Insights Into the Metabolic Responses of Bacillus Subtilis to the Absence of the Key Signal Metabolites Pyruvate and UDP-Glucose

Insights into the metabolic responses of Bacillus subtilis to the absence of the key signal metabolites pyruvate and UDP-glucose

Inauguraldissertation zur Erlangung des akademischen Grades eines doctor rerum naturalium (Dr. rer. nat.) der Mathematisch-Naturwissenschaftlichen Fakult¨at der Universit¨atGreifswald

vorgelegt von Joana Manuela Dias de Sousa geboren am 13.10.1986 in Braga, Portugal

Greifswald, den 24.10.2018 Dekan: Prof. Dr. Werner Weitschies

1. Gutachter : Prof. Dr. Michael Lalk

2. Gutachter: Prof. Dr. Wolfgang Eisenreich

Tag der Promotion: 29. April 2019

Contents

List of Tables...... v List of Figures...... vii

List of abbreviations1

1 Introduction7 1.1 Metabolomics...... 7 1.1.1 Workﬂow and analytical approaches in metabolomics...... 7 1.1.2 Analytical tools in metabolomics...... 8 1.1.2.1 1H-Nuclear Magnetic Resonance: optimization of quantiﬁca- tion...... 9 1.1.2.2 Mass Spectrometry...... 12 1.2 B. subtilis ...... 13 1.2.1 B. subtilis strains and integrated databases...... 15 1.3 Metabolic control of bacterial division dynamics...... 16 1.4 Cell division and elongation mechanisms...... 17 1.5 The cell wall of B. subtilis ...... 18 1.5.1 Peptidoglycan...... 18 1.5.2 Teichoic acids...... 20 1.5.2.1 Role of teichoic acids in bacterial physiology...... 21 1.6 Pyruvate a regulator of division...... 23 1.6.1 Pyruvate catabolism and the transport mechanism in Bacillus subtilis 25 1.7 UDP-glucose as signal of nutrient availability...... 27 1.8 Main objectives and general research plan...... 31

2 Metabolic responses to pyruvate kinase mutation when glucose and pyruvate are the main carbon sources 33 2.1 Results...... 33 2.1.1 Concomitant consumption of glucose and pyruvate by ∆pyk – Carbon catabolite repression of pyruvate in the presence of glucose was relieved 33

i 2.1.2 Intracellular metabolome revealed several central metabolic pathways altered in ∆pyk...... 37 2.1.3 Increased amino acids and shikimate 3-phosphate levels in ∆pyk under glucose medium cultivation...... 41 2.2 Discussion...... 42 2.2.1 Relieve of carbon catabolite repression of pyruvate in ∆pyk cells... 42 2.2.2 Pyk mutation leads to impairment of glucose uptake when glucose is the only carbon source...... 43 2.2.3 FBP accumulation in ∆pyk could have led to dismiss of gap operon repression by CggR...... 44 2.2.4 Possible carbon reroute to PPP...... 46 2.2.5 TCA cycle might be less active in ∆pyk...... 46 2.2.6 Overﬂow metabolites: increased of acetoin and 2,3-butanediol secretions by ∆pyk...... 46 2.2.7 Lower secretion of BCAA by ∆pyk...... 48 2.2.8 Amino acids pool: proline, ornithine, citrulline, and arginine metabolism altered...... 48 2.2.9 Shikimate metabolism...... 49 2.3 Conclusion...... 49

3 Metabolic responses to pyruvate kinase mutation in a complex medium 51 3.1 Results...... 51 3.1.1 Similar metabolomic uptake and secretion by wt and ∆pyk...... 51 3.1.2 Changes in the central carbon metabolism pathways...... 54 3.1.3 Alteration in the metabolites levels of diﬀerent cell wall structures and possibly the membrane precursors...... 56 3.2 Discussion...... 62 3.2.1 Sugar consumption alterations...... 62 3.2.2 Unexpected secretion of the same acetate levels...... 63 3.2.3 Carbon distribution in the TCA and PPP...... 64 3.2.4 Lower shikimate 3-phosphate level in ∆pyk...... 64 3.2.5 Membrane and cell wall assembly altered...... 65 3.3 Conclusion...... 66

4 Study of lipoteichoic mutants in a complex medium 69 4.1 Results...... 69 4.1.1 B. subtilis 168 and BSB1 - a similar metabolome proﬁle between both strains...... 69 4.1.2 Consumption and secretion of metabolites: higher secretion of acetate by ∆pgcA...... 74

ii 4.1.3 Mutations in the lipoteichoic acid alter cell wall metabolites and possibly the peptidoglycan assembly...... 75 4.1.4 Cells lacking PgcA showed expressive alterations in the central carbon metabolism...... 78 4.1.5 Accumulation of the amino acids glutamine and histidine...... 84 4.2 Discussion...... 84 4.2.1 The partition of glucose 6-phosphate in ∆pgcA altered glycolysis pathway...... 84 4.2.2 Increased malonyl-CoA levels...... 85 4.2.3 Cell wall alterations...... 87 4.2.4 Glutamine and histidine regulation...... 89 4.3 Conclusion...... 91

5 Concluding marks and outlook 93

6 Material and Methods 95 6.1 Material...... 95 6.1.1 Chemicals...... 95 6.1.1.1 Chemicals for media preparation...... 95 6.1.1.2 Other chemicals...... 96 6.1.2 Growth media and solutions...... 97 6.1.3 Strains collection...... 100 6.2 Methods...... 101 6.2.1 Strains maintenance...... 101 6.2.2 Isolation of chromosomal DNA...... 101 6.2.3 Transformation...... 102 6.2.4 Polymerase chain reaction and restriction endonuclease digestion... 102 6.2.5 Agarose gel electrophoresis...... 103 6.2.6 Cultivation conditions...... 103 6.2.6.1 Growth cultivation for metabolic studies of the cell wall precursors mutants...... 103 6.2.6.2 Growth cultivation for metabolic studies of the pyruvate kinase mutant...... 103 6.2.7 Sampling of extracellular metabolites...... 104 6.2.8 Sampling of intracellular metabolites...... 104 6.2.9 Analytical methods...... 105 6.2.9.1 1H-NMR spectroscopy measurement and data analysis of extracellular metabolites...... 105 6.2.9.2 GC-MS measurement and data analysis of intracellular metabolites...... 106

iii 6.2.9.3 LC-MS measurement and data analysis of intracellular metabolites...... 106 6.2.9.4 LC-MS/MS measurement and data analysis of intracellular metabolites...... 107 6.3 Incorporation and release of labelled D-alanine (NADA) in the peptidoglycan and ﬂuorescent microscopy...... 109 6.4 Transmission electron microscopy...... 109 6.5 Statistical analysis and visualization...... 110

Literature 111

7 Supplementary information 129

Affidavit - Eigenständigkeitserklärung 159

Curriculum Vitae 161

Publications 163

Acknowledgements 165

iv List of Tables

1.1 Quantiﬁcation of a standard sample of glutamate by 1H-NMR...... 12

2.1 Glucose and pyruvate concentrations present in the supernatant during cultivation of wt and ∆pyk (M9GlcPyr)...... 35 2.2 Physiological parameters of wt and ∆pyk grown in the three chemically de- ﬁned media...... 35

3.1 Result of untargeted metabolomics analysis...... 61

4.1 EC of B. subtilis 168 and BSB1...... 69 4.2 Extracellular concentrations of acetate, isobutyrate and histidine...... 74

6.1 M9 stock solution (5x)...... 97 6.2 Trace elements stock solution...... 98 6.3 M9 medium composition...... 99 6.4 SMM composition...... 99 6.5 Competence medium composition...... 99 6.6 Starvation medium composition...... 100 6.7 TAE buﬀer solution...... 100 6.8 B. subtilis strains...... 101 6.9 Gradient elution method in LC system...... 108 6.10 Gradient elution method in LC-MS/MS system...... 108 6.11 MS source parameters used during amino acids measurements...... 109

7.1 OD determined for wt and ∆pyk in chemically deﬁned media...... 129 7.2 Intracellular metabolome data of Bacillus subtilis (B. subtilis) wt and ∆pyk in chemically deﬁned media conditions...... 136 7.3 OD determined for wt and ∆pyk in LB medium...... 140 7.4 Intracellular metabolome data of B. subtilis wt and ∆pyk in LB medium.. 141 7.5 OD determined for 168 and BSB1 wt and respective mutants in LB medium. 144 7.6 Consumption of aminoacids in LB medium...... 147

v 7.7 FC of metabolites of cell wall mutants in LB medium...... 149 7.8 OD determined for wt and ∆pgcA in chemically deﬁned media...... 152 7.9 Intracellular metabolome data of B. subtilis wt and ∆pgcA in chemically deﬁned medium...... 156

vi List of Figures

1.1 Experimental design of metabolic data analysis...... 10 1.2 1H-NMR spectrum of glutamate...... 11 1.3 Stepwise assembly of peptidoglycan precursors...... 19 1.4 Lipoteichoic synthesis machinery and lipid turnover...... 21 1.5 Proposed Role for PDH E1α in nutrient-dependent regulation of B. subtilis cell division...... 24 1.6 PTS and CCR mechanisms in B. subtilis ...... 26 1.7 Roles of PftAB and LytST in pyruvate homeostasis...... 27 1.8 Metabolic pathways...... 28 1.9 Wt and mutant strains in complex medium...... 30

2.1 Growth curves and extracellular concentrations of glucose and pyruvate cultivated in M9GlcPyr...... 34 2.2 Growth curves and extracellular concentrations of pyruvate in M9Pyr.... 34 2.3 Growth curves and extracellular concentrations of glucose cultivated in M9Glc 34 2.4 Time-resolved extracellular metabolites under M9GlcPyr medium cultivation 36 2.5 Time-resolved extracellular metabolites of wt under M9Glc medium cultivation 38 2.6 Heat-map of intracellular metabolites...... 39 2.7 FC of some intracellular metabolites from central metabolism...... 40 2.8 Altered amino acids levels under M9Glc cultivation...... 41 2.9 Relative amount of shikimate 3-phosphate in wt and ∆pyk in M9GlcPyr, M9Pyr, and M9Glc media cultivation...... 42 2.10 PTS and CCR mechanisms in Bacillus subtilis ...... 45

3.1 Growth curves of B. subtilis wt (circle) and ∆pyk (square)...... 52 3.2 Time-resolved extracellular metabolites of wt and ∆pyk under LB medium cultivation...... 53 3.3 Volcano-plot of intracellular metabolome data of B. subtilis wt and ∆pyk in LB medium...... 54 3.4 Absolute concentration of glycolytic metabolites in LB medium...... 55

vii 3.5 Absolute concentration of metabolites from TCA cycle in LB medium.... 56 3.6 Relative amount of PPP and shikimate phathway metabolites in LB medium 57 3.7 Heat-map of detected intracellular amino acids in wt and ∆pyk cultivated in LB medium...... 58 3.8 PCA analysis and cluster separation of intracellular data of wt and ∆pyk.. 59 3.9 Relative amount of intracellular CDP-glycerol and some PG precursors in LB medium...... 60 3.10 TEM micrographs of B. subtilis wt and ∆pyk...... 60

4.1 Time-resolved extracellular metabolites of BSB1 wt and lipoteichoic mutants under LB medium cultivation...... 70 4.2 Heatmap of extracellular metabolites of B. subtilis 168 and BSB1 in LB medium 71 4.3 Time-resolved extracellular metabolites of BSB1 wt and lipoteichoic mutants under LB medium cultivation...... 72 4.4 PCA analysis of intrametabolome data of B. subtilis 168 and BSB1 in LB medium...... 73 4.5 Volcano plots of intrametabolome of ∆pgcA, ∆gtaB and ∆ugtP in LB medium 76 4.6 Relative intracellular concentrations of peptidoglycan precursors in complex medium...... 77 4.7 Time-lapse of NADA incorporation in peptidoglycan...... 78 4.8 Time-resolved of BSB1∆dacA and ∆ugtP∆dacA incubated in LB medium.. 79 4.9 TEM micrographs of B. subtilis wt, ∆pgcA, ∆gtaB and ∆ugtP in LB medium 80 4.10 Relative intracellular concentrations of glycolytic metabolites and other relevant metabolites in complex medium...... 82 4.11 Multivariable analysis of metabolic responses of B. subtilis BSB1 in LB medium 83 4.12 Relative amount of intracellular UDP-glucose and UDP-glucuronate in LB medium...... 83 4.13 Relative intracellular concentrations of glutamine and histidine in complex medium...... 84 4.14 Schematic illustration of extracellular acetate and relevant intracellular metabolites in wt and ∆pgcA...... 86

6.1 Experimental workﬂow...... 104

7.1 Time-resolved extracellular metabolite concentrations of ∆pyk under M9GlcPyr medium cultivation...... 131 7.2 Time-resolved extracellular metabolite concentrations of wt and ∆pyk under M9Pyr medium cultivation...... 132 7.3 Time-resolved extracellular metabolite concentrations of wt under M9Glc medium cultivation...... 133

viii 7.4 Time-resolved extracellular metabolite concentrations of ∆pyk under M9Glc medium cultivation...... 134 7.5 Relative amount of intracellular PG precursors in wt and ∆pyk in minimal media...... 135 7.6 Volcano-plots of intracellular metabolome data in M9GlcPyr, M9Pyr, and M9Glc media cultivation...... 139 7.7 Time-resolved extracellular metabolite concentrations of asparagine, aspartate and alanine in 168 and BSB1 strains under LB medium cultivation... 145 7.8 Time-resolved extracellular metabolite concentrations of 2-methylbutyrate and isovalerate in LB medium...... 146 7.9 1H-NMR spectrum of a sample in LB medium...... 147 7.10 Growth curves of wt and ∆pgcA in M9GlcMalGlut medium cultivation... 148 7.11 Time-resolved extracellular metabolite concentrations of wt and ∆pgcA under M9GlcMalGlut medium cultivation...... 153 7.12 Extracellular concentration proﬁles of branch chain amino acids (BCAA) metabolism of wt and ∆pgcA under M9GlcMalGlut medium cultivation..... 154 7.13 Heat-map of intracellular metabolites...... 155

ix x List of abbreviations

AMBER Advance Multidisciplinary training in Molecular Bacteriology...... 69 B. subtilis Bacillus subtilis ...... v BCAA branch chain amino acids...... ix CCA carbon catabolite activation ...... 25 CE capillary electromigration ...... 9 CcpA catabolite control protein A ...... 5 CCR carbon catabolite repression...... 5 COSY Correlated Spectroscopy ...... 9 DAG diacylglycerol ...... 20 dQC daily quality control ...... 106 EC energy charge...... 51 E. coli Escherichia coli ...... 44 EI electron-impact ...... 13 ESI electrospray ionization ...... 13 FC Fold Change ...... 37 FBP fructose 1,6-bisPhosphate ...... 6 FDA Food and Drug Administration ...... 13 FT-IR fourier transform - infrared spectroscopy ...... 9 GC-MS gas chromatography coupled to mass spectrometry ...... 9 GDH glutamate dehydrogenase ...... 48 GlcNAc N-acetylglucosamine ...... 18 GlcNAc-1P N-acetylglucosamine-1-phosphate ...... 20 GOGAT glutamate synthase ...... 89 GRAS Generally Recognized As Safe...... 13 GS glutamine synthetase ...... 89 GtaB UTP-glucosephosphate uridylyltransferase ...... 28 HPr histidine-containing phosphocarrier HPrK ATP-dependent HPr kinase/phosphorylase ...... 25 HSQC Heteronuclear Single-Quantum Correlation ...... 9 ISTD internal standard ...... 9

1 LC-MS liquid chromatography coupled to mass spectrometry...... 9 LC-MS/MS liquid chromatography tandem-mass spectrometry ...... 9 LytE DL-endopeptidase lytE ...... 88 LTA lipoteichoic acid ...... 6 LtaS LTA synthase...... 20 m/z mass/charge ratio...... 58 ManNAc N-acetylmannosamine ...... 20 MS mass spectrometry ...... 9 MurNAc N-acetylmuramic acid...... 18 NADA NBD-amino-d-alanine...... 76 NIST National Institute of Standards and Technology...... 106 1H-NMR Proton Nuclear Magnetic Resonance ...... 9 NMR Nuclear Magnetic Resonance ...... 9 NOESY Nuclear Overhause Eﬀect Spectroscopy...... 9 NRPSs non-ribosomal peptide synthetases ...... 15 OD optical density ...... 33 PBPs penicilin binding proteins ...... 18 PBP5 DD-carboxypeptidase ...... 76 PC principal component ...... 75 PCA principal component analysis...... 7 PdH pyruvate dehydrogenase...... 24 PEP phosphoenolpyruvate ...... 5 PG peptidoglycan...... 6 PgcA alpha-phosphoglucomutase ...... 28 Pgi glucose-6-phosphate isomerase ...... 29 PKSs polyketide synthases ...... 15 PPP pentose phosphate pathway ...... 14 PTS phosphotransferase transport system Pyk pyruvate kinase ...... 5 S. aureus Staphylococcus aureus ...... 20 SD standard deviation...... 51 TA teichoic acid ...... 18

2 TCA tricarboxylic acid ...... 5 TEM Transmission Electron Microscopy ...... 57 TOF time of ﬂight ...... 106 TSP 3-trimethylsilyl-[2,2,3,3-D4]-1-propionic acid ...... 97 UDP-GlcNAc UDP-N-acetylglucosamine ...... 20 UgtP UDP-glucose diacylglycerol glucosyltransferase ...... 20 WTA wall teichoic acid ...... 20

3 4 Summary

Metabolomics is referred as the systematic identification and quantification study of large numbers of metabolites in cells, tissues, or biofluids at a given time. In a broadly concept, metabolomics aims to comprehend the regulatory mechanisms and its coordination to the changes occurring in an organism. Recently, the development of viable protocols and the advances on the analytical instruments and metabolic data analysis, have spurred metabolomics to a wider and accessible range of research areas. One of these is the Microbial metabolomics. The environmental conditions and the nutrition availability play a crucial role on the size of bacteria cells. The study of the appropriate metabolic responses to the nutritional changes and the adjustment of the cell cycle has emerged as a fundamental topic in microbiology. In this study, B. subtilis, the best characterized member and referenced model of the Gram- positive bacteria, was investigated for a better knowledge of metabolic sensing mechanisms in different conditions. The metabolites pyruvate and UDP-glucose, and their respective mechanisms have been associated to cell division machinery localization and assembly to the nutritional information in B. subtilis. Thereby, this study comprehensively monitored metabolic changes in the extracellular space and the cytosol of B. subtilis triggered by the absence of these two possible signaling metabolites. To do so, biochemical analytical platforms were applied such as 1H-NMR, HPLC-MS, and GC-MS. B. subtilis lacking pyruvate kinase (Pyk), the enzyme that catalyzes the production of pyruvate in the final step of glycolysis, (∆pyk), was subject of metabolic investigation. Metabo- lites from several pathways and its possible changes were monitored in diverse media. In chemically defined media, it was shown that the perturbation created in the pyruvate node drove an adaptation to new conditions by altering the nutritional compounds’ consumption. The results indicate that the carbon catabolite repression (CCR) of pyruvate uptake in the presence of glucose subject in wt, was relieved in ∆pyk. This is consistent with the current knowledge of pyruvate transport system PftAB, which is under the control of the pyruvate influx into and the master CCR regulatory protein, catabolite control protein A (CcpA). Moreover, when ∆pyk was cultivated in glucose as single carbon source, the growth rate and the glucose consumption was drastically slow. In this condition, the accumulation of glycolytic metabolites, the low pyruvate level, and the inability of B. subtilis to redirect the accumulated phosphoenolpyruvate (PEP) to the tricarboxylic acid (TCA) cycle flux was possibly compromising the glucose uptake. It can also be speculated that the high levels of

5 fructose 1,6-bisPhosphate (FBP) could have had an impact on the glucose transport system. Other metabolic alterations were observed when cultivated in LB medium. With the application of the untargeted approach on the identiﬁcation of compounds, several features were altered in the mutant and proposed as being precursors of the cell membrane. It was previously proposed that proteins involved in the synthesis of the lipoteichoic acid (LTA) glucolipid anchor, such as PgcA, GtaB, and UgtP, link nutrient availability to cell division. However, the impact of these proteins on the cell wall and other metabolic pathways is still unclear. Thus, the metabolome of B. subtilis in the absence of these enzymes were monitored. The mutant cells had increased levels of several peptidoglycan (PG) precursors. The combination of these results with other studies, suggest that the balance betweenPG synthesis and hydrolysis is altered in cells lacking the LTA glucolipid precursors. Altogether, the presented work contributed to unravel the two metabolic networks proposed as key signals for the adaptation to environmental conditions in B. subtilis.

6 Chapter 1

Introduction

1.1 Metabolomics

Metabolomics is described as the study of the complete set of metabolites synthesized through a series of biochemical pathways at a given time by a biological system under defined conditions [1,2,3]. Metabolomics has become an increasingly large field, where high resolution analytics techniques are used together with chemometric statistical tools, such as principal component analysis (PCA), that help to determine consistencies and differences in large data sets [4,5]. The neologism Metabolomics was first suggested by Oliver Fiehn in 2001, having nowadays around 21 000 entries available on PubMed database [6]. The increasingly popularity of this “omics” approach is due to the possibility of identify end-products of genomics, proteomics, epigenetic and environmental interactions and, consequently, having a more comprehensive picture of the organism’s dynamic. Thereby, in a post-genomic era, the metabolome analysis became the most robust, reproducible and cost- effective tool to assess the physiological status of living cells in a high-throughput manner [4,7,8]. Nowadays, it is applied in different areas of life science’s research, like phenotypic characterization of microbial strains, and gene-knockout mutants in diverse field such as agriculture and personal medicine [3,9, 10].

1.1.1 Workﬂow and analytical approaches in metabolomics

Regardless the purpose of the metabolomics’ studies, the experiments are set up to capture a representative snapshot of the metabolome at a given time. The selection of a suitable metabolomics workﬂow, from the sample collection, extraction of metabolites, the analytic technique, and the data analysis, is not always straightforward [11, 12].

