pro-(3K processing and/or O’K stability. The reduced level 171 of OK lowered the expression of all 03-dependent genes tested. Maintaining the proper levels of SpoIIID and UK is crucial for normal sporulation because the overproduction of SpoIIID and/or the resulting reduction in the level of OK reduced the production of heat-resistant spores. This is probably why the production of both SpoIIID and OK is normally subject to multiple levels of regulation. The problem of SpoIIID overproduction reducing the level of (5'K during sporulation was circumvented by engineering cells to produce (3’K during vegetative growth. Expressing SpoIIID from Pspac-spoIIID in this strain increased sigK promoter activity and reduced cotD promoter activity, but did not affect cotA or gerE promoter activity, in agreement with the effects of SpoIIID on transcription of these promoters in vitro. Clearly, cotD is expressed about one hour later than gerE and cotA during sporulation, but measuring the relative contributions of a falling level of SpoIIID and a rising level of GerE to the timing of cotD expression, and understanding the biological significance of this finely tuned regulation, will require further investigation. CHAPTER VI Conclusions “The end justifies the means only when the means used are such as actually bring about the desired and desirable end.” John Dewey 172 173 The goal of my research was to characterize the role of SpoIIID during Bacillus subtilis sporulation. Experiments were performed to determine the transcriptional properties and fate of SpoIIID. Genetic studies suggested that SpoIIID affects the transcription by 03 RNA polymerase (reviewed by Errington, 1993). To determine whether this was a direct effect, in vitro transcription assays were performed. SpoIIID stimulated spoIVCA and sigK transcription, but represses bofA transcription, by GE RNA polymerase. DNase I footprinting revealed that SpoIIID binds to specific sequences in the promoter regions and open reading frames (ORFs) of these genes. Other studies demonstrated that SpoIIID stimulates sigK transcription, but represses cotD transcription, by'cm RNA polymerase in vitro (Kroos et al., 1989). Consistent with this observation, the presence of SpoIIID reduced the level of B-galactosidase generated from a cotD-lacz fusion in cells engineered to produce (3'K during vegetative growth, indicating that SpoIIID can repress cotD expression in vivo. SpoIIID binds to the promoter region and ORF of both sigK and cotD. However, SpoIIID binding in the promoter region is sufficient to mediate the transcriptional effects observed in vitro. Comparison of the sequences in the SpoIIID binding sites of (5'E and OK-dependent genes revealed a putative consensus, 174 WWRRACAR-Y. This consensus sequence is centered at -28 and -27 in spoIVCA and sigK, respectively, but at -33 in cotD, suggesting that SpoIIID binds on opposite sides of the DNA helix (relative to RNA polymerase) in promoters where it exerts opposite effects on transcription. The spatial orientation of SpoIIID relative to RNA polymerase would be expected to affect the interaction between these proteins. In the case of spoIVCA and sigK, SpoIIID may interact with the a or 6 subunits of RNA polymerase, as has been observed for other transcriptional activators. In the case of cotD, SpoIIID may block the interaction between the -35 region of the cotD promoter and (5'K RNA polymerase. Additional genetic and biochemical experiments may provide further insight into how SpoIIID acts as a transcriptional activator and repressor. Based on the observation that SpoIIID activates and represses transcription by 0'K RNA polymerase, it was proposed that the inactivation and/or sequestering of SpoIIID establishes a switch in the mother-cell pattern of gene expression. Western blot analysis demonstrated that the level of SpoIIID does decrease at the appropriate time during sporulation to produce such a switch. This decrease is dependent upon the production of OK. Northern blot analysis revealed that the production of (3'K reduced the synthesis 175 and/or stability of spoIIID mRNA. This could be a direct or an indirect effect. OK may compete with CE for binding to core RNA polymerase and thereby reduce the level of 03 RNA polymerase. Consistent with this idea, OE appears to be a labile protein. Alternatively, 0K RNA polymerase may transcribe a gene whose product represses spoIIID transcription. These possibilities can be distinguished by mutating sigK'such that the 0* produced can bind to core RNA polymerase but is transcriptionally inactive, or it can bind to DNA but is unable to bind to core RNA polymerase. The parallel between the decrease in the levels of spoIIID mRNA and SpoIIID protein, suggests that a mechanism exists for degrading SpoIIID in sporulating cells. Consistent with this idea, SpoIIID appears to be converted to a less stable 9 kDa form by removing 7 amino acids from its C-terminus, based on N-terminal sequencing and mass spectral analysis. However, additional mass spectral experiments are required to determine the C-terminus of the 9 kDa protein conclusively and a pulse-chase experiment is required to demonstrate a precursor-product relationship between SpoIIID and the 9 kDa protein. The conversion of SpoIIID to the 9 kDa protein is developmentally regulated. Interestingly, mutations in genes required for the formation of the spore cortex inhibit the production of the 9 kDa protein, 176 suggesting the conversion may be coupled to spore morphogenesis. For example, the conversion is blocked by a mutation in spoIVA, which encodes a novel 50 kDa protein that is required for cortex and coat formation. It remains to be determined if the SpoIVA protein is responsible for the conversion. A mobility shift system is ready to monitor the levels of SpoIIID and the 9 kDa protein during attempts to purify the convertase. Specific proteolysis plays a role in controlling gene expression in other systems. The proper level of SpoIIID is crucial for normal sporulation. Cells engineered to produce an elevated level of SpoIIID throughout sporulation produced fewer heat- resistant spores. This effect appeared to result from a reduction in the level of (3'K and OK-dependent gene expression. Several experiments were performed to determine how the overproduction of SpoIIID reduced the level of OK. Unfortunately, the results were inconclusive. Sigma factors, anti-sigma factors, and DNA-binding proteins are present in other organisms. Continued characterization of the role of SpoIIID during B. subtilis sporulation may reveal basic mechanisms by which DNA—binding proteins regulate gene expression during development. APPENDICES APPENDIX A Processing of the Mother-Cell 0 Factor, 0“, May Depend On Events Occurring in the Forespore During Bacillus subtilis Development 177 Proc. Natl. Acad. Sci. USA Vol. 87. pp. 9722-9726. December 1990 Genetics Processing of the mother-cell a- factor, a 178 K, may depend on events occurring in the forespore during Bacillus subtilis development (gene ”Instant/INA palm/MIMI“) Sure Lu. chrrxao HALaaao. mo Lee Kaoos‘ Department of Biocheaustry. Mich." State University. East um Ml 48824 Cmrrnrunlrvrrcd by Dale Kaiser. September l7. l9” ABSTRACT Dru-lagsporulatleaeftheGr-an-posltlve bee. terlumBacillassabrilis.traaaerlptlenofgeaeseneodlrrgspere eoatprotelnslatheasother-celeorapartraeateftheape- raaglumlseoatrolledhyRNApolyraeraaeeeataHagtherr anhualtealledrr".8asedeaeemparlseaoftbeN-termhal aminoacldseqeeneeefc"wlththeaacleotidesequeaeeefthe geaeencodlagc"(sigl().theprlrnarypreduetefsigl(wu hferredtobeapre-protein(pro—o")w1th20ertraamleeaelds attheNtermlaas.Uslngaatbodiesgeneratedagalastpro—rr". wehavedetectedlprw‘beglaalagatthethlrdhouref sporulatloaaada heglnalagaboutlhrlaterJ-Zveawhea pro-«“hexpressedartlflcidydurlaggrewthaedtbroughed qrorulatlon.o‘appearsstthenormaltlmeaadexpreaslosef a rr"-eoatrolled geaeoecers normally. These resulssuggeet thatpro-rr"haalaactiveprecursorthstbpreteolytleally precessedteacttvec‘hadevelop-eatallyregalatedfashles. Mststlonsthatblockferesporegeaeespredoahleehateuraa- letlonefc‘butaotaecuraalatleaefpre-c".aauesthgthst pro-c‘prece-lnglsaregdatorydevleethateaquesthe WNWWWhMQNh the nearer-eel at more east protein that w! eaeaae the fercpere. Upon starvation the Gram-positive bacterium Bacillus sub- rr'lr's undergoes a series of morphological changes that result in endospore formation (1). The first easily observed mor- phological change is asymmetric septum formation. Which divides the cell into two compartments. the mother-cell and the forespore. each receiving a capy of the genome. A complex regulatory circuit ensures the correct temporal and spatial pattern of gene expression during sporulation. Critical to this regulatory circuit are the synthesis and activation of 0 subunits of RNA polymerase that direct the enzyme to transcribe different gene sets (2). Two a factors are compart- ment-specil'rc. 47°. the product of the spoIIIG gene. is pro- duced predominantly. if not exclusively. in the forespore and controls the expression of forespore-specific genes (3-5). The counterpart to 0° in the mother-cell is c" (6). which controls the expression of mother-cell-specifrc genes such as cotA (7). cotD (8). and gcrE (9) (referred to as the MM regulon). The cotA and cotD genes encode spore coat proteins that assem- ble on the forespore surface (10). and gcrE encodes a regulator of spore coat synthesis (11). o" is encoded in a composite gene (sigK) generated by a mother-cell-specific chromosomal rearrangement that joins two loci. spolVCB (encoding the N-terminal portion) and spollIC (encoding the C-terminal portion) (12. 13). Transcrip- tion of the sigK promoter is also confined to the mother-cell (14) and compartmentalization of both the DNA rearrange- 1hepeblieuioaeostsofthisanieleweredefrayediapanbypagecharge paymeatJhisarticlearustrhereforebeherebymarted “Mmismnr” is aeeortaee with 18 U.S.C. 11734 solely to indicate this fact. 9722 meat and sigK transcription appear to result from mother- ceil-specifrc expression of spoIIID (15). A third possible regulatory mechanism for 0" was inferred from a comparison of the N-terminal amino acid sequence of 0" and the nucle- otide sequence of sigK (6. 12) and by analogy to 0‘. a sporulation-specific B. subtilis a factor that is activated by proteolytic processing (16-18). The primary product of sigK was predicted to be a pro-protein (pro-o") bearing 20 extra amino acids at the N terminus. Here we present evidence that 0" is first made as an inactive precursor and is processed to the active 0 factor in a developmentally regulated fashion. Furthermore. mutations in forespore regulatory genes (e.g.. spoIIIG. encoding the forespore a factor. 0°) appear to block processing of pro-0" to 0". suggesting that the previously noted dependence of mother-cell-specifrc gene expression on forespore events (7-9. 14) is mediated at the level of prote- olytic activation of the mother-cell a factor. MATERIALS AND METHODS Bacterfl Stu-flu. Escherichia coli strain A6115 (morn. Mara. Icai7697. AlacX74. galU ’. galK ‘ . Irsr‘. lisnr’. ser. (F'. proAB. lacl'ZzzTnSH was'obtained from A. Grossman. B. subtilis strains were obtained from R. Losick. liubrr’lls cells were made competent (19) and transformants were selected on Luria-Bertani (LB) W (20) with kanamycin sulfate at 5 ug/rnl. Use of the specialized transducing phage SPfixcorD-lacl (obtained from L. Zheng and R. Losick) has been described (8). Construction of Moulds. All plasmids were derived from pSKS. which contains sigK (13). and from pDGl48 (18). which permits isopropyl B-o-thiogalactopyranoside (IPTG)- inducible expression of an inserted gene from the P.., promoterlZl) in E. coli or B. subtilis. To fuse sigK expression to the P..... promoter. a 1.4-kilobase-pair (kbp) Ssp l—Hindlll fragment from pSKS (including 141 bp upstream and 556 bp downstream of the sigK open reading frame) was ligated to Hindlllcdigested pDGl48. the unligated end of the vectorwas made blunt using the till-in reaction of Klenow enzyme. and ligation was continued (20). Ampicillin-resistant E. coli trans- formants were obtained (20) and the structure of pSLl. a plasmid containing the insert in the proper orientation to fuse sigK transcription to P . was verified by restriction map- ping. pSLZ and pSlA were derived from pSLl and pDGl48. respectively. by deletion of the EcoRI fragment containing the origin of replication and the kanamycin-resistance gene that function in B. subtilis. Pr-eductlea ef Pro-a" and of Antibodies. To produce pro-a" in E. coli. strain ESLZ (strain A6115 con- taining pSL2) was induced with 1 mM IPTG during the late logarithmic phase of growth at 37°C in LB medium (22). One hour after the addition of IPTG. cells were harvested by Abbreviations: IPTG. isopropyl fi-o—thiogalactopyranoside; '1'... hour a of sporulation. 'To whom reprint requests should be addressed. Genetics: Lu er al. centrifugation (6 min. 7500 x g). The cell pellet was resus- pended in 0.05 volume of sample buffer (0.125 M Tris-HCI. pH 6.8/296 SDS/5% 2-mereaploethanol/10% glycerol/0.1% bromophenol blue) and the sample was boiled for 5 min to produce a whole-cell extract. From 0.5 ml of extract. ~1Nug of pro-0" was purified by preparative SDS/PAGE (IO-15% polyacrylamide gradient) and electroelution. Pro-or" (65 us) was precipitated with acetone. dissolved in phosphate- bulfered saline (23). emulsified with Freund‘s complete adju- vant (BRL). and injected into or near the popliteal gland of a New Zealand White rabbit. Three weeks later a booster injection [30 pg of pro-r7" emulsified with Freund's incom- plete adjuvant (BRLH was given at the same site. The rabbit was bled 1 week after the boost and serum was prepared (23). and Western Blot Analysis. Sporulation of B. subtilis was initiated by nutrient exhaustion in Difco sporu- lation (08) medium (24) as described (7). Cells were har- vested by centrifugation (5 min. 16.0“) x g) and whole-cell extracts were prepared (22). Extract protein was quantitated by the Bradford method (25). After addition of 0.5 volume of 3x sample buffer. proteins were separated by SDS/12.5% PAGE and electroblotted to a poly(viny1idene difluoride) membrane (26). The membrane was incubated in TBS (20 mM Tris-HCI. pH 7.5/0.5 M NaCl) with 2% nonfat dry mill: for 4 hr at room temperature with shaking to block nonspe- cific antibody binding and then incubated overnight at room temperature with shaking in polyclonal antiserum diluted 1:2000 into TBS/2% nonfat dry milk/0.05% Tween 20. lmmunodetection using a goat anti-rabbit alkaline phospha- tase conjugate was performed according to the manufactur- er‘s instructions (Bio-Rad). RESULTS Antbodlestehe-c" Deteetho-o‘"andc"h8pordatlag B. srrhrllls. To purify pro—0" for the generation of a polyclonal antiserum. the protein was expressed in E. coli. Transcription of sigK (encoding pro-r7") was fused to an IPTG-inducible promoter in plasmid pSL2. Whole-cell extracts of IPTG induced and uninduced E. r all containing pSL2 or a control plasmid (pSU. which does not contain sigK) were analyzed by gel electrophoresis (Fig. 1). A protein of the expected mobility for pro-0" [-29 kDa. since pro-c" is predicted to contain 20 amino acids at its N terminus that are absent from rr". which migrates at 27 kDa (6)] increased upon IPTG induction of cells containing pSLZ. but not upon IPTG induction of cells containing the control plasmid. This protein wasassumedtobepro-o". the predictedprirnarytranslation product of sigK. since oSLZ contains no other 29-kDa. 1234 .A -W ”1'11 1‘ ’ - ' - — p 1" FIG. 1. Production of pro-o" in E. coli. Proteins in whole-cell extractstloul)ofll"l‘G-induced(lanes2and4landuninduced (lanes 1 and 3) E. coli were separated by SDS/PAGE(10—15% polyacryl- amide gradient) and visualized by Coomassie blue staining. Strains 881.4le 1 andZiand ESLZlIanes3and4)wereeonstnrctedby transformation of strain A0115 with the control plasmid (951.4) and the Pw-sigK fusion plasmid (pSL2). respectively Only the 35- to 25-ltDaregionofthegelisshown: thepositionsofaMDamarlter proteintcarbonic anhydraselandtheproteinassurnedtobepro-e" are i 179 Proc. Natl. Acad. Sci. USA 87 ((990) 9723 protein-encoding open reading frames downstream of the IPTG-inducible promoter. Antibodies to “gel-purified pro-c“ were generated in a rabbit and used in Western blot analyses. The antibodies detected pro-or" and one larger protein in a whole-cell extract of IPTG-induced E. coli containing pSL2 (Fig. 2A. lane 2). while only the larger protein was detected for cells containing the control plasmid without sigK (lane 1). The antibodies. referred to hereafter as "anti-prov" antibodies." easily detected 15 ng of pro-0'" gel-purified from E. coli (Fig. 28. lane 1). The anti-prov" antibodies also recognized o" (6) gel-purified from B. subtilis (Fig. 28. lane 2) and these antibodies detected either pro-tr" or a" with similar sensi- tivity. The antibodies detected proteins that comigrated with " and o" in a whole-cell extract of sporulating B. subtilis (Fig. 28. lane 3). while these proteins were not detected in extracts of growing B. subrilis (lane 4) or in extracts of sigK mutants (i. e.. spolVCB or spolllC mutants: see below). Thus. Western blot analysis using the anti-prov“ antibodies provides a sensitive assay for the level of pro-0" and or“ in B. subtilis. Levels of I’m-c" and tr" Are Developmentaly Registed. To examine the levels of pro-r7" and o" in B. subrilis at various times during sporulation. cells were harvested at hourly intervals during growth and sporulation in OS me- dium. Under these conditions. the end of exponential growth defines the initiation of sporulation (T4,). prespores that appear gray in the phase-contrast microscope begin to appear 4 hr later (T4). and phase-bright free spores (released by mother-cell lysis at the end of sporulation) begin to appear at T.. Whole-cell extracts were subjected to Western blot analysis using the anti-pro-rrK antibodies and the result for the Spo° strain PY79 (27) is shown in Fig. 3. A similar result was obtained for the Spo° strain 8638 (28) (data not shown). Pro-r7" was first observed at 3 hr rnto the sporulation process (1‘ ,) reached a maximum atT ri' and then decreased to a barely detectable level by To it was first observed at T. (1 hr later than pro-o "). increased to a maximum 51.13.“ decreased thereafter. These results demonstrate that the levels of pro-0" and a“ are regulated durirg sporulation. Since the appearance ofpro-a" precedes the appearance of 0" and since the“ N terminus of 0 “corresponds to codon 21 of suit (6 12). a "maybederived from pro-0" by proteolytic "W335!“- Mutations b Many Sperdstloa Genes Bloch AM efrr". MutationsatmanydifferentlociintheB. subtilis A B 1 2 1 2 3 4 . g "h '. :2 ‘3 3." :3..’.‘~';.( rev: 1'9: . . W fat". 51;” ‘, ‘- ’7." ,1;~::"“..£. . . .é. - J P -- — Frc. 2. Characterization of the anti-pro-o" antiserum by West- ern blot analyses. (A) Whole-cell extracts (1 pl ofa 1:11!) dilution) from IPTG-induced E. coli strains ESL4 (lane 1) and ESL2 (lane 2). containir' the control plasmid (981.4) and the PfinsigK fusion plasmid (pSL2). respectively. were prepared as described for the production of pro-o“ (see Materials and Methods). (B) Pro-tr" (15 u) from E. mlr‘ (lane 1) and or" (15 ng) from sporulating B. sfirllis (lane 2) were gel-purified. Whole-cal extracts (lougof protein) were fromB.srrbrilirharvestedduringyowth(lane4)andat6hrlnto sporulation (lane 3) in 05 medium. ..- 9724 Genetics: LII (t al. T..g T0 T. T! T, T‘ T. T. T, T. 'tzvuefw s. 1'! bf P4 ' 4 f t 12345678010 Fro. 3. Pro-or" and a“ in sporulating B. subtilis. Wild-type strain PY79 (27) was harvested at hourly intervals durir' growth and sporulation in 08 medium. Whole-cell extracts (10 pg of protein) were subjected to Western blot analysis using the anti-prov“ antibodies. Lanes 1-10. samples harvested at hourly intervals be- ginning 1 hr before the end ofexponential growth (T4) and ending 8 hr into sporulation (T .l. Pro-tr" gel-purified from E. coli served as amatkerontheblotandthe inferred positionofo" is indicated. genome block or reduce expression of genes in the tr"- controlled cotA regulon (7-9). To further investigate these ell‘ects. 16 mutants with altered cotA regulon expression were examined for pro-0" and <7" by using the anti-prov“ anti- ' bodies in Western blot analyses. Samples collected at hourly intervals from T. through T1 in 08 medium were tested for each mutant. but only the result at T.. the time when both pro-tr" and a“ are abundant in the Spo' strains (see above). is shown in Fig: 4 for each mutant. Five mutants accumulated neither pro-a nor 0" (lanes 3. 5. 6. 12. and 13). These mutants have either a mutation in the sigK gene [spolVCB artd spolllC (12)] or a mutation in a gene whose product is essential for the chromosomal ment that generates sigK [spoIIGB. spoIIID. and spell/CA (12. 13)]. Nine mu- tants accumulated pro-0" but not 0" (lanes 1. 2. 4. 7. 8. 11. 14. 15. and 16). Interestingly. this group includes 4 strains with mutations in genes required for forespore-specific gene expression [spolllA. spolllE. and spoIIIG (32. 33)] and/or in genes expressed predominantly. if not exclusively. in the forespore [spoIIIG (4. 5) and spoIVB (S. Cutting and R. Losick. personal communication". suggesting that accumu- lation of a" in the mother-cell compartment of the spo- rangium depends on events occurring in the forespore corn- partrnent. In addition. this group includes 2 mutants (spoIIB and spollD) blocked early in sporulation at the stage of asymmetric septum formation and 3 strains with mutations in the spalVF locus. which is not required for expression of a 180 Proc. Natl. Acad. Sci. USA 87 (I990) forespore-specific gene (32). Finally. 2 strains with mutations in spoil/A accumulated a normal amount of pro-tr" but accumulated much less (7" (lanes 9 and 10) than the wild-type strain. The spolVA mutants express the cotA regulon at a reduced level. whereas all the other mutants examined in this study fail to express the cotA regulon (7—9). Thus. for all the mutants examined. the impaired cotA regulon expression observed previously (7-9) may be due to impaired accumu- lation of a“. If a“ is derived from pro-0" by proteolytic processing as suggested above. at least eight loci (spollB. spallD. spoIIIA, spoIIIE. spoIIIG. spolVA. spoIVB. and spalVF) may be directly or indirectly involved in processing pro-0" and/or stabilizing a". ProceBlngofPro-o"toa"lsReqolredtoProduceanActlve tr Factor and Is Developmentally Regulated. In vitro and in viva approaches were used to address whether pro-er" can . direct transcription of a"-controlled promoters. For the in ‘ 't't‘tm approach. pro-0" gel-purified from E. coli was tested for its ability to direct transcription of the sigK [previously called spoIVCB (6. 14)] and cotD (8) promoters upon addition to B. subtilis core RNA polymerase (6). As a positive control. tr" (60 ng) partially purified from sporulating B. subtilis was eluted from a gel. renatured (34). and added to core RNA polymerase (60 ng). The reconstituted enzyme produced a ntn-off transcript from the sigK promoter in the presence of the SpolllD protein (120 ng) and from the cotD promoter in the absence of SpoIIID (data not shown). as shown previ- ously (6). Under these conditions. pro-o“ (300 ng) failed to direct transcription of the sigK promoter in the presence of SpoIIID (120 ng) and also failed to direct transcription of the cotD promoter in the absence of SpollID (data not shown). These results suggest that pro-0" is inactive as a a factor. To determine whether pro-0" could direct transcription of a c"-controlled gene in viva. we used a multicopy plasmid bearing sigK fused to an IPTG-inducible promoter to express pro-0" in B. subtilis during growth and sporulation and a cotD-larZ fusion (8) to monitor‘lhe transcriptional activity of a 0"-controlled promoter. A sigK mutation prevented pro- duction of pro-a“ or a“ from the chromosome in this exper- iment. Production of pro-or" from the plasmid was induced with IPTG ~2 hr before the end of exponential growth. and samples collected at hourly intervals were tested for B-ga- laetosidase production from the cotD-lacZ fusion and were also subjected to Western blot analysis using the anti-pro-a" antibodies (Fig. 5). Even though a large amount of pro-tr" was present 1 hr prior to the end ofexponential growth (T-.) and throughout the early stages of sporulation (Western blot. Inset). cotD-directed B-galactosidase activity remained low 3101112131415 10 ‘ . 3“.- . ‘ - - ‘o:r . - ‘ t i ‘ “A . " 3‘ r . " . r. - ' r I ,..:._ . 44.»..w , '- a l -' ' "-‘ l ‘l or.- l". 1 ' - ' I.'.' ‘ 0. "£ a I 't.‘ - , j ‘ . ‘ -- . , ‘.~ -2 ..--9‘- - .. - I - 'rs . . l- .3 . ' . ..rft' .33 J ‘ I ' '. . Fro. 4. Pro-0" and 0" in B. subtilis sporulation nartants harvested6 hrafler the end of exponential growth in 08 medium. Whole-cellextracts (lougofproteinlweresubjectedtoWestemblotanalysisusirtgtheuni-pro-a‘antibodies.Arrowsindieatethepositionolpro-o".whichserved as a marker on the blots: the inferred position of 0" is also indicated. Lanes: 1. strain 131.5 (spoIIBIJI. trpCI): 2. K82” (sprtllbzt Ta9l7flflU298): 3. K8440 (spoIIGfl ): 4. K813 (spolllA ::Tn9l7(1HUI3): 5. BK410(spollIC94). 6. BK395 (spolllwlz 7. SC622(spollIE16): 8. BK338 (spolllGAl ): 9. K8194 (spolVA :tTn9l7flllUl94); 10. strain 67 (spalVA67. trpC2); 11. BK750 (sprerBizermG); 12. BKSSB (spell/CA I”): 13. BK556 (spoIVCBZJ): 14. SC834 (spoIVFISZ): 15. K8301 (spull'FzzTn9l7flIlU30I): 16. K8179 (spolVFzthr9l7flllUI79). These strains are isogenic to PY79 (27). except 131.5 and 67 are isogenic to 8038 (28). These strains have been described (7. 14. 15. 29). except BK750 was constntcted by transformation of DNA prepared from 1H12719 (30) into PY79 with selection for the a,“ ,-L.. ' ‘ pne (ermG) inserted in the spulVB gene (B. Kunkel and R. Losick. personal communication) and K8301 has Tn917 inserted in the sprerF locus (8. Cutting artd R. Losick. personal communication). fire Tn9l7 insertions HUIW and ”U!” were thought todefrne new loci designated spaVP and sanL. respectively (31). but are now assigned to the indicated loci (8). Genetics: Lu et al. T-t To Tr Ta Ta Tr 1's Ta Ti 3 « ...--—---— W0“ "\ an N d a p-Galactosldaaa activity. Mlllar units it 10" O 2 4 B 8 Tlmo.hr Fro. 5. Effect ofprodueing pro-0" frornaplasmidduring growth and sporulation of B. subtilis. Strain BK410 (sputum: ref. 15) was transformed with the P..-sigK fusion plasmid (pSLl ) or the control plasmid (pDG 148). and both resulting strains were lysogenized with phage SPBScutD—larl. resulting in strains BSL3 and BSL4. respec- tively. The Spo' strain PY79 (27) was also lysogenized with SPBIIrutD-lm-Z. resulting in strain BSLS. Cells were grown and sporulated in 08 medium with the addition of 1 mM IPTG -2 hr before the end of exponential growth. Samples were harvested at hourly intervals and B-galactosidase activity was determined (35) using the substrate o-nitrophenol B—o-galactoside. One unit of en- zyme hydrolyzes l umol of substrate per min perODM unit of initial cell density. Background activity (ranging from 0.5 to 6 units) of PY79 at each time point was subtracted from the values obtained for strains containing the (‘rrlD—ldl'l fusion. cotD-directed B—galacto- sidase activity was determined for strains BSL3 (o). BSL4 (o). and 851.5 (A). (Inset) Western blot analysis of whole-cell extracts (10 pg of protein) of strain BSL3. using the anti-prov“ antibodies. until T. (o). Pro-0" produced in B. subtilis appears to be inactive as a a factor. unless the presence or absence of another regulatory factor(s) prevents cotD transcription dur- irtg growth and early in s lation. Beginning at T.. and rrtore noticeably at T,. a was observed by Western blot analysis and cotD-directed B-galactosidase activity increased significantly compared to the level observed in a control strain harboring a plasmid without sigK (Fig. 5. O). The increase in cotD-directed B—galactosidase activity paralleled that observed from the cotD-lacZ fusion in wild-type B. subtilis (A). Thus. a" was first detected at T. and increased through T. in wild-type cells (Fig. 3) or in cells expressing pro-a from a plasmid (Fig. 5). and in both cases the increase in the a" level coincided with the increase in cotD-directed fi-galactosidase activity (Fig. 5). The finding that a" accu- mulated at the normal time in cells expressing pro-0" from a plasmid during growth and early in sporulation demonstrates that production of pro-a“ is not the limiting factor in the production of 0". This suggests that if tr" is derived from pro-0" by proteolytic processing. the processing step itself may be a developmentally regulated event that begins at I130!" Tg. I M 0 DISCUSSION The primary product of sigK was inferred to be a pro-protein (pro-0") with 20 extra residues at the N terminus based on a comparison of the N-terminal amino acid sequence of a“ (6) with the nucleotide sequence of sigK (12). Using anti-prov“ antibodies in Western blot analyses of whole-cell extracts of sporulating B. subtilis. we detected proteins that we believe are pro-0" and a“ for the following reasons: (i) the proteins comigrated with gel-purified pro-trx and a" (Fig. 2). (ii) the proteins were not observed in Western blot analyses of whole-cell extracts prepared from growing wild-type cells 181 Proc. Natl. Acad. Sci. USA 87 (I990) 9725 (Fig. 2) or from developing cells of five strains that were expected to be unable to produce pro-tr" and or" due to a mutation either in the sigK structural gene or in a gene whose product is required to generate the composite sigKfene (Fig. 4). and (iii) the protein that comigrated with pro-a was first observed in Western blot analysis of wild-type cells at T, (Fig. 3). which is consistent with the timing of sigK expres- sion (14). while the protein that comigrated with or" was first observed at T. (Fig. 3). which is consistent with the timing of expression of the a“-controlled cotA regulon (7-9). Proteolytic processing has been shown to control the activity of or" (16. 17). another sporulation-specific a factor in B. subtilis. By analogy to at and based on the finding that the N terminus of 0" corresponds to codon 21 of sigK. it was proposed that pro-tr" may be an inactive precursor that is proteolytically processed to active o" (6. 12). Several of our results are consistent with this model. First. the appearance of pro-tr" preceded the appearance of 0" during sporulation of wild-type B. subtilis (Fig. 3). and the timing of appearance of cotD-directed B-galactosidase activity (Fig. 5. A) coincided with the appearance of a“. not ". Second. mutations in eight sporulation loci (.rpalIB. sprrllD. spolllA. spolIlE. sluiIIIG. spol VA. spoIVB. and spa! VF) blocked or reduced accumulation of or". but not accumulation of pro-0" (Fig. 4). and the impaired ('rrlA regulon expression in strains with these mutations (7-9) correlates with the impaired accumu- lation of a". not with the level of pro-a". which was normal in all these mutants except the spollB mutant (see below). Third. when pro-a“ was gel-purified from E. coli and rena- tured under the same conditions that permit recovery of activity of a“ gel-purified from B. subtilis. it failed to promote transcription of a"-conttolled promoters in vitro (data not shown). Fourth. production of pro-a" from a plasmid in a B. subtilis sigK mutant resulted in production of a“ during sporulation (Fig. 5 laser). and. just as in wild-type cells. the timing of appearance of cotD-directed B-galactosidase activ- ity (Fig. 5. o) coincided with the appearance of or"..not with the level of pro-tr". which was high during growth and throughout sporulation. Our data do not rule out the inter- pretation that o" is produced by translational initiation at an alternative site: however. this possibility is unlikely since no apparent ribosomc-binding site or initiation codon exists at the appropriate position in the sigK mRNA. Nevertheless. it may be possible to use a pulse-chase experiment to demon- strate directly a precursor-product relationship between pro-0"anda".ashasbeendoneinthecaseofaeandits precursorll7). Proof that pro-0" is an inactive precursorthat can be proteolytically processed to active 0' will require reconstitution of the processing reaction in vitro. The accumulation of 0" is a developmentally regulated event that begins at about T. in wild-type cells (Fig. 3) or in sigK mutant cells expressing pro-0" from a plasmid (Fig. 5). This event directly or indirectly requires proper functioning of the products of at least eight sporulation loci since. as noted above. mutations in eight loci blocked or reduced accumulation of 0" but not accumulation of pro-or". If a“ is derived from an inactive precursor by a developmentally regulated proteolytic processing event. what purpose might this regulatory device serve? In the case of 0". processing has been suggested to be a mechanism for coupling formation of the sporulation septum to activation of a5 and the subse- quent pattern of gene expression (17. 18). Our finding that spolllA. spoIIIE. s IMO. and spa! VB mutants accumulate pro-0". but not a . artd the results of Cutting er al. (29). discussed below. suggest that pro-or" processing may couple activation of the mother-cell a factor to events occurring in the forespore compartment. , A regulatory mechanism connecting mother-cell-specific gene expression to forespore events was inferred (7-9. 