7 The ﬁrst steps of sampling are crucial to obtain reliable metabolic data. Ideally, the sampling method should:

be simple and fast, to prevent metabolite loss and/or degradation during the preparation procedure;

be eﬃcient in the metabolic extraction, either in the study of speciﬁc metabolites or the global metabolome;

include a metabolism-quenching step to represent the true metabolome composition at the time of sampling;

be reproducible.

Moreover, depending of the aim of the investigation, different sample preparation methods can be applied. The common approaches in metabolomics are the targeted analysis, which aims a quantitative measurement of a select group of metabolites and requires an a priori knowledge of the compounds of interest; and untargeted analysis, which aims a global metabolic profiling, allowing the detection of a broad range of metabolites including unknown metabolites. In targeted analysis, selective methods are carefully utilized and optimized, not only for the identification but also for quantification of the desired compounds. On the other hand, in untargeted analysis, usually a non-biased towards any certain groups of metabolites sampling method is used in order to reach a global screening of the metabolome [13, 14, 15]. Nevertheless, both approaches can be combined, which allow the identification and quantification of compounds in order to investigate specific metabolic pathways, but also for the detection of features for hypothesis-generating studies.

1.1.2 Analytical tools in metabolomics

Metabolites present a high variability in their chemical structure and chemical properties. They can be hydrophilic carbohydrates, volatile alcohols, hydrophobic lipids, amino and organic acids. Due to this diversity, the analysis of an entire set of metabolites results in a hard and challenging task [2, 16, 17]. In an ideal metabolomics study, one analytical system would permit sufficient coverage of a global metabolic profile, allowing simultaneous analysis of a wide concentration range of metabolites. However, until now, no single analytical tool has yet been devised that can measure all metabolites in a single organism [12, 18]. This brought the necessity of developing and optimize better and more specialized methods in the field, that lead to more efficient metabolite’s separation or powerful detectors with higher resolution. Furthermore, the selection of the techniques is usually a compromise among sensitivity, selectivity and speed.Because of the diverse chemical nature and the variety of metabolites

8 amount in a single cell, is frequently common to utilize several technologies in the same study. This complementary approach allows the identification and quantification of more metabolites than if just a single technique would be used. Presently, the techniques applied in metabolomics for separation and identification of compounds are based mainly on chromatographic methods, such as liquid chromatography (LC) and gas chromatography (GC) combined with mass spectrometric (MS) detection, but also Nuclear Magnetic Resonance (NMR), fourier transform - infrared spectroscopy (FT-IR), and capillary electromigration (CE) techniques have also been utilized to some extent [12, 19, 20]. In this sense, four techniques were applied in the study: Proton Nuclear Magnetic Reso- nance (1H-NMR), for the investigation of extracellular metabolome; and liquid chromatography coupled to mass spectrometry (LC-MS), gas chromatography coupled to mass spectrometry (GC-MS), and liquid chromatography tandem-mass spectrometry (LC-MS/MS), all for the intracellular metabolome analyses (figure 1.1).

1.1.2.1 1H-Nuclear Magnetic Resonance: optimization of quantiﬁcation

NMR spectroscopy is one of the technologies commonly associated with the metabolome research. There are many advantages of using NMR: the ability of analysing several metabolites at once since there is no compound separation during measurement; the direct quantification of compounds by comparing the analyte’s protons signal with the ones from the internal standard (ISTD); the possibility of determining the structures of unknown compounds and, the high reproducibility [19, 21, 22, 23]. Thus, NMR is generally accepted as the gold standard in metabolite structural elucidation [19]. NMR can also offer an alternative for compounds that are difficult to ionize or require derivatization for mass spectrometry (MS). Moreover, sample’s preparation is usually simple, a small volume is needed and the measurement can be repeated since is a non-destructive technique.On the other hand, the major drawback of NMR is its high limit of detection, making it inappropriate for the analysis of low-abundance metabolites. Another disadvantage is the overlap of signals due to the lack of separation of metabolites [19, 21, 23]. There are several spectroscopic methods available: one dimensional methods such as 1H- NMR, 13C-NMR, 15N-NMR, 31P-NMR; two dimensional methods like Correlated Spec- troscopy (COSY), Nuclear Overhause Effect Spectroscopy (NOESY) or Heteronuclear Single- Quantum Correlation (HSQC), and several adaptions of these [24]. In this study, NMR was used for the investigation of the exometabolome of B. subtilis - usually named as metabolic footprinting - since most of the metabolites excreted from cells and that constitute the media are in high concentrations allowing their detection by this technique. A modified 1D-1H-NOESY pulse sequence method was applied (being refereed in this work for simplification as 1H-NMR), where the identification and quantification were based on the proton analysis and the solvent signal was suppressed - water signal at 4.79 ppm [25, 26].

9 Extracellular metabolites Intracellular metabolites

1H-NMR

metabolome extraction

Abundance

TIC: KD_19H.D 2.8e+07 LC-MS 2.6e+07 GC-MS 2.4e+07

2.2e+07

2e+07

1.8e+07

1.6e+07

1.4e+07

1.2e+07

1e+07

8000000

6000000

4000000

2000000

0 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 55.00 60.00 Time-->

LC-MS/MS

data process and analysis

Figure 1.1: Experimental design of metabolic data analysis. The metabolome of B. subtilis was studied in detail. For the extracellular metabolites, NMR was chosen, while different chromatographic techniques coupled to mass spectrometry were combined for the study of the intracellular metabolites. This complementary approach allows the identification and quantification a wide of metabolites.

Concerning the quantiﬁcation, diverse approaches can be done [27]. In 1H-NMR, the signal intensities are directly proportional to the number of protons. Knowing that one molecule

10 produces several peak signals in different chemical shifts in accordance to hydrogen bonding and the atoms surround, the quantification is done by integrating the peak area of just one region with known theoretical number of protons. Then, the integration is normalized to a external standard peak with a pre-determined concentration and proton number. In the platform used, an artificial signal peak formed at 15 ppm region (QuantRef signal) is used for these calculations. During the development of this project, there was a need to optimize the quantification method used so far since it was detected a consistent inaccuracy when determining the concentration of few metabolites. Instead of using a theoretical number of protons of the region of interest for quantification, a proton weight was given to each peak. As an example, for glutamate (figure 1.2), it is produced three peak signals: a double doublet in the region of 3.73-3.78 ppm (A) related to the proton bonded at carbon number 4 and two multiplets at 2.31-2.42 ppm (B) and 2.03-2.18 ppm (C) related to the two protons bonded at carbons number 6 and 7, respectively. Using the theoretical number of protons for each

Figure 1.2: 1H-NMR spectrum of glutamate standard. The peaks were integrated at δ [ppm]: 3.73-3.78 doublet of doublet (A), 2.31-2.42 multiplet (B) and 2.03-2.18 multiplet (C). region, the calculation of the glutamate concentration in the sample with ﬁnal concentration of 5 mM resulted in 4.76 mM when integrated at region A, 5.12 mM with region B and 5.06 mM in region C (table 1.1). On the other hand, if a relative weight is done by summing the integration area of each peak‘s region and divided by the number of protons in glutamate it gives the integration

11 area per H. Then, it’s possible to determine the number of protons related to each region of the spectrum. By using this quantiﬁcation method, the quantiﬁcation of glutamate in any region of the spectrum gave more accurate results (table 1.1).

1.000 + 2.121 + 2.106 = 5.236 (1.1)

5.236 = 1.047 perH (1.2) 5H

2.121 ∗ 1H RegionB = = 2.03 H (1.3) 1.047

Table 1.1: Quantiﬁcation of a standard sample of glutamate by 1H-NMR using the theoretical and weighted method. The weighted method for the calculation of the number of protons of each region was more accurate and precise.

Method Region A Region B Region C Theoretical H 1 2 2 Concentration (mM) 4.76 5.12 5.06 Weighted H 0.95 2.03 2.01 Concentration (mM) 4.98 5.03 5.03

1.1.2.2 Mass Spectrometry

MS is a powerful qualitative and quantitative analytical technique that has been widely selected for metabolomics studies in clinical and research laboratories [17, 19, 28]. The main reason for the common usage of theMS techniques is due to the good combination of selectivity and sensitivity parameters.MS brings a high specific chemical information that is provided by an accurate mass determination, isotope distribution pattern for elemental formula determination, or even from characteristic fragment ions for structural elucidation. Another advantage of theMS is their low limit of detection, sensitivity, which allows the measurement of metabolites in a wide range of concentrations [17, 19, 28]. In contrast to NMR, where usually no compound separation is done,MS is frequently coupled to chromatographic techniques. The hyphenation of these greatly aids advantages that are: the separation of isomeric substances, higher accuracy in the quantification, separation and quantification achieved at the same time; and additional information valuable in the metabolites annotation and metabolomics data base (ex. retention time) [19]. The most popular hyphenated techniques in use and applied in this study are GC-MS, LC- MS, and LC-MS/MS.

12 Being the first and best-established coupling of methods toMS for metabolite analysis, GC- MS is suited either for identification and quantification of compounds [19, 29]. Thermo-stable samples are vaporized and then ionized by either electron-impact (EI) or chemical-ionization. In this process, not only the analytes are ionized but also the molecules are broken into frag- ments. These are obtained by the detector and their different relative abundances can be interpreted and compared with spectra libraries [19, 29, 30]. Moreover, since a prerequisite in GC-MS is the high volatility and thermal stability of the analytes, it is frequent to include derivatization as part of the sample preparation analysis. With this step, compounds that would formerly exhibit non-reproducible peak areas, heights, and shapes such as poor volatile, poor thermo-stable compounds or would be ad- sorbed in the injector, can now be efficiently detected [30]. Although the derivatization step enhances the selectivity of metabolites, it also prolongs the overall analysis time. Also, a lyophylization step needs to be added prior to derivatization for the elimination of molecules of water. Some of these analytical problems were solved with the development of other choromato- graphic techniques such as the LC-MS. The LC-MS, and in bigger extent LC-MS/MS, doesn’t require derivatization prior to injection, saving on reagent costs, time, and sampling preparation. Moreover, it enables the analysis and detection of compounds with high polarities and/or high molecular mass, which are incompatible with GC-MS [31, 32]. LC-MS and LC-MS/MS also combine the chemical separating power, where impurities and degradation products can be separated, with the ability of anMS to selectively detect and confirm molecular identity [19, 30, 32]. Currently, the most widely employed method for LC-MS is the electrospray ionization (ESI), where a soft ionization of the metabolites is made with little fragmentation. A big disadvantage and concern using LC-MS (and LC-MS/MS) is the possibility of occurring the ion suppression effect. This phenomenon results in the early stages of the ionization process, when a component eluted from the column influences the ionization of a co-eluted analyte. Hence, the ionization efficiency of a target analyte can be altered and the reproducibility of method compromised [31, 33]. Since in LC-MS just the molecular ion is observed, the structural information of a compound is quite poor. To overcome this drawback, the coupling a tandem mass spectrometry (ESI-MS/MS) has been developed. In this technique, the ions formed in the first stage of MS are separated and fragmented through collision-induced dissociation [30, 34].

1.2 B. subtilis

B. subtilis is a rod-shaped bacterium and Generally Recognized As Safe (GRAS) by Food and Drug Administration (FDA) agency [35].

13 In the early 90s, an international consortium involving diverse working groups from Eu- rope, Japan and United States was set up for the sequencing and annotation of the whole B. subtilis genome. This achievement makes B. subtilis nowadays the best characterized member of the gram-positive bacteria and is currently the reference model system for diverse physiological studies and cell differentiation processes, e.g. competence, cell wall synthesis, sporulation and related pathways [36, 37, 38, 39]. B. subtilis is ubiquitous in nature, living primarily in the soil and associated water sources, and can successfully adapt to various changes in the environment. In extreme stressful conditions, a master transcription regulator, Spo0A, is phosphorylated by a multi-component phosphorelay. This Spo0A activation affects the expression of more than 10% of the genome of B. subtilis and, consequently, triggers adaptive mechanisms such as sporulation, motility, competence and biofilm formation. In sporulation, a metabolically inactive dormant cell (endospore) is formed as a microbial response, enabling the survival in stressful environment until favorable growth conditions are restored. Sporulation has been subject of very active studies since it has provided insight into fundamental processes e.g. gene expression, metabolism coordination, and chromosome partitioning during cell division. Also, this mechanism has assumed great importance in human health field due to the occurrence of highly resistant spores in food production compromising its preservation and in waste discarded in the environment leading eventually to human diseases [40, 41, 42]. Another adaptive mechanism developed in stressful environment is the biofilm production, an extracellular matrix commonly constituted of polysaccharides, proteins and DNA. As the biofilm is developed, a chemical concentration gradient of nutrients is established and signaling and waste compounds are secreted. This way, the cells living in community, are repleted with nutrients and receive protection from fluctuations in environmental conditions. Several human diseases are diagnose as biofilm-associated infections, which make this process a field of research [43, 44]. B. subtilis is also a focus of interest in biotechnology and industrial fields because of its’ easy genetic manipulation and the availability of physiological and biochemical data. It has the capacity to produce and secret large quantities of enzymes with various industrial applications such as proteases, that can convert waste into biomass, heal skin ulceration process or being used in detergent production [45]. Moreover, secondary metabolism of B. subtilis is also in great expansion of investigation due to the economically valuable microbial products. As an example, Duan and co-workers improved the riboflavin production yield by 25%, an important vitamin in food, cosmetic and animal feed business, by over-expressing glucose 6-phosphate dehydrogenase and, consequently, increasing the carbon flux to pentose phosphate pathway (PPP)[46, 47]. In pharmaceutical area, B. subtilis has been used for several years in the production of an- tifungal and antibiotics compounds. The best understood anti-microbial active metabolites

14 include the generation of peptides either by non-ribosomal peptide synthetases (NRPSs) and polyketide synthases polyketide synthases (PKSs) or ribosomal synthesis and post- translational modification [48, 49, 50]. There is also an interest in the pharmaceutical industry for aromatic compounds production, i.e. aromatic amino acids, since they can be used in monoamine neurotransmitters drugs. Therefore, successful attempts have been made to increase the production of these metabolites in B. subtilis by manipulating several step reactions of the shikimate pathway [123]. Furthermore, the use of modern tools in functional genomics and the control of metabolic pathways has allowed the development of system and synthetic biology studies and the recognition of B. subtilis as an efficient microbial cell factory [51, 52]. In this context, the metabolome flux knowledge, the strategies for engineering pathways and, consequently, the successful over-expressing of metabolites, have proven that B. subtilis is a potent applicant organism in a wide industrial fields.

1.2.1 B. subtilis strains and integrated databases

In the past decades, the majority of academic and industry research were made in a tryptophan auxotroph B. subtilis, namely 168. The consensus of its origin is that derived from B. subtilis Marburg after two researchers in Yale University exposed Marburg strain to sub- lethal doses of UV and X-ray emission and, consequently, produced auxotrophic survivors. Several years later, Spizizen and co-workers provided a full study of 168 strain in which they demonstrate the highly transformable skill and the easy acquirement of prototrophy when exposed to DNA of other strains. This landmark achievement led to the desire of studying B. subtilis in 168 in several working groups all over the world [38, 53]. Recently, a prototrophic variant of 168 has been isolated by transformation with W23 DNA strain, called BSB1. BSB1 strain has emerged nowadays as a strain of interest and is increasingly replacing 168 in systemic research studies [54, 55]. The dissemination of the same strain to working laboratories, the development of genomic technology and, subsequently, the acquisition of genetic diversity, led in the 70s, to the necessity of creating the Bacillus Genetic Stock Center, a centralized repository of characterized strains and cloning vectors [53, 56]. With the novel omic technologies advent, e.g. as transcriptomics and metabolomics, the genetic and physiology investigation of B. subtilis have been complemented to a systemic biology analysis level. This led to the generation of a vast information and the need to col- lect all knowledge in databases. The ﬁrst reference database developed was SubtiList, which contains relevant annotations and functional assignment of 168 strain and is now integrated in a multi-genome framework, GenoList, to allow the correlation of other bacterial species databases [57, 58, 59].

15 Several databases have followed with different specific purposes: SporeWeb focused on sporulation cycle; BsubCyc based on the updated B. subtilis 168 genome sequence and annotation, and DBTBS, which is specialized in transcriptional regulation [38, 60, 61, 62]. Recently, with the aim of collecting all available information on B. subtilis and to make it easily accessible to the scientific community, a main database was created. In this database, Subtiwiki, gene expression, metabolic and regulatory pathways and protein-protein interactions are graphically presented. It has also become one of the most popular databases dedi- cated to a living organism in one resource with 50 000 pages accessed per day [38, 58, 63, 64].

1.3 Metabolic control of bacterial division dynamics

In bacteria, cell-cycle processes like replication and division, must be spatially and tem- porarily coordinated to ensure the viability of the newborn cells. Moreover, they must have adaptive responses to modulate cell events during nutritional changing conditions [65]. Although being a fundamental aspect of bacteria cell cycle, their size control in diﬀerent environmental conditions is still poorly understood. In nutrient-rich environment, when cell growth rate is faster, B. subtilis achieves approximately twice the length of cells grown in a nutrient-poor medium [66]. For decades, it was assumed that bacteria has a conserved mechanism responsible for coordinating size and cell composition (i.e. nucleic acids, proteins) with growth rate. Addition- ally, it was thought that cell mass achieved in a certain nutrition condition was dependent of growth rate, rather than the composition of the media. Nowadays, studies on this ﬁeld suggest that cell size and composition are primarily dependent of nutrient availability rather than growth rate and that, in fact, growth rate is a consequent parameter of the ability of bacteria to assimilate the nutrients from the media. Furthermore, increasing the cell size during a growing-fast condition can also be a way to ensure that division is coordinated with DNA replication and nucleoid segregation [66, 67, 68]. The capability of bacterial cells to rapidly adjust their size in response to media changing conditions, shows that nutritional status can be directly transmitted, via metabolic pathways, to the division apparatus. In this view, metabolic sensing mechanisms have emerged as key roles in the nutritional regulation of cell division. Two metabolites and mechanisms have been described with direct connection of cell division machinery localization and assembly to the nutritional information: pyruvate (chapter2 and 3), and UDP-glucose (chapter4).

16 1.4 Cell division and elongation mechanisms

When B. subtilis cells reach a critical cell mass, a division septum is formed in the midpoint of the rod-shape cell. The cell division process is driven by a divisome, a dynamic machinery constituted by a protein complex structure. The divisome assembly is initiated by polymerization of FtsZ protein, a highly conserved tubulin-like GTPase, in a ring structure (the Z ring). The formation of the divisome structure continues with recruitment of early-assembly proteins to the Z ring such as FtsA, a dimer with ATP depending stability required for initiation of cell division; EzrA, an integral membrane protein, that prevents the assembly of the later at the cell poles; and ZapA, a protein required for the correct septal morphology. Afterwards, later-assembling proteins are recruited to the divisome site, the Z ring is constricted ahead of the new newly synthesized division septum and the daughter cells are separated [69, 70, 71]. The efficient arrangement of this structure is ensured by two regulatory systems, the nucleoid occlusion and the Min system, which guide Z ring positioning and prevent the inappropriate localization in the cell. In the nucleoid occlusion system, a DNA-binding protein Noc with recognition sites dis- tributed all over the chromosome, except in the replication terminus region, preventing the septum formation in the region where DNA is present. The Min system comprises four proteins, MinC, MinD, MinJ and DivIVA, present in higher concentration in the poles and inhibit the Z ring formation close to the later [72, 73]. Recent studies revealed that the synergistic action of these two negative regulatory systems is not essentially required for identification of the division site but ensures its efficient utilization. Rodrigues and co-workers analyzed the Z ring formation in minCD noc double- mutant strain, which detected the assembly of Z ring precisely at the mid cell. However, their assembly was significantly delayed, forming longer cells and resulted in much fewer Z rings at mid cell compared to wild-type cells. This way, it was considered the existence of an additional mechanism for identification of the cell centre in bacteria [65, 74, 75]. Despite most of the division proteins in B. subtilis have been identified, the assemble of the divisome machinery and septum synthesis knowledge remains incomplete. In cell elongation, an internal cytoskeleton constituted by actin-like MreB family system is formed around the periphery of the cell, allowing the maintenance and determination of the cell shape. The MreB family is a widely conserved system among rod-shaped bacteria and is constituted by three actin isoforms (MreB, Mbl and MreBH) with partial redundant function. These MreB proteins undergo ATP-dependent polymerization into in a single helical structure which spatially directs the synthesis of newPG structures [76, 77]. Moreover, MreB has also been implicated in the localization of the replication machinery to the cell centre, becoming aberrant soon after depletion of MreB. Some studies showed that MreB is recruited to mid cell directly interacting with FtsZ [77, 78, 79].

17 1.5 The cell wall of B. subtilis

A fundamental aspect in the cell’s integrity during growth and division cycle processes, is the coordination of those with the cell envelope synthesis machinery. The cell envelope of bacteria separates the external and internal cell environment being vital for the cell’s protection and physiology. Furthermore, cell envelope is the main target of numerous antibiotics, making the assembly mechanisms and organization focus of extensive studies [80, 81]. In gram-positive bacteria like B. subtilis, cell envelope is a complex structure constituted of a cell membrane, that is formed by a phospholipid layer, and a cell wall, composed by a complex network ofPG with teichoic acid (TA) and a variety of proteins attached, surrounding the cytoplasmic membrane. The dynamic of these structures is crucial for the adaptation and survival of bacteria in diﬀerent environments. The cell wall plays a major role in bacterial physiology since it maintains cell integrity during growth, regulates osmotic pressure and is the surface of a multitude reactions. It is also responsible for the sustain of a deﬁned cell shape and is intimately involved in the cell‘s growth and division processes. Moreover, the cell wall synthesis and the proteins involved, such as penicilin binding proteins (PBPs), are seen as a good target for antibacterial action with selective toxicity since the enzymes and the metabolic pathways are unique in bacteria [82].