14) from the observation that mutations in spolllA. spalllE. and 9726 Genetics: Lu er al. spoIIIG that impair forespore-specific gene expression (32. 33) also impair mother-cell-specific gene expression. Al- though little is known about the functions of the spolllA and spollIE gene products. spoIIIG is expressed predominantly. if not exclusively. in the forespore compartment and it encodes a a factor. or6 .that directs forespore- -specific gene expression (4. 5). Recently. spoIVB (30) has been shown to be expressed specifically in the forespore. yet mutations in this gene impair mother-celI-specific gene expression (5. Cutting and R. Losick. personal communication). Cutting er al. (29) isolated mutants (called 601' mutants for bypass 0f forespore) that bypass the dependence of cotA regulon expression on spoIIIA, spolllE. spoIIIG. and spa! VB muta- tions. Using the anti~pro-a" antibodies described here. it was shown that lmf mutations restore production of a" in spolllA and spoIIIG mutant cells (29). Thus. lmf mutations appear to uncouple mother-cell-spccif'tc gene expression from fore- spore events by perrnitting pro-tr" processing. Furthermore. replacement of sigK with a deletion-mutated version lacking codons 2—20 (so that the protein produced. 0“". would differ from a" only by a methionine residue at its N terminus) relieved the dependence of cotA regulon expression on the spoIIIG gene product (29). In this case the proposed coupling between forespore events and pro-a" processing appears to be circumvented by producing the truncated. active 0“" instead of pro-or". A protein that was presumably 0“". since it comigrated with a" in Western blot analysis using the anti-pro-a" antibodies. was detected beginning at T, in a spoIIIG mutant containing the deletion~mutated sigK gene (data not shown). This finding suggests that the failure of the spoIIIG mutant to accumulate a' when it contains an intact sigK gene (Fig. 4. lane 8) results from a failure to process pro-o“ rather than from instability of 0". unless 0“" is significantly more stable than tr" in the spoIIIG mutant. Cells containing the deletion-mutated sigK gene also began ex. pressing a cotA-lacz fusion at T. 1 hr earlier than normal (29). as would be expected if a“ ’but not pro-0‘" were able to function as a a factor. The results presented here and the results of Cutting et al. (29) strongly suggest that pro-0“ processing is a regulatory device that couples mother-cell gene expression to forespore morphogenesis. The sprrlllA. spolllE. spoIIIG. and spa! VB mutations are inferred to block forespore morphogenesis at a stage that is incompatible with pro-tr" processing. Mutations in the spollB. spallD. and spoIVF loci also blocked accumulation of 0". but not pro-or" (Fig. 4). The spollB mutant used in this study was shown previously to express only 6% of the B—galactosidase activity normally expressed from a sigK-lacZ fusion during sporulation (14). This may explain the reduced amount of pro-r7" detected by the anti-prov" antibodies in this mutant (Fig. 4. lane 1). Production of pro-o" was unimpaired in the spolID mutant (Fig. 4. lane 2). Since spolID mutations have been shown to impair forespore-specific gene expression (5. 32. 33). perhaps these mutations also block forespore morphogenesis at a stage that is incompatible with pro-0" processing. The spolVF locus is the best candidate to encode a protein directly involved in the pro-0" processing reaction. Muta- tions in spolVF block expression of the cotA regulon (7-9) and our results show that these mutants accumulate pro-tr" but not 0" (Fig. 4. lanes 14-16). Like a spoIIIG mutant. a spa! VF mutant engineered to produce truncated. active 0“" expresses the cotA gene (29). However. a spolVF mutation does not block the expression of a forespore- specific gene (32). Furthermore. a bofA mutation does not bypass the dependence of cotA expression on a spalVF mutation. and the 6018 mutations are alleles of the spa! VF locus (29). These results have led to the proposal that the spoIVF gene productls) governs processing of pro-0" to aK and that 601B mutations alter the spoIVF gene product(s) so 182 Proc. Natl. Acad. Sci. USA 87 (I990) as to relieve its dependence on the products of spoIIIA, spalIlE. spoIIIG. and spa! VB (29). As far as we know. we (17. 18) and 0" are the only transcription factors thought to be synthesized as inactive precursor proteins and activated by specific proteolytic cleavages. In both cases. proteolytic processing may couple completion of a morphogenetic step to the subsequent. new pattern of gene expression. but in each case the molecular mechanism of the coupling remains to be elucidated. We thank .l. Healy. B. Kunkel. S. Cutting. L. Zheng. V. Oke. R. Losick. P. Stragier. and A. Grossman for providing bacterial strains and plasmids and for helpful advice. We thank R. Losick. 8. Cutting. and H. Douthit for critical reading of the manuscript. This research was supported by the Michigan Agricultural Experiment Station and by Grant GM43585 from the National Institutes of Health. 1. Smith. I.. Slepecky. R. A. R Setlow. P. (1N9) Regulation of fritters-uric Development (Am. Soc. Microbiol.. Washiqtoa. DC). 2. Moran. C. P.. 3r. (1”) in Regulation o/Pmcarynrir Derelt'ment. eds. Smith. l.. Slepeeky. R. A. R Setlow. P. (Am Soc. Microbiol.. Washington. DC). pp. 167-184. 3. Masada. E.. Anaguchi. H.. Yamada. K. R Kobayashi. Y. (1”!) Prat: Natl. Ar‘ud. Sci. USA U. 7637-7641. 4. Sun. D.. Stragier. P. R Setlow. P. (1”) Genes Dev. 3. 141-149. 5. Karmazyn-Campelli. C.. Bonalny. C.. Savehi. B. R W. P. (1”) Genes Dev. 3. 150-157. 6. Kroos. L.. Kunkel. B. R Losick. R. (1”915rience 243. 326-528. 7. Sandmn. K.. Kroos. L.. Cuttir'. 8.. Youqnun. P. R Losick. R. (1”!) 1. Hal. Biol. ass. 461-473. 8. Zheng. L. R Losick. R. (19”) J. Mal. Biol. 111. 645-660. 9. Cuttiu. 8.. Panzer. S. R Losick. R. (1'9) 1. Ital. Biol. 87. 393-404. to. Donovan. W.. Zheng. l... Sandman. K. R Losick. R. "”711. Mal. Biol. 19‘. 1-10. 11. Holland. 8. K.. Cutting. S. R Mandelstam. 3. (1”?) J. Gen. Microbial. 133. 2381-2391. 12. Strwfer. P.. Kunkel. B.. Kroos. L. R Losick. R.ll”9)Scienre343. 507-512. 13. Kunkel. B. .Stragier. P. RLosick. R. (lNGenesDev. 4.525-535. l4. KurtkelMBSaadtnanK mans- YoungmaaHPRLosicle. (lull). Bacterial. 170. 3513-3522. 15. Kunkel. B.. Kroos. L.. Poth. H.. Youngnua.P.RLosick.R.(1”9) Genes Dev. 3. 1735-1744. 16. Trempy. l. E.. Morrison-PhrmmerJ. R Haldenwan. W. 0. (1”5) 1. Bacteriol. 161. 340-346. 17. LaBell. T. L.. Trempy. l. E. R Haldenwan. W. G. (1”?) her. Natl. Acad. Sci. USA 84. 1784-1788. 18. Stnrgicr. P.. Bonamy. C. R Karmazya-Campehi. C. (1%!) Ce! 52. 697-704. 19. Dtbnau. D. R Davidolf-Abclsoa. R. (1971) 1. Hal. Biol. 54. ”-221. so. ManiatisHT. Fritsch. EF.RSantrook.l.(1fl2)llaler-ulor Chain; A Lahaarary Manatd(ColdSpr'i~llar'borLab..Cold Sprig Harbor. NY). 21. Yansura. D. G. R Henner. D. 1. user) Plot. Natl.Acad. Sci. USA 81. 439-443. 22. Healy. l. (1989) Ph.D. thesis (Harvard Univ.. Cartridge. MA). 23. Harlow. E. R Lane. D. (lulAnrihaaes (Cold SprquarborLab" Cold Spring Harbor. NY). 24. Youngrnan. P.. Mins. J. B. R Losick. R. (1‘3) Proc. Natl. Acad. Sci. USA I. 2305-2303. 25. Bradford. M. (1976) Anal. Biochem. 12. 248-254. 26. Matsudaira. P. (1931) J. Biol. Client. ‘2. llll35-1N38. 27. Youngman. P.. Perkins. l. B. R Losick. R. (1”) Plasmid 11. 1-9. 28. Errington. J. R Mandelstam. l. (1%) J. Gert. tram. 131. 2967-2976. 29. Cutting. 8.. Oke. V.. Driks. A.. Losick. R.. Lu. 8. R Kroos. L (1990) Cell 61. 239-250. 30. Van Hoy. B. E. R Hoch. l. A. (19%)]. Bacterial. I72. lib-1311. 31. Sandman. K.. Losick. R. R Youw P. (1'?) Genetics 117. 603-617. 32. Errinton. l. R Mandelstam. .1. (1%) J. Gen. MM. 131. 2977-2985. 33. Mason. 3. M.. Hackett. R. H. R Setlow. P. (1'81). Bacterial. I”. 239-244. 34. Hager. D. A. R Burgess. R. R. (1%”) Anal. Birxhent. I”. 76-“. 35. MilerJ. 11.11972) ExperimentslnhlalerwlorGeneticstGoldSpriq Harbor Lab" Cold Spring Harbor. NY). APPENDIX B Sporulation Regulatory Protein GerE from Bacillus subtilis Binds To and Can Activate or Repress Transcription from Promoters for Mother-Cell-Specific Genes 183 J. Hal. Biol. (1992) 226. 1037-1050 184 Sporulation Regulatory Protein GerE from Bacillus subtilis Binds to and Can Activate or Repress Transcription from Promoters for Mother-cell-Specific Genes Llangbiao Zheng‘, Richard Halberg’, Steven Roels', Hiroshi Ichikawa2 Lee Kroos’ and Richard Lodck' |Department of Cellular and Developmental Biology - ’l’lie Biological Laboratories Harvard University Cambridge. Massachusetts 021.38. U.S.A. ’Department of Biochemistry Michigan State University East Lansing. Michigan 48824. U .S.A. (Received 26 December 1991. accepted .3 April 1992) The mother-cell line of gene expression during sporulation in Bacillus subtilis is a hierarchical cascade consisting of at least four temporally controlled gene sets. the first three of which each contain a regulatory gene for the next gene set in the pathway. gerB'. a member of the penultimate gene set. is a regulatory gene whose product is required for the transcriptional activation of genes (coat protein genes cotB and cotC) in the last gene set. The gerE' product also influences the expression of other members of the penultimate gene set (coat protein genes cotA and cotD appear to be repressed and activated. respectively). We now report that the purified product of gerB (GerE) is a DNA-binding protein that adheres to the promoters for cotB and cotC. We also show that GerE stimulates cotB and w‘ cotC transcription in vitro by RNA polymerase containing the mother-cell sigma factor a". These findings support the view that GerE is a positively acting. regulatory protein whose appearance at a late stage of development directly activates the transcription of genes in the last known temporal class of mother-cell-expreaeed genes. In addition. GerE stimulates cotD transcription and inhibits cotA transcription in vitro by 0" RNA polymerase. as expected from in vivo studies. and. unexpectedly. profoundly inhibits in vitro transcription of the gene (sigK) that encodes a". The effects of GerE on cotD and sigK transcription are just the opposite of the effects exerted by the earlier-appearing. mother-cell regulatory protein SpoIIID. suggesting that the ordered appearance of first SpoIIID. then GerE. ensures proper flow of the regulatory cascade controlling gene expreuion in the mother cell. K eyurords: sporulation; sigma factor; regulatory protein; Bacillus subtilis Llltl'odllcdon Following the formation of a transverse septum at morphological stage 11. gene expression during the process of sporulation in Bacillus subtilis is regulated differentially between the forespore and mother-cell chambers of the developing sporangium (De Lencastre R Piggot. 1979; Losick R Stragier. 1992). Thus. the transcription of certain genes is restricted to the forespore. whereas the transcrip. tion of other genes is limited to the mother cell. Although some heterogeneity in the time of induc- tion of gene expression in the forespore has been reported (Panser et al.. 1989; Bis-man R Setlow. 1991). most forespore-expressed genes are switched on coordinately and only a single regulatory gene. spoIIIG. which encodes the forespore sigma factor 6° (Karmazyn-Campelli et al.. 1989; Sun a al.. 1989). is known to be exclusively transcribed in the forespore chamber of the sporangium. Gene expression in the mother cell. in contrast. in rela- tively complex. involving the expression of at least four coordinately controlled gene acts. which are switched on successively during the course of sporulation (Cuttingdal.. I989; Kunkel etal.. 1988. 1989; Sandman et al.. 1988; Zheng R Losick. 1990). 1037 mu-ssssm/rstosr-tr sum/o OlMAcademiePruw I038 These gene sets constitute a hierarchical cascade in that the first three gene sets each contain a regula- tory gene that governs the expression of the next gene set in the pathway (Zheng t Losick. I990). spoIIID. a member of the earliest regulon. is regulatory gene whose product (a small. DNA- ~binding protein; B...“ & L..K unpublished results) turns on the transcription of the next set of genes (Halberg & Kroos. I992; Kroos et al.. I989; Kunkel etal.. I989; Stevens a Errington. I990). One of the genes in the spa] I lD-depeudent class of co- ordinately controlled genes is sigK (Kroos et al.. I989; Kunkel et al.. I988. I989). a composite gene (generated from two truncated genes by a DNA rearrangement in the mother-cell chromosome; Kunkel et al.. I990; Stragier et al.. I989) that encodes the mother-cell sigma factor a“ (Kroos et al.. I989). The sigK gene product (after a ngulatory step involving the conversion of its primary product. pro-6‘. to the mature sigma factor; Cutting et al., I990; Lu et al.. I990) then turns on the penultimate class of genes. This set of genes includesthesporecoat proteingenescaM andcatD (Sandman et al.. I988; Zheng & Losick. I990) and the ngulatory gene gerE (Cutting et al.. I989; Holland et al.. I987). which encodes a product (GerE) that is. in turn, required for the expression of genes in the last known gene set in the mother-cell line of gene expression (Zheng & Losick. I990). Two members of this pert-dependent gene set are the coat protein genes cotB and cotC (Donovan et al.. I987). Hereweareconcernedwiththeroleofgerl'inthe hierarchical cascade of mother-cell gene expression. gerB is inferred to be a transcriptional regulatory gene because: (I) a pert nonsense mutation (gerBJG; Cutting, I986) has highly pleiotropic effects on sporulation. causing the production of spores that are Iysosyme-sensitive. germination-defective and aberrant in cost ultrastructure and protein com-g position (Feng s Aronson, I986; Jenkinson a Lord. I983; Moir. I98I); (2) ”£36 partially inhibits expression of cotD (Zheng & Losick. I990) and causes overexpression of cotA in rich sporulation medium (Cutting et al.. I989; Sandman et al.. I988); (3) as indicated above. ”336 prevents the expression of cot genes B and C (Zheng a Losick. I9”); and (4) the predicted product of per! (GerE). an 86 kDa polypeptide (Cutting & Mandelstam, rm; Hasnain et al.. I985). contains a region of similarity to the a-helix—fl-turn—a-helix motif of many procaryotic transcriptional regulatory pro- teins (Holland d al.. I987) and exhibits high overall similarity to the CODE-terminal region of certain regulatory members (the ngA sub-group) of the family of two-component sensor-regulator systems in bacteria (Gross et al., I989; Kahn in Ditta. I99I). which inclmles the B. subtilis regulatory proteins DegU and ComA (Henner et al.. I988; Kunst et al.. I988; Weinrauch et al.. I989). GerE is also similar to the COOK-terminal region of the Escherichia coli regulatory protein MaIT (Gross d al.. I989) and to the COOK-terminal region of sigma factors, which is 185 L. Zhenget aI. involved in the recognition of the -35 region of bacterial promoters (Kahn & Ditta. I99I). 0n the basis of DNase I footprinting experiments with purified GerE. we now report that the peril gene product is a DNA-binding protein that adheres to the regulatory regions for cotB and cotC. We also show that GerE greatly stimulates cotB and cotC transcription in vitro by a‘ RNA polymerase. a finding rn support of the view that the appearance of GerE at a late stage of sporulation directly acti- vates transcription of these genes. Moreover. we show that GerE affects the in vitro transcription of several other mother-cellexpressed genes. either positively or negatively. findings that suggest a functional analogy between GerE and the mother- cell regulatory protein SpoIIID (Kroos et al.. I989). However. the effects of GerE on cotD and sigK transcription in vitro are just the opposite of the effects exerted by SpoIIID. Because production of a" RNA polymerase appears to cause a decrease' In the level ofSpolIII) (Halberg a Kroos. I992) and rs required for the transcription of gerE (Cutting et al.. I989; and this work). we propose that a declining level of SpoIIID and a rising level ofGerE produce a reinforced switch in the patwrn of mother-cell gene expression during sporulation. 2. Materials and Methods (a) Strains’ and plassuds’ B. coli K38 (HfrC trp thi 11’) carrying plasmid pGPI-2 (Tabor I Richardson. I985) was maintained at ”'C in LB medium containing 26 pg of kanamycin/ml. The source of the gerl' open reading framemas pSGlIUIOI (provided by J. Errington of Oxford University; Cutting & Mandelstam. I986). The pert open reading frame was released as a 0'6 kbf BeoRV/szl fragment and was cloned into the Sinai/Xbal site of pT'tI3 (Bethesda Research Laboratories) to create p123“. The BeoBV site is 67 bp upstream from the garl' coding sequence. whereas the Xbal site is a polylinlter sits adjacenttoaB. subtilis Ubol site locatedm bpdown- stream from the pert open reading frame. Plasnrirh pm and pT7I3 were introduced into K38 cells containing pGPI-2 by transformation and selection on LB medium containing 60 pg of ampicillin/ml and ”pg of kanamycin/ml at the permimive temperature (30"?) Plasmids pLBKIOO and pBKI6 containing the cstD andsingromoterrqrons.respeetively.ssrvedas templatssforinsitretranseriptionandhavebesn doeribed previously (Kroos at al.. I989). A 0-6 kb Becki/Pea" fragment (Fig. I) containing the estB promoter (Zheng & Losick. I990) was subcloned into Becki/SassI-digested replicative form MI3mpI8 (Yanisch-Psrron et al.. INS) and replicative form DNA of the recombinant phage was used for in vitro transcription. Plasmid pBDfiflI containing the cotC promoter region was constructed as follows: the fueled-containing Barn!" fragment of pSGMIJaI (Errington. I986) was cloned into the Bell site of pBD95 (Zheng & Losick. I990). creating anin-framefusionafcatCtolacZianDm.tbenthe tAbbrevistioru used: kb. ro’ bases 4.. base-pairs; bp. bass-pairs; nt. nucleotides. 186 Sporulation Reputatory Protein Gert I039 sonar ssssa mu cats I / 1< P .'.. '-.. -.°. 4. ' ... I 9 I 9 9 P r—J r————r 1 ‘ 2 Gate fi I f 4 '1‘"! “at! M 8.8 ”I! I“!!! I“. Figural.“re0er8bindingsitesirrtlie5'rcgiorrsofect8urdcot0.1heFigureshowsrcstrietionmapsofDNAinthe vicinityofcctBandcotC.basedcnapreviousreport(Donovandal.. lm)andcthcrunpublbhedanslyais.(Notethat themap ofSau3A sites is incompleteandonlyasingle site isshown.)Thepositions cftheopcn reading frames forcaeh gene are shown by the filled bars (the map includes only part of the cotB open reading frame). Also shown are the nudwfidamqumcudthepmmbrmghnsofbothgemflhuartdtmdmfipthnmiwiaudbythearrowa. RegionsofDNAthstwereprotectedbyGerEfromthesctionofDNaseIareindieatedsboveandbelowcachstrand. 'l'thGcrl-IbindiugsitesincctCaredesignstedlandzintheFigure. cotC-lacz-cat-containing BamlII/Bplll fragment of pBDm was cloned into the BarnIlI site of me (Bolivar d al.. I977) to construct p802.” in which the Hindi" its upstream from the cotC promoter is proximal to the Hindi" its ofthe vector. and finally pBDQ59 was digested with Hindi" and recircularised to construct pBD26I in which the small HindIII fragment was deleted. pH" was constructed by subcloning a 06 kb HindIlI/szl fragment from pBD26I (extending from thslfindlll site upstream fromcatCtoan Xbal site in the polylinker downstream from the former Bell site of cotC (Fig. I) into IliadIII/XbaI-digcsted pUCI9 (stischo chrron d al.. I”). Restriction fragments from pHII werepurifiedafterclectroplmesisonanagarosegeland uscdastemplatesforincitrotranscriptionofcctc.‘l‘he pert promoter-containing plasmid. pSCl46. was constructed by 8. Cutting as follows: the 266 bp Alal fragment from pSGHUIOI (encompaming the pert promoter region; Cutting a Itandelstam. I986) was subcloned into the Basal site in the polylinbsr of pSGIUSI (Errington. I9“). then excised as a KpnI/Baer-II fragment and . inserted Into KpnI/BamHI-digested pUCI9 (Yanisch-Pcrron et al.. I35). Three cotA promoter-containing plasmids were constructed by K. Sandman as follows: (I) pKS22 was contracted by ligating Iliad"! linkers to a 08 kb llinclI/Asal fragment from pKSII (encompassing the cotA promoter region; Sandman et al.. I988). cleaving the linkers with HindIII. and subcloning the fragment into HindIII-digested pIBI76 (International Biotechnologiss. Inc); (2) pK823 was constructed by digesting pKSI9 (Sandman at al.. I988) with PstI and ligating to delete B.sabtilis DNA beyond 65 bp upstream from the cotA trarueriptional startsite; (8) pK824 was constructed by digesting pKSI9 (Sandman 4 al.. I988) with tcoRV and ligatirg to delete B. subtilis DNA beyond II5 bp upstreamfromthecctAtrar-scriptionalstsrtsite. (bl Pradadsan' oIGert in E. coli Cultures of K38 cells containing pGPl-2 (bearing the phage T7 RNA polymerase gene) alone. pGPI-2 and pT'II3. or pGPI-2 and pm were grown at ”'0 to an A... of 03. Cells were induced by a temperature shift to 42'C for 90 min. Rifampiein was added to a final concen- tration orsoo ulml. and the cells were incubated at 42‘C for I0 min and then at N’C for 30 min. Cells were collected bycsntrifugationanddimolvedinsamplcbaflcr (Lsemmli. I970). The sample was denatured at 90‘C for 2 min and fractionated by electrophoresis through an SDS/polyacrylamide gel containing "5% acrylamide. For the preparation of (1ch. protein from induced K38 cells containing pGPI-2 and pLZ304 was subjected to 'andtheputativchrEbandwaseutfrom thegel.0erEwasthcncIutedfromthegeIalicsand renatured as described (Hager & Burgcm. I”). For the preparation of control protein. protein from induced K38 cells containing pGPI-2 and p’l'lI3 was subjected to cIeetrophoresis.Aslicefromthepositioncorrespondingto dratchchwaseatfromthegeI.Proteinwasthar elutedandrensturedfromthegclsliceasdescribedfcr GerE. (e)PreparaticndDNA prabsslabaladatcalpcnsend Porthe ofradioactivcccthrobas.s868bp IliafI/Bcll fragment (Fig. I) whose Hinfl termirun hsdbeenrenderedfiuhbyasecftthlenowfr-agmsnt of DNA polymsrtss I was cloned hto flincIIlBamHI-digested pUCI8 (Yanisch-Perron d al.. I040 I985). This created plasmid pIZI275 in which the Hindi" site of the vector was upstream from the cotC promoter. Plasmid plZl275 was linearized with Ilindlll and then treated with alkaline phosphatase. Next. the catCocontaining fragment was released from the pUCIIl vector by digestion with tle. which cuts at the end of the polylinlrer next to the Barn!“ site. A probe labeled at the Hindi" site in the non-transcribed strand was prepared using phage T4 polynucleotide kinase and [y-’ PIATI’. To prepare a probe labeled at the Hindlll site in the transcribed strand. cctf' was released as a HindIII/Srnal fragment (Srnsl also cuts in the polylinker next to Barnlil site) and was labeled by end-filling the Hindi" terminus using the Klenow fragment of DNA polymerase I and (a-’ I’ldATP. Additional cat(.‘ probes were prepared by digesting pI-lll (see above) with Earl. which cleaves in the 7th codon of cotC (Donovan at al.. I987). labeling in the non-transcribed strand using the fill- in reaction of the Klenow fragment of DNA polymerase I and [ao’zl‘ijTI’ or labeling in the transcribed strand by treatment with alkaline phosphatase followed by phage T4 polynucleotide. kinase and [y-ul’lATl’. then diguting with tcoRV and tle. and purifying the 239 bp teaRV/I'Jarl fragment encompassing the cotC promoter regionaflerelectro ' cnanon-denaturing.poly- acrylamide gel containing 8% acrylamide (llanistis et al.. I982). For preparation of radioactive cotB probes. we took advantage ofa plasmid pUCIB derivative called pBDI36 (constructed by W. Donovan. unpublished results). which contains a 08 kb Sau3AI/Sau3AI fragment that includes the promoter and the N H z-terminal coding region of the cotB open reading frame cloned into the Band" site in an orientation such that the polylinker 87le site was proxi- mal to and upstream from the promoter. pBDI36 was digested with teoRI. treated with alkaline phosphatase. and the cotB-containing fragment was released by digeso tion with HindIII. which cuts at the opposite end of the polylinlrer. The non-transcribed strand probe was labeled at the tle site by the T4 polynucleotide kinase reac- tion.To preparethetranscribedstrand probe.acotB- containing. tcoRI/PrvuII fragment was purified (see Fig. I) and was then labeled by end-filling using the Klenow fragment of DNA polymerase I and (a-“PldATI’. (d) DNase I [wtprintinp Twodifiercntmethodswereusedtocarrycutthe DNase I footprinting experiments. In method (I). the conditions for the binding ofGerIt' were the same as described by Strauch et al. (I989). except that poly(dI-dC) was added to a final concentration of mpg/ml. DNA fragments labeled at one end were incubated in separate experiments without protein. with control protein. or with different amounts of GerE protein in so pl reaction mixtums for I5 min. Then Ipl ofa 00I mglml DNase I solution (BRL) was added to each reaction for I min. The digests were stopped by adding 5 pl of stop solution (O-I u-EIYI‘A and 05% SDS) and chilled on ice. The DNA in each reaction was precipitated with I ml of ethanol and with 02 pg of poly(dI-dC) as carrier. The precipitates were dissolved in formsmide loading buffer ()Isniatis et al.. I982) and denatured at “PC for 90 s. The samples were then subjected to electrophoresis in an 8 Homes- containing polyacrylamide gel. In method (2) the conditions for the binding of GerE were the same as in the in vitro transcription experiments (see below). except that poly(dI-dC) was added to a final concentration ,of l-2 pg/ml. DNA fragments labeled at one end were incubated at 37'C in separate experiments 187 L. Zheng et .1. without protein. with control protein. or with different amounts of GerE in 42 pl reaction mixtures for I0 min. Then 8 pl of 0m mg DNase I/ml (Brn-hringer. Mannheim) solution (prepared by diluting a stock solution (Davis et al.. I980) with buffer (20 mlr-Tris-HCI. pH 8-0. 20 MI’MgNI. 20 I'D-ml), "II “(10“ to each reaction. After I ruin. the digests were stopped by adding 50 pl of buffer (IOO mn-Tris-HCI. pH 80. 50 mn-EDTA. 200 pg yeast tRNA/ml) and incubating for 2 min at 65°C. The DNA in each reaction was precipitated with 250 pl ethanol. The precipitates were dissolved in formamirlr- loading buffer (Maniatis et al.. I982) and denatured by boiling. The samples were then subjected to electro- phoresis in an ll shuns/polyacrylamide. gel containing 6% acrylamide. (e) DNA sequencing End-labeled DNA probes were subjected to the chemical cleavage reactions of Iaxam & Gilbert (I90!) with a kit from New England Nuclear or as described previously (Maniatis et al.. I982). (I) In vitro transcription c‘ RNA polymerase was partially purified from B. subtilis strain SCI04 (tr-p02 pert.” SPfixeatAolacl) as described (Kroos et al.. I989). This enzyme was compar- able in protein composition and in cotD- and sigK-tran- scribing activities to fraction 24 shown in Fig. 2 of Kroos d al. (I989). 0‘ RNA polymerase was reconstituted from gel-purified. renatured c‘ and B. subtilis core RNA poly- merase as deucribcd previously (Kroos et al.. I989). Transcription reactions were performed as described previously (Kroos d al.. I989) except that heparin (8 pg) was added 2 min after the addition of nucleotidu to prevent reinitiation. and after the1eactions were stopped l0 pl ofthe reaction mixture was subjected to electlb?‘ phoresis. [rs-”PET? was the labeled nucleotide unlem indicated otherwise. After gel electrophoresis. transcripts were detected by autoradiography and the signals were quantitated using a Visage ”0 Image Analyser (BioImage). (g) l’rirner estensrcn‘ analysis RNA was from sporulating cells as described by Cutting et al. (I99Ia). In vitro synthesised cotC tran- scripts were gcnerated as described above (but without radiolabeled nucleotide) and then precipitated with ethanol and suspended in 25 pl of diethylpyrocarbrmate- treated water. In preparation for primer extension analy- sis. a sample of the in vitro synthesised RNA (4 pl) was treated with 5 units of DNase I (Pharmacia) in buffer (20 mnoTris-HCI. pH 76. I0 mn-MgCIJ) in a total reaction volume ofli pl. The reaction was incubated at 37°C for I0 min and then at 90°C for I0 min. Primer extension was carried out by use of the mt!!- speeific oligonucleotides PH and Pr2 (Zheng & Losick. I990). The oligonucleotides were Send-labeled using [y-"I’IATI’ and T4 polynucleotide kinase (BRL) as described by Sambrook et al. (I39). For analysis of in viva synthesised RNA. from 2 to 6 pmol of 5'-end labeled oligonucleotide was incubated with 5 pg of total RNA. andreactionswcrecarriedoutasdescribed by Roelsetal. (I992). For analysis of in vitro-generated cotC transcripts. I2 pmol of 5'-end-labeled Pr2 oligonucleotide was incu- bated with 4 pl in vitro synthesised RNA (from above) in I0 pl of annealing buffer (50 mn-Tris-HCI. pH 7'6. I00 mn-KCI) at “PC for 2 min and then at 47’C for Sporutation Regulatory Protein GerE 30 min: 5 pl ofthe primcrzRNA hybrid solution was then incubated with 5 units of phage Il- MuLV reverse tron- scriptsse (Pharmacia) at 47°C for 45 min in a final volume of IO pl of reverse transcriptase bufl'er (50 mIr- -Tris HCI. pH 7'6. 60 IIIII- KCI. I0 mII- MgCh. I mil-dithiothreitol (DTT). 25 our of each dNTP. I unit placenta IRNase inhibitor/pl (Pharmacia)) alter which 7 pl of 95% form- amide loading dyew ded. The 5'-errd labeled oligonucleotides were also used to generate sequence ladders by the dideoxy chain termina tion method of Sanger d cl. (I977) The prulucts of primer extension we ere subjected to electrophoresis In 6% polyacrylamide slab gels containing 8 Ir- urea 3. Results (a) Purification of GerE GerE was purified by engineering It‘. coli cells to express a cloned copy of the gerlr‘ gene using the T7 RNA polymerase/1‘7 promoter system of Tabor A Richardson (I985). A DNA fragment containing the yerE open reading frame was placed under the control of a phage T7 promoter by inserting into the xpremron vector pT7I3 a gulf-containing segment of B. subtilis DNA that extended 67 hp upstream and 227 bp downstream from the qerlr.‘ open reading frame to create plssmid plz304 (see Materials and Methods). Cells containing pLZ304. the vector pT7l3. or neither plasmid. were grown at 30°C and were then shifted to 42°C to induce transcription from the phage T7 promoter. Total cellular proteins from induced and uninduced cells were displayed by electrophoresis through an SDS/polyacrylamide gel (Fig. 2). In addition to normal cellular proteins. induced cells of the pLZ304vcontaining strain (lane 1“) produced a protein of 6 to 8 kDa. which was Flori! 2. Production ofGerE in E. coli. Total cellular proteins were extracted from cells grown at 30‘C (lanes A toC)orfromcellsthathad beenshiftedto42'C(IancsD to?) andwcre then resolved by electrophoresis in a smll2% polyacrylamide gel. The cells were derivatives fiwlistrar nK38 containing pGPI-2 alone (lane A D). pGPI 2 and p’l'7l3 (lsrre B and E). or pGPI- -2 and pLZMtflansC .(-pGI2containsthephsgr-17 RNA merase mgcne.) The positions of the molecular ma- markers (in kDa) are shown on the 188 l04l absent in induced cells that lacked the gerE-bearing plasmid (lane D) or that contained the plasmid vector. p'l'll3 (lane E). This protein was also absent in cells that were not heat-induced (lanes A to C). The sire of the protein was consistent with that (85 kDa) deduced from the nucleotide sequence of the peril open reading frame (Cutting & Mandelstsm. I986). Because there was no other open reading frame in the peril-bearing If. subtilis insert in pl].304 that could encode a protein of this rise. we presume that the 6 to 8 kDa protein was GerE. To purify GerE. the 6 to 8 kDa protein band was excised from an SDS/polyacrylamide gel displaying proteins from pLZ304-containirrg bacteria. Protein was eluted from the gel slice and renatured as described by Hager A Burgess (I980). As a control. a gel slice corresponding' In position to GerE was cut from an SDS/polyacrylamide gel displaying total cellular proteins from i u cells of pT7I3- containing bacteria. Protein was eluted from the gel slice and renatured and is referred to as “control protein (b) Mapping the 5' terminus of cotC mRNA To study the interaction of GerE with cotB and cotC. it was first necessary to know the precise sites from which transcripts from these genes originate. Previously reported primer extension experiments established the location of the 5' terminus of cotB mRNA (see Fig. l) but only provided a tentative assignment for the 5’ terminus of cotC mRNA (Zheng & Losick. I990); smextension product of I27 at that would correspond to an apparent‘li' terminus located I20 bp upstream from the cotC initiation codon was obtained with an oligonucleo- tide primer called Prl but no extension products were observed with two other oligonucleotides called Pr2 and Pr3 (see Fig. I of Zheng & losiclr (I990) for s description of the primers). Alerted from the results of in vitro transcription experi ments (below) that the truedtsrtsrte of transcrip- tion was actually very close to the beginning of the cotC open reading frame (and hence that. relatively short extension products were to be expected). we repeated the primer extension analysis and found that Prl and M extension products of 33 and 68 nt. respectively. that corresponded (as ' Prl and nucleotide requencing ladders; Fig. 3(a) and 3(b)) with a 5’ terminus that preceded the initiation codon by only 26 bp (Fig. I). These extension products were specific to gerE‘ cells at a late stage of sporulation In that Pr] and M ge-ner ated little or no extension products with RNA:3 from wild-type cells at an early stage of development or