1.5.1 Peptidoglycan

ThePG, also known as sacculus, is the main component of the cell wall and is constituted by glycan chains of alternated N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) which are linked by β-1,4-glycosidic bonds. In the N-terminus of the lactyl group of MurNAc, peptidic chains are covalently linked and cross-linked, generating a three- dimensional meshwork which ensures bacterial integrity. In B. subtilis, the amino acid sequence of these peptidic chains are L-alanine, D-glutamate, meso-diaminopimelic acid, D-alanine, and D-alanine [83, 84]. The synthesis ofPG is coordinated by synthases and hydrolases in a multiple step process, where sugar linked precursors such as UDP-MurNAc-pentapeptide and UDP-GlcNAc are formed in the cytoplasm. The UDP-MurNAc-pentapeptide is synthesized from UDP- GlcNAc and is mediated by MurA to MurF synthetases (ﬁgure 1.3). Then, in the cytoplasmic membrane, lipid I is formed by the transfer of phospho-MurNAc-pentapeptide moiety of UDP-MurNAc-pentapeptide to the membrane acceptor undecaprenyl phosphate. After- wards, GlcNAc from UDP-GlcNAc is added to lipid I leading to the lipid II synthesis. Lipid II is transferred by lipid II ﬂippase from cytoplasmic side to outer side, where glycosyl- transferases catalyze the synthesis of linear glycan chains and transpeptidases catalyze the

18 Fructose 6-phosphate

Glucosamine 6-phosphate

Glucosamine 1-phosphate acetyl-CoA N-acetylglucosamine 1-phosphate UTP UDP-GlcNAc PEP UDP-GlcNAc-enolpyruvate NADPH UDP-MurNAc L-Ala UDP-MurNAc-L-ala D-Glu UDP-MurNAc-dipeptide A2pm UDP-MurNAc-tripeptide D-Ala-D-Ala UDP-MurNAc-pentapeptide undecaprenyl-phosphate Lipid I UDP-GlcNAc Lipid II

Peptidoglycan

Figure 1.3: Stepwise assembly of peptidoglycan precursors.

formation of peptide cross-linkages using energy from D-ala-D-ala bond and leading to the binding nascent of thePG[83, 84]. During cellular growth,PG covalent bonds are cleaved by hydrolases for the insertion of newPG (lipid II) in the structure. This allows the elongation of cells without compromising the sacculus thickness. The activity of synthases and hydrolases must be coordinated and continuously adjusted to the requirements of cell division and cell elongation processes so that the cell integrity is not compromised and that the assembly activity is restricted to speciﬁc zones on the cell surface. Moreover, PBPs, the enzymes responsible forPG polymerization and cross-linking, must be under tight spatio-temporal control [83, 85, 86].

19 1.5.2 Teichoic acids

TA are anionic polymers found in gram-positive bacteria. Despite being seen as a critical component of the cell wall, their complete function remains unclear. They are mainly classi- fied into two groups: LTA, which are linked to the cytoplasmic membrane, and wall teichoic acid (WTA), which are covalently attached to thePG. In B. subtilis, both groups of TAs are present. Albeit their structural similarity, LTA and WTA are synthesized in distinct pathways. LTA is a glycerol phosphate polymer, constituted by a glycolipid anchor and polyglycerolphosphate chain (figure 1.4). The glycolipid anchor is a diglucosyldiacylglycerol synthesized from diacylglycerol (DAG) present in the inner side of the cytoplasmatic membrane, and two glucose moieties transferred from UDP-glucose by UDP-glucose diacylglycerol glucosyltransferase (UgtP)[87, 88]. In Staphylococcus aureus (S. aureus), the glycolipid anchor is exported to the outside of the membrane by the LtaA enzyme but no homologous has been found in B. subtilis. The LTA synthase (LtaS) protein elongates the TA by polymerizing glycerol-phosphate chain using the lipid phosphatidylglycerol as substrate [87, 89]. After synthesis, the LTA can suffer chain modifications, such as glycosylation, where GlcNAc or galactose residues are attached to the polymer chain and through d-alanylation of the glycerol-phosphate subunits. The later process is carry out by enzymes encoded by the dltABCD operon, which provides protonated amino groups to LTA and therefore, changes the negative net charges of LTA derived from the phosphate groups. The D-alanylation can be further modulated in response to osmolarity, pH and temperature changes [90, 91, 92]. WTA are composed by a polyglycerolphosphate chain and a disaccharide unit constituted by N-acetylmannosamine (ManNAc) and GlcNAc covalently attached to MurNAc of the PG. The WTA synthesis starts with the action of TagO which catalyze the transfer of N-acetylglucosamine-1-phosphate (GlcNAc-1P) residue from UDP-N-acetylglucosamine (UDP-GlcNAc) to the lipid carrier undecaprenyl phosphate, a common metabolite ofPG. Then, TagA adds ManNAc, followed by addition of glycerol 3-phosphate by TagB, and finally TagF links 45 to 60 glycerol units to the nascent chain assembling the polymer. The glycerol moiety is synthesized from actived CDP-glycerol by TagD. The polymer is exported to the cytoplasmatic membrane of the cell by TagGH and subsequently the chain is covalently linked to MurNAc of thePG[87, 88]. In B. subtilis strains, WTA can also undergo D-alanylation and glycosylation, where glucose derived from UDP-glucose is attached by TagE to the TA before exportation to the cell surface [93, 94, 95]. When B. subtilis is grown under phosphate-limited conditions, WTA are replaced by teichuronic acids which are repetition chains of glucuronic acid synthesized by TuA proteins. In this media conditions, the transcription of the tag operon is repressed, while the transcription of the tua operon is induced. In this way, the anionic nature of glucuronic acid

20 ? LtaS

Glc -DAG FtsZ UgtP 2 UDP

D-alanine ester UgtP GtaB PgcA Glycerol phosphate UDP-glucose glucose 1-P glucose 6-P Glc2-DAG

Figure 1.4: Lipoteichoic synthesis machinery and lipid turnover in B. subtilis. Glycolipid synthesis precursors are produced in the cytoplasm of the cell. The glucose is transfer to DAG of the membrane by UgtP. The anchor moiety is transported outside of the membrane and glycerol-phosphate subunits derived from the membrane lipid phosphatidylglycerol are added to the anchor. The chain of glycerol-phosphate is elongated and LTA is formed. The LTA can suﬀer chain modiﬁcations, such as glycosylation and d-alanylation of the glycerol- phosphate subunits. Adapted from Reichmann et al [88].

can satisfy the cellular requirement for an anionic polymer. Moreover, the cell releases the existing TA and reutilizes the liberated phosphate for other cellular processes such as nucleic acid synthesis [96, 97, 98].

1.5.2.1 Role of teichoic acids in bacterial physiology

Although TA are commonly present in bacteria, their role in the cell is not completely clear. For a long time it has been assumed that TA are vital for bacterial growth. However, recently, this idea has been refuted. Schirner and co-workers suggested that both WTA and LTA are dispensable for viability but essential for cell morphogenesis and division and that their spatial distribution determines their speciﬁc functions [71]. In WTA studies, it was reported that the viability of B. subtilis cells were compromised

21 when late-acting genes such as tagB, tagD, and tagF were deleted. On the other hand, the same didn’t happen when accompanied by deletion of the early gene tagO. D’Elia and co-workers concluded that the apparent lethality was due to accumulation of WTA synthesis intermediates. Their physiologic experiments revealed that with the depletion of late-acting genes, thePG synthesis was decreased but that wasn’t seen when TagO was missing. This phenotypic difference was explained by the shared metabolite pathways of WTA andPG. The lesions in late steps of TA synthesis lead to the sequestration of undecaprenyl-phosphate in lipid-linked TA intermediates. Since undecaprenyl-phosphate is a shared metabolite withPG and is known to be a limiting metabolite of thePG synthesis, the reduction of free undecaprenyl-phosphate pool leads to the decrease ofPG assembly. On the other hand, upon tagO depletion, undecaprenyl-phosphate was no longer used for the synthesis of WTA allowing the increase of undecaprenyl-phosphate pool for the process ofPG synthesis [97, 99]. Albeit WTA aren’t essential for cell viability, cells lacking this TA exhibit morphological abnormalities. In B. subtilis, the absence of WTA is associated with loss of the characteristic rod shape, becoming spherical, a nonuniform thickness of thePG layer, and an aberrant placement of the division septum [71, 99, 100, 101]. Furthermore, it is hypothesized that WTA participates in the elongation process in B. subtilis since the lack of WTA affected their shape. Cells can still divide but lose the ability to maintain a rod shape and become almost round [71, 95]. Formstone and co-workers reported that WTA interacts with the MreB system during elongation [100]. In the absence of the WTA, the cell elongation was deregulated. Also, the analysis of the Tag proteins localization, revealed a localization over the lateral wall, a strik- ingly similar pattern as seen in the MreB helix-like structure. Concerning the LTA, it has been proposed to have several important roles, including the regulation of cell wall autolysins, the control of the cations availability within the cell envelope, and the increase of antibacterial resistance. Studies showed that with the lack of LtaS, no LTA is synthesized in B. subtilis and cells experienced shape defects and impairment of the septum and cell division. Moreover, FtsZ assembly was either missing, delocalized or with abnormal morphology. This deregulation of the divisome machinery led to the increase of cell length and a filamentous phenotype [71, 92, 102]. It has also been presumed that, due to the presence of D-alanine molecules in the structure of the LTA, cells are able control the surface charge. The D-alanylation tailoring of LTA modulates interactions between the cell envelope and the environment, which has been implicated in several roles: allow the maintenance of cation homeostasis, provides a reservoir of ions close to the cell surface that might be required for enzyme activity, and regulates the autolysins localization preventing their binding in the surface except to the cell septum [90, 101, 102]. Moreover, the modifications in the content of these amino acids showed an increased of the

22 bacterial sensitivity to lytic activity of enzymes produced by neutrophils during host infection and antimicrobial peptides [90, 101, 102]. In the absence of LTA, the autolytic activity at the septum is delayed which may be due to the absence of the anionic environment, impairing the cation homeostasis that is crucial to autolytic enzymes. Moreover, it has been seen a sporulation defects when ltaS and homologues were deleted [71, 92, 102]. Studies on antimicrobial peptides revealed that their mode of action is strongly related to cellular envelope constitution. The alteration of D-alanylation content on cell surface is seen as a bacterial adaptive response. With the increase of D-alanine in TA it reduces the anionic charges of cell surface which, in turn, decreases the interaction of cationic peptides. That was observed by Peschel and co-workers when the dlt operon was overexpressed in S. aureus and became resistant to cationic peptides [103]. The absence of D-alanylation also attenuates the binding of the bacterial cells to artificial surfaces as well as host tissue [90, 101, 102]. LTA and WTA have also been proposed to be involved in biofilm formation, immunogenic- ity and innate immune recognition. In pathogenic bacteria, during host infection, LTA is thought to be perceived by receptors on the immune cells, triggering innate responses in an effort to defend host tissues [104, 105]. Furthermore, studies revealed that the loss of both TAs in B. subtilis is lethal. This phenomenon may be explained by the loss of anionic polymers in the cell surface, affecting the cation homeostasis or that the effect provoked in the elongation and division processes impaired the cell viability. This confirms the idea that anionic polymers are essential for B. subtilis [71].

1.6 Pyruvate a regulator of division

A mechanism recently discovered that connects cell metabolism and division processes in B. subtilis is the pyruvate dependency for normal FtsZ assembly. Pyruvate is a key intermediate of various central carbon metabolism pathways, acting as a branching point of glycolysis and TCA cycle, and substrate of fatty acids and amino acids syntheses. In glycolysis, the ﬁnal reaction step is catalyzed by Pyk, that converts PEP to pyruvate (ﬁgure 1.8). Recently, it was reported by Monahan and co-workers that the intimate link between glycolysis and cell division could be disrupted with the loss of Pyk[65]. When pyk mutant cells (∆pyk) were cultivated in a rich medium, one-third had a defect in localization and time assembly of the Z ring. The latter was positioned near the cell poles and more than one FtsZ assembly was present in each cell. Interestingly, the defects were alleviated when exogenous pyruvate was added to the medium. These data indicated that pyruvate, or possibly a downstream metabolite, plays a crucial role in coordinating the central metabolism and bacterial division [65, 106].

23 Further in the study of other proteins from glycolysis and connected pathways whose activity could be aﬀected by pyk mutation, it was discovered a correlation between pyruvate levels, the localization of pyruvate dehydrogenase (PdH) - the enzyme responsible for the conversion of pyruvate to acetyl-CoA - and the nucleoid region [65]. A subunit of PdH, namely E1α, displayed a similar pattern localization as the nucleoid region in wt cells. In a ∆pyk background condition, where pyruvate accumulation is low, the colocalization of E1α and the nucleoid was absent. Furthermore, when exogenous pyruvate was added, E1α was shifted to nucleoid-like localization rather than accumulating in the cell poles. It was hypothesized that PdH could act as a positive regulator of FstZ assembly in a pyruvate/nutrient-dependent manner. In nutrient-rich conditions, when pyruvate levels are high, increased proportion of E1α of PdH are localized over nucleoid in the central region of the cell. On the other hand, in low pyruvate level conditions, E1α is unable to associate in the nucleoid region and accumulates in cell poles, which may lead to defects in Z ring assembly (ﬁgure 1.5). Since pyruvate is a key intermediate of various central carbon metabolism pathways, makes

Figure 1.5: Proposed Role for PDH E1α in nutrient-dependent regulation of B. subtilis cell division [65]. PDH E1α function as a positive regulator of Z-ring formation in a nutrient- dependent manner linked to pyruvate synthesis. In a low-nutrient availability condition (left image), PDH E1α (blue dot) shows a weak association with the nucleoid (grey circle). On contrary, in nutrient-rich medium (centre image) exhibits a strong localization over the chromossome. Moreover, in ∆pyk cells, where synthesis of pyruvate is blocked, the colocalization of PDH E1α and the nucleoid is absent. Instead, PDH E1α accumulates in the cell poles (right image). Accumulation of PDH E1α at the nucleoid-free cell poles under these conditions could trigger polar Z-ring formation by overcoming the inhibitory eﬀects of the Min system to generate a net positive signal for FtsZ assembly. it as an ideal candidate as a marker for nutritional status information and regulation of accurate division and growth rate in the cell [65, 68, 106, 107]. Besides the knowledge of the association of PdH and pyruvate during the division process, the interaction of E1α with the nucleoid and FtsZ assembly is still unclear.

24 1.6.1 Pyruvate catabolism and the transport mechanism in Bacil- lus subtilis

The control of pyruvate homeostasis and its fate is also of great importance for the insurance of cells’ robustness and viability during environmental changing conditions. A poor coordination between glucose consumption and precursor synthesis in the TCA cycle has been seen as one of the responsible causes for the incomplete oxidation of glucose and production of alternative metabolites - namely overflow metabolites. The reroute of the carbon overflow to the production of the latter, such as acetate, acetoin and 2,3-butanediol, is mainly happening in the pyruvate done [108, 109, 110]. It is well-known that in B. subtilis, the preferred carbon sources, such as glucose, are able to repress the transport and catabolism of alternative substrates. The CCR effect is achieved through the transcriptional control of genes required for the utilization of secondary carbon sources, imposing a hierarchy in the use of the available nutrients and preventing a waste of resources. The CCR is mediated by the CcpA, a master regulatory protein, which can also function as a carbon catabolite activation (CCA) in certain genes (figure 1.6)[110, 111, 112]. Different mechanisms exist in bacteria for the signal-transduction pathways that lead to the CCR. In B. subtilis, the CCR is mediated by the PTS. PTS is a multiprotein phosphorelay mechanism that catalyzes the phosphorylation of incoming sugar substrates and their simultaneous translocation across the cell membrane [113, 114]. In B. subtilis, during sugar uptake, the autophosphorylation of the enzyme EI with the phosphoryl group of PEP starts the reaction chain system which, consequently, leads to the production of pyruvate. The phosphoryl group is subsequently transferred to the His15 residue of the HPr. HPr-His-P transfers the phosporyl group to the enzymatic transporter EII, allowing the uptake of glucose [110, 115, 111]. When the intracellular concentrations of FBP and ATP are high, reflecting the presence of preferred carbon sources, HPr can also be phosphorylated at Ser46 by the ATP-dependent HPr kinase/phosphorylase (HPrK). This reaction allows the binding of HPr-Ser-P to CcpA. The HPr-Ser-P-CcpA interaction permits the binding of CcpA to specific sites on the DNA, and thereby, represses or activates the transcription of determined genes. Furthermore, the interaction between CcpA and HPr-Ser-P is enhanced by glucose 6-phosphate and FBP [110, 115, 111]. Although the extensive reports made on the regulatory phenomenon, the enlightening of the glucose-mediated repression of pyruvate uptake is still poorly understood. Recently, a pyruvate transport system was identified and characterized in B. subtilis [109, 116, 117, 118]. Charbonnier and co-workers demonstrated that the preferred carbon sources in B. subtilis, glucose, and malate, trigger the binding of CcpA upstream of pftAB gene, which encodes a pyruvate facilitated transporter, resulting in the repression of pyruvate utilization (figure 1.7). When these preferred sources are absent, CcpA repression is diminished [109]. Furthermore, a two-component regulatory system named LytST is also involved in pyruvate

25 Figure 1.6: PTS and CCR mechanisms - adapted from G¨orke et al.[111]. Glucose is transported by the PTS and, concomitantly phosphorylated to glucose 6-phosphate to feed glycolysis. The PEP produced during glycolysis is used in the phosphorylation of EI of the PTS. Afterwards, the phosphoryl group is transferred to His15 residue of HPr. HPr-His-P can provide the phosphoryl group to the enzymatic transporter (EII), allowing the transfer of glucose inside the cell. HPr is also phosphorylated at Ser46 residue by the HPr when the intracellular concentrations of FBP and ATP are high, reﬂecting the presence of preferred carbon sources. HPr-Ser46-P forms a complex with CcpA, which can bind to cre sites on the DNA, repressing the transcription of catabolic genes. The interaction of HPr-Ser-P and CcpA is enhanced by glucose 6-phosphate and FBP.

metabolism and uptake through the membrane. LytST induces the expression of pftAB in the presence of external pyruvate. However, when the pyruvate ﬂux into the cells is high, pftAB transcription is retro-inhibited via LytST. It was also shown that LytST activity works in a pyruvate dose-manner, which permits the balancing of intracellular pyruvate le-

26 Figure 1.7: Roles of PftAB and LytST in pyruvate homeostasis [109]. LytST senses the extracellular pyruvate concentration and responds by inducing pftAB transcription (pink arrow). The accumulation of intracellular pyruvate (or of an intermediate of overflow metabolism) reduces the level of induction of pftAB via LytST. Malate (or glucose) triggers the CCR of pftAB via CcpA. There is at least one other pyruvate transporter yet to be identified (grey circle). vels and adaptation to environmental changes [109]. Also in this study, the consumption of pyruvate was delayed and strongly reduced in ∆pftAB cells cultivated in a glucose and pyruvate medium (i.e. higher external pyruvate concentration than inside the cells). Consistently, cells overexpressing pftAB showed a drop of external pyruvate concentration from the beginning of the culture and reached zero prior the entry into stationary phase. On the contrary, control cells had a slight increase of external pyruvate, followed by a drop as soon as cells entered stationary phase. Moreover, it was reported that cultivating in a glucose medium, wt and ∆pftAB cells exhibited the same pyruvate secretion amount followed by its assimilation, while overexpressed pftAB cells increased by three times their pyruvate secretion. These, and the fact that there was no differences in the specific glucose consumption between wt and overexpressing pftAB cells, revealed that the gradient of pyruvate drove the PftAB-mediated transport of pyruvate across the cell membrane [109]. Although part of pyruvate homeostasis mechanism and the adaptation to new gradient concentrations by B. subtilis is better understood, the complete regulatory system remains unclear.

1.7 UDP-glucose as signal of nutrient availability

Another mechanism that links the division and the information of the nutrient availability to the cell is mediated by UgtP enzyme and its substrate, UDP-glucose. As referred, UgtP is a β-glucosyltransferase that synthesize the last step of diglucosyldia-

27 cylglycerol anchor of LTA by transferring two glucose moieties from UDP-glucose to DAG. UDP-glucose itself, is produced in two reversible steps from glucose 6-phosphate from the beginning of glycolysis: glucose 6-phosphate is converted to glucose 1-phosphate by alpha-phosphoglucomutase (PgcA), followed by the synthesis of UDP-glucose by UTP- glucosephosphate uridylyltransferase (GtaB) (ﬁgure 1.8)[104]. Besides the role of UgtP in the glycolipid synthesis, a landmark study done by Weart and co-workers in B. subtilis showed that UgtP interacts directly with FtsZ of the division apparatus, modulating its polymerization and delaying cytokynesis. Moreover, it also revealed a close connection between UDP-glucose amount and the division process [66]. Previously, it was reported by Lazarevic and co-workers that cells lacking PgcA, GtaB or

PgcA glucose

EIIGlc

EIIGlc A B

A B

P P glucose 6- P glucose 1- P UDP-glucose PgcA GtaB UgtP fructose 6- P

glycolysis

phosphoenolpyruvate diglucosyldiacylglycerol glycosylation of WTA Pyk synthesis anchor of LTA pyruvate Pck PdH Pyc

acetyl-coA

oxaloacetate

TCA cycle

Figure 1.8: Schematic representation of PTS glycolysis, the TCA cycle, and the glycolipid biosynthesis pathway, highlighting 4 enzymes for which mutants were considered in this study. PgcA, phosphoglucomutase; GtaB, uridine-diphosphoglucose pyrophosphory- lase; UgtP, uridinediphosphate glucosyltransferase; Pyk, pyruvate kinase.

UgtP exhibited deformed cell shape and defects in the bioﬁlm production [104].