with RNA from gerE mutant cells at a late stage of development (Fig. 3(d)). Consistent with our earlier results (Zheng & Losick. I990). Prl generated the previously observed I27 nt extension product in addition to the 33 nt product (Fig. 3(a)). but direct nucleotide sequencing of this extension p not by the use of dideoxynucleotides in the primer exten- 189 1042 L.Zhengetsl. (a) (b) (c) GATCP: GATC Pr 1 2 CTAG HO’HQw 51112-3900” (d) § 0 Sporulation Regulatory Protein GerE sion reection esteblished thet the I27 nt product wee not copied from 0010 mRNA but rethcr from the trenscript of enother. similerly reguleted gene (dete not shown). Finelly. the previous failure to observe en extension product with Pr3 is now expleined by the feet thet this primer corresponds to e sequence (see Fig. I of Zheng & Losick. WOO) thet is loceted upstreem from the 5' terminus of cotC mRNA. (c) GerE binds to specific sequences Preliminery gel reterdstion experiments indiceted thet GerE binds to e [liedIIl/Srpl freg- ment of 500 bp conteining the 5' region of cotC end to en EcoRI/Pvull fragment of 635 bp conteining the 5' region of cotB (Fig. I; end Zheng. l990). To locelixe the binding of Got!) to cotB end cotC more precisely, DNese I protection experiments were cerried out with redioective DNA probes seperetely endolebeled on one or the other DNA strend. The redioectively lebeled DNAs were incubeted with GerE end then mildly treeted with DNeee I to generete e spectrum of fregments. After the enzyme digestion step, the DNA fregmente were frection- eted by gel electrophoresis. Figure 4(e) displeys the petter'n of fregments genereted by enzyme treet- ment of GerE-bound cotC DNA thet hed been Iebeled on the non-trenscribed strend (the upper strend in Fig. I) et e Iliad!" site loceted upstreem from the promoter. The Figure shows thet GerE ceused protection from the ection of DNese I elong en epproximetely l6 bp stretch of DNA extending from position — 126 to position - ltl reletive to the 5' terminus of cotC RNA. No protection wee observed with control protein (lene 1). Likewise. Figure 4(b) shows thet GerE protected e 19 bp stretch of DNA extending from position -l29 to position - I47 on the trenscribed strend (the lower strend in Fig. l) ofcotC end thet no protection wee observed with control protein. A second GerE binding site wes mepped in the cotC promoter region using DNA probes Iebeled et the Earl site loceted downstreem from the promoter. Figures 4(c) end (d) show thet GerE protected en epproxirnetely 22 bp stretch of DNA extending from position -56 to position -77 on 190 l043 the non-trenscribed strend (the upper strend in Fig. I) end en epproximetely 2| bp stretch of DNA extending from position -58 to position -78 on the trenscribed strend (the lower strend in Fig. l). respectively. In both ceses. no protection wee observed with control protein. The regions upstreem from cotC thet were protected from DNese I digestion by GerE binding ere indiceted in Figure I. Anelogous experiments showed thet GerE protected e wide region of cotB from DNese l diges- tion. Figure 4(e) shows thet the GerE-protected region on the non-trenscribed strend (the upper strend in Fig. l) wes 40 to 50 bp in length. with the strongest protection occurring between position —4l end position -8|. Figure 4(f) shows thet s similer region wss protected on the trenscribed strend. the protected ngion being 47 bp in length end extending from position -36 to position —82. The region upstreem from cotB thet wee from DNese l digestion by the binding of GerE is indiceted in Figure I. (d) GerE dimslata cotB and cotC transcription in vitro To test for effects of Ger-E on trenscription of 0018 end 00:0 in vitro. lineerized DNA templetes were trenscribed in the presence ofGerE or control pro- tein with a" RNA polymereee pertielly purified from s gerE mutent (see Meteriels end Methods). 0" RNA polymereee produced "IQ-OE trenscripts of the expected sizes from cotB in the presence of OCR (Fig. 6(e). lenes 3 end 4). The signel wss sevenfold weeker in the presence of control protein (Fig. 6(e), lene 2) or with no eddition (Fig. 5(e). lene l). a" RNA polymereee reconstituted from gel-purified 6‘ end purified B. subtilis core RNA polymereee wee elso stimuleted by GerE to trenscribe cotB (dete not shown). . Pertielly purified 6" RNA polymereee feiled to trenscribe e lineerised plesmid (pBDZfll) beering the cotC promoter. even in the presence of GerE. epperently beceuse sequences loceted in the vector portion of the plesmid compete with the sequences loceted upstreem from cotC for binding to GerE (dete not shown). However. when e 05 lrb restric- MJ. leppir‘thefi'terrninr-ofchmRNA. (e)end(b) Resultsofhigb-resolution meppingoftbsb'terminmof cotC mRNA ring the oligonucleotide primers Prl (s) end Pr! (b) to prime cDNA synthesis from totel RNA from rm (spo‘) cells hervested to h efter the crust of sporuletion (1"). (c) Results of high resolution mepping of the 5' terrninm of cotC mRNA using the oligonucleotide primer M to prime cDNA synthesis from RNA trenscribed in vitro with e‘opolymerese in the ebssnoe (lens I) or presence (lene 2) of GerE protein. The products of primer extension were mbjecudwebamphondsdmgsidenuckofidenqumdngleddmgemwdwitheitherdn Prl orMprimer.The errews indicete the position end site of the principel extension product(s) obteined with eech oligonucleotide primer. The opsnerrowin(e)indicetesthepositionofenertifect bend.seenwith Prl butnotPrflJhetisnottheresultofextenm‘on of out!) mRNA (see the text). The sequences shown in the vicinity of the 5' terminus correspond to the non-trenscribed DNA etrend end the filled circles indicete the nucleotide corresponding to the 5' terminus. (d) The time course end geneticdependenceoftheeppeerenceofcotCmRNA.’I‘heprimers Prl (lenes l toO)end Pr2(lenes7to l0)wereusedto gensrete extension products from totel RNA from sporuleting PY'IO (spo‘) cells hervested et 1, (lenes 4 end 8). t. (lenes 5end0).ort..(lenes6end l0)orfrom Maui”) cellshervestedett, (lene l).t. (lene 2).ort..(lenes$end7). M4 Ahmad (0) (cl (e) G G AG123456 A G 1 2 3 I 5 6 G123456 - ".5 : mgr ; r 3., :35 ’- : 41> as = ==srs= 4 fi.ll ' 'r- :: =: -125>!:EIE§§§ -ss> I 2: I C 3... .- 2 -2: -1u>"’7 f“ .3 2' l :3888: '- - ‘ . ‘I. I: ' - 81’..--- c h-::-:! '- :!-.. . i! ll ' é.... .. .. . I 1553s; g :::::: '{ --: i—:SS“I I’ 2 ;;§§ii 3' 5- -3-.‘&. -- .. ‘0. O i 27:22: slee' ° u 6123456 is 1 2 3 r s s cfizarsa ;' I ' e _ . '5' ,A 3' 5' ; .. ' no - w- E: 2 as» iggggg “1+9 "°’ 3 2 == , .eses :- : 3 ': -129>' —sa> \ g I 'E '3 ¥ .2 w! ' -.-. ‘ eg g 5 -1r1> - -s2>ls _ _ _ _: Seen! ‘3 ‘[ . ‘3":-§ ..--.. o -. eseee: . - - - .2'f'?!’ 3'seeees ’ "l! .: ’ moo... - - _-_‘ 2 -... Figure 4. GerE footprints in cotB end calf DNAs. deioeetive DNA fregments seperetely end lebelrd on the trenscn'bed or non- -trenecribed bedstrends werei ncu'bsted In sepsrete reections without protein (lene 6) with control protein (lene l). orwith l pg(lene2).05ug(lsne3). 0| “(lene 4) end005pg(lene5)ofGerI-I. Altertreetment with DNeseI. the pertielly digested DNAs were-epsreted bv electrophoresis through en 8 .0 uree/polyeery emide gel slang-ide s sequencing lsdder genereted by chemicsl clcsvege of the respective end lebeled DNAs. (s) end (b)l Footprints of the non- trenscr'ibcd (upper strend in Fig. I) endt hetrenscribed (lower) etrends ofratC. respectively. using probes lebeled st the HisdIII site loceted upstreemrom from the promoter (c) end (d) Footprints ofthe non-truism trenscribed (upper) end transcribed (lower)etren etrsnds ofch. respectively. using probes lebelcd et the Earl site loceted downstreem from the promoter. (s) end (0 Footprints of the non- -trsnscribed (upper) end trenscribed (lower-"trends strenrhcdof Brespectively. using probes Isbcled st the Eco RIsi loceted upstreemrom from the promoter. The experiments of(e). (b). (s) end (0 were cerried out using method ' (I) In Ifeterisls end Methods end the experiments of (e) end (d) were cerricd out using method (2). elthough in other experiments (dete not shown) similer footprinting results to those seen in (s) end (b) were obteined using method (2). Sporulation Regulatory Protein GerE lel (hi I 2 3 4 l 2 3 4 .-{ '-‘ "ti” ; < >ifl.fl'\;' > .' L- :‘-, Figure 5. GerE stimuletes cotB end cotC trenscription is vitro. Templste DNA (04 pmol) wes trenseribed with pertielly purified a‘ RNA polymereee (02 pg) elone. or with control protein or Geri-I (0.4 pg) edded immedietely efter the eddition of RNA polymereee. Run—oll' tren- scripts were electrophoreeed in 5% polyurylemide gels contsining 8 it-ures end were detected by eutorsdio- grsphy. Arrowhesds denote the positions of run-06' tren- scripts of the expected sites in eech penel. es judged from the migretion ol' end-lsbeled DNA fregments of Hopi-digested pBR322. (e) cotB trenscription from repli- cetive form DNA of e recombinent phsge containing the cotB promoter region. The phsge DNA wss Iineerised with BostHI (lenes l to 3. l77-bese trenscript) or Hindi" (lens 4. ”Thus trenscript) end trenscribed with 6‘ RNA polymereee elone (lene l). or with control protein (lens 2) or GerE (lenes 3 end 4) edded. (b) cotC trenscription from restriction f ts isolsted from pH“. The HindIII/szl fregment (lenes I to 3. l74-bese trenscript) or the HiredIlI/Soll f ment (lene 4. lw-bese trenscript) were trenscribed with RNA polymereee elone (lens I). orwithcontrolprotein (lene2)orGerl-3(lenes3endt) edded. is) to) tel 192 l045 tion frsgment wes ueed es the templete for in vitro trenscription. 6‘ RNA polymereee produced run-off trenscripts of the expected sizes from cotC in the presence of GerE (Fig. 5(b). lenes 3 end 4). A very week signsl we: observed in the presence of control protein (lene 2) or with no eddition (lene l) in e longer eutorediogrephic exposure then thet shown in Figure 5(b). Primer extension enelysis of the GerE-stimuleted cotC trenscript produced by a" RNA polymereee in vitro demonstrsted thet it her! the seme 5’ terminus es cotC mRNA produced in viva (Fig. 3(c)). othermotlrer-cell-etpraeedgem a“ RNA polymereee pertielly purified from egerE mutent hes been shown to trenscribe from the cotD end sigK (previously celled spa! VCB) promoters in vitro (Kroos et al., l989). GerE stimuleted cotD trenscription two- to threefold (Fig. 6(e)) end com- pletely inhibited sigK trenscription (Fig. 6(b)) by pertielly purified 0" RNA polymereee. Although the effect of GerE on cotD trenscription wes modest. e two- to threefold stimuletion wes consistently observed in four independent experiments (dete not shown). The level of stimuletion wss not further onhenced by the use of twice es much GerE es thet employed in the experiment of Figure 6(e) (dete not shown). The pertielly purified 0" RNA polymereee produced run-off trenscripts of the expected sires from gerE (Fig. 6(c). ienes l ind 2) end from acid (Fig. 6(d). lsnos I end 2). These trenscripts were elso produced by 0‘ RNA polymereee reconstituted (e) Eject: of GerE on in vitro transcription of 3 4 s s '- "2""..7’3' . I .r L .- . B . r O ' more 6. Efl'ects ofGerB on cotD. sigK. p8 end cotA trenscription is vitro. Lineerised plesmid DNA (2 pg) wes trenscribed with pertislly purified a“ RNA polymereee (02 pg) slone. or with control protein or GerE (04 fig) edded immedistely efler the eddition of RNA polymereee. Run-otl' trenscripts were electrophoresed in 5% polyecrylemide gels conan 8 u-uree end were detected by eutorediogrephy. Arrowheeds denote the positions of run-oi? trenscripts of the expected sires in eseh penel, es judged from the migretion ol'end-leboled DNA frsgments of Kept-digested pBR322. (e) not!) trenscription from ”infill-digested pLRKIOO (225-bees trenscript) with a‘ RNA polymereee end control protein (lene l) or GerE (lens 2). (b) sigK trenscription from Xbol-digested pBK16 (I70-bese trenscript) with control protein (lens I) or GerE (lens 2). (c) pert trenscription from pSCHd digested with Born"! (lens I. I'M-bees trenscript) or Hindi“ (lenes 2 to 4. flint-bees trenscript) with 0‘ RNA polymereee elone (lenes I end 2). or with control protein (lene 3) or GerE (lens 4) edded. (d) cotA trenscription from pK823 digested with Neal (lene l. l3l-bese trenscript) or Ele (lens 2. l49-bese trenscrilfil. from Naoldigested pKS22 (lenes 3 end 4. l3l-bese trenscript). end from Neal-digested pK824 (lenes 5 end 0. l8l-bue trenscript) with rr‘ RNA polymereee elone (lenes I end 2). or with control protein (lenes 3 end 5) or GerE (lenes 4 end 6) edded. [s-"PlUTP wee the lebeled nucleotide in the experiments shown in (c) end (d). l04~6 from gel~purified 6‘ and B. subtilis core RNA poly- merase (data not shown). GerE protein had no efi‘ect on gerE transcription in vitro by partially purified a" RNA polymerase (Fig. 6(c). lane 4). The effect of GerE on cotA transcription in vitro varied depending on the particular DNA template used (Fig. 6(d)). GerE had little effect on cotA transcrip- tion from a template containing “5 bp of DNA upstream from the cotA transcriptional startsite (lane 6); however. GerE inhibited cotA transcription approximately twofold (in 2 independent experi- ments) from a template containing approximately 430 bp of upstream DNA (lane 4). Discussion (a) Transcriptional adimlion by GerE We have identified binding sites for GerE at the 5' ends of cotB and 0010. coat protein genes whose transcription depends on the appearance of GerE during sporulation. We have also shown that GerE stimulates cotB and cotC transcription in vitro by a" RNA polymerase and that the part? gene itself can be transcribed in vitro by 0‘ RNA polymerase. These results support the view (Zheng & Losick. I990) that In the mother cell. 6‘ RNA polymerase first directs the transcription of perE. then acts in conjunction with the product of gerE' to direct the transcription of cotB. calf) and. perhaps. other late- activated sporulation genes. Interestingly. the region of GerE-conferred protection from DNase I action in cotB (4| to 47 bp) was approximately twice the length of the two separate protected regions in cotC (l6 to [9 bp for binding site I and 2| to 22 bp for binding site 2). Our interpretation of this obnrvation is that cotB contains tandem GerE binding sites and that cotC contains two separate GerE binding sites. Inspection of the sequences in the prowcted regions reveals three similar 5 bp sequences. two (TGGGT and TAGGC) in cotB and one ('I‘GGGC. found in binding site I) in cotC. If these are recognition sequences for GerE. then a possible consensus sequence for the GerE binding site is TPuGGPy. The closest match to this consensus in cotC binding site 2 is the sequence TGGAC. Nevertheless. binding site 2 appears to be sufficient to mediate GerE-stimulated transcription of on“). since a DNA template with less than half of binding site I (produced by cleavage with final" at position - I33) retains the ability to be transcribed in vitro by 0‘ RNA polymerase in the presence of GerE (R. Pl. & L. K.. unpublished results). The downstream boundaries of the GerE binding sites in cotB (position -36) and in cotC (position -56 for binding site 2) are immediately adjacent to or near regions of DNA that are expected to interact with RNA polymerase (that is. the promoters). By analogy with the positioning of the binding sites for several well-characterised. positively acting regula- tory proteins. the DNA.bound GerE can be con- sidered to be appropriately positioned to stimulate RNA polymerase to transcribe from the cotB and 193 L. Zheng et al. cotC promoters. For example. the binding region (the tandem operator sites 0., and O") for the phage l. repressor is located between 34 and 74 bp upstream from the transcriptional startsite for the promoter (P...) from which the repressor stimulates transcription of its own structural gene (cl) by E coli RNA polymerase (Johnson at al.. I979; Meyer & Ptashne. “380). As another example. the catabolite gene activator protein (CAP) binds to sequences (typically protecting about 25 bp) centered from about 4| to l0? bp upstream from the transcrip« tional startsite of genes whose transcription it stimulates (de Crombrugghe et al.. I984). An added significance of our demonstration in vitro that GerE c'an'bind to and stimulate transcrip- tion from the regulatory regions of genes under its control is that GerE is the prototypical representa- tive of the putative regulatory domain of a large and diverse group of procaryotic transcriptional activator proteins. Thus. GerE exhibits high overall similarity to the mOH-terminal regions of certain mgulator members (the ngA sub-group. which includes DegU. ComA and FixJ) of the family of two-component sensor and response regulator systems in bacteria (Gross at al.. l989; Kahn & Ditts. l99l). The NIL-terminal region of the response regulator proteins contains a site for phos- phorylation by the sensor component of the two- component system. and its phosphorylation state influences the activity of the ODOR-terminal domain with to transcriptional activation (Kofoid a Parkinson. l988; Nixon at al.. “86). Strikingly. GerE. which lacks the NH,- terminal domain. Is highly similar along its entire 72 amino acid length to the COOH- terminal domain of the ngA subgroup of response regulator proteins. No member of the ngA sub-group has (to our know- ledge) been shown to bind to DNA or activate transcription in vitro. but our results reinforce the view that the GerE-like. DOOR-terminal domain of these proteins is directly responsible for the tran- scriptional activation of genes under the control of this sub- -group of response regulator proteins. Likewise. the similarity of GerE to the COOH terminus of E. coli MalT. a large regulatory protein whose Nflz-terminal region is dissimilar to the phosphorylation domain of the two-component regulator proteins (Gross et al.. l989). once again suggests that the GerE like region of MalT could be responsible for DNA- -binding and transcriptional activation by this regulatory protein (Richet at al.. l99l; Vidal-lngigliardi et al.. IOOI). Finally. the similarity of GerE to the OOOH-terminal domain (region 4) of sigma factors (Kahn & Ditta. l99l) reinforces the view that this domain mediates the recognitionofthe —35regronofpromoters (Gardella at al.. l989; Siegele et al.. l989). (b)Cmm/wpmmsmized a‘ RNA polymerase has been shown to transcribe from the 0010 and sigK (previously called spa! VCB) Sporrrlation Regulatory Protein GerE - 35 - to asset—vs ac - 1 7 hp - ”-u-. s 1 7K mtscaeacmaqacsecct ccceetcacatncatftscatatsgee cotA attttttetaACcat caeqt ccttattetcn‘l'taac‘lnaetaecaat cotD ttecstcaeaACataesceecttatttr.tCA‘l’Aaet'l'Mtsttctast gar: tetaaacqt CACctcctecqccctt cttaCA‘nt ns‘l'Atctccac-tst cor. ttgastta'ttCaaeaaataaatqteacaO'tAt atatqcaataegc cotC aactqt ccaancqcaaaatc tactc'cCQ‘tAt armaments Fine 1. Alignment of promoters transcribed by a" RNA polymerase. The nucleotide sequences of the n‘gK. cotA. cotD and gerfl' promoter regions (see the text for references) are aligned with respect to conserved nucleo— tides(capital letters) in the - l0 and -35 regions relative to the transcriptional startsites (underlined). Shown above are the consensus -l0 and -35 sequences. separated by l7 bp. and shown below are the cotB and cathromoterregionswithmatchestothswnsensus indicated by capital letters (a l bp gap was introduced intothecotCsequencebetweenthe-ltland -35 "sion-)- ' promoters in vitro (Kroos d al.. I989). Efficient transcription of sigK by a“ RNA polymerase also required a small. DNA-binding protein that is the product of the spoIIID gene (Kroos d al.. I989; Kunkel el al.. I989). Using a“ RNA polyrnersse reconstituted from gel-purified It" and B. subtilis core RNA polymerase. we find that perE and cotA. like cotD. are transcribed efficiently by a" RNA polymerase in the absence of SpoIIID. findings that confirm and extend the results of studies on the regulation of these genes in viva (Cutting er al.. I989; Sandman at al.. l988). As shown in Figure 7 and as noted previously (Foulger a Errington. l99l; Zheng & Losick. 1990). the promoter regions of cotD (ZMng & Losick. I990). gerE (Cutting at al.. I989). 00“ (Sandman et al.. I988) and sigK (Kunkel at al.. 1988) each contain sequences similar to CATA---TA at about position —l0 relative to their transcrip- tional startsites. Figure 7 also shows that the cotB and 00:0 promoters. which were transcribed weakly by a" RNA polymerase in the absence of GerE. display some similarity to the CATA-«TA sequence. Interestingly. the putative - l0 consensus uquence for c‘-recognised promoters is similar to the sequence CATACA—T. which is conserved in the - l0 region of promoters transcribed by RNA poly- merase containing the related sporulation sigma factor. a" (see Roels at al. (l992) and Foulger a Errington (l99l) for recent compilations of promoters recognised by I!'5 RNA polymerase). Unlike promoters recognised by 05 RNA poly- merase. however. promoters recognised by 0" RNA polymerase that have been characterised to date display only a limited region of similarity to each other in their —35 regions. The sequence AC is. however. found l7 bp upstream from the -l0 region in the four promowrs transcribed by a" RNA 194 l047 polymerase in the absence of GerE. but only the C is found at the corresponding position in cotB and cotC. We note that three other promoters that are inferred to be transcribed by 6" RNA polymerase in the absence of GerE. namely. spoVJP2 (Foulger & Errington. l99l). calm (Zheng t. Losick. I990). and the promoter for the newly discovered coat protein gene cotf' (Cutting at al.. l99lb). contain —35 and - l0 sequences that strongly conform to the sequences AC and CATA---TA. respectively. at a spacing of l5 to l7 bp. Mutational analyses will be needed to determine whether the AC and- CATAmTA sequences are important for promoter recognition by a" RNA polymerase. (c) Eject o] GerE on the trenscription of cotD and sigK GerE stimulated cotD transcription in vitro by 0‘" RNA polymerase two- to threefold (Fig. 8(a)). This result is in qualitative agreement with the finding that calD-locZ expression is reduced about sevenfold in gerE mutant cells (Zheng & Losick. I990). Inspection of the cotD promoter region (Zheng t Losick. I990) reveals a l2 bp sequence (AAAA- TAGG'IC'I'I‘) at positions -43 to -54 with ten matches to a sequence (positions -89 to -80) protected by GerE in the cotB promoter region (Fig. l). Within the l2 bp sequence in the call) promoter region is a 5 bp sequence (TAGGT) that conforms to the putative consensus binding sequence for GerE described above. It will be inter- esting to determine whetheF'CerE binds to this sequence. since it would appear to position GerE appropriately to stimulate RNA polymerase. as discussed above. GerE completely inhibited sigK transcription by partially purified 6‘ RNA polymerase (Fig. 5(b)). The partially purified 0‘ RNA polymerase used in this experiment contained a small amount of SpoIIID. thus permitting adequate transcription of sigK. Inhibition of sigK transcription by GerE was unexpected. since expression of a sigK-locZ fusion was about normal in peril mutant cells (Kunkel at al., l988). Inspection of the sigK promoter region (Kunkel et al.. l988) reveals a l5 bp sequence (ACATATAGGC'I'I‘TI‘G) at positions -4 to +ll with l2 matches to a sequence (positions -tl to —55) protected by GerE in the call? promoter region (Fig. l). Within the l5 bp sequence in the sigK promoter region is a 5 bp sequence (TAGGC) that conforms to the putative consensus GerE binding sequence. In addition. this 5 bp sequence is repeated in inverted orientation at positions + II to 4» l5. Thus. there may be two GerE binding sequences near the startsite of sigK transcription and GerE bound at these sites may prevent I!" RNA polymerase from transcribing sigK. If these sites do mediate repression of sick transcription by GerE. it could explain why a sigK-locZ fusion was expressed equally in wild-type or gerE mutant cells. since the fusion was created by insertion of a transposon (Tn9l7lac) 4 bp downstream from the sigK tran- I048 w \W ,. SE Oar 'X ‘ \EotA cotB 359“ a {2:3, I . .... < tc / FIJI“! 8. Regulatory efi’ects of SpoIIID and (lerI-I during stages IV and V of sporulation. The efi‘ecta of SpoIIID and Gerl') on transcription by 6‘ RNA poly- merase in vitro are illustrated. As noted in the text. studies is vice also support some of the regulatory reflects depicted. (a) During stage IV of sporulation. SpoIIID stimulates transcription of sigK and inhibits transcription ofcotD. The perE gene is transcribed by a1 RNA poly- merase. (b) During stage V of sporulation. Geri-2 inhibits transcription ofss'gK and cold. and stimulates transcrip- tion cfcotl). cotB and cotC. scriptional startsite (Kunkel et al.. l988). and this would have presumably disrupted the putative GerE binding site. (d) Opposite ejects ofOerE and 81301110 help to drive the mother-cell regulatory cascade GerE and SpoIIID exert opposite effects on a‘directed transcription of both cotD and sigK (Fig. 8). It has been shown that SpoIIID stimulates sigK transcription and inhibits cotD transcription in vitro (Kroos et al.. I989; and Fig. 8(a)). These properties of SpoIIID led to the suggestion that inactivation or sequestering of SpoIIID during sporulation causes a switch from transcription of sigK (and perhaps other stage IV sporulation genes) to transcription of cotD (and perhaps other stage V genes) (Kroos et al.. I989). Recently. it has been shown thet the level of SpoIIID decreases at the appropriate time during sporulation to cause such a switch (Halberg s Kroos. l992). Furthermore. the decrease in the level of SpoIIID correlates with the pearance of 0", suggesting that the appearance of :2 initiates the switch. We have shown here that 6" RNA polymerase transcribes gerE (Fig. 6(c)). Thus. the appearance of a“ RNA polymerase beginning at about hour 4 of sporulation (Cutting et al.. I989; Lu et al.. I990) would result not only in a declining level of SpoIIID, but also in a rising level of GerE. We have also shown here that GerE inhibits sigK tran- scription (Fig. 6(b)) and stimulates cotD transcrip- tion (Fig. 8(a)) by 0" RNA polymerase in vitro 195 L. Zheng et sl. spoIIID—db o" —>oore /\/ Stags w—esugo v 9. latory interactions controlling the levels of h‘polllll. and (IerI-I govern the stage IV to \' transition in the mother cell. SpoIIID stimulates sigK transcription. leading to 6" production (Kroos at al.. l989). 0‘ RNA polymerase transcribes grrlt'. leading to (:‘erl'I production. The sppmranu- of 0‘ ram a rlr'r‘n'rus' in the level of S'sillll) (Halberg & Knssr. I992). (ierl-I inhibits transcription of sigK . down-regulating e‘ produc- tion. Thus. 0‘ RNA polymerase functions during both stage IV and stage V. but a declining level of SpoIIID and a rising level of Geri-I switch the pattern of s‘slin-r-tr-rl gene expansion from the stage I\' pattern to the stage \' pattern. (Fig. 8(b)). Because GerE exerts the oppodte cfl'ects of SpoIIID on o"-directed transcription of sigK and call). the appearance of Gerl-I would reinforce the switch in the pattern of mother-cell gene cxprersrion previously postulated to be brought about by inactivation or sequestering of Spolll I). The reason for the apparent redundancy in the switch is unclear. Perhaps SpoIIID prevents premature expmsion of oath (and perhaps other stage V genes) during the stage (IV) of spare cortex formation so that the c‘ produced initially directs exprosrion of sigK (autogenous regulation): perlt‘. and W involved in cortex formation (Halberg & Kroos. l992). Accumulation of GerE would terminate this period by diverting 6‘ RNA polymerase away from transcription of sigK (and perhaps other stage IV genes) and would initiate the stage (V) of spore mst formation by directing a“ RNA polymerasc to tran- scribe genes encoding spore coat proteins. According to this model. the regulatory interactions illustrated in Figure 9 co-ordinste the levels of SpoIIID. a“ and GerE so as to produce a molecular switch governing the transition from the stage IV pattern of mother-cell gene exruession to the stage \' pattern. The effects of GerE on transcription ofgerlt' and cotA by 6" RNA polymerase in vitro are consistent with the effects ofa gerlt‘ mutation on expression of these genes in viva. GerE had no effect. on transcrip« tion of the gerb' gene in vitro (Fig. 6(c)) and expression of a gerE-loez fusion is normal in a prrls' mutant (Cutting et al.. I989). The effect ofGerI'I on cotA transcription in vitro varied from little effect with a template containing ”5 bp of DNA upstream from the transcriptional startsite to s modest (but reproducible). twofold inhibition with s template containing approximately 430 bp ol upstream DNA (Fig. 8(d)). If GerE inhibits cotA transcription by binding to DNA. this result suggests that it must do so by binding to a site(s) more than llti bp upstream from the startsite ot transcription. In a previous study (Cutting et al. Sporulation Regulatory Protein GerE I989). a aolA-lacZ fusion containing as little as 300 bp of DNA upstream from the cotA transcrip- tional startsite was found to be expressed about threefold higher in a gerE mutant relative to wild- type cells. In summary. we have presented biochemical evidence that GerE is a regulatory protein capable of either stimulating or inhibiting transcription of particular genes in the mother-cell line of gene expression. In this respect. GerE appears to be analogous to Spollll); however. the ordered appearance of first SpoIIID. then GerE. and the opposite effects of these two proteins on the tran- scription of genes like sigK and call) presumablv ensures proper flow of the regulatory cascade (7 heng a Losick. I990) controlling mother-cell gene expressioan the cases of cotB and cod". Geri" binds to specific sequences immediately adjacent to the promoter and stimulates transcription by a" RNA polymerase. This finding an the previous pro- posal (Zheng a Losick. I990) that GerE directly activates the. expression of genes in the terminal temporal clan of mother-cell expressed genes. We thank Simon (‘utting and Kathleen Sandman for providing plasmids. This work was supported by the Michigan Agricultural Experiment Station. by NIH grant (“143585 to L. K. and by NIH grant (“"8568 to R. L. References Bolivar. F.. Rodrigues. R. I... Greene. P. .I.. Betlach. H. C.. Heyneker. H. L. & Boyer. H. W. (I977). Construction and characterisation of new cloning vehicles: II a multipurpose cloning system. 0m 2. 95—I I3. Cutting. 8. (IM). Genetics and properties of germination mutants of Bacillus subtilis. D.Phil. thesis. Oxford University. Cutting. S. & Mandi-Istarn. J. (I986). The nucleotide sequence and the transcription during sporulation of the cart? gene of Bacillus subtilis. J. Gas. Microbial. I32. film-3024. Cutting. 8. M. & Vander Horn. P. 8. (I990). Genetic analysis. In Mdecslar ' ' ”Mfor Bacillus. pp. 27-74. John Wiley & Sons Ltd. Chichester. UK. Cutting. 8.. Panzer. S. & Losick. R. (I989). Regulatory studies on the promoter for a gene governing synthe- sis and ...!!!"ny of the spore coat in Bacillus subtilis. J. Mol. Bid. 207. 393-404. Cutting. 8.. Oke. V.. Driks. A.. Losick. R.. Lu. 8. & Kroos. L. ( I990). A forespore checkpoint for mother cell gene expremion during development in B. subtilis. Cell. 62. 239-250. Cutting. 8.. Reels. S. & Losick. R. (I99Ia). Sporulation operon spol VP and the characterization of mutations that uncouple mother-cell from forearm geneexpression in Bacillus subtilis. J. Nd. Bid. 22I. I237-I256. Cutting. 8.. Zheng. L. & Losick. R. (I99lb). Gene encoding two alkali-soluble components of the spore coat from Bacillus subtilis. J. Baderid. I73. 29l5—29I9. Davis. R. W.. Hotstein. I). & Roth. .I. R. (Will). Advasred Barterinl ”radian. (‘old Spring Harbor Laboratory I’m-s. (‘old spring Harbor. NY. 196 l0-l9 de Crombrugghe. B.. Busby. S. & IIuc. H. (I984). Cyclic AMP receptor protein: role in transcription activa- tion. Science. 224. Bill-838. De Lencastre. H. & Piggot. I'. (I979). Identification of different sites of expression for spa loci by trans- formation of Bacillus subtilis. J. (lea. Alicrabiol. I14. 377-389. Donovan. W.. Zheng. L.. Sandman. K. & Losick. R. (I987). (lenes encoding spore coat polypeptides from Bacillus subtilis. J. Mol. Biol. I96. I-Itl. Errington. .l. (I986). A general method for fusion of the Extericlria coli larZ gene to chromosomal genes in Bacillus subtilis. J. lies. Microbial. I32. 2953-2966. Feng. I‘. & Aronson. A. I. (I986). Character-infirm of a Bacillus subtilis germination mutant with pleiotropic alterations in spore cost structure. l' arr. Microbial. I3. 2‘.’.l —2"6. Foulger. I). & Errington. .I. (I99I). Sequential activation of dual promoters by different sigma factors main- tains npol’J expression during succesrrive develop- mental stages-of Barillss subtilis. Md. Microbial. S. I383-l373. ' (‘.ardella T.. Hoyle. H. & Su-kind. ll. N. (I989). A mutant E. coli 6" subunit of RNA polymerase with altered promoter specificity. J. Mal. Bid. 206. 579-5“). (iron. R.. Arico. B. a Rappuoli. R. (I989). Families of bacterial signal-transducing proteins. Nd. Microbial. 3. l66I-l067. Huger. I). & Burgess. R. (IM). I-3lution of proteins from SDS polyacrylamide gels. removal of SDS. and , «maturation of enzymatic activity: results with sigma subunit of E. coli RNA polymerase. wheat- germ topoisomerase. and other enzymes. Anal. Biocsers.109.76-80. Halberg. R. & Kroos. L. (I992). Fate of the SpoIIID switch protein during Bacilrls sslrliliss rulation depends on the mother-cell sigma factor 0:01“ Bid. In the press. Hasnain. 