28 In ∆pgcA, cells were one-third shorter than the wt, although they could still maintain a normal FftZ localization. Since both phenotypes showed the same doubling time, it was suggested that PgcA had a possible role in communicating the nutrient availability to the division site for cell size regulation. Furthermore, it was concluded that PgcA could modulate direct or indirectly FtsZ assembly dynamics. Additionally, since PgcA is a link between glycolysis and glycosylation, cells lacking glucose-6-phosphate isomerase (Pgi), that converts glucose 6-phosphate to fructose 6-phosphate in glycolysis, were also cultivated in same conditions. This mutation had no effect on the assembly of FtsZ, which supported the idea that cell division defect is associated with glycosylation pathway [104]. Years later, when Weart and co-workers cultivated ∆gtaB and ∆ugtP in a rich-medium, it was observed shorter cells and consequences in FtsZ assembly dynamics. The phenotypes of these LTA mutants in B. subtilis 168 and BSB1 strains were also reported by collaborators of this project (figure 1.9). Further in the study of glycolipid biosynthesis and nutrient availability, it was shown that UgtP inhibits FtsZ assembly through direct interaction and that the later is dependent of its substrate, UDP-glucose. Under nutrient-poor conditions, cells are deficient in UDP-glucose, which causes a reduced expression of UgtP. The remaining protein self-interacts and is sequestered in randomly position in the cell. Hence, UgtP doesn’t interact with FtsZ, which consequently allow the formation of the Z ring. On the other hand, during growth in nutrient rich-medium, UDP-glucose levels are abundant, inhibiting the oligomerization of UgtP which in turn, concentrates at the Z ring and interacts with FtsZ. The UgtP-FtsZ interaction delays cell division and, consequently, cells increase their size. In this mechanism, UgtP is able to detect the nutrition status of the cell via the accumulation of UDP-glucose, that is directly coupled to central carbon metabolism. With the inactivation of PgcA, GtaB and UgtP cells are unable to incorporate diglucosyldiacylglycerol in the LTA anchor. However, B. subtilis is still able to synthesize LTA with an unclear anchor composition. In S. aureus, it is known that ∆ugtp contains a diacylglycerol anchor instead of the glycolipid, and the glycerol chain is shorter than the wt [66]. In Lazarevic studies, the normal morphology of these three LTA mutant cells was reestab- lished with the addition of Mg2+, which can functionally replace the missing membrane glycolipid. In this view, it has been speculated that LTA anchor could also affect the cell phenotype in a indirect way by impeding the access to Mg2+ to the envelope synthetic enzymes [104]. Thus, it is also speculated that one function of LTA could be the maintainance of the correct ionic environment to the cation-dependent membrane systems [120, 92, 121]. The phenotypes observed in PgcA and GtaB lacking mutants support the model of subtract- localization dependency. In both mutants, UgtP accumulation was still present although this enzyme was localized diffusely in the cytoplasm. Thus, this change in UgtP localization might be dependent on the intracellular concentration of its substrate UDP-glucose. PgcA

29 Figure 1.9: Wt BSB1, ∆pgcA, ∆gtaB, and ∆ugtP cultivated in LB medium. The mutant strains presented shorter cells. Measurements are based on 100 cell count. Sassine J. et al. [119]. and GtaB can be seen as a way of informing carbon availability and growth rate information to next protein step of glycosylation pathway, UgtP. The mechanism discovered in these studies revealed that glycolipid biosynthesis pathway act as a link between metabolism and cell division in a nutrient-dependent manner.

30 1.8 Main objectives and general research plan

This project seeks to increase insights into the metabolic responses of B. subtilis to the absence of key signal metabolites - pyruvate and UDP-glucose - in the described mechanisms that mediate division and the nutritional status of the cells. First, B. subtilis lacking Pyk was subject to metabolomics’ studies, where target metabolites from several pathways and its possible changes were monitored. Cells were cultivated in chemically defined media, with pyruvate and glucose a carbon sources. Furthermore, it was also studied their metabolomics changes when cultivated in a complex media such as LB. Second, it was explored the impact of LTA mutations on the metabolome of B. subtilis. The presence of LTA in Gram-positive bacteria it has long been known. However, their precise function in the cell have remained elusive and studies have been hampered by the absence of mutants lacking LTA [122]. Furthermore, no reports have been made concerning the interaction of the metabolic precursors of the LTA and the central metabolic pathways of the cell. Here, a first insight was done into the metabolic profile of B. subtilis lacking the three lipoteichoic enzymes PgcA, GtaB, and UgtP.

31 32 Chapter 2

Metabolic responses to pyruvate kinase mutation when glucose and pyruvate are the main carbon sources

2.1 Results

2.1.1 Concomitant consumption of glucose and pyruvate by ∆pyk – Carbon catabolite repression of pyruvate in the presence of glucose was relieved

For the analysis of B. subtilis wt and ∆pyk, cells were cultivated in three chemically defined media: a mixture of glucose and pyruvate (M9GlcPyr), pyruvate (M9Pyr), and glucose (M9Glc) as single carbon sources. The growth curves arising from these conditions are represented in figures 2.1, 2.2, and 2.3 (black lines). Under cultivation in M9GlcPyr (figure 2.1), ∆pyk showed a growth delay during exponential phase when comparing to wt, even though both strains reached the same optical density (OD) maximum (OD=2.5). For wt, a diauxic growth curve was visible with an intermediate stationary phase between 360 and 420 min atOD 1.6. Using 1H-NMR spectroscopy, changes in the concentration of extracellular metabolites were monitored in a time dependent manner. These results showed that, in the wt cultivation, pyruvate consumption was initiated after 360 min of cultivation, when glucose was depleted from the medium (fig 2.1 and table 2.1). On the contrary, pyruvate uptake in ∆pyk started at 360 min, when 6.4 ± 0.8 mM of

33 Figure 2.1: Growth curves and extracellular concentrations of glucose (red) and pyruvate (blue) of B. subtilis wt and ∆pyk cultivated in M9GlcPyr. Black lines illustrate the growth curves. Data shown as mean values ± SD of four biological replicates.

Figure 2.2: Growth curves and extracellular concentrations of pyruvate (green) of B. subtilis wt and ∆pyk cultivated in M9Pyr. Black lines illustrate the growth curves. Data shown as mean values ± SD of four biological replicates.

Figure 2.3: Growth curves and extracellular concentrations of glucose (red) of B. subtilis wt and ∆pyk cultivated in M9Glc. Black lines illustrate the growth curves. Data shown as mean values ± SD of four biological replicates.

glucose was still available, showing a concomitant consumption of both carbon sources. Interestingly, acetate was accumulated in similar concentrations in both strains and no uptake was detected until the end of cultivation (ﬁgure 2.4 and table 2.2). Other typical overﬂow metabolites produced by B. subtilis, such as acetoin and 2,3-butanediol, were de-

34 Table 2.1: Glucose and pyruvate concentrations (mM) present in the supernatant during cultivation of wt and ∆pyk (M9GlcPyr). Data are shown as mean values ± SD of four biological replicates.

Table 2.2: Physiological parameters of wt and ∆pyk grown in the three chemically deﬁned media. Growth rate calculated in the beginning of exponential phase. The concentrations of acetate, acetoin, and 2,3-butanediol correspond to the maximum concentration achieved for each strain during the cultivation time course.

tected in both strains, however, these were secreted in notably higher amounts in ∆pyk, than in the wt. In both strains, these metabolites were taken up when external glucose was exhausted. The wt secreted signiﬁcantly larger amounts of 2-oxoglutarate, valine and BCAA degradation products such as 2-ketoisovalerate, isobutyrate, 2-methylbutyrate and isovalerate. Moreover, the concentration of valine and 2-ketoisovalerate started to drop when glucose was completely taken up in both strains.

In M9Pyr cultivation, cells grown slower if compared to the previous medium. Until the end of experimental growth, strains remained in exponential phase reaching aroundOD=0.5 at 540 min (ﬁgure 2.2 and table 7.1). In both phenotypes, pyruvate was taken up in slow rates. While at the initial of cultivation, 36.30 ± 2.49 mM (wt) and 36.41 ± 2.64 mM (∆pyk) of pyruvate were available in the medium, after 600 min of growth, pyruvate concentrations were 32.8 ± 2.34 mM and 33.02

35 Figure 2.4: Time-resolved extracellular metabolites under M9GlcPyr medium cultivation. Absolute concentrations of consumed and secreted metabolites by wt (black) and ∆pyk (red) are displayed. Data are represented as mean concentrations ± SD (shaded) of four biological replicates.

± 0.43 mM in wt and ∆pyk, respectively (ﬁgure 2.2). Notably, acetate was secreted in great amounts for both strains, reaching 3.2 ± 0.7 mM in wt and 3.0 ± 0.1 mM in ∆pyk at the last time point of cultivation (ﬁgure 7.2)). Moreover,

36 with the exception of the later metabolite, all secreted metabolites were detected in lower abundance when compared to M9GlcPyr determined metabolites.

Growing on glucose as a single carbon source (M9Glc), wt was able to reach a maximum OD of 4.9 ± 0.39 at 480 min, when utilization of glucose ended (ﬁgures 2.3 and 2.5). Im- mediately after reaching the stationary phase, cell mass decreased and cells suﬀered lysis. During exponential phase, several metabolites were secreted. Between 480-540 min, when growth reached the maximum OD, also the concentrations of acetolactate, acetate, pyruvate, 2,3-butanediol, acetoin and succinate were maximum in the medium. Both acetate and acetoin amounts were notably high: 10.57 ± 0.49 mM and 3.94 ± 0.80 mM, respectively. Similar to the other carbon surplus conditions, these metabolites were taken up again when glucose was depleted. The deletion of pyk resulted in impairment growth in M9Glc. In the end of the cultivation experiment, ∆pyk reachedOD=0.5 (table 2.1). At this time point, cells had taken up 2.17 mM of glucose from medium (table 7.1). Similar growth and glucose consumption rate was previously reported [123]. The metabolic footprint of ∆pyk also indicates that only minor amounts of metabolites were secreted by the cells.

2.1.2 Intracellular metabolome revealed several central metabolic pathways altered in ∆pyk

Studying the intracellular metabolome of B. subtilis wt and ∆pyk, several metabolites showed altered concentration levels when both strains were compared. The major common differences found when cells were cultivated in M9GlcPyr and M9Glc was the notably increased levels of glycolytic metabolites in ∆pyk (figures 2.6 and 2.7). In M9GlcPyr, some metabolites upstream pyruvate accumulated, being PEP the uppermost altered metabolite with a Fold Change (FC) of 10, followed by glucose 6-phosphate (FC 5.0), fructose 6-phosphate (FC 4.6), 3-phosphoglycerate (FC 3.1), and 2-phosphoglycerate (FC 3.2) (figures 2.6 and 2.7). Considering the PPP, only sedopheptulose 7-phosphate presented altered concentration levels (FC 3.2). In the TCA cycle, although 2-oxoglutarate secretion during late exponential and stationary phase was lower in ∆pyk, the intracellular concentration at 0.5 OD - when intracellular sampling was conducted - was similar in both phenotypes (FC 1.0). Among the nucleotides, nucleoside triphosphate UTP and CTP were determined in significantly lower amounts (p-value≤0.05) in ∆pyk (FC 0.40 and 0.38, respectively), as well as IMP (FC 0.40) and the deoxy nucleotides dCTP (FC 0.27) and dTMP (FC 0.47). On the contrary, AMP was the only nucleotide with statistically higher levels in ∆pyk (FC 2.1) (figure 2.6).

As far as the cultivation under M9Glc, 28 metabolites diverged between strains (table 7.2).

37 Figure 2.5: Time-resolved extracellular metabolites of wt under M9Glc medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis wt are displayed. Data are shown as mean concentrations ± SD (shaded) of four biological replicates.

Almost all the glycolytic metabolites detected in this condition showed drastically higher concentrations in ∆pyk: PEP, followed by 2-phosphoglycerate (FC 22.0), 3-phosphoglycerate (FC 15.6), fructose 6-phosphate (FC 11.7), glucose 6-phosphate (FC 8.3), and FBP (FC 5.3) (ﬁgures 2.6 and 2.7).

38 Figure 2.6: Heat-map of intracellular metabolite levels of glycolysis, TCA cycle and PPP (A), amino acids (B), intermediates of purine and pyrimidine metabolism and cell wall precursors (C) in wt and ∆pyk under M9GlcPyr, M9Pyr, and M9Glc. Color-code represent the log2 FC ratio between mutant and control cells, whereas increased levels are indicated in orange and lower levels in purple. Data are shown as the mean of four biological replicates.

For the PPP, sedoheptulose 7-phosphate was determined with a FC of 4.3. Thirteen metabolites were found in lower levels such as pyruvate (FC 0.48) and some nucleotides. Another altered metabolic pathway was the peptidoglycan synthesis. Peptidoglycan precursors like UDP-MurNAc-ala-glu and UDP-MurNAc-ala-glu-pm-ala-ala had a quantity de-

39 / 1.31 / Glucose 0.65 / 1.04 / 2.33 1.43 / 1.38 / 0.78 5.02*/ 1.98 / 8.32* Glucose 6-P Glucono-1,5-lactone 6-P 6-phosphogluconate

4.57*/ 1.47 / 11.69* Fructose 6-P Pentose Phosphate Pathway 1.85/ 1.00 / 5.27* Fructose 1,6-bisP 3.21*/ 1.20 / 4.26* Glyceraldehyde 3-P Sedoheptulose 7-P

2.30 / 1.03 / 7.06* 1.88 / 1.22 / 1.45 4.57*/ 1.47 / 11.69* Dihydroxyacetone P Glyceraldehyde 3-P Erythrose 4-P + Fructose 6-P

1,3-bisphosphoglycerate

3.07* / 1.06 / 15.63* 3-phosphoglycerate

3.19 / 1.56 / 21.95* 2-phosphoglycerate 10.0* / 1.90 / 90.0* Phosphoenolpyruvate 3-deoxy-D-arabino-heptulosonate 7-P / / 0.48* 0.75 / 1.19 / 0.73 Pyruvate Lactate Shikimate Pathway 1.47 / 2.00 / 0.33* Acetyl-CoA 0.95 / 0.77 / 0.75 Shikimate Oxoloacetate Citrate 2.72 / 1.06 / 0.87 Malate cis-Aconitate 1.44 / 1.02 / 0.84 0.69 / 0.84 / 2.23* 0.90 / 1.38 / 2.74* Isocitrate 0.75 / 1.59 / 1.52 Fumarate Tyrosine Phenylpyruvate Δpyk 1.02 / 0.70 / 0.58 Tryptophan FC = 0.84 / 0.78 / 0.86 wt Succinate 2-oxoglutarate 0.73 / 0.90 / 1.27 M9GlcPyr/M9Pyr/M9Glc Succinyl-CoA Phenylalanine

Figure 2.7: FC of some intracellular metabolites from central metabolism. Relative levels of metabolites from central metabolism for ∆pyk as compared with wt under M9GlcPyr (first value), M9Pyr (second value), and M9Glc (third value).FC determined using the relative quantification of four biological replicates. Significant alterations (p-value≤0.05) are marked in bold and asterisk.

crease in ∆pyk with a FC of 0.37 and 0.39, respectively (figure 2.6). Both purine and pyrimidine synthesis metabolites were affected, being the most significant changes seen in deoxynucleotides like dCMP, dCDP, dCTP, and dATP.

Out of 92 identified metabolites, only 2 were significantly altered when cultivated in M9Pyr: dCMP (FC 0.36) and FAICAR (FC 2.5), which reveals an identical metabolome profile between both strains under this condition (figures 2.6 and 2.7).

40 2.1.3 Increased amino acids and shikimate 3-phosphate levels in ∆pyk under glucose medium cultivation

The cultivation of wt and ∆pyk strains under glucose as single carbon source triggered several alterations in the amino acids concentration pool. As seen previously, with the lack of Pyk enzyme, pyruvate concentration was significantly decreased. Albeit this phenomenon, and except for 2-oxoglutarate, concentration of TCA cycle metabolites marginally differ between strains (figure 2.7). Interestingly, amino acids which are synthesized via glutamate were highly accumulated in the mutant as they are: ornithine (FC 5.9), citrulline (FC 4.0), arginine (FC 2.0) and proline (FC 8.0) (figure 2.8). Besides the feed of the TCA cycle through acetyl-CoA, pyruvate can also be directly con-

TCA cycle

Figure 2.8: Altered amino acids levels under M9Glc cultivation. Relative amount of 3- phosphoglycerate, pyruvate, 2-oxoglutarate, and some amino acids in wt (black column) and ∆pyk (grey column) showing an increase of almost all amino acids displayed in ∆pyk. Data are shown as mean values ± SD of four biological replicates. Signiﬁcant alterations levels (p-values≤0.05) between strains are marked with asterisk. verted to oxaloacetate and the later be used for the synthesis of aspartate, asparagine and other amino acids. Likewise, aspartate and asparagine had FC of 2.8 and 2.5, respectively. Surprisingly, alanine and valine, both metabolized from pyruvate, showed no statistically diﬀerences between strains (p-value≤0.05). Surprinsingly, shikimate 3-phosphate, which is synthesized from PEP and erythrose 4- phosphate for the production of aromatic amino acids, had lower concentration levels in

41 ∆pyk in all media conditions (ﬁgure 2.9).

Shikimate 3-phosphate 0.05

t 0.04 n u

o 0.03 m a e v

i 0.01 t

a ** l e r * *** 0.00 t k t k t k w y w y w y r p r p p y y lc P y P y G lc c P 9 P 9 G l lc 9 9 G G M M M 9 9 M M M

Figure 2.9: Relative amount of shikimate 3-phosphate in wt and ∆pyk under M9GlcPyr, M9Pyr, and M9Glc media cultivation. Data are shown as mean values SD of four biological replicates. Statistical diﬀerences between control and mutant were considered signiﬁcant for p-values≤0.05 (marked with asterisk).

2.2 Discussion

2.2.1 Relieve of carbon catabolite repression of pyruvate in ∆pyk cells

An interesting outcome of this study is that pyruvate was subject to glucose CCR in wt cells, while this repression wasn’t observed in ∆pyk. Under M9GlcPyr, the rise of pyruvate inﬂux into wt cells initiated at 360 min of growth, only after external glucose was depleted, indicating that pyruvate was susceptible to glucose dependent CCR. This phenomenon is also perceived by the diauxic growth curve obtained, not described so far. On the other hand, in ∆pyk, 6.4 ± 0.8 mM of glucose was still available in the medium when pyruvate consumption was rapidly initiated. The CCR derepression seen in ∆pyk can be speculated as a cellular metabolic response for the immediate need of other carbon source since the central metabolic pathways could be altered due to accumulation of glycolytic metabolites (discussed below). Although it’s well-known that the primary carbon choice of B. subtilis is glucose, the CcpA-

42 dependent catabolite repression mechanism on pyruvate uptake was just recently clarified. Charbonnier and co-workers reported the discovery of a new pyruvate transport mechanism mediated by CcpA. The CcpA controls the expression of pftAB, that encodes a pyruvate facilitated transporter, in accordance to the absence/presence of glucose or malate [109]. Knocking out and overexpressing pftAB, it was concluded that the pyruvate gradient between outside and the inside of the cell drives the facilitated PftAB-mediated transport of pyruvate across the cell membrane. In this study, when ∆pyk cells are growing in M9GlcPyr and intracellular pyruvate pool is presumably low, the early influx of pyruvate can be consequence of LytST response to the pyruvate gradient. LytST might senses the high pyruvate gradient concentration and responds by inducing pftAB transcription for a rapidly increase of pyruvate influx. In wt, where pyruvate concentration is under physiological levels inside the cells and pyruvate gradient is lower, LytST induction is probably impaired and/or the repression of pftAB by CcpA is highly active. When glucose is completely consumed, CCR effect on pftAB is relieved and wt cells are able to take up pyruvate as a nutrient source. Moreover, when cultivating wt cells in M9Glc, pyruvate was exported during exponential phase, when glucose was still available, followed by its assimilation. In this way, it is en- couraged to conclude that another mechanism independently of PftAB transporter might result in the pyruvate efflux. As described Charbonnier study, ∆pftAB cells were able to secrete and taken up pyruvate from medium. These results indicated that there exists at least another pyruvate transport system allowing the import and export of pyruvate, yet to be identified [109]. Additionally, both strains were cultivated in a chemically defined medium with pyruvate and single carbon source. In this condition, wt and ∆pyk showed the same uptake rate of pyruvate, suggesting that the lack of Pyk didn’t affect pyruvate transport mechanism. Although part of pyruvate homeostasis mechanism and the adaptation to new gradient concentrations by B. subtilis is better understood, part of the regulatory system remains unclear.

2.2.2 Pyk mutation leads to impairment of glucose uptake when glucose is the only carbon source

The fact that, when glucose is the only carbon source, the latter is slowly taken up from medium into the cytoplasm in ∆pyk cells, and that, except for pyruvate, all detectable glycolytic metabolites risen, reveal that the absence of Pyk is creating a bottleneck eﬀect downstream of glycolysis. One speculation for the diminish glucose consumption in the mutant strain is that with the absence of Pyk, cells cannot accomplish physiological levels of pyruvate which, in turn, could relieve the surplus of carbon and energy from glycolysis. This could be achieved

43 by proceeding the carbon flux towards the TCA cycle, for reducing power and anabolic intermediates production, or via overflow metabolism. The knockout of Pyk led to a substantial increase of PEP concentration and a significant reduction of pyruvate during the growth on glucose (M9Glc). On the contrary to Escherichia coli (E. coli), where the anaplerotic reaction that replenish the TCA cycle with oxaloacetate is made by PEP carboxylase, B. subtilis relies on pyruvate carboxylase [124]. Thus, the accumulation of glycolytic metabolites and the lack of ability to reroute the accumulated PEP to the TCA cycle flux could have compromised the glucose uptake. Another possible cause for the low glucose consumption rate seen in ∆pyk is the impairment of the glucose uptake by the PTS system, due to the accumulation of high levels of FBP. The concentration of glycolytic intermediates, predominantly FBP, have been proposed to regulate the PTS system at the level of EII-mediated uptake [125, 126, 114]. High FBP concentration stimulates directly the phosphorylation of HPr through HPrK, favoring the formation of HPr-Ser-P-CcpA complex. Consequently, less HPr is available for phosphoryl group transfer from PEP to the sugars during the taken up via the EII of the PTS[125, 126, 114]. In ∆pyk cells, the increase of FBP might have led to an efficient phosphorylation of HPr at Ser-46 which, in turn, could have compromised the PTS system and reduced the uptake of glucose (figure 2.10). To note that in M9GlcPyr, the glucose consumption in ∆pyk wasn’t impaired and FBP levels weren’t statistically different between both strains.

2.2.3 FBP accumulation in ∆pyk could have led to dismiss of gap operon repression by CggR

Another function identiﬁed for FBP in the carbon metabolism regulation is its activity in inhibiting CggR, a repressor of part of the glycolytic pathway [127]. The genes encoding the enzymes of lower part of glycolysis (gap-pgk-tpi-pgm-eno) along cggr, form the gapA operon [127, 128]. The increase of FBP negatively modulates the action of CggR by diminishing the ability of the later to bind the cis-sequence of gapA operon, which encodes enzymes of lower part of glycolysis [127, 128]. Doan and co-workers showed that the increased FBP levels relieves the gapA operon from repression and allows the enhancement of the glycolysis ﬂux from dihydroxyacetone-phosphate to PEP[129]. Since in this study ∆pyk accumulated great amount of FBP in M9Glc, is speculated that CggR binding to DNA was strongly inhibited. In fact, 3-phosphogycerate and 2- phosphogycerate concentrations were 16 and 22 times higher, respectively, than in the control strain.

44 Figure 2.10: PTS and CCR mechanisms - adapted from G¨orke et al. [111]. Glucose is transported by the PTS and concomitantly phosphorylated to glucose 6-phosphate to feed glycolysis. The PEP produced during glycolysis is used in the phosphorylation of EI of the PTS. Afterwards, the phosphoryl group is transferred to His15 residue of HPr. HPr- His-P can provide the phosphoryl group to the enzymatic transporter (EII), allowing the transfer of glucose inside the cell. HPr is also phosphorylated at Ser46 residue by the HPrK when the intracellular concentrations of FBP and ATP are high, reﬂecting the presence of preferred carbon sources. HPr-Ser46-P forms a complex with CcpA, which can bind to cre sites on the DNA, repressing the transcription of catabolic genes. The interaction of HPr- Ser-P and CcpA is enhanced by glucose 6-phosphate and FBP. The accumulation of FBP in M9Glc condition, could have directly stimulated the phosphorylation of HPr through HPrK, favoring the formation of HPr-Ser-P-CcpA complex. Consequently, less HPr was available for phosphoryl group transfer from PEP to the glucose be taken up via the EII.

45 2.2.4 Possible carbon reroute to PPP

As previously described, the growth of ∆pyk in M9Glc and M9GlcPyr resulted in accumulation of glucose 6-phosphate. It has been reported that the lack of Pyk in E. coli and B. subtilis under glucose medium increases the flux of glucose 6-phosphate through PPP[130, 131, 132, 133]. Although just few metabolites of this pathway were identified, the results are consistent with this finding, since sedoheptulose 7-phosphate showed higher amounts in ∆pyk when grown under M9PyrGlc and M9Glc. The perturbation seen in glycolysis in the mutant could have caused the reroute of flux metabolism to compensate the required amount of NAD(P)H and syntheses of nucleotide precursors [130, 134].

2.2.5 TCA cycle might be less active in ∆pyk

In the initial exponential growth of wt cells with available glucose and pyruvate, no TCA cycle metabolites were detected inside the cells, which might due to the low activity of this pathway since is predicted to be weakly active during this growth phase [135]. In B. subtilis, when preferred carbon sources are available in the medium, the activity of the first TCA cycle enzymes are inhibited by several regulatory proteins [115, 136]. Moreover, when the external glucose was depleted, an efflux of 2-oxoglutarate in wt is initiated (figure 2.4), suggesting an increase of TCA cycle activity in response to the cells’ requirements. At this time point (360 min), pyruvate consumption also begin, which can be used for replenishment of TCA cycle. As previously discussed, during the late exponential growth phase of ∆pyk in M9GlcPyr, pyruvate uptake is initiated when glucose is still available. Notably, at this time, 2-oxoglutarate secretion begins, reaching in the end of cultivation 30% of the concentration seen in wt (0.23 ± 0.03 mM and 0.07 ± 0.01mM for wt and ∆pyk, respectively). Moreover, in ∆pyk cells cultivated in M9Glc, the intracellular acetyl-CoA was 3-fold lower and less 2-oxoglutarate was secreted, which suggests that the TCA cycle was probably less active than in wt. Nevertheless, in stationary phase, TCA cycle repression was probably relieved for the use of alternative carbon sources in the central metabolic pathways. Thus, pyruvate that is cross- ing into the cell is oxidized to acetyl-CoA production or catabolized by pyruvate carboxylase to oxaloacetate, and immediately directed to the TCA cycle.

2.2.6 Overﬂow metabolites: increased of acetoin and 2,3-butanediol secretions by ∆pyk

A common phenomena during B. subtilis cultivation with high availability of glycolytic carbon sources are the production and secretion of overﬂow metabolites even in aerobic conditions. Studies suggest that this onset of overﬂow metabolism during surplus of carbon and

46 energy is driven by saturation of the electron transport chain that generates the demanding ATP in fast growing cells, and the imbalance of intracellular NADH/NAD+ ratio observed prior to the beginning of fermentation [137, 138]. As typical in B. subtilis, acetate was the overflow compound with the highest secretion amount [134]. The expression of genes encoding phosphotransacetylase and acetate kinase, which are involved in the synthesis of acetate, are activated by CcpA in the presence of glucose (CCA effect). Furthermore, acsA, which encodes acetyl-CoA synthetase and permits utilization of acetate, is subject to CCR by glucose [115, 139, 140]. Consistently, in M9GlcPyr, acetate was secreted as soon as glucose was imported and metabolized by both strains. Besides providing a carbon source storage, acetate enables the generation of ATP needed to fuel the high growth of the cells. Studies on electron transport chain kinetics and energy efficiency suggested that the high demand for surface-inefficient electron transport chain by NADH production in the TCA cycle, and its membrane limitation forces the synthesis of the required ATP through acetate fermentation, a more surface-efficient mean [138]. This could be the reason for the secretion of the same amount of acetate in both strains, which might be the optimal/limit value in this condition. Furthermore, since acetate accumulation results in acidification of the cells and its growth environment, it probably imposed a limit on its production. Although acetate could be reimported as an alternative carbon source after depletion of glucose due to dismiss of the CCA and CCR mechanism, cells preferred to utilize pyruvate. Thus, acetate efflux continued until the end of cultivation. When stationary phase was reached, acetate uptake wasn’t noticed, as commonly reported [111, 141]. However, it is speculated that acetate influx could be observed if cells would have been cultivated for longer time on stationary phase. In fact, when cultivating wt in M9Glc, acetate assimilation was observed as soon as external glucose ended. Furthermore, it is also hypothesized that acetate uptake could have been subject of catabolic repression by pyruvate since the latter was consumed as alternative carbon source during the exponential phase. Acetoin and 2,3-butanediol secretions were also detected in M9GlcPyr cultivation. Notably, the effluxes were much higher in ∆pyk than in wt. The reason for these differences remain unclear. The upper high efflux of acetoin was previously observed in ∆pyk [123]. The production of these overflow metabolites is perceived as a preventive mechanism of the environment acidification due to acetate accumulation [140, 142]. Acetoin is synthesized from pyruvate and can either be secreted or converted into 2,3- butanediol in order to generate NAD+. These compounds can also be reimported during stationary phase and, in this way, being used as an energy-storing strategy. The fact that ∆pyk secreted more acetoin and 2,3-butanediol than in wt lead us to think that the perturbation occurring in pyruvate node, could result in an imbalance of NADH/NAD+

47 ratio, which is known to be potentially toxic to the cells [137, 143]. Consistently, the higher synthesis of acetoin and subsequent 2,3-butanediol, result in more NAD+ molecules regenerated, which can help the reducing power ratio balance. In agreement to the fact that acetoin synthesis can also be induced by acetate, the secretion of acetoin, and in later extent 2,3-butanediol, occurred alongside acetate accumulation [142, 144, 145]. On the contrary to acetate, acetoin and 2,3-butanediol uptake was observed when stationary phase was reached, suggesting that they were used as alternative carbon sources as soon as glucose and pyruvate were exhausted.

2.2.7 Lower secretion of BCAA by ∆pyk

Besides common overﬂow metabolites, valine and several intermediates of BCAA metabolism were secreted to the supernatant. Since BCAA are among the most abundant amino acids in proteins, the maintenance of their intracellular pool is fundamental for the regulation of high protein synthesis level. For this, degradation of BCAA and synthesis of branched-chain keto acids may occur [136, 146]. During growth on M9GlcPyr, valine and 2-ketosisovalerate were accumulated until depletion of both carbon sources. Afterwards, they were reimported into the cell, probably to feed necessary metabolic pathways like the TCA cycle. Moreover, the other detected branched-chain keto acids, isovalerate, isobutyrate, and 2- methylbutyrate, were successively exported over time until the end of the growth. All these metabolites reached higher levels in wt when compared to ∆pyk. This result seem coherent since intracellular pyruvate, the precursor of their synthesis and degradation, is possibly lower in the mutant. Except for valine, all metabolites presented the same behavior in wt under M9Glc. The fact that only secretion and no uptake was observed for valine, reveals that B. subtilis prefers other alternative carbon sources, like acetate, than this amino acid.

2.2.8 Amino acids pool: proline, ornithine, citrulline, and arginine metabolism altered

Several amino acids presented changed concentration levels in ∆pyk when compared to wt on M9Glc. The intracellular change observed for proline-ornithine-citrulline-arginine metabolism in ∆pyk remains unclear. It can only be speculated that rocR regulon, involved in nitrogen metabolism and arginine degradation pathway, could be altered [147]. The rocG gene, member of rocR, encodes the catabolic glutamate dehydrogenase (GDH), which reversibly converts 2-oxoglutarate and glutamate. Moreover, RocR also controls the rocABC and

48 rocDEF operons, whose products catalyze the degradation of arginine to glutamate. The production of 2-oxoglutarate from glutamate through GDH is viewed as the ﬁnal step in the utilization of arginine, ornithine, and proline, providing a metabolizable carbon or nitrogen- containing compounds for biosynthesis [147, 148]. Since rocG is also subject to the direct CcpA-dependent glucose repression, the higher stimulation of the HPr-Ser-P-CcpA complex previously speculated, could have led to an inhibition of rocG [148]. Consequently, this directly contributed to increased intracellular glutamate, followed by an impairment of arginine degradation. Furthermore, despite the low pyruvate concentration in ∆pyk, aspartate and asparagine didn’t diminish as it could be expected. Nevertheless, since the conversion of oxaloacetate to aspartate is coupled to glutamate degradation, the accumulation of aspartate could be associated to the higher glutamate levels.

2.2.9 Shikimate metabolism

Surprisingly, shikimate 3-phosphate, a metabolic intermediate of aromatic amino acids synthesis, was detected in lower amounts in ∆pyk in all three chemically defined media (figure 2.9). Due to their pharmaceutical values,the intermediates of shikimate pathway are commonly investigated. [149, 150, 151]. For the increase of shikimate 3-phosphate precursors, such as 3-deoxy-arobino-heptulosonate 7-phosphate and shikimate, several approaches have been developed. One of these strategies has been the inactivation of pyk and genes downstream of shikimate pathway, resulting in the accumulation of the metabolite of interest. The shikimate pathway starts with the condensation of erythrose 4-phosphate and PEP. Considering that the flux partitioning at the PEP node has been identified as the major determinant of the yield of aromatics from glucose, the accumulation of PEP seen would made predictable the accumulation of shikimate 3-phosphate [123].

2.3 Conclusion

The metabolic responses of B. subtilis lacking Pyk under different nutritional environments were inspected. It was highlighted relevant differences between wt and ∆pyk cells, when monitored in three chemically defined media. When cultivated with glucose and pyruvate as carbon sources, it was observed a concomitant consumption of both metabolites in ∆pyk cells, whereas in wt, the uptake of pyruvate was perceived after the complete consumption of glucose. This distinct behavior is in accordance to the pyruvate transport mechanism, recently discovered in B. subtilis. While in wt, the pyruvate utilization suffers CCR by the CcpA-dependent glucose repression, in ∆pyk, this effect might be relieved due to the induc-

49 tion of the pyruvate facilitated transporter, by LytST. The lack of Pyk and consequent low intracellular pyruvate level, could be involved in this induction, since LytST works in a pyruvate dose and gradient-manner. Although the metabolic results helped in the elucidation of the regulatory transport of glucose and pyruvate, the complete knowledge of the pyruvate transport mechanism and the adaptation to new gradient concentrations by B. subtilis is yet to be done. Another outcome of this study was the diminishing glucose uptake by ∆pyk, when glucose was the only carbon source. It is hypothesized that the accumulation of glycolytic metabolites, as a result of the bottleneck created in the pyruvate node, and the impossibility of continuing to the complete glucose oxidation through TCA cycle, led to the great decrease of glucose influx in the mutant cells. Moreover, it was speculated that the PTS system could be impaired, since ∆pyk presented a significant increased level of FBP. It is well known that high levels of FBP stimulates directly the phosphorylation of HPr, favoring HPr-Ser-P- CcpA formation. Consequently, less HPr would be available to proceed the glucose uptake through the PTS. Further experiments are necessary to clarify the slow glucose consumption and perhaps the impairment of the PTS system. These would include the inspection of other PTS-sugars transport by ∆pyk, the quantification of HPr-Ser-P, and the HPrK activity levels. Other prominent pathways were affected by pyk mutation such as the overflow metabolism, amino acid synthesis, TCA cycle, and the peptidoglycan assembly. Taken together, these evidences that the perturbation created in the pyruvate node had consequences in several central metabolism of bacteria. Furthermore, with the unexpected decrease of shikimate 3-phosphate in ∆pyk, the regulation of shikimate pathway, especially downstream of shikimate, is worth of further investigation. Considering that pyruvate is a link of essential pathways and its fate is important for cells’ robustness and viability, the metabolic approach of this study helps to unravel the control of pyruvate homeostasis and B. subtilis’ adaptation to the environmental challenges.

50 Chapter 3

Metabolic responses to pyruvate kinase mutation in a complex medium

3.1 Results

3.1.1 Similar metabolomic uptake and secretion by wt and ∆pyk

B. subtilis wt and ∆pyk metabolic profiles were also analysed using LB as a complex medium. Wt achieved a well reproducible growth curve having a higher yield growth than ∆pyk (figure 3.1). This good growth behavior was also reflected by the high adenylate energy charge (EC) ratio of 0.76. For the calculation of theEC, the ratio of intracellular AMP, ADP, is used and given by the equation 3.1. Besides being addressed for the determination of efficient quenching of metabolism during sampling - experimental quality-control - it is an acceptable parameter that reflects the energy status of the cell [152, 153].

[AT P ] + 1/2[ADP ] EC = (3.1) [AMP ] + [ADP ] + [AT P ]

It is assumed that during normal growth, theEC of intact cells during active metabolism is stabilized in the vicinity of 0.8 [152]. For ∆pyk, the growth curve revealed a high standard deviation (SD) during the exponential phase. Moreover, the EC was lower than in control strain. Under this rich medium, cell mass of both strains remained constant when stationary phase was reached and no cell lysis was observed. Whilst LB is the preferable medium for cultivating in a complex medium, the exact composi-

51 10 wt pyk mutant m n 0 0

6 1 D O

0.1 0 100 200 300 400 500 600 t (min)

Figure 3.1: Growth curves of B. subtilis wt (circle) and ∆pyk (square) are presented. Data are shown as mean values ± SD of three biological replicates.

tion is roughly known [154]. The major component are oligopeptides, that can be degraded by oligopeptide permeases and peptidases present in B. subtilis, enabling the recovery of catabolizable amino acids [155]. In less extent, glucose, trehalose and some free amino acids are also present in LB medium. Due to measurement limitations, the oligopetides present in the extracellular environment aren’t detectable by 1H-NMR. Nevertheless, 22 metabolites were identified, 12 of which correspond to amino acids that constitute the LB medium (figure 3.2). Lysine and phenylalanine were the amino acids consumed in lower amount by both strains. While for wt, 16% of lysine was consumed during the whole time of cultivation, in ∆pyk, the consumption was of 5.8%. For phenylalanine, the consumption was 12% and 8.1% for wt and ∆pyk, respectively. Although the main nutritional source in LB medium are oligopeptides, the small amount of glucose detected in the beginning of cultivation (0.42 ± 0.02 mM) was fully taken up by wt and ∆pyk. Concomitantly, trehalose was also completely consumed from the external medium. Only some secreted metabolites were detected, mainly BCAA intermediates and acetate. The notably differences observed in the extrametabolome profiles between both strains were the lower secretion rate of BCAA precursors 2-methylbutyrate, isobutyrate and isovalerate by ∆pyk. Furthermore, ∆pyk secreted lower amounts of 2-methylbutyrate and isobutyrate. After 540 min of cultivation, when cells were in stationary phase for at least 3 hours, 2- methylbutyrate reached a maximum concentration of 1.65 mM and 1.87 mM in wt and ∆pyk, respectively. The same happened for isobuytrate which reached a secretion maximum of 2.0 mM and 2.4 mM for ∆pyk and wt, respectively. Acetate was secreted in notably high concentration. After 180 min of cultivation, wt pre-

52 Figure 3.2: Time-resolved extracellular metabolites under LB medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis wt (brown) and ∆pyk (orange) are displayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

sented 5.2 ± 0.07 mM of acetate outside the cells, reaching similar concentration value in ∆pyk one hour later (4.9 ± 0.13 mM). Afterwards, acetate was fully taken up, reaching

53 minimal values after 540 min of cultivation. Other commonly overﬂow metabolites in B. subtilis, such as acetoin and 2,3-butanediol, weren’t detected in this condition.

3.1.2 Changes in the central carbon metabolism pathways

Using GC-MS and LC-MS, intracellular metabolites from different metabolic pathways were identified. By plotting in a volcano-type plot the log2FC between both strains as function of the t-test results, it is visible divergences in the metabolic profiles (figure 3.3).

From the metabolites identiﬁed inside the cells, 10 were statistically higher (log2FC≥1) in

0.0001

0.001 e u l

a 0.01 v - p

0.1

1 -4 -2 0 2 4

Log2 Fold Change

Figure 3.3: Volcano-plot of intracellular metabolome data of B. subtilis wt and ∆pyk in LB medium. In the volcano-plot is presented the FC given by the relative amount of wt and ∆pyk intracellular metabolites (in logarithmic scale) as function of unpaired t-tests (p- value results). Metabolites with signiﬁcant changes (p-value≤0.05) and -1>log2FC>1 are displayed in the upper left (red) and right (green) region of the plot.

∆pyk, while 12 showed lower concentration amounts (log2FC≤-1). The knockout of pyk resulted in the alteration of the glycolysis pathway. The metabolites of the lower glycolytic pathway were accumulated in the ∆pyk being 3-phosphoglycerate the most prominent altered metabolite with a concentration of 1.18 ± 0.26 and 5.57 ± 0.15 nmol/OD in wt and ∆pyk, respectively; followed by PEP (0.36 ± 0.02 and 1.68 ± 0.5 nmol/OD) and 2-phosphoglycerate (0.21 ± 0.07 and 0.45 ± 0.13 nmol/OD) (ﬁgure 3.4). On the contrary, it was also detected lower amounts of pyruvate in mutant cells with 4.4 ± 1.2 nmol/OD detected in wt and 0.60 ± 0.08 nmol/OD in ∆pyk.

54 Figure 3.4: Absolute concentration of glycolytic metabolites in wt (black column) and ∆pyk (grey column). Data are shown as mean values ± SD of four biological replicates. Signiﬁcant alteration levels (p-values≤0.05) between strains are marked with asterisk.

Lactate, produced from pyruvate, was also found in lower concentrations inside ∆pyk cells (3.61 ± 0.60 and 1.95 ± 0.53 nmol/OD in wt and ∆pyk, respectively), although their consumption from the medium was similar during the time course. Moreover, the intracellular amounts of some TCA metabolites were also absolutely quantified (figure 3.5). These metabolites presented lower concentrations amount in the mutant, in particular 2-oxoglutarate with 8.38 ± 0.06 nmol/OD and 2.98 ± 0.01 nmol/OD in wt and ∆pyk, respectively. Additionally, the perturbation in the pyruvate node resulted in alteration of metabolites levels from PPP (figure 3.6). The metabolite 6-phophogluconate was significantly higher in ∆pyk (FC 5.8, p-value=0.01), followed by sedoheptulose 7-phosphate (FC 4.2, p-value=0.01). Moreover, a metabolite with the molecular mass of ribulose, xylulose or ribose 5-phosphate was also detected in higher concentration levels in the mutant cells. Like observed in the chemically defined media study, shikimate 3-phosphate showed decreased level in the ∆pyk (FC 0.27). Concerning the amino acids detected inside the cells, only two showed altered concentra-

55 Figure 3.5: Absolute concentration of metabolites from TCA cycle in wt (black column) and ∆pyk (grey column). Data are shown as mean values ± SD of four biological replicates. Signiﬁcant alteration levels (p-values≤0.05) between strains are marked with asterisk.

tion levels when both strains were compared, as they were asparagine with a FC of 0.4 and citrulline, 2.7-fold higher in the mutant (ﬁgure 3.7).

3.1.3 Alteration in the metabolites levels of diﬀerent cell wall structures and possibly the membrane precursors

To annotate the possible separation between strains according to their metabolome, a PCA of the relative quantification of all metabolites was performed. In the PCA, the resulted score vectors can be thought as a good representations of the whole metabolite data set. It was assessed a cluster separation between strains, although with a total variance of 47.5% given by PC1 (figure 3.8 A). Furthermore, with the inspection of the impact of metabolites in the cluster separation, cell wall metabolites like CDP-glycerol andPG precursors were shown to be the main responsible compounds for this clustering distance (figure 3.8 B). CDP-glycerol, a constituent of TA, accumulated in ∆pyk, having a FC of 4.7. The same

56 Figure 3.6: Relative amount of PPP and shikimate phathway metabolites in LB medium in wt (black column) and ∆pyk (grey column). Data are shown as mean values ± SD of four biological replicates. Signiﬁcant alteration levels (p-values≤0.05) between strains are marked with asterisk.

happened to thePG precursors, UDP-MurNAc-ala and UDP-MurNAc-ala-glu which both had a FC of 10.7 (figure 3.9). Two otherPG precursors had also a strong impact in the cluster’s separation showing a decrease concentration levels in the mutant strain as they were UDP-GlcNAc (FC 0.22) and UDP-GlcNAc-enolpyruvate (FC 0.28). These results prompt to speculate thatPG synthesis and/or TA assembly could be deregulated in ∆pyk. Thus, cells were examined by Transmission Electron Microscopy (TEM) during the early exponential phase. As seen in figure 3.10, both strains had a smooth cell wall and presented the same thickness structure, showing no morphological differences. Moreover, no alterations in the septum formation was observed in this condition.