8.. Sammons. R.. Roberts. I. 6 Thomas. I‘. )I. (I985). Cloning and deletion analysis of a genomic segment of Bacillus subtilis coding forthesdk A. B. (‘ (succinate dehydrogenase) and gerfl' (spore germina- tion) loci. J. (lea. Microbial. I3I. 2269-2279. Henner. I). .I.. Yang. M. a Ferrari. E. (l988). Localization of Bacillus subtilis sarUflII’) mutations to two linked genes with similarities to the conserved pro- caryotic family of two-component signalling systems. J. W. I70. 5I02-5l09. Holland. 8. K.. Cutting. 8. s Mandelstam. J. 0987). Possible DNA-binding nature of the regulatory pro- teins encoded by rpolll) and gut. involved in sporulation of Bacillus subtilis. J. (lea. Nicrolrid. I33. 238I-239l. Jenkinson. H. F. & Lord. H. (I983). Protease deficiency and itsa-ocratronwrthdefectainaporecoatstrue ture, germination and resistance properties in a mutantofBacillussubtilis. J. Gert. Hierobid. I29. 2727-2737. Johnson. A. D.. Meyer. 8. J. I. I’tashne. H. (I979). Interactions between DNA bound repre-ors govern regulationby thelphagerepreuor. Proc. Nat. Arad. Sci" 0.8..A 76. bail-5065. Kahn. I). & Ditta. G. (I99I). Modular structureoffixJ: homology of the transcriptional activator domain with the -35 domain of sigma factors. Mal. Microbial. S. 987-997. Karmazyn-Campelli. C.. Bonamy. C.. 8avelli. B. & Stragier. P. "9%). Tandem genes encoding s-factors I050 for consecutive steps of development in Bacillus subtilis. Genes Develop. 3. ISO-I57. Kofoid. E. C. & Parkinson. J. 8. (l988). Transmitter and receiver modules in bacterial signalling proteins. Proc. Nat. Acad. Sci. USA. 85. “SI-4985. Kroos. I... Kunkel. B. s Losick. R. ("89). Switch protein elteru specificity of RNA polymerase containing a compartment-specific sigma factor. Science. 243. 526—529. Kunkel. I3.. Sandman. K.. Panzer. 8.. Youngman. l’. a Lusiclt. R. (I988). The promoter for a sporulation genein thespoll'ClocusofBacillussubtilisand its use in studies of temporal and spatial control of gene expression. J. Bacteriol. I70. 35l3-3522. Kunkel. 3.. Kroos. I... Poth. H.. Youngman. P. & Losick. R. 0989). Temporal end spatial control of the mother-cell regulatory gene spoIIID of Bacillus subtilis. (lenes Develop. 3. I735—l744. Kunkel.ll..l41siclt. R. & Stragier. P. "990). The Bacillus subtilisafene for the developnnntel transcription generated by excision of a dispensable DNA element containing a sporulation recombinase gene. Genes Develop. 4. 525—535. Kunst. P.. Deberbouille. H.. Msadelt. T.. Young. M.. Msuel. C.. Karemata. D.. Klier. A.. Repoport. G. & Dedonder. R. (l988). Deduced polypeptide encoded bytheBacillussubtilisschlocusehare with two-component sensor-regulatory systems. J. Bacteriol. I70. 5093-5IOI. Leernmli. U. K. (”70). Cleavage of structural proteins during the a-embly of the heed ofbecteriophage T4. Nature (London). 227. Gal—685. Losick. R. & Stragier. P. (I992). Orinoco- reguletion of cell-type-specrfic gene expression during development in Bacillus subtilis. Nature (London). 355. 00l-604. Lu. 8.. Halberg. R. tKroos. b.0990). Processingofthe mother-cell a factor. 6‘. may depend on events occurring in the forespore during Bacillus subtilis development. I’roc. Nat. Add. Sci. U.S.A. 87. 9722—9726. Maniatis. T.. Pritsch. E. F. & Sembrook. J. "982). Molecular Cloning. A Laboratory Nenual. Cold Spring Harbor Laboratory Press. Cold Spring Harbor. NY. Ilexam. A. in Gilbert. W. (IND). Sequencing end-labeled DNA with base-specific chemical cleavagu. JIM anylnd. 63. 499-500. Meyer. 8. J. & Pteshne. M. (I980). Gene regulationatthe right operator (0.) of bacteriophage A. III. 1 repre-or directly activates gene transcription. J. Nd. Bid. I39. nos-m. Mair. A. (l98l). Germination properties ofa spore coat. defective mutant of Bacillus subtilis. J. Bacteriol. I46. ll06-lll6. Nixon. T.. Roneon. C. W. & Ausubel. P. l. (I986). Two-component regulatory systems responsive to environmental stimuli share strongly conserved domains with the nitrmn e-imilation genes ntrB end ntrC. Proc. Nd. Acad. Sci.. USA. 83. 7850-7854. Panzer. 8.. Losick. R.. Sun. I). I. Setlow. P. 0989). Evidence for an additional temporal class of gene expression in the forespore compartment of lating Bacillus subtilis. J. Bacteriol. I7I. 5mm. Richet. 8.. Videl-Ingigliardi. D. I. Reihaud. 0. (l99l). Anew mechanism for coactivation of transcription initiation: repositioning of an activator triggered by the binding of a second activator. Cell. 66. ll85—Il95. 197 L. Zhenget al. Reels. 8.. Driks. A. t. Losick. R. (1992). Characterization of spoll’A. e sporulation gene involved in cost morphogenesis in Bacillus subtilis. J. Bacteriol. I74. 575—585. Sembroolt. .I.. Fritsch. E. F. a Maniatis. 1'. (I989). Molecular Cloning: A laboratory Manual. Cold Spring Harbor Laboratory Press. Cold Spring Harbor. NY. Sandman. K.. Kroos. I... Cutting. 8.. Youngman. P. & Losick. R. (l988). Identification of the promoter for a spore coat protein gene in Bacillus subtilis and studies on the regulation of its induction at a late stage of sporulation. J. Mal. Biol. 200. “I473. Sanger. P.. Nicklen. S. t Coulson. A. R. ("77). DNA sequencing with chain-terminating inhibitors. I’m. Nat. Acad. Sci.. USA. 74. “63-5467. .D.A.. Hu. J. C.. Welter. W. A. t0ro-.C. A. (I989). Altered ”promoter recognition by mutant forms of the a” subunit of Escher-His coli RNA polymerase. J. Nd. Biol. 206. 59l-603. Stevens. C. M. s Errington. J. (I990). Difl'erentiel gene expression during sporulation in Bacillus subtilis: structure end regulation of the spoIIID gene. llol. Microbial. 4. 543—552. Stragier. P.. Kunkel. I3.. Kroos. I.. & IAIICh. R. (l989). Chromosomal rearrangement generating a composite gene for a developmental transcription factor. Science. 243. 507—5”. Streuch. I. A.. Spiegelman. 0. I3.. Perego. IL. Johnson. W. C.. Burbulys. D. & Hoch. J. A. (IW). The transition state transcriptional regulator aer of Bacillus subtilis is a DNA binding protein. EMBO J. 3. l6l5-I62l. Sun. D.. Stragier. I'. s Setlow. P. 0989). Identification of a new o-factor involved in compertmentsliaed gene expression during sporulatisnrof Bacillus subtilis. Genes Develop. 3. l4l-I49. W- Su-man. II. D. s Setlow. P. (I99I). Cloning. nucleotide sequence. and regulation of the Baciuus subtilis 'pr gene which codes for the protease that initiates degradation of smell. acid-soluble. proteins during spore germination. J. Bacteriol. I73. tor-300. Tabor. S. A. Richardson. C. C. 0985). A bacteriophage 17 RNA polymerase/promoter system for controlled exclusive expre—ion of specific genes. Proc. Nat. And. Sci.. US.A. 81. l074—I078. Videl-Ingiglierdi. D.. Richet. E. & Reibaud. 0. (IBM). Two MeIT-binding sites in direct repeat: a structural motif involved in the activation of all the promoters ofthemaltoseregulonsinBscAeridioroliend Kldsiells pneumonia. J. Nd. Bid. 2I8. 323-334. Weinreuch. Y.. Guillen. N. a Dubnau. D. A. "989). Sequence and transcription mapping of Bacillus subtilis competence genes cornB and canal. one of which is related to a family of bacterial regu determinants. J. Bacterid. "I. 5362-5375. Yenisch-Perron. C.. \‘ieire. J. & lie-ing. J. (I985). Improved III: phage cloning vectors and host strains: nucleotide sequence-i of the Il3mpl8 and pUCI9 vectors. Gene. 33. Illa-l l9. Zheng. L "900). Spore coat protein genes and their regulation in Bacillus subtilis. Ph.D. thesis. Harvard University. Zheng. Ltlnsick. R. (ION). Caecederegulationofspore coetgeneexpreseionin Bacillussubtilis. J. Nd. Bid. 211645—660. Edited by M. Gottesmae BIBLIOGRAPHY 6‘- BIBLIOGRAPHY Alper, 8., Duncan, L. & Losick, R. (1994). An adenosine nucleotide switch controlling the activity of a cell type-specific transcription factor in B. subtilis. Cell, 77, 195-205. Antoniewski, C., Savelli, B. & Stragier, P. (1990). The spoIIJ gene, which regulates early developmental steps in Bacillus subtilis, belongs to a class of environmentally responsive genes. J. Bacteriol., 172, 86-93. Bach, M. L. & Gilvarg, C. (1966). Biosynthesis of dipicolinic acid in sporulating Bacillus megaterium. J. Biol. Chem, 241, 4563-4564. Banuett, P., Hoyt, M. A., MacFarlane, L., Echols, H. & Hershowitz, I. (1986). hle, a new Escherichia coli locus regulating lysogeny and the level of bacteriophage lambda cII protein. J. Mbl. Biol., 187, 213-224. Bradford, M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 72, 248-254. Burbulys, D., Trach, K. A. & Hoch, J. A. (1991). Initiation of sporulation in B. subtilis is controlled by a multicomponent phosphorelay. Cell, 64, 545-552. Carlson, H. C. & Haldenwang, W. G. (1989). The 03 subunit of Bacillus subtilis RNA polymerase is present in both forespore and mother cell compartments. J. Bacteriol., 171, 2216-2218. Carter, H. L., III & Moran, C. P., Jr. (1986). New RNA polymerase sigma factor under spoO control in Bacillus subtilis. Proc. Natl. Acad. Sci. USA, 83, 9438-9442. Cheng, H. H., Muhlred, P. J., Hoyt, M. A. & Echols, H. (1988) . Cleavage of the cII protein of phage lby purified HflA protease: Control of the switch between lysis and 198 199 lysogeny. Proc. Natl. Acad. Sci. USA, 85, 7882-7886. Church, B. D. & Halvorson, H. (1959). Dependence of the heat resistance properties of spores on their dipicolinic acid contents. Nature, 183, 124-125. Collado-Vides, J., Magasanik, B. & Gralla, J. D. (1991). Control site location and transcriptional regulation in Escherichia coli. Microbiol. Rev., 55, 371-394. Coote, J.G. (1972). Sporulation in Bacillus subtilis: Characterization of oligosporagenous mutants and comparison of their phenotypes with those of asporagenous mutants. J. Gen. Microbiol., 71, 1-15. Cutting, S., Driks, A., Schmidt, R., Kunkel, B. & Losick, R. (1991a). Forespore-specific transcription of a gene in the signal transduction pathway that governs pro-0K processing in Bacillus subtilis. Genes Dev., 5, 456-466. Cutting, S. & Mandelstam, J. (1986). The nucleotide sequence and the transcription during sporulation of the gerE gene of Bacillus subtilis. J. Gen. Microbiol., 132, 3013- 3024. Cutting, S., Oke, V., Driks, A., Losick, R., Lu, S. & Kroos, L. (1990). A forespore checkpoint for mother-cell gene expression during development in Bacillus subtilis. Cell, 62, 239-250. Cutting, 8., Panzer, S. & Losick, R. (1989). Regulatory studies on the promoter for a gene governing synthesis and assembly of the spore coat in Bacillus subtilis. J. M01. Biol., 207, 393-404. Cutting, S., Roels, S. & Losick, R. (1991b). Sporulation operon spoIVF and the characterization of mutations that uncouple mother-cell from forespore gene expression in Bacillus subtilis. J. M01. Biol., 221, 1237- 1256. Cutting, s., Zheng, L. & Losick, R. (1991c). Gene encoding two alkali-soluble components of the spore coat from Bacillus subtilis. J. Bacteriol., 173, 2915-2919. Cutting, S. M. & Horn, P. B. V. (1990). Genetic analysis. In Molecular Biological Methods for Bacillus. (Harwood, C. R.& Cutting, S. M.), John Wiley & Sons, New York, NY. Daniel, R. A., Drake, S., Buchanan, C. E., Scholle, R. & Errington, J. (1994). The Bacillus subtilis spoVD gene 200 encodes a mother-cell-specific penecillin-binding protein required for spore morphogenesis. J. Mbl. Biol., 235, 209- 220. Daniel, R. A. & Errington, J. (1993). Cloning, DNA sequence, functional analysis and transcriptional regulation of the genes encoding dipicolinic acid synthetase required for sporulation in Bacillus subtilis. J3 Mbl. Biol., 232, 468-483. Dombroski, A. J., Walter, W. A. & Gross, C. A. (1993). Amino-terminal amino acids modulate O-factor DNA-binding activity. Genes & Dev., 7, 2446-2455. Donovan, W., Zheng, L., Sandman, K. & Losick, R. (1987). Genes encoding spore coat polypeptides from Bacillus subtilis. J. Mbl. Biol., 196, 1-10. Driks, A. & Losick, R. (1991). Compartmentalized expression of a gene under the control of sporulation transcription factor CE in Bacillus subtilis. Proc. Natl. Acad. Sci. USA, 88, 9934-9938. Dubnau, D. & Davidoff-Abelson, R. (1971). Fate of transforming DNA following uptake by competent B. subtilis. J. Mbl. Biol., 56, 209-221. Dubnau, E., Weir, J., Nair, G., Peters, H. L., III, & Moran, C.P., Jr., Smith, I. (1988). Bacillus sporulation gene spoOH codes for 030 (OH) . J. Bacteriol., 170, 1054-1062. Duncan, L. & Losick, R. (1993). SpoIIAB is an anti-O factor that binds to and inhibits transcription by regulatory protein 0? from Bacillus subtilis. Proc. Natl. Acad. Sci. Echols, H. & Green, L. (1971). Establishment and maintenance of repression by bacteriophage lambda: The role of the cI, cII and cIII proteins. Proc. Natl. Acad. Sci. USA, 68, 2190-2194. Eiglmeier, K., Honore, N., Iach, 8., Lin, B. C. C. & Cole, S. T. (1989). Molecular genetic analysis of FNR- dependent promoters. .MOl. Microbiol., 3, 869-878. Errington, J. (1986). A general method for fusion of the Escherichia coli lacz gene to chromosomal genes in Bacillus subtilis. J. Gen. Microbiol., 132, 2953-2966. Errington, J. (1993). Bacillus subtilis Sporulation: 201 Regulation of Gene expression and Control of Morphogenesis. Microbiological Review, 57, 1-33. Errington, J., Fort, P. & Mandelstam, J. (1985). Duplicated sporulation genes in bacteria: implication for simple developmental systems. FEBS Lett., 188, 184-188. Errington, J. & Mandelstam, J. (1986). Use of a lacz gene fusion to determine the dependence pattern of sporulation operon spoIIA in spa mutants of Bacillus subtilis. J. Gen. Microbiol., 132, 2967-2976. Errington, J., Wootten, L., Dunkerley, J. C. & Foulger, D. (1989). Differential gene expression during sporulation in Bacillus subtilis: regulation of the spoVU'gene. .Mbl. Microbiol., 3, 1053-1060. Fan, C. & Maniatas, T. (1991). Generation of p50 subunit of NF-kappa B by processing of p105 through ATP- dependent pathway. Nature, 354, 395-398. Feng, P. & Aronson, A. I. (1986). Characterization of a Bacillus subtilis germination mutant with pleiotropic alterations in spore coat structure. Curr. Microbiol., 13, 221-226. Fort, P. & Piggot, P. J. (1984). Nucleotide sequence of the sporulation locus spoIIA in Bacillus subtilis. J. Gen. Microbiol., 130, 2147-2153. Freese, B., Freese, E. B., Allen, B. R., Olempske-Beer, Z., Orrego, C., Varma, A. & Wabiko, A. (1985). Metabolic initiation of spore development. In Mblecular Biology of Microbial Differentiation. (Hoch, J. & Setlow, P.), Am. Soc. Microbiol., Washington, DC. Fujiwara, S., Zielinski, N. & Chakrabarty, A. M. (1993). Enhancer-like activity of algRl-binding site in alginate gene activation: positional, orientational, and sequence specificty. J. Bacteriol., 175, 5452-5459. Gholamhoseinian, A. & Piggot, P. J. (1989). Timing of spoII gene expression relative to septum formation during sporulation of Bacillus subtilis. J. Bacteriol., 171, 5747- 5749. Ghosh, S. (1990). Cloning of the p50 DNA subunit of NF-KB: Homology to rel and dorsal. Cell, 62, 1019-1029. Grossman, A. D. & Losick, R. (1988). Extracellular control of spore formation in Bacillus subtilis. Proc. Nat. 202 Hager, D. & Burgess, R. (1980). Elution of proteins from SDS polyacrylamide gels, removal of SDS, and renaturation of enzymatic activity: results with sigma subunit of E. coli RNA polymerase, wheat-germ topoisomerase, and other enzymes. Anal. Biochem., 109, 76-86. Halberg, R. & Kroos, L. (1992). Fate of the SpoIIID switch protein during Bacillus subtilis sporulation depends on the mother-cell sigma factor, OK. J. Mbl. Biol., 228, 840-849. Helmann, J. D. & Chamberlin, M. J. (1988). Structure and function of bacterial sigma factors. Ann. Rev. Biochem., 57, 839-872. Higgins, M. L. & Piggot, P. J. (1992). Septal membrane fusion - a pivotal event in bacterial spore formation? Mbl. .Microbiol., 6, 2565-2571. Ho, Y. S., Wulff, D. L. & Rosenberg, M. (1983). Bacteriophage kguxmein cII binds promoters on the opposite face of the DNA helix from RNA polymerase. Nature, 304, 703- 708. Hoch, J. (1994). The Phosphorelay Signal Transduction Pathway in the Initiation of Sporulation. In Regulation of Bacterial Differentiation. (Piggot, P. J., Moran, C. P.