57 Figure 3.7: Heat-map of detected intracellular amino acids in wt and ∆pyk cultivated in LB medium. Color-code represent the log2 ratio between ∆pyk and wt samples, whereas increased levels are indicated in orange and lower levels in purple. Data are shown as the mean of four biological replicates.

The intracellular metabolome data were also explored in an untargeted metabolomic approach. For this, the data set obtained by LC-MS was analyzed in the XCMS on-line platform using METLIN, an open-access repository data base of metabolite information, that can assist in metabolite research and putative identiﬁcation [156, 157]. By comparing both strains, it resulted in 533 features with signiﬁcant altered FC. Although some of the features can be attributed to artifacts, several mass/charge ratio (m/z) are suggested to be fatty acid metabolites or adduct ions of the later, formed during LC-MS analysis, such as phosphatidic acid and phosphatidylglycerol, precursors of phospholipids and the LTA anchor, diglucosyldiacylglycerol, but also phosphatidylethanolamine, phosphatidylserine, and phosphatidylcoline, constituents of the cell membrane (table 3.1).

A B

Loadings 2 Loadings PC 2 (27.5%) 2 PC

PC 1 (47.5%) Loadings 1

Figure 3.8: PCA analysis and loading plot of the selected PC. Intracellular dataset of quadriplicate samples of wt (green) and ∆pyk (pink). The loading plot reveals that cell wall metabolites are the main responsible for the clustering distance observed. The percent- ages of total variance are 47.5% for PC1 and 27.5% for PC2. Single values are represented. UDPMurNalagl, UDP-MurNAc-ala-glu; UDPMurNala, UDP-MurNAc-ala; UDPGlucNepyr, UDP-GlcNAc-eneolpyruvate; UDPGlucN, UDP-GlcNAc.

59 Figure 3.9: Relative amount of intracellular CDP-glycerol and some PG precursors in wt (black) and ∆pyk (grey) in LB medium. Data are presented as mean values ± SD of four biological replicates. Statistical diﬀerences between wt and ∆pyk were considered signiﬁcant (∗) for p-values≤0.05.

Figure 3.10: TEM micrographs of B. subtilis wt and ∆pyk. Cells were cultivated in LB o medium at 37 C and collected for microscopy when OD600nm=0.5 was reached. Images show no apparent morphological diﬀerences on the cell wall.

60 Table 3.1: Some features resulted of XCMS untargeted metabolomics analysis. In the table are presented some of unknown m/z that are predicted to be fatty acid metabolites. FC correspond to the ratio of the mean of ∆pyk to the mean of wt samples.

m/z FC pvalue Possible metabolites Adduct Formula Unknown_1 592.402 0.196 0.028 Lysophosphatidylethanolamine M+Hac-H C27H52NO7P LysoPE(0:0/22:2(13Z,16Z)) Diacylglycerol M-H C37H72O5 DG(21:0/13:0/0:0) Unknown_2 595.674 11.496 0.000 Diacylglycerol M-H C37H72O5 DG(12:0/22:0/0:0) Branched fatty acid esters M+TFA-H C34H64O4 FAHFA(16:1(9Z)/9-O-18:0) Unknown_3 649.464 0.113 0.028 Diacylglycerol M+Cl C39H66O5 DG(14:0/22:5(7Z,10Z,13Z,16Z,19Z)/0:0) Phosphatidylethanolamine M+K-2H C41H78NO7P PE(18:1(11Z)/P-18:1(9Z)) Unknown_4 663.470 10.471 0.002 Diacylglycerol M+Cl C40H68O5 DG(15:0/22:5(7Z,10Z,13Z,16Z,19Z)/0:0) Unknown_5 668.477 29.422 0.008 Cardiolipin M-2H C72H140O17P2 CL(i-14:0/i-17:0/i-17:0/i-15:0) Unknown_6 675.481 17.402 0.027 Cardiolipin M-2H C73H142O17P2 CL(16:0/16:0/16:0/16:0) Phosphatidylethanolamine M-H C36H72NO8P PE(15:0/16:0) Unknown_7 676.485 63.336 0.039 Phosphatidylserine M-H C34H64NO10P PS(14:0/14:1(9Z) Phosphatidylcholine M-H C36H72NO8P PC(14:0/14:0) Diacylglycerol M+Na-2H C38H64O5 DG(11M3/9D3/0:0) Unknown_8 679.453 0.086 0.018 Phosphatidic acid M+FA-H C35H63O8P PA(10:0/21:0) Unknown_9 680.459 0.110 0.015 Phosphatidylethanolamine M+FA-H C33H66NO8P PE(14:0/14:0) Phosphatidylglycerol M-H C37H73O10P PG(a-13:0/i-18:0) Phosphatidic acid M+K-2H C37H67O8P PA(14:0/20:3(5Z,8Z,11Z)) Unknown_10 707.488 0.059 0.009 Phosphatidic acid M+FA-H C36H71O8P PA(15:0/18:0) Diacylglycerol M+Br C42H70O7 DG(15:0/22:5(7Z,10Z,13Z,16Z,19Z)/0:0) Phosphatidylethanolamine M+K-2H C37H70NO7P PE(14:1(9Z)/P-18:1(9Z)) Unknown_11 708.491 0.063 0.012 Phosphatidylcholine M+Na-2H C38H74NO7P PC(14:1(9Z)/P-16:0) CL(16:0/16:0/18:2(9Z,12Z)/ Unknown_12 723.480 0.042 0.007 Cardiolipin M-2H C81H142O17P2 22:6(4Z,7Z,10Z,13Z,16Z,19Z)) Diacylglycerol M+Na-2H C44H74O7 DG(11D3/11M5/0:0) Unknown_13 735.518 0.098 0.024 Diacylglycerol M+Na-2H C44H74O7 DG(11D5/11M3/0:0) Phosphatidic acid M+FA-H C38H75O8P PA(15:0/20:0) Phosphatidylethanolamine M+FA-H C37H74NO8P PE(16:0/16:0) Unknown_14 736.520 0.111 0.019 Phosphatidylcholine M+Na-2H C40H78NO7P PC(14:0/P-18:1(11Z)) Unknown_15 937.611 0.116 0.006 Phosphatidylglycerol M+TFA-H C46H81O10P PG(20:4(5Z,8Z,11Z,14Z)/20:1(11Z))

61 3.2 Discussion

3.2.1 Sugar consumption alterations

A complex medium like LB is widely applied in B. subtilis studies due to the fast growth and good growth yields. Nevertheless, their usage in carbon distribution and metabolomics investigation can be a complex task because of the diverse nutrient sources and their unknown concentrations [154]. The cultivation of B. subtilis lacking pyk in LB medium suggests that inactivation of Pyk had a moderate effect on the growth rate, as reported by Monahan and co-workers [65]. Still, the strains presented similar nutritional consumption behavior of the amino acids and sugars detected in the medium, showing no metabolic transport impairment in ∆pyk (figure 3.2). The intracellular metabolome data provided a complementary knowledge and insight of the metabolic responses to the perturbation occurred in the pyruvate node. In this condition, the detected glycolytic compounds were accumulated in ∆pyk being the main altered metabolites the PEP, 2-phosphoglycerate, 3-phosphoglycerate and dihydroxyacetone phosphate. Moreover, a significantly decrease of pyruvate was also observed in the mutant cells. Once again, the lack of Pyk created a bottleneck downstream of glycolysis, which led to the accumulation of these metabolites. Trehalose, one of the main constituents of LB medium, can be used as carbon source in glycolysis. The metabolic regulation of this metabolite in B. subtilis is made by the tre operon, which is subject of independent-CcpA CCR. Trehalose is translocated into the cells by the specific trehalose permease, EIItre, encoded by treP gene of tre operon, and is simultaneously phosphorylated to trehalose 6-phosphate. The latter can be further hydrolyzed to glucose 6-phosphate and glucose. While glucose 6-phosphate can be immediately used in pathways such as glycolysis, glucose is phosphorylated by glucose kinase. Since EIItre doesn’t possess the A domain necessary for the phosphoryl transfer of PEP via EI, the A domain of the EII of glucose from the PTS system, EIIglc, transfers the phosphate to EIItre complex, showing a synergistic cross talking between transport mechanisms [158]. Furthermore, treR, located downstream of the operon, encodes the repressor TreR. TreR is able to interact with tre, downregulating its transcription. Studies showed that trehalose 6-phosphate plays a transcriptional inducer effect since it inhibits tre-TreR interaction in a concentration-dependent manner, and, in turn, leads to the transcription of the tre operon [158, 159]. On the other hand, when adding high concentration levels of glucose 6-phosphate, it was reported an abolish of trehalose 6-phosphate effect [159]. In wt, at the time of intracellular sampling extraction (90 min of cultivation), half of the initial trehalose was already consumed (0.49 mM), while for ∆pyk cells, only 13% of the initial concentration was taken up from medium (0.3 mM). At the same time, even if present in small quantities, glucose was also being taken up from the medium by both strains. The

62 concomitant consumption of both sugar is in line with what previously described, since trehalose transport and phosphorylation is accompanied by the transfer of phosphoryl group from EIIglc from PTS. Additionally, in ∆pyk, intracellular glucose 6-phosphate was present in higher abundances than in wt, which can be hypothesized as one of the causes for the lower trehalose uptake in this strain. With the increase of glucose 6-phosphate, tre-TreR interaction may not be inhibited and, consequently, tre transcription is downregulated. During exponential growing phase in LB medium, enzymes of the lower part of glycolysis are usually highly expressed in bacteria [135]. Thus, despite the fact that the carbon ﬂux through glycolysis is probably low, at the time of intracellular exponential analysis, the metabolites of the lower part of the pathway were produced and accumulated inside the cells. The high discrepancy in pyruvate concentration amounts in this condition is also justiﬁed by the diminish of both mechanism that synthesize pyruvate from PEP: through glycolysis, since Pyk is absent, and via PTS system, since few glucose molecules are present as substrate.

3.2.2 Unexpected secretion of the same acetate levels

During the rapid growth of cells in LB medium condition, the enzymatic levels of the lower glycolytic pathway, as well as pyruvate dehydrogenase are known to be high, while in the TCA cycle enzymes they are usually low. This leads to potential bottleneck and a carbon flux to overflow metabolites like acetate production [135]. In line, acetate was secreted by wt during exponential phase. Unexpectedly, the acetate production, synthesized from acetyl-CoA, wasn’t affected by the lower pyruvate amounts in ∆pyk. This results remains unclear. It can only by speculated that carbon sources present in the medium, such as the ketogenic amino acids leucine, isoleucine and threonine, could have been degraded directly to acetyl-CoA. Also, this result is inconsistent to other studies on B. subtilis and E. coli cells lacking Pyk, which show a decrease in the formation of acetate [131, 123]. On the other hand, a study reported that although acetate accumulation was lower in ∆pyk than wt cells when cultivated in a glucose medium, the transcript levels of the pta and ackA genes were only 1.3- and 1.4-fold higher, respectively, suggesting the existence of an alternative transcriptional regulatory mechanism [123]. Nevertheless, acetate was further taken up similarly by both strains during cultivation, which may serve as alternative carbon source to fill up the central metabolic pathways. On the other hand, with the decrease of pyruvate in ∆pyk, it is expected a lower concentration of overflow metabolites. In fact, lactate, that is produced directly from pyruvate, was detected in lower amounts inside the mutant cells. On the other hand,

63 3.2.3 Carbon distribution in the TCA and PPP

The five metabolites detected from the TCA cycle were in lower concentration amounts in ∆pyk. This suggests that the bottleneck created in the pyruvate node, might have lead to a less active TCA due to the lower replenishment of the cycle through citrate synthase, as observed in chemically defined media (chapter2)[123, 141]. Contrarily, the elevation of several metabolites of the PPP suggest that this metabolic pathway is enhanced in ∆pyk cells. As previously described, the phenomenon was also referred when cells lacking Pyk were cultivated in minimal media. Similar responses was seen in other bacteria such as E. coli and Corynebacterium glutamicum [131, 160]. The glycolytic metabolite glucose 6-phosphate is possibly highly utilized in ∆pyk as precursor of PPP, causing a reroute of the metabolism to this pathway. As a result, cells are still able to supply NAD(P)H and nucleotidic resources. This may partially contribute to the loss of NADH generally produced when TCA is active in normal physiological conditions. Also, in a study using succinate and glutamate as nutrient sources, significantly less glucose 6-phosphate is converted to fructose 6-phosphate and byosynthetic metabolic purposes by B. subtilis, than when cultivated in a glucose medium. Rather, the glucose 6-phosphate oxidation is increased [134]. The high activity of PPP can also be the case for LB medium cultivation of normal condition cells, due to the availability of diverse amino acids [134]. In line with the finding of good activity rates of PPP is that both purine and pyrimidine metabolic pathways weren’t affected by the mutation and deregulation created, showing that no compromise in the production of most of the nucleotidic compounds.

3.2.4 Lower shikimate 3-phosphate level in ∆pyk

Remarkably, less shikimate 3-phosphate was produced by ∆pyk. As seen previously, the same happened when cells were cultivated in chemically deﬁned media (chapter2). The primary challenge for engineering the shikimate pathway is to improve the availability of the two pathway precursors PEP and erythrose 4-phosphate. Although erythrose 4-phosphate had similar amount levels, PEP accumulated in ∆pyk [150, 151]. Considering that PEP highly accumulated and that cells seem to reroute their carbon metabolism to the PPP, the resulted low concentration of shikimate 3-phosphate seem unexpected. Moreover, the aromatic amino acids resulted from shikimate pathway (i.e. phenylalanine and tyrosine) had similar abundances when comparing both strains, revealing possibly no impairment in the synthesis of these amino acids. The intracellular pool of amino acid marginally diﬀer between wt and ∆pyk. This suggests that cells are able to maintain a stable amino acid metabolite pool independently of the metabolic disturbance, which is in line with studies in B. subtilis and other bacteria

64 [81, 146, 141]. The BCAA valine and leucine, which are synthesized from pyruvate, didn’t present statistical diﬀerences between both strains. Moreover, the catabolic products of BCAA degradation, like BCAA keto acids isobutyrate, isovalerate, 2-ketoisovalerate and 2-methylbutyrate, were secreted similarly during late exponential and stationary phase, which evidences the use of BCAA for byosynthtetic purpose during this cultivation time. This also suggests that the supplement of amino acids from medium provided suﬃcient BCAA concentration sources.

3.2.5 Membrane and cell wall assembly altered

The cell envelope synthesis in bacteria such as in B. subtilis, is a highly regulated process. A striking result of this study was the dynamic changes of the cell wall metabolites, especially solublePG precursors, as well as WTA precursor CDP-glycerol. PG assembly was affected in diverse ways. While early precursors like UDP-GlcNAc (FC=0.21) and UPD-GlcNAc-enolpyruvate (FC=0.27) were significantly lower in ∆pyk, the later metabolites ofPG synthesis UDP-MurNAc-ala and UPD-MurNAc-ala-glu were detected in higher amounts (FC=10.7). The remarkable differences inPG amount metabolites pointed out for a possible alteration in PG assembly in consequence of the deregulation in the pyruvate node. If thePG alteration is related to the direct nutritional information via signaling pathways or the deregulation of other protein activities with a direct effect onPG synthesis it is yet not clear. CDP-glycerol, the substrate for WTA elongation, showed significantly increased concentration (FC 4.7) in ∆pyk. In line with this finding, the amount of dihydroxyacetone phosphate as a precursor of CDP-glycerol was also increased. As previously described (section 1.5.2.1),PG and WTA have a genome context interaction since they share the same anchor, undecaprenyl-phosphate. Hence, the shared lipid intermediate is proposed to influence the regulation of both biosynthetic pathways [99, 101]. It is reasonable to think that an alteration of one of these pathways, could influence the synthesis of the other, as reported by D’Elia and co-workers [99]. Also, if the increase of CDP-glycerol would correspond to increase of the WTA or longer polymers in ∆pyk, that would consequently alter the charge homeostasis on the cell surface. One speculation is that that may lead to alternative strategies on cell envelope pathways to overcome the increase of anionic polymers. The untargeted metabolic approach, revealed possible alterations in phosphatidylglycerol content, the major substrate of cellular fatty acid membrane, and the precursor of glycerol moieties of LTA. Phospholipids like phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine, constituents of the B. subtilis membrane, are also suggested to be in alter amounts in ∆pyk. Lipid composition of the membranes suffer alterations during the wide variety of devel-

65 opmental changes. That can occur during cell cycle processes like competence, division and sporulation, but also for adjusting to the environmental conditions. These syntheses adaptations are crucial for the insurance of the biophysical properties of the membrane [161, 162]. In B. subtilis, the production of phospholipids, are mostly done in the septal regions of the cells [163]. In the septum membrane, the main lipid constituents are phosphatidylethanolamine and phosphatidylcholine, since other phospholipids, such as phosphatidylglycerol and lysophosphatidylethanolamine, are driven to the lateral membrane [163]. As recently suggested by Monanah and co-workers, the lack of pyruvate inhibited the colocalization of E1α of pyruvate dehydrogenase during cell division. In this condition, Z-ring were positioned near the cell poles, more than one FtsZ assembly was present in each cell, and were observed minicells [65]. Thus, the alteration of Z-ring formation seen in ∆pyk, and consequent deregulation of septum formation, it could be a plausible reason for the alter content of membrane precursors in the mutant cells. No alterations in the cell envelope morphology was detected, as cell surfaces of both strains investigated by TEM microscopy were similar and no difference in the cell envelope thickness was observed, despite the metabolic differences detected in the study. Nevertheless, the results suggest that the absence of Pyk influences B. subtilis cell envelope metabolism. The data can provide additional information for further insights into bacterial physiology, particularly, the cell wall assembly and regulation.

3.3 Conclusion

Although the LB medium is a widely used and accepted medium in microbiological studies, its distinctive carbon and nitrogen sources, and also its diverse nutrient concentrations creates an extra complexity in metabolomics ﬁeld. Nevertheless, it was possible to characterize the metabolic proﬁles of B. subtilis wt and ∆pyk when cultivated in this medium. The analysis showed that the inactivation of pyk didn’t cause a reduction or increase in the consumption of amino acids present in the medium. Nevertheless, altogether, cells lacking Pyk revealed an altered distribution of intermediate metabolites of central pathways that could be resulted of a deregulation of the central metabolism, but also alterations speculated as strategies of the cells to overcome metabolic disturbance caused by the pyk deletion. The variations included the rise of phosphorylated metabolite pools the accumulation of metabolites of PPP, and decrease of TCA metabolites, similar to the results reported when cultivated in minimal media (chapter2). Also, ∆pyk secreted the same amount of acetate as the wt cells. This result is surprisingly since it is reported lower production of acetate by cells lacking Pyk in several organisms and

66 media conditions [124, 131, 123]. To comprehend this result, it would be advantageous to determine the concentration levels of other metabolites of the pyruvate node such as the intracellular amount of acetyl-CoA, and the overflow metabolites 2,3-butanediol and acetoin. Cells lacking Pyk also revealed lower amounts of shikimate 3-phosphate. Strategies developed for the increase of shikimate 3-phosphate production in biotechnology have been primarily focused in inhibiting the utilization of PEP in other pathways (such as the deletion of pyk) with concomitant inhibition of genes downstream of the shikimate pathway. On contrary to E. coli, of which the condensation of PEP and erythrose 4-phopshate is the rate-limiting reaction in the shikimate production and is subjected to extensive feedback regulations, shikimate dehydrogenase, that converts 3-dehydroshikimate into shikimate, was proposed to be the rate-limiting step for shikimate accumulation in B. subtilis [150]. This suggest that B. subtilis and E. coli have different responses to genetic manipulations of genes involved in the shikimate and other pathways - like the gene tkt from PPP and pyk. Moreover, the diminish of shikimate 3-phosphate in ∆pyk in this study conditions brings the awareness of the necessity of further investigation of the shikimate pathway regulation on B. subtilis, specially in the feedback regulation and enzymatic levels. It is known that cells present a transcriptional response consistent with nutritional, environmental challenges and gene modifications [123, 135]. Although an extensive transcriptional study was done by Cabreras-Valladares in cells lacking Pyk, it was only conducted in a chemically defined medium [123]. Thus, transcriptional analyses should be performed in a complex medium condition such as LB, to provide further insight into gene expression and metabolic responses in mutant B. subtilis with deletion of pyk and ultimately, to further comprehend the roles of Pyk. The results also suggest that cell envelope is altered in ∆pyk. The detection of altered amounts of thePG precursors and CDP-glycerol, points out for a possible alteration in thePG assembly and WTA. Moreover, the complementary of the untargeted with the targeted metabolomics approaches revealed very helpful in the analysis and detection of other metabolites with altered levels in mutant cells. The majority of the identified features in the untargeted approach are suggested to be fatty acids metabolites or precursors of the cell membrane. It can be speculated that one of the possible reasons for the altered lipidic content in cells lacking Pyk is the deregulation of the septum formation observed previously in ∆pyk cells, since the chemical constitution of the septum membrane is distinct. As recently observed, the loss of Pyk and decrease of pyruvate levels interferes with the normal function of the FtsZ and cell cycle processes [65]. Ultimately, this connection revelead that nutritional information can be transmitted directly from metabolic pathways to the cell cycle machinery, serving as a mechanism for fine-tuning cell cycle processes in response to changes in environmental conditions. During the untargeted approach, a diverse of unknown features with significantly altered amounts were detected. Despite the availability of more and diverse metabolic databases,

67 the identification of these is still a challenging task, and ultimately,a promising effort in the field of metabolomics. The metabolic approach of the study of B. subtilis lacking Pyk provides information that should be useful in future efforts that aim to understand if pyruvate is a marker nutritional status and if it helps in the coordination bacterial division.