& Youngman, P.), American Society for Microbiology, Washington, DC. Hoyt, M. A., Knight, D. M., Das, A., Miller, H. I. & Echols, H. (1982). Control of phage 1 development by stability and synthesis of cII protein: role of the viral cIII and host hflA, himA and himD genes. Cell, 31, 565-573. Hughes, K. T., Gillen, K. L., Semon, M. J. & Karlinsey, J. E. (1993). Sensing Structural Intermediates in Bacterial Flagellar Assembly by Export of a Negative Regulator. Science, 262, 1277-1280. Igo, M. M. & Losick, R. (1986). Regulation of a promoter region that is utilized by minor forms of RNA polymerase holoenzyme in Bacillus subtilis. J; Mbl. Biol., 191, 615-624. Illing, N. & Errington, J. (1991a). Genetic regulation of morphogenesis in Bacillus subtilis: roles of GE and (3'F in prespore engulfment. J. Bacteriol., 173, 3159-3169. Illing, N. & Errington, J. (1991b). The spoIIIA operon of Bacillus subtilis defines a new temporal class of mother- 203 cell-specific sporulation genes under the control of the OE form of RNA polymerase. Mbl..Microbiol., 5, 1927-1940. Ireton, K. & Grossman, A. (1992a). Interaction among mutations that cause altered timing of gene expression during sporulation in Bacillus subtilis. J. Bacteriol., 174, 3185- 3195. Ireton, K. & Grossman, A. D. (1992b). Coupling between gene expression and DNA synthesis early during development in Bacillus subtilis. Proc. Nat. Acad. Sci. USA, 89, 8808-8812. Ishihama, A. (1993). Protein-protein communication within the transcription apparatus. J. Bacteriol., 175, 2483-2489. Jonas, R. M., Weaver, E. A., Kenney, T. J., Moran, C. P., Jr. & Haldenwang, W. G. (1988). The Bacillus subtilis spoIIG operon encodes both of and a gene necessary for GE activation. J. Bacteriol., 170, 507-511. Kahn, D. & Ditta, G. (1991). Modular structure of FixJ: homology of the transcriptional activator domain with the -35 domain of sigma factors. Mbl..Microbiol., 5, 987— 997. Karmazyn-Campelli, C., Bonamy, C., Savelli, B. & Stragier, P. (1989). Tandem genes encoding O-factors for consecutive steps of development in Bacillus subtilis. Genes & Dev., 3, 150-157. Kenney, T. J., Kirchman, P. A. & Moran, C. P., Jr. (1988). Gene encoding GE is transcribed from a Om-like promoter in Bacillus subtilis. J. Bacteriol., 170, 3058- 3064. Kenney, T. J., York, K., Youngman, P. & Moran, C. P., Jr. (1989). Genetic evidence that RNA polymerase associated with 0* uses a sporulation-specific promoter in Bacillus subtilis. Proc. Nat. Acad. Sci. USA, 86, 9109-9113. Kieran, R. (1990). The DNA binding subunit of nf-KB is identicel to KBFl and homologous to the rel oncogene product. Cell, 62, 1007-1018. Kirchman, P. A., DeGrazia, H., Kellner, E. M. & Moran, C. P., Jr. (1993). Forespore-specific disappearance of the sigma-factor antagonist SpoIIAB: implications for its role in determination of cell fate in Bacillus subtilis. Mbl. Microbiol., 8, 663-671. 204 Kroos, L., Kunkel, B. & Losick, R. (1989). Switch protein alters specificity of RNA polymerase containing a compartment-specific sigma factor. Science, 243, 526-529. Kroos, L. and Cutting, S. (1994). Intercellular and Intercompartmental Communication during Sporulation in Bacillus subtilis. In Regulation of Bacterial Differentiation. (Piggot, P. J., Moran, C. P.& Youngman, P.s, eds), 155-180, American Society for Microbiology, Washington, D.C. Kunkel, B., Kroos, L., Poth, H., Youngman, P. & Losick, R. (1989). Temporal and spatial control of the mother-cell regulatory gene spoIIID of Bacillus subtilis. Genes & Dev., 3, 1735-1744. Kunkel, B., Losick, R. & Stragier, P. (1990). The Bacillus subtilis gene for the developmental transcription factor UK is generated by excision of a dispensable DNA element containing a sporulation recombinase gene. Genes & Dev., 4, 525-535. Kunkel, B., Sandman, K., Panzer, S., Youngman, P. & Losick, R. (1988). The promoter for a sporulation gene in the spoIVC locus of Bacillus subtilis and its use in studies of temporal and spatial control of gene expression. J. Bacteriol., 170, 3513-3522. Kuroda, A., Rashid, M. H. & Sekiguchi, J. (1992). Molecular cloning and sequencing of the upstream region of the major Bacillus subtilis autolysin gene: a modifier protein exhibiting sequence homology to the major autolysin and the spoIID product. J. Gen. Microbiol., 138, 1067-1076. Kuspa, A., Plamann, L. & Kaiser, D. (1992). A- signalling and the cell density requirement for Myxococcus xanthus development. J; Bacteriol., 174, 7360-7369. Kutsukake, K., Ohya, Y. & Iino, T. (1990). Transcriptional analysis of the flagellar regulon in Salmonella typhimurium. J. Bacteriol., 172, 741-747. Kutsukake, K., Ohya, Y., Yamaguchi, S. & Iino, T. (1988). Operon structure of flagellar genes in Salmonella ty phimurium. Mbl. Gen. Genet., 214, 11-15. LaBell, T. L., Trempy, J. E. & Haldenwang, W. G. (1987) . Sporulation-specific sigma factor, 029, of Bacillus subtilis is synthesized from a precursor protein, P31. .Proc. N at. Acad. Sci. USA, 84, 1784-1788. Lazarevic, V., Margot, P., Soldo, B. & Karamata, D. 205 (1992). Sequencing and analysis of the B. subtilis lthABC divergon: a regulatory unit encompassing the structural genes of the N-acetylmuramoyl-L—alanine amidase and its modifier. J. Gen. Microbiol., 138, 1949-1961. Lonetto, M., Gribskov, M. & Gross, C. A. (1992). The (NO family: sequence conservation and evolutionary relationships. J. Bacteriol., 174, 3843-3849. Losick, R. & Stragier, P. (1992). Crisscross regulation of cell-type-specific gene expression during development in B. subtilis. Nature, 355, 601-604. Losick, R. & Kroos, L. (1989). Dependence Pathways for the Expression of Genes Involved in Endospore Formation in Bacillus subtilis. In Regulation of Prokaryotic Development. (Smith, I., Slepecky, R. A.& Setlow, P.), American Society for Microbiology, Washington, DC. Lu, 8., Halberg, R. & Kroos, L. (1990). Processing of the mother-cell 0 factor, 0“, may depend on events occurring in the forespore during Bacillus subtilis development. Proc. Natl. Acad. Sci. USA, 87, 9722-9726. Mandal, N., Su, R., Haber, S., Adhya, S. & Echols, H. (1990). DNA looping in cellular repression of transcription of the galactose operon. Genes Dev., 4, 410-418. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982). In Mblecular Cloning, a Laboratory.Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Margolis, P., Driks, A. & Losick, R. (1991). Establishment of cell type by compartmentalized activation of a transcription factor. Science, 254, 562-565. Martin, K., Huo, L. & Schlief, R. (1986). The DNA loop model for ara repression: AraC protein occupies the proposed loop sites in vivo and repression-negative mutations lie in some of these sites. Proc. Nat. Acad. Sci. USA, 83, 3654- 3658. Mason, J. M., Hackett, R. H. & Setlow, P. (1988). Regulation of expression of genes coding for small, acid- soluble proteins of Bacillus subtilis spores: studies using lacz gene fusions. J. Bacteriol., 170, 239-244. Masuda, E. S., Anaguchi, H., Yamada, K. & Kobayashi, Y. (1988). Two developmental genes encoding sigma factor homologs are arranged in tandem in Bacillus subtilis. Proc. Nat. Acad. Sci. USA, 85, 7637-7641. 206 Matsudaira, P. (1987). Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J. Biol. Chem., 262, 10035-10038. McMacken, R., Manteki, N., Butler, B., Joyner, A. & Echols, H. (1970). Effects of mutations in the cII and cIII genes of bacteriophage lcxlnmcromolecular synthesis in infected cells. J. Mal. Biol., 49, 639-655. Miller, J. (1972). In Experiments in Mblecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Ohnishi, K., Kutsukake, K., Suzuki, H. & Iino, T. (1990). Gene fliA encodes an alternative sigma factor specfic for flagellar operons in Salmonella typhimurium. Mbl. Gen. Genet., 221, 139-147. Oke, V. & Losick, R. (1993). Multilevel regulation of the sporulation transcription factor OK in Bacillus subtilis. J. Bacteriol., 175, 7341-7347. Otwinowski, Z., Schevitz, R. W., Zhang, R. G., Lawson, C. L. & Joachimiak, A. (1988). Crystal structure of trp repressor/operator complex at atomic resolution. Nature, 335, 321-329. Partridge, S. R., Foulger, D. & Errington, J. (1991). The role of CF in prespore-specific transcription in Bacillus subtilis. Mbl. Microbiol., 5, 757-767. Perego, M., Cole, S. P., Burbulys, D., Trach, K. & Hoch, J. A. (1989). Characterization of the gene for a protein kinase which phosphorylates the sporulation-regulatory proteins SpoOA and SpoOF of Bacillus subtilis. J. Bacteriol., 171, 6187-6196. Perego, M., Spiegelman, G. B. & Hoch, J. A. (1988). Structure of the gene for the transition state regulator aer: regulator synthesis is controlled by the spoOA sporulation gene in Bacillus subtilis. Molec..Microbiol., 2, 689-699. Peters, H. K., III & Haldenwang, W. G. (1991). Synthesis and fractionation properties of SpoIIGA, a protein essential for pro-OE processing in Bacillus subtilis. J. Bacteriol., 173, 7821-7827. Piggot, P.J. & Coote, J.G. (1976). Genetic aspects of bacterial endospore formation. Bacteriol. Rev., 40, 908-962. 207' Piggot, P., Moir, A. & Smith, D. A., (1981). Advances in the genetics of Bacillus subtilis differentiation. In Sporulation and Germination. Am. Soc. Microbiol., Washington, DC, Piggot, P. J., Curtis, C. A. & DeLancastre, H. (1984). Use of integrational plasmid vectors to demonstrate the polycistronic nature of a transcription unit (spoIIA) required for sporulation. J. Gen..Microbiol., 130, 2123- 2126. Plamann, L., Kuspa, A. & Kaiser, D. (1992). Proteins that rescue A-signal-defective mutants of Myxococcus xanthus. Popham, D. L. & Stragier, P. (1992). Binding of the Bacillus subtilis spoIVCA product to the recombination sites of the element interrupting the OK-encoding gene. Proc. Natl. Acad. Sci. USA, 89, 5991-5995. Rather, P. N. & Moran, C. P., Jr. (1988). Compartment- specific transcription in Bacillus subtilis: identification of the promoter for gdh. J. Bacteriol., 170, 5086-5092. Rattray, A., Altuvia, S., Mahajna, G., Oppenhiem, A. B. & Gottsman, M. (1984). Control of bacteriophage lambda cII activity by bacteriophage and host functions. J; Bacteriol., 159, 238-242. Reichardt, L. & Kaiser, A. D. (1971). Control of l repressor synthesis. Proc. Nat. Acad. Sci. USA, 68, 2185- 2189. Ricca, B., Cutting, S. & Losick, R. (1992). Characterization of bofA, a gene involved in intercompartmental regulation of pro-(3K processing during sporulation in Bacillus subtilis. J. Bacteriol., 174, 3177- 3184. Roels, S., Driks, A. & Losick, R. (1992). Characterization of SpoIVA, a sporulation gene involved in coat morphogenesis in Bacillus subtilis. J. Bacteriol., 174, 575-585. Rong, S., Rosenkrantz, M. S. & Sonenshein, A. L. (1986). Transcriptional control of the Bacillus subtilis spoIID gene. J. Bacteriol., 165, 771-779. Rudner, D. Z., LeDeaux, J. R., Ireton, K. & Grossman, A. D. (1991). The spoOK locus of Bacillus subtilis is homologous to the oligopeptide permease locus and is required 1208 for sporulation and competence. J. Bacteriol., 173, 1388- 1398. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). In Molecular Cloning, A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sandman, K., Kroos, L., Cutting, S., Youngman, P. & Losick, R. (1988). Identification of the promoter for a spore coat protein gene in Bacillus subtilis and studies on the regulation of its induction at a late stage of sporulation. J..Mbl. Biol., 200, 461-473. Sato, T., Harada, K., Ohta, Y. & Kobayashi, Y. (1994). Expression of the Bacillus subtilis spoIVCA gene, which encodes a site-specific recombinase, depends on the spoIIGB product. J. Bacteriol., 176, 935-937. Sato, T., Samori, Y. & Kobayashi, Y. (1990). The cisA cistron of Bacillus subtilis sporulation gene spoIVC encodes a protein homologous to a site-specific recombinase. J. Bacteriol., 172, 1092-1098. Satola, S., Kirchman, P. A. & Moran, C. P., Jr. (1991). SpoOA binds to a promoter used by 63 RNA polymerase during sporulation in Bacillus subtilis. Proc. Natl. Acad. Sci. Satola, S. W., Baldus, J. M. & Moran, C. P., Jr. (1992). Binding of SpoOA stimulates spoIIG promoter activity in Bacillus subtilis. J. Bacteriol., 174, 1448-1453. Savva, D. & Mandelstam, J. (1986). Synthesis of spoIIA and spoVA mRNA in Bacillus subtilis. J. Gen Microbiol., 132, 3005-3011. Schmidt, R., Margolis, P., Duncan, L., Coppolecchia, R., Jr., C. P. M. & Losick, R. (1990). Control of developmental transcription factor OTIby sporulation regulatory proteins SpoIIAA and SpoIIAB in Bacillus subtilis. Proc. Natl. Acad. Sci. USA, 87, 9221-9225. Schultz, S. C., Shields, G. C. & Steitz, T. A. (1991). Crystal structure of a CAP-DNA Complex: The DNA is bent by 90 degrees. Science, 253, 1001-1007. Shimatake, H. & Rosenberg, M. (1981). Purified 1 regulatory protein cII positively activates promoters for lysogenic development. Nature, 292, 128-132. Shorenstein, R. G. & Losick, R. (1973). Purification 209 and properties of the sigma subunit of Bacillus subtilis RNA polymerase. J. Biol. Chem., 248, 6163-6169. Singh, R. P., Setlow, B. & Setlow, P. (1977). Levels of small molecules and enzymes in the mother cell compartment and the forespore of sporulating Bacillus megaterium. J. Bacteriol., 130, 1130-1138. Smith, I. (1989). Initiation of Sporulation. In The Regulation of Procaryotic Development. (Smith, I., Slepecky, R. A. & Setlow, P), Amer. Soc. for Microbiol., Washington, DC. Sterlini, J. M. & Mandelstam, J. (1969). Commitment to sporulation in Bacillus subtilis and its relationship to development of actinomycin resistance. Biochem. J., 113, 29- 37. Stevens, C. M. & Errington, J. (1990). Differential gene expression during sporulation in Bacillus subtilis: structure and regulation of the spoIIID gene. Mblec. Microbiol., 4, 543-552. Stock, J. B., Ninfa, A. J. & Stock, A. M. (1989). Protein phosphorylation and regulation of adaptive response in bacteria. Microbiol. Rev., 53, 450-490. Stragier, P. (1986). Comment on "Duplicated sporulation genes in bacteria". FEBS Lett., 195, 9-11. Stragier, P., Bonamy, C. & Karmazyn-Campelli, C. (1988). Processing of a sporulation sigma factor in Bacillus subtilis: How morphological structure could control gene expression. Cell, 52, 697-704. Stragier, P., Kunkel, B., Kroos, L. & Losick, R. (1989). Chromosomal rearrangement generating a composite gene for a developmental transcription factor. Science, 243, 507-512. Stragier, P., Margolis, P. & Losick, R. (1994). Establishment of compartment-specific gene expression during sporulation in Bacillus subtilis. In Regulation of Bacterial Differentiation. (Piggot, P. J., Moran, C. P.& Youngman, P.s, eds), 139-154, American Society for Microbiology, Washington, D.C. Sun, D., Stragier, P. & Setlow, P. (1989). Identification of a new O-factor involved in compartmentalized gene expression during sporulation of Bacillus subtilis. Genes & Dev., 3, 141-149. 210 Tatti, K. M., Jones, C. H. & Moran, C. P., Jr. (1991). Genetic evidence for interaction of OEvdth.the spoIIID promoter in Bacillus subtilis. J. Bacteriol., 173, 7828- 7833. Thomas, J. D. & Kornberg, R. D. (1978). The study of histone-histone associations by chemical cross-linking. Methods Cell Biol., 18, 429-440. Trach, K. & Hoch, J. A. (1989). The Bacillus subtilis spoOB stage 0 operon encodes an essential GTP-binding protein. J. Bacteriol., 171, 1362-1371. Trach, K. A. & Hoch, J. (1993). Multisensory activation of the phosphorelay initiating sporulation in Bacillus subtilis: identification and sequence of the protein kinase of the alternate pathway. Mol..Microbiol., 8, 69-79. Trempy, J. E., Bonamy, C., Szulmajster, J. & Haldenwang, W. G. (1985a) . Bacillus subtilis sigma factor, 029, is the product of the sporulation-essential gene spoIIG. Proc. Nat. Acad. Sci. USA, 82, 4189-4192. Trempy, J. E., Morrison-Plummer, J. & Haldenwang, W. G. (1985b). Synthesis of Oflb an RNA polymerase specificity determinant, is a developmentally regulated event in Bacillus subtilis. J. Bacteriol., 161, 340-346. Wireman, J. W. & Dworkin, M. (1975). Morphogenesis and developmental interactions in myxobacteria. Science, 189, 516-523. Youngman, P., Perkins, J. B. & Losick, R. (1984). Construction of a cloning site near one end of Tn917 into which foreign DNA may be inserted without affecting transposition in Bacillus subtilis or expression of the transposon-borne erm gene. Plasmid, 12, 1-9. Zaluzec, E. J., Gage, D. A., Allison, J. & Watson, J. T. (1994). Directed matrix-assisted laser desoption ionization mass spectrometric analysis of proteins immobilzed on nylon- based membranes. J: Am. Soc. Mass Spectrom., 5, 230-237. Zhang, J., Ichikawa, H., Halberg, R., Kroos, L. & Aronson, A. I. (1994). Regulation of the transcription of a cluster of Bacillus subtilis spore coat genes. J. Mbl. Biol., In press. Zheng, L., Donovan, W., Fitz-James, P. C. & Losick, R. (1988). Gene encoding a morphogenic protein required in the assembly of the outer coat of the Bacillus subtilis 211 endospore. Genes & Dev., 2, 1047-1054. Zheng, L., Halberg, R., Roels, S., Ichikawa, H., Kroos, L. & Losick, R. (1992). Sporulation regulatory protein GerE from Bacillus subtilis binds to and can activate or repress transcription from promoters for mother-cell-specific genes. J. Mol. Biol., 226, 1037-1050. Zheng, L. & Losick, R. (1990). Cascade regulation of spore coat gene expression in Bacillus subtilis. J; Mbl. Biol., 212, 645-660. Zuber, P., Healy, J., Carter, H. L., III, Cutting, S., Moran, C. P., Jr. & Losick, R. (1989). Mutation changing the specificity of an RNA polymerase sigma factor. J: Mbl. Biol., 206, 605-614. "IllllllllMilli?