68 Chapter 4

Study of lipoteichoic mutants in a complex medium

4.1 Results

4.1.1 B. subtilis 168 and BSB1 - a similar metabolome proﬁle between both strains

In accordance to the Advance Multidisciplinary training in Molecular Bacteriology (AMBER) consortium which this work is integrated, both strains 168 and BSB1 were investigated. Thus, using the extra and intrametabolome data when cells were cultivated in LB medium, metabolic proﬁles of the lipoteichoic mutants ∆pgcA, ∆gtaB, and ∆ugtP of B. subtilis 168 and BSB1 strains were inspected for a global perspective. In all conditions, cell growth were reproducible reaching a stationary phase after 6h of incubation (ﬁgure 4.1). Additionally, either 168 and BSB1 strain, ∆pgcA showed the highestEC, with 0.90 and 0.88, respectively (4.1). With 1H-NMR spectroscopic analysis, it was possible to monitor changes in the concentra-

Table 4.1:EC of B. subtilis in wt and mutants of 168 and BSB1 determined for each biological sample using absolute concentrations of intracellular AMP, ADP and ATP. Displayed are the mean values ± SD of three biological replicates.

Energy charge wt ∆pgcA ∆gtaB ∆ugtP 168 0.76 ± 0.09 0.90 ± 0.04 0.73 ± 0.04 0.72 ± 0.05 BSB1 0.75 ± 0.04 0.88 ± 0.04 0.83 ± 0.05 0.74 ± 0.05 tion of extracellular metabolites in a time dependent manner.

69 Figure 4.1: Time-resolved extracellular metabolites under LB medium cultivation. Abso- lute concentrations of consumed and secreted metabolites by B. subtilis BSB1 wt (black), ∆pgcA (green), ∆gtaB (blue), and ∆ugtP (red) are desplayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

In ﬁgure 4.2, the metabolite concentrations at each time point are displayed as a color coded heat map (as red for high and green for low concentration levels). It is possible to inspect that there aren’t considerable extrametabolome diﬀerences in the consumption and secretion of metabolites, when 168 and BSB1 are compared. Moreover, the glucose and trehalose sugars present in the medium were rapidly taken up

70 (a) 168

(b) BSB1

Figure 4.2: Heatmap visualization of extracellular metabolites of B. subtilis wt, ∆gtaB, ∆ugtP and ∆pgcA in 168 (a) and BSB1 (b) in LB medium. FC represented as log2 ratio of a metabolite concentration at each time point and the average concentration of the respective metabolite at all time points. Green: lower concentration limit and red: higher concentration limit. during exponential phase by all strains. Also, a fast consumption of amino acids such as arginine, aspartate, and alanine, was observed at 180 minutes (ﬁgures 4.3 and 7.7). A similar amino acid preference has been already described in a chemical deﬁned medium containing several amino acids [164] . Interestingly, a slower consumption rate of arginine was observed

71 for 168 ∆ugtP. It was also detected a low rate consumption of lysine and aromatic amino acids like pheny-

Figure 4.3: Time-resolved extracellular metabolites under LB medium cultivation. Abso- lute concentrations of consumed and secreted metabolites by B. subtilis BSB1 wt (black), ∆pgcA (green), ∆gtaB (blue), and ∆ugtP (red) are desplayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

72 lalanine, tryptophan and tyrosine during all time of cultivation and in all strains (table 7.6). The intracellular metabolome profiles of wt and mutants of 168 and BSB1 were also evalu- ated. At OD=0.5, cells were harvested in biological triplicates. By combining LC-MS and GC-MS techniques, 120 metabolites were identified from several metabolic pathways such as glycolysis/gluconeogenesis, TCA, PPP, nucleotide metabolism, cell wall synthesis and amino acid metabolism, providing a global view of the metabolic fingerprint of B. subtilis. The intracellular metabolite data were subject to PCA that revealed no clear separation was observed between 168 and BSB1 strains (figure 4.4).This metabolic similarity is not

Figure 4.4: PCA analysis of intrametabolome data of B. subtilis control and mutants strains from 168 and BSB1 in LB medium. The colors grey (wt), pink (∆ugtP), dark blue (∆pgcA), and red (∆gatB) correspond to 168 samples while the colors black (wt), yellow (∆ugtP), light blue (∆pgcA), and green (∆gatB) correspond to BSB1 samples. Single values of three biological replicates are represented. surprising since BSB1 is a tryptophan prototrophic variant of 168 transformed with DNA from strain W23 [55]. Noticeably, there is a cluster separation of samples according to the genotype. This separation has a statistically low variance (PC1 of 30.7%). Nevertheless, is also interesting to observe a cluster proximity between wt and ∆ugtP and in less extent ∆gtaB, and ∆pgcA for both strains.

Since there is no evidence for obvious metabolic diﬀerences between 168 and BSB1, one

73 of the strains was selected for further experiments. In agreement with the AMBER consortium, BSB1 was the chosen strain to continue the metabolic studies. The intracellular metabolome proﬁles of ∆pgcA, ∆gtaB, and ∆ugtP in BSB1 were studied in detail.

4.1.2 Consumption and secretion of metabolites: higher secretion of acetate by ∆pgcA

Altogether, 22 different metabolites were identified and quantified from the extracellular medium (figure 4.3). Although glucose was present in low concentrations (0.5 mM in the beginning of cultivation), the results showed that for all strains, this metabolite was initially assimilated in a low rate followed by a concentration drop from 60 to 120/180 minutes. Trehalose and succinate consumption started at 180 minutes. In line with previous studies done in B. subtilis, the main metabolite secreted was acetate. It was secreted during all exponential phase reaching its maximum when entering the stationary phase. Afterwards, acetate was completely taken up from medium being used as alternative carbon source (figure 4.3). Notably, in ∆pgcA, acetate maximum concentration was higher than in wt and other mutants, and this peak was observed with one hour of delay (table 4.2). Other common overflow products secreted by B. subtilis like 2,3-butanediol, acetoin or

Table 4.2: Extracellular concentrations of acetate, isobutyrate and histidine in B. subtilis BSB1 wt, ∆pgcA, ∆gtaB, and ∆ugtP. Values are shown as mean concentrations (mM) ± SD of three biological replicates. For acetate and isobutyrate, the concentration correspond to their maximum of secretion, while histidine concentration correspond to the amount still present in the end of the cultivation (540min).

metabolite (mM) wt ∆pgcA ∆gtaB ∆ugtP acetate 6.64 ± 0.09 7.83 ± 0.11 5.86 ± 0.32 6.08 ± 0.32 isobutyrate 2.04 ± 0.08 1.03 ± 0.02 2.11 ± 0.06 1.63 ± 0.21 histidine 0.55 ± 0.00 0.73 ± 0.00 0.50 ± 0.04 0.83 ± 0.03 lactate, weren’t detected in this study. Metabolites of BCAA pathway, including isobutyrate, 2-methylbutyrate, 2-ketoisovalerate, and isovalerate were detected outside the cells. Isobutyrate was the metabolite with the biggest concentration diﬀerence between strains (table 4.1). The secretion was initiated at 240 min of growth, when cells entered stationary phase, and was continuously accumulated in the medium until the end of cultivation. Wt

74 and ∆gtaB reached the highest concentration secretion of isobutyrate (2.04 ± 0.08 and 2.1 ± 0.06 mM, respectively). Remarkably, ∆pgcA showed significantly less secreted isobutyrate (50% lower) after 540 min of cultivation (p-value≤0.001). A metabolite with different consumption profile between strains was histidine. While in wt and ∆gtaB, histidine was similarly taken up reaching 0.5-0.6 mM at the end of cultivation, in ∆pgcA and ∆ugtP, histidine was consumed in slower rates (table 4.2).

4.1.3 Mutations in the lipoteichoic acid alter cell wall metabolites and possibly the peptidoglycan assembly

In order to evaluate the metabolic changes of ∆pgcA, ∆gtaB, and ∆ugtP when cultivated in LB medium, the relative amount of all identified intracellular metabolites were analyzed. Several metabolites exhibit altered concentration levels when wt and mutant strains were compared. By plotting in a volcano-type plot the log2FC as function of the unpaired t-test results of each mutant, it is visible that many metabolites aren’t affected by the respective genotype (figure 4.5). The majority of the significant altered metabolites have increased concentration levels in the mutants than in wt. Moreover, ∆pgcA showed higher number of altered metabolites concentration than ∆gtaB and ∆ugtP. A statistical PCA analysis was also performed for the four metabolomic data set groups to annotate their separation. By inspecting the PCA projection, it is possible to observe a clear cluster separation for wt and ∆pgcA sample groups with a total variance of 59.1%, represented by principal component (PC)1 (figure 4.11). It also revealed a bigger distance separation between wt and ∆pgcA and less extent ∆gtaB, whilst ∆ugtP exhibit a bigger proximity to wt. The loading plot result shows that wt and ∆ugtP proximity is strongly associated with UDP-glucose and UDP-glucuronate levels (figure 4.11). This is in agreement with the relative quantification of metabolites: UDP-glucose was absent in ∆pgcA and ∆gtaB, reflecting a low FC(FC <0.001) for both mutants, while ∆ugtP showed the same amount of this metabolite as wt (FC=0.99, figure 4.12). UDP-glucuronate was also absent in ∆pgcA and ∆gtaB (FC<0.001 for both mutants), while in ∆ugtP presented a lower concentration level when compared to wt (FC=0.48 and p-value=0.01).

A striking result in the study of the intracellular metabolism of LTA mutants was also the accumulation of severalPG precursors like as NAcGlucosamine-P, UDP-GlcNAc, UDP- MurNAc-ala, UDP-MurNAc-ala-glu and UDP-MurNAc-pentapeptide in ∆pgcA, ∆gtaB and ∆ugtP (ﬁgure 4.6 and table 7.7). The predominant diﬀerences were observed in later steps ofPG assembly whereby UDP-MurNAc-ala with a FC=2.5 in ∆pgcA, FC=3 in ∆gtaB, and FC=5.5 in ∆ugtP comparing to the wt, and UDP-MurNAc-ala-glu that doubled the amount

75 ΔpgcA ΔgtaB

ΔugtP

Figure 4.5: Volcano plots of intracellular metabolites of three biological replicates. Metabo- lites with p-value≤0.05 and -1>log2FC>1 are colored in green (∆pgcA), blue (∆gtaB) or red (∆ugtP).

in all mutants. These changes set the hypothesis thatPG synthesis could be upregulated. In this way, it was speculated that cells may be synthesizing higher amounts ofPG. The physiochemical alterations provoked in glycosylation lacking LTA anchor mutants (like potential alteration of cell wall charges) could lead to the production of a thicker cell wall. To check this hypothesis, cell wall of wt and ∆ugtP were labelled with NBD-amino-d- alanine (NADA) and analyzed using fluorescent microscopy. NADA is a fluorescent amino acid that emits in a green wavelength, and mimics D-alanine present in the pentapeptide ofPG with minimal perturbation to the cell. Since DD-carboxypeptidase (PBP5) removes terminal D-alanine for cross-linking, dacA, the gene that encodes PBP5, was deleted in both strains (BSB1∆dacA and ∆ugtP∆dacA) allowing the accumulation of more NADA molecules. The first attempt in observing the possible differences inPG thickness was to analyze the incorporation of NADA at different time points of incubation: 0, 15, 30, and 60

76 Figure 4.6: Relative intracellular concentrations of peptidoglycan precursors in wt, ∆pgcA, ∆gtaB and ∆ugtP in LB medium. Data are presented as mean values ± SD of biological three biological replicates. Statistical diﬀerences between control and mutants were considered signiﬁcant (∗) for p-values≤0.05.

minutes (figure 4.7). Strains were grown in LB medium containing NADA (material and methods section for more detail). After 15 minutes of incubation, BSB1∆dacA and ∆ugtP∆dacA were fully labelled. SincePG turnover occurs at a very high rate this time-course microscopy failed to discriminate any differences between both cells. Since both strains showed a complete fluorescence throw out the experiment, another approach was done where, instead of analyzing the incorporation of labelled amino acids, it was monitored their release. BSB1∆dacA and ∆ugtP∆dacA were incubated together in LB medium containing NADA for full labeling ofPG. Moreover, to distinguish ∆ugtP∆dacA from the parental strain, BSB1 was also labelled with a membranar fluorescent protein (m.cherry protein with a rede wavelength emission). Afterwards, cells were transferred to a new NADA-free medium andPG assembly observation was done by monitoring the decreased of NADA fluorescent during time (figure 4.8). In the beginning of the experiment,

77 Figure 4.7: Time-resolved of NADA incorporation in peptidoglycan (0, 25, and 80 min after incubation in LB medium containing NADA. Both BSB1∆dacA and ∆ugtP∆dacA have a very high cell wall turnover showing already a fully incorporation of NADA after 15 minutes.

∆ugtP∆dacA appeared with a higher intensity, however this effect seemed false fluorescence intensity since UgtP mutant was usually in a bulk/stack disposition. This was confirmed in the later time-points, where both strains presented the same fluorescence intensity in all cell collection times. A second approach was done to analyze a possible alteration in cell wall structures. Wt, ∆pgcA, and ∆ugtP were examined by TEM during early exponential phase. Although no obvious increase ofPG thickness was observed, TEM images showed distinct integrity of PG. While wt presented an even and smooth cell wall, ∆pgcA and ∆ugtP showed a fuzzier appearance and alteration inPG structure along the cell periphery (figure 4.9). Another cell wall modification identified was the accumulation of CDP-glycerol, especially in ∆pgcA, the main percursor of WTA. Taken together, these results ofPG and WTA precursor concentrations show that the alteration of LTA synthesis influenced other cell wall metabolic pathways in a key manner.

4.1.4 Cells lacking PgcA showed expressive alterations in the central carbon metabolism

The lack of PgcA had a strong impact on the central carbon metabolism, especially in the glycolysis pathway. Except for pyruvate, all glycolytic intermediates determined, such as glu-

78 Red channel Green channel Red channel 0 min 0 min

BSB1

ΔugtP

25 min

BSB1

ΔugtP

Red channel

80 min 80 min 100 min min

BSB1 ΔugtP

Figure 4.8: Time-lapse microscopy of BSB1∆dacA and ∆ugtP∆dacA incubated in LB medium. Images captured at red (left images) and green wavelength emission (right images) at 0, 25 and 80 minutes incubated in a fresh LB medium without NADA molecules.

79 A

Figure 4.9: TEM micrographs of B. subtilis wt (A), ∆pgcA (B) and ∆ugtP (C). Cells were o cultivated in LB medium at 37 C and collected for microscopy when OD600nm=0.5 was reached. Images show distinct integrity of the cell wall in ∆pgcA and ∆ugtP.

80 cose 6-phosphate, fructose 6-phosphate, FBP, dihydroxyacetone-phosphate, 3-phosphoglycerate, 2-phosphoglycerate and PEP, were significantly increased (p-values≤0.05, figure 4.10). For ∆gtaB and ∆ugtP, these metabolites were also found in higher amounts but without statistical relevance. Also, when performing a PCA of the intracellular data of the wt and mutants, a cluster distance was observed for wt and ∆pgcA, which was mainly consequence of the different amounts of glucose 6-phosphate and fructose 6-phosphate levels (figure 4.11). Glucose 6- phosphate and fructose 6-phosphate ere determined with a FC of 1795 and 796, respectively (figure 4.12). In ∆gtaB and ∆ugtP, these metabolites were also present in a higher level but with no statistical significance. Enhanced amounts of acetyl-CoA was also observed for ∆pgcA. The fatty acid precursor malonyl-CoA, synthesized from the later, accumulated in all the three mutants with a FC of 4.7, 3.0 and 2.1 for ∆pgcA, ∆gtaB and ∆ugtP, respectively.

81 Glyceraldehyde-3-phosphate

1,3-bisphosphoglycerate

Figure 4.10: Relative intracellular concentrations of some glycolytic metabolites, acetyl-CoA and malonyl-CoA in wt, ∆pgcA, ∆gtaB and ∆ugtP in LB medium. Data are presented as relative amount ± SD of three biological replicates. Statistical diﬀerences between control and mutants were considered signiﬁcant (∗) for p-values≤0.05.

82 uat.PAaayi lf mg)adlaigpo rgtiae fitaeaooedata intrametabolome of image) (right plot loading and image) of (left analysis PCA mutants. rvae igrcutrn itnebtenw n ∆pgcA. and 6- group wt the glucose between for observed. metabolites 6-phosphate, distance responsible distance fructose the clustering clustering as UDP-glucuronate, bigger L-glutamine and UDP-glucose, a 2-phosphoglycerate shows revealed phosphate, plot PCA loading The The medium. LB in of responses metabolic of analysis Multivariable 4.11: Figure ilgclrpiae.Saitcldffrne ewe tadec uatwr considered were values mutant mean each as and represented wt, wt Data between in differences ( UDP-glucuronate medium. significant Statistical and LB UDP-glucose in replicates. intracellular ∆ugtP biological of and ∆gtaB amount ∆pgcA, Relative 4.12: Figure .subtilis B. ∗ o p-values for ) S1w lgtbu) pc gen,∆tB(e)ad∆gP(akblue) (dark ∆ugtP and (red) ∆gtaB (green), ∆pgcA blue), (light wt BSB1 ≤ 0.05. 83

Loadings 2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 UDPGlucose UDPGlcA -0.4 .subtilis B.

-0.2 Loadings 1 Loadings CDP-Glucose IMP dGMP AMP dTDP UDP dADP CDP GDP UMP Pyruvate dCDP XMP ADP dGDP FAD UDPGlcNAc-en CoA CMP D-glucose GAR Asparagine Adenosine UDPMurNAc-al Trehalose6P Hypoxanthine Gluconate 0.0 L-Isoleucine Uridine Deoxycytidin CTP Uracil Erythrose4P Adenylsuccin L-Proline Deoxythymidi GluconateP FADH2 Deoxyuridine Sedoheptulos Succinate PRA cCMP L-Lysine Guanosine FGAR L-Tyrosine 5-Oxoproline Pseudouridin PRPP cUMP Citrate L-Threonine FGAM GMP dCMP L-Methionine SuccinylCoA 5-methylurid Serine 2-Oxoglutari NADH cGMP Glycine L-Leucine Acetyl adeny Acetyl CAIR cdiAMP Glyceric aci Glyceric cdiGMP UDPGlcNAc Lactate Guanine cAMP AICAR NADPH dUTP SAICAR NAD dATP DL-Ornithine L-Cysteine NADP UDPMurNAc-al dTTP dCTP Cytidine UDPMurNAc NAcGlucosami UDPMurNAc-5a Myo-inositol Aspartate L-Valine GTP UTP ITP ATP XTP Malate L-Glutamate Acetyl-CoA Fumarate UDPMurNAc-al Adenine D-Ala-D-Ala 3-Pglycerate tadlipoteichoic and wt Frutose16bis CDP-Glycerol MalonylCoA Hydroxybutyr L-Histidine D-Alanine PEP

0.2 DHAP 2-Pglycerate L-Glutamine ± Glucose6P o 3 of SD 0.4 Fructose6P 4.1.5 Accumulation of the amino acids glutamine and histidine

Surprisingly, the mutation on the LTA had a strong impact in the amino acids glutamine and histidine. The intracellular amounts of both amino acids were strongly accumulated in the exponential growing phase. For glutamine, ∆pgcA was 22.8-fold higher than wt, followed by ∆gtaB with a FC of 18.7 and 8.7 for ∆ugtP (ﬁgure 4.13). As far as histidine, the concentrations were also signiﬁcantly higher in the mutants, with a FC of 8.7, 4.5 and 4.9 for ∆pgcA, ∆gtaB and ∆ugtP, respectively.

Figure 4.13: Relative intracellular concentrations of of glutamine and histidine in wt, ∆pgcA, ∆gtaB and ∆ugtP in LB medium. Data are presented as mean values ± SD of three biological replicates. Statistical diﬀerences between control and mutants were considered signiﬁcant (∗) for p-values≤0.05.

4.2 Discussion

4.2.1 The partition of glucose 6-phosphate in ∆pgcA altered glycolysis pathway

During growth in rich medium like LB, cells can catabolize macromolecules required for their proliferation, instead of diverse the ﬂux towards the synthesis of precursor molecules, such as amino acids, since they are present in the medium [135]. Although it’s concluded that neither glycolysis nor gluconeogenesis are very active in LB medium, a small concentration of glucose was still present in the medium (0.5 mM). Thus, glucose was consumed via glycolysis in the beginning of cultivation, even though it is repleted of other nutrient sources like the amino acids [39, 165]. A reasonable result in the lack of PgcA is the accumulation of glucose 6-phosphate inside the cells since this enzyme converts glucose 6-phosphate in glucose 1-phosphate. Glucose 6-phosphate is an intracellular metabolite that is simultaneously metabolized by more than

84 one enzyme in important metabolic pathways. Under unperturbed growth conditions and in minimal media containing glucose, one-third of glucose 6-phosphate is catabolized to ribulose 5-phosphate via PPP, just around 2.5% is used for biomass synthesis such as cell wall, and the rest is consumed via glycolysis [134, 166]. In the ∆pgcA results, glucose 6-phosphate, fructose 6-phosphate and FBP were present in higher levels when compared to the wt. The increased amount of glucose 6-phosphate is probably being synthesized to fructose 6-phosphate and FBP in a higher rate via glycolysis, since there is also an accumulation of the later inside the cells. This metabolic reroute change is quite striking since a metabolic flux study of B. subtilis knockout mutants revealed its general network rigidity and metabolic robustness at the glucose 6-phosphate branch point [166]. Also, since just a small amount of glucose 6-phosphate is thought to be derived for cell wall synthesis, the accumulation of this metabolite and the possible change is glycolysis flux is an outstanding result. Moreover, whilst glucose 6-phosphate can also be utilized in the PPP, the metabolites identified from pentose shunt such as glucono 1,5-lactone 6-phosphate, gluconate 6-phosphate, erythrose 4-phosphate, and sedoheptulose 7-phosphate showed similar levels in both strains (table 7.7). Thus, PPP seemed to be unaltered.

4.2.2 Increased malonyl-CoA levels

Another important issue to consider when growing cells in LB is that, during exponential phase, the free amino acids present in the medium additionally contribute to the overload on the lower glycolytic pathway and the pyruvate node. Thus, the highly active lower glycolytic part and pyruvate dehyxdrogenase, and the fact that TCA intermediates are presumably down-regulated in this medium condition, creates a bottleneck in the junction of acetyl-CoA, leading to a carbon flux to overflow metabolites production [144, 167]. Previous studies using E. coli showed that enzymes catalyzing the production of acetate from acetyl-CoA, encoded by ptA and ack, are present at elevated levels in cells growing exponentially in rich medium [135]. This is in accordance to the high secretion of acetate seen in wt and ∆pgcA strains (figure 4.14). This metabolite, considered as an overflow metabolite when cells are in a surplus of carbon and energy environment, was secreted during all exponential phase reaching its maximum when entering the stationary phase. The secretion of acetate was aggravated in ∆pgcA through the accumulation of the metabolites from glycolysis. Interestingly, these concentrations are considered very high if compared to the initial amount of sugars, which reveals that free amino acids, are being converted to pyruvate, and, thus, to acetate and ATP. Afterwards, when glucose and trehalose sugars were completely exhausted, acetate started to be consumed, indicating that acetate was used as alternative carbon source after depletion of the sugars. Since the accumulation of acetate changes the pH medium, this uptake

85 Figure 4.14: Schematic illustration of extracellular acetate and relevant intracellular metabolites in wt and ∆pgcA. Data are presented as relative amount ± SD of three biological replicates.

could also be a realcalization strategy of the cells. Other perturbations were observed in the acetyl-CoA node such as the significant higher concentration levels of malonyl-CoA (figure 4.14). Malonyl-CoA is synthesized by acetyl- CoA carboxylase and is the first step in fatty acid synthesis, more specifically the straight fatty acids. The increased malonyl-CoA concentration perceived in ∆pgcA could reflect alterations of fatty acids synthesis since in B. subtilis, this is the only known cellular process that utilizes malonyl-CoA [168]. The fatty acids play an essential role in B. subtilis as components of the cellular membrane and source of metabolic energy [169]. The precise transcriptional regulation of homeostatic mechanisms maintaining the concentration of fatty acids at particular levels and the lipidic

86 synthesis are still not completly understood [169, 170, 171]. The synthesis and elongation of fatty acids involves repeated cycles of condensation, reduction, dehydration of carbon–carbon linkages catalyzed by the type II fatty acid synthase (FASII cycle). In E. coli, it has been postulated that acetyl-CoA carboxylase is the essential rate-controlling step in fatty acid synthesis [172]. Malonyl-CoA concentration is tightly regulated to be very low, so as to coordinate the rate of fatty acid synthesis with phospholipid production, macro- molecule synthesis, and cell growth [171, 172]. Furthermore, it was reported that in gram-positive bacteria, besides being the intermediate of fatty acids, malonyl-CoA is a signaling molecule to FapR, a global transcriptional repressor, that controls the expression of almost an the entire set of genes involved in the FASII cycle [162, 171]. Studies reported that, when fatty acids synthesis is inhibited, B. subtilis is able to detect the decrease of the activity of FASII, and increases the transcription of fap regulon, the genes involved in this system. Moreover, the inhibition of the FASII cycle augmented the intracellular levels of malonyl-CoA, which in turn, is thought to be the key metabolite for the release of the FapR regulator from its binding sites, increasing the expression of the fap regulon [170, 172, 173]. Thus, it is speculated that a transient increase of malonyl-CoA levels could be consequence of altered rate of fatty acid synthesis in ∆pgcA. Since the fatty acid production is tightly regulated according to cell’s need and also that the membrane’s composition is modiﬁed in response to environmental changes, the perturbation of the LTA structures in the mutants, could may lead to the alteration of membrane’s properties, such as the permeability and viscosity, so that a cell envelope homeostasis and integrety could be achieved. [169, 170, 171].

4.2.3 Cell wall alterations

As described, B. subtilis with mutations in glycosylation pathway of LTA presents shorter cells mainly due to the alteration in UDP-glucose levels and UgtP and consequently perturbations in Z-ring dynamics. Whilst it’s known that LTA anchor mutants are still able to form LTA with a modified anchor composition and display numerous prototypical alterations whereby cell morphology, other physiochemical effects are still not well understood. UDP-glucose was absent in ∆pgcA and ∆gtaB, reflected by the low FC<0.001 for both mutants, while ∆ugtP showed the same amount of this metabolite. This result is not surprising since PgcA and GtaB proteins are involved in the synthesis of UDP-glucose from glucose derivatives (glucose 1-phosphate and glucose 6-phosphate). UDP-glucuronate was also absent in ∆pgcA and ∆gtaB (FC<0.001), while in ∆ugtP presented a lower concentration level when compared to wt (FC=0.48 and p-value=0.01). In fact, the only known reaction for the synthesis of UDP-glucuronate is via UDP-glucose catalyzed by UDP-glucose 6-dehydrogenase.

87 UDP-glucuronate is a precursor of teichuronic acid, a cell wall polymer attached toPG that is present in B. subtilis under conditions of phosphate-limitation. When B. subtilis transits from phosphate-repleted to phosphate-limited growth, teichoic acids are degrade and recov- ered for teichuronic acid synthesis. This way, cells can reduce the requirement of phosphate and are able to re-utilize the released phosphate for other cellular processes such as nucleic acid synthesis [96, 174]. This result suggests that B. subtilis don’t possess a backup mechanism for the synthesis of UDP-glucuronate and an existence of a pool of this metabolite even in conditions of phosphate availability. The cytosolic pool of UDP-glucuronate is considerably lower under phosphate-replete conditions compared with phosphate starvation, but the purpose of maintaining this pool is still not clear [175]. The altered amounts ofPG precursors, and possiblyPG structure, suggests thatPG hydrolases activity may be higher or uncontrolled. This observation is in accordance with previous studies made in S. aureus cells lacking YpfP, the homologous of UgtP, which result in increased autolytic activity [120, 176]. Recently, TA have been implicated in the expression, localization and activity of DL-endopeptidase lytE (LytE) which, along with CwOl, are hydrolases localized in the periphery of cell and responsible for lateralPG hydrolysis during cell elongation. In the absence of LTA or anchor mutants such as ∆pgcA, ∆gtaB and ∆ugtP, LytE activity is significantly increased [177]. Analysis done by collaborators in double mutant cells ∆ugtP∆lytE revealed severe growth and shape defects such as shorter and twisted morphology and a high number of mini-cells [119]. Moreover, cells lacking UgtP and PBP1, a protein with transglycosylase and transpep- tidase activities in thePG assemby, encoded by ponA gene, showed a lower thickness of cell wall than observed in single mutants ∆ugtP and ∆ponA, which later on caused lethality. Interestingly, the triple mutant ∆ugtP∆lytE∆ponA, exhibited normal rod shape cells and a moderate chaining morphology [119]. The same results were reported for ∆ugtP∆lytE cells lacking SigM, the later to be responsible for upregulation of proteins expression involved in cell wall synthesis and division during stress conditions [178]. In the absence of PBP1 synthase in ∆ugtP∆lytE cells partially recover their normal morphology. These data suggest that cells balancePG synthesis and hydrolysis in cell lacking glucolipid LTA anchor. Taking these together, the high levels ofPG precursors are in agreement with these observa- tions. In ∆ugtP, when hydrolase activity is high, cells might also increasePG synthesis to avoid a weak point in the the sacculus that would eventually leas to burst by turgor pressure. Thereby,PG synthesis rate is still high although with no difference inPG thickness. This can be confirmed by TEM analysis and metabolomics data that of ∆pgcA and ∆ugtP where cells present irregular cell wall structure and accumulatePG precursors.

88 4.2.4 Glutamine and histidine regulation

The catabolism of the oligopeptides is the main nutritional source in LB medium. Although neither glutamate and glutamine were detected in the extracellular medium, it was previously determined that LB medium is repleted of both amino acids [154]. Glutamate is the principal nitrogen donor for most amino acids and nucleotides biosyn- theses, and the most abundant metabolite in the cytoplasm, accounting for 40% of the total intracellular metabolome measured [179]. Therefore, bacteria attempt to maintain the glutamate pool in homeostatic and high levels to drive the reactions forward [146, 180]. Also, glutamate is a very important metabolic intersection between carbon and nitrogen metabolism since its regulation is directly related to the inter-conversion of glutamate and 2-oxoglutarate from the TCA cycle [136]. In this way, the synthesis and the degradation of glutamate are tightly regulated by signals derived from carbon and nitrogen metabolism to ensure a sufficient supply of this amino acid to the cells [179, 181]. An outcome of this study was the accumulation of glutamine in the mutants cells, which can rapidly fill up the intracellular glutamate pool. All mutants presented increased amounts of glutamine, being the highest accumulation perceived in ∆pgcA, followed by ∆gtaB and ∆ugtP. The control of glutamine and glutamate metabolism is a complex mechanism regulated by several global proteins. Glutamine is an optimal source of nitrogen and it can be converted to glutamate by glutamate synthase glutamate synthase (GOGAT). Also, glutamine synthetase (GS) catalyses the synthesis of glutamine from ammonium and glutamate, being the only pathway for glutamine synthesis in the cell. Their activity is tightly regulated in response to nitrogen availability. In B. subtilis, the synthesis ofGS is mediated by the repressor, GlnR. When preferred nitrogen sources are present,GS binds to GlnR and represses the expression of the operon glnRA encoding forGS and GlnR. Furthermore, GlnR also traps and inhibits another regulatory protein, TnrA which, when active, improves the nitrogen availability inside the cells. In conditions of nitrogen limitation, TnrA activates the expression of ammonium transporters and weakly represses glnRA. This allows cells to be flexible to environmental changes and enhances their viability [182, 181]. Furthermore, during the synthesis ofPG precursors, two enzymes require glutamine as the ammonium donor. First, glutamine-fructose 6-phosphate aminotransferase (GlmS), that catalizes the conversion of fructose 6-phosphate and glucosamine 6-phosphate; and a putative enzyme, that catalyzes the glutamate of thePG monomer precursor.GS supplies glutamine by assimilating ammonia into glutamate [183]. A correlation between glutamine amount and the cell wall thickness in S. aureus was reported by Cui and co-workers [183]. In a strain with a thicker cell wall, the consumption of labelled glucose was proven to be bigger, with a concomitant increase of GlmS activity. Moreover, thePG of this strain is characteristically higher in proportion of nonamidated muropeptides at the glutamate residue. It is suggested that, with the overactivity of GlmS in

89 disproportion to the supply of glutamine, the intracellular glutamine pool likely falls. Thus, with the a low glutamine pool, muropeptides are noamidated. When cells were supplied with excess of glutamine, the proportion of murein nonamidated substantially decreased. There is no clear evidence why the LTA mutants have an accumulation of glutamine inside the cells. Also, since it wasn’t possible to have an absorption and secretion profile of both amino acids, it is not possible to know if the higher glutamine pool levels are related to the alteration of glutamine transporter [184]. Nevertheless, Cui study shows that there is a close relation between cell wall synthesis activity and the intracellular glutamine pool. In B. subtilis, glutamine functions as a gauge of the nitrogen supply in the cell. Therefore, with the high levels of glutamine,GS can be subject to feedback inhibition by glutamine, leading to the repression of glnRA. Besides glutamine, the intracellular histidine concentration at exponential phase was significantly higher in all three mutants. Furthermore, in the stationary growing phase, only 15% of the histidine was consumed after 9h of cultivation in ∆pgcA cells - whereas more than 30% was taken up by wt cells at the same time window. These results may point out to an alteration of histidine metabolism. The central carbon metabolites of cells lacking PgcA were also inspected when cultivated in a minimal medium with glucose and glutamate as carbon sources. In this condition, the concentrations of the amino acids glutamate, glutamine, histidine inside the cells are similar between both strain in three time points of cultivation inspected (exponential, transition and beginning of stationary phase). Moreover, 2-oxoglutarate and other TCA metabolites but also BCAA precursors, were secreted to the medium in lower amounts by ∆pgcA (figures 7.11, 7.12 and 7.13). These results entail the need of further research on these mutations in a broader media conditions since they seem to display different alterations in the central carbon pathway and BCAA in accordance to the medium used. Also, studies should be conducted to comprehend if GlmS or GS activity or its transcription are altered in B. subtilis LTA mutants. The monitoring of antibiotic resistance and the characterization of the muropeptides of the PG of ∆ugtP, ∆pgcA, and ∆pgcA using digestive enzymes could contribute to the clarifi- cation of cell wall structure in these conditions. Albeit histidine can also be degraded to glutamate, in B. subtilis, the histidine levels aren’t regulated by the nitrogen availability like glutamine. In B. subtilis, the histidine transport and utilization is done by Hut enzymes, encoded in the hut operon. The mechanism of induction of Hut expression is regulated in response to the presence of histidine, the amino acid status by the global regulatory protein CodY and, the carbon availability by CcpA [185, 186]. The global regulator CodY, represses the hut operon when cells are growing in media containing amino acids [187]. By binding to a gene located downstream of the promoter, CodY inhibits histidine transport and represses the transcription of hut operon. Moreover, the ac-

90 tivity of CodY is known to be stimulated by GTP and BCAA in an additive manner, which reflects the general nutritional state of the cell and, in this way, permits the adaptation to possible nutrient limitations [187, 136, 186]. Nevertheless, GTP and BCAA concentrations weren’t altered in ∆pgcA, ∆gtaB and ∆ugtP. In cells growing with preferred carbon sources, such as glucose, hut operon is also repressed by the CcpA. CcpA binds to a DNA region in hut operon similar to cre sites, as previously described, leading to the repression of histidine utilization [187, 186]. It was also confirmed that the degree of repression reflects the quality of the carbon source and that the phosphorylated compounds like glucose 6-phosphate and FBP, are potentially the physiological link that connects the binding of CcpA at cre sites [188, 186]. The regulation of the Hut enzymes in LB medium, can result in the repression of hut their synthesis during exponential growth. This regulation is seen as a way of B. subtilis cells ensure that not all of the available nutrients are consumed during rapid growth and, that adequate supplies of nitrogen, carbon, and phosphate compounds are present during other growth phases [189]. It is not clear why histidine accumulated in all mutants in exponential phase. It can only be speculated that alteration of the general nutritional state could also have led to the inhibition of histidine import and degradation. If the accumulation of glycolytic metabolites in the mutant cells like glucose 6-phosphate, fructose 6-phosphate, seen as a potential stimulus for CcpA activity, had any influence in the histidine regulation is hard to conclude. On the other hand, as seen for cells lacking Pyk, where accumulation of phosphorylated metabolites were determined, there weren’t observed alterations in the glutamine levels.

4.3 Conclusion

For the first time, a global perspective of qualitative metabolites of B. subtilis lacking lipoteichoic enzymes was monitored. Here, it was aimed to comprehend in a better way the interactions of metabolic precursors of LTA and the central carbon metabolism in B. subtilis. First, the metabolic profiles of B. subtilis strains 168 and BSB1 wt and the mutants ∆pgcA, ∆gtaB, and ∆ugtP were inspected to determine possible differences between strains. The extracellular nutrient consumption and secretion patterns in both wt and mutants showed no metabolic variations between both strains. Moreover, insights of the intracellular metabolites were further inspected using chemometric tools. This complementary approach showed no evidences of altered metabolites of 168 and BSB1 strains. Thus, BSB1 was selected to continue the desired investigation. B. subtilis BSB1 lacking either PgcA, GtaB, or UgtP showed affected metabolic pathways. The prominent alterations were the cells wall components, specifically the accumulation of UDP-glucose, UPD-glucuronate, andPG precursors. The hypothesis that an interaction

91 between the LTA andPG assembly was questioned and worth of further research. Although no cell wall thickness was observed when cells were microscopically inspected, ∆pgcA and ∆ugtP presented a fuzzier cell wall structure. These results emphasized the assumption of the affected hydrolytic activity occurring in thePG assembly of the cell wall mutants, as proposed to ∆ugtP by collaborators. With altered LTA composition, cell may alter the balance ofPG synthesis and hydrolysis. This study also implies that interactions between the central carbon metabolism and LTA must exist in B. subtilis. Besides the alterations in the cell wall metabolic content, several other metabolites were accumulated, as they were glutamine, histidine, and glycolytic metabolites. Many studies on the synthesis, activities and importance of TA in B. subtilis cells and other microorganisms are extensively being made [190, 191]. Nevertheless, still little is known concerning the central carbon pathways, the metabolism, and the TA com- munication. Moreover, the selection of the nutritional environment seem to affect and modulate the metabolic alteration of LTA in different ways. As refereed, while glutamine and histidine are increasingly accumulated inside mutant cells when cultivated in a rich medium, the same didn’t occurred in a minimal medium. Taken together, valuable data source was gained, that can provide information related to the central pathways and cell wall of gram-positive bacteria like B. subtilis.

92 Chapter 5

Concluding marks and outlook

Using validated analytical tools and optimized sampling procedures, it was possible to detect a vast number of metabolites from the extracellular space but also from the cytosol of B. subtilis. The results indicate that the complement of the analytical methods was suitable in the monitoring of the metabolome since it allowed a great coverage of physicochemical diverse metabolites. However, a wide number of unknown metabolites/features were also detected. Although broad databases exist that can help in the annotation of metabolites, further investigation is needed in their identiﬁcation. In unpredictable changing conditions, bacterial cells possess appropriate adaptation strategies for a successful bacterial growth. These rely on sensing mechanisms that globally adjust gene expression via transcription and feedback regulations. The metabolic sensing mechanisms have emerged as key roles in the nutritional information and regulation of cell cycle processes. In this work, a new quality of information regarding the metabolism and adaptation to the absence of key signal mechanisms in B. subtilis was provided. Investigations of cells lacking Pyk uncovered alterations in the import of glucose and pyruvate from the nutritional media. These results gives insights to the pyruvate homeostasis mechanism but also brought new questions concerning the regulation of the CCR. Pyruvate wasn’t susceptible to the glucose dependent CCR in ∆pyk. The earlier inﬂux of pyruvate in these cells is in accordance to the newly discovered pyruvate transport mechanism. Also, it was speculated that the lower consumption of external glucose could be a consequence of the impairment of the PTS system in the mutant cells due to the accumulation of glycolytic metabolite FBP. In future studies, insights of the PTS system mechanism should be done in these conditions, that could comprise the determination of HPr phosphorylation and the HPrK activity. This study also arose new questions that should be address, that include the higher secretion

93 of acetoin and 2,3-butanediol, and the lower accumulation of shikimate 3-phosphate by the mutant cells. In an untargeted metabolomic analysis, a vast number of altered features were suggested to be fatty acids metabolites, precursors of phospholipids and LTA. Complementary approaches should be done for the confirmation of these metabolites and the inspection of possible alterations in the membrane structure. In the study of LTA mutants, the accumulation ofPG precursors provided a hint of altered cell wall assembly. Although by fluorescence microscopy no clear changes were detected, the metabolic results emphasized the previous assumption of the affected hydrolytic activity occurring in thePG. For comprehensive knowledge of the cell wall it would be important to detect and identify more metabolites of the LTA anchor using optimized cromatographic method. These results could be complemented with other omics data sets studies which would help in the elucidation of these key regulatory systems mechanisms.

94 Chapter 6

Material and Methods

6.1 Material

6.1.1 Chemicals

6.1.1.1 Chemicals for media preparation

Agar (Sigma-Aldrich, St. Louis, USA)

Amonium chloride (Sigma-Aldrich, St. Louis, USA)

Calcium chloride dihydrate (Merck, Darmstadt, USA)

Citric acid monohydrate (Sigma-Aldrich, St. Louis, USA)

Cobalt (II) chlorate dihydrate (Merck, Darmstadt, USA)

Copper (II) chloride dihydrate (Merck, Darmstadt, USA)

D-Glucose (CAS 50-99-7, Merck, Darmstadt, USA)

Iron (III) chloride hexahydrate (Sigma-Aldrich, St. Louis, USA)

LB Broth (Sigma-Aldrich, St. Louis, USA)

L-Malate (CAS 97-67-6, Sigma-Aldrich, St. Louis, USA)

Magnesium sulfate heptahydrate (Merck, Darmstadt, USA)

Manganese (II) chloride tetrahydrate (Merck, Darmstadt, USA)

Potassium phosphate monobasic (Merck, Darmstadt, USA)

Sodium chloride (Merck, Darmstadt, USA)

95 Sodium molybdate dihydrate (Sigma-Aldrich, St. Louis, USA)

Sodium phosphate monobasic dihydrate (Merck, Darmstadt, USA)

Sodium pyruvate (CAS 113-24-6, Sigma-Aldrich, St. Louis, USA)

Zinc chloride (Sigma-Aldrich, St. Louis, USA)

Erythromycin (Roth, Karlsruhe, Germany)

Tetracycline (Sigma-Aldrich, St. Louis, USA)

Kanamycin (Sigma-Aldrich, St. Louis, USA)

Spectinomycin (Sigma-Aldrich, St. Louis, USA)

Boc-D-2,3-diaminopropionic acid (NADA) (Alfa Aeser, Thermo Fisher Scientiﬁc, Mas- sachusetts, USA)

Nuclei lysis solution (Promega, Wisconsin, USA)

6.1.1.2 Other chemicals

4-chlorophenylalanine hydroxide (Bachem, Bubendorf, Switzerland)

Camphorsulfonic acid (Sigma-Aldrich, St. Louis, USA)

Chloroform (Sigma-Aldrich, St. Louis, USA)

D-norvaline (Fluka Chemie, Buchs, Switzerland)

Ethanol MS grade (Fluka Chemie, Buchs, Switzerland)

Glutaraldehyde (Sigma-Aldrich, St. Louis, USA)

MeOx (Fluka Chemie, Buchs, Switzerland)

Methanol LC-MS grade (Sigma-Aldrich, St. Louis, USA)

MSTFA (Chromatographie Service GmbH, Langerwehe, Germany)

N,N-dimethyl-L-phenylalanine (Sigma-Aldrich, St. Louis, USA)

Paraformaldehyde (Science Services GmbH, Munich, Germany)

Picric acid (VWR, Radner, USA)

Formic acid (Sigma-Aldrich, St. Louis, USA)

96 Acetonitrile (Sigma-Aldrich, St. Louis, USA)

Ammonium formate (Sigma-Aldrich, St. Louis, USA)

Pyridine (Sigma-Aldrich, St. Louis, USA)

Ribitol (Merck, Darmstadt, Germany)

Tributylamine (Sigma-Aldrich, St. Louis, USA)

3-trimethylsilyl-[2,2,3,3-D4]-1-propionic acid (TSP) (Sigma-Aldrich, St. Louis, USA)

Isopropanol (Sigma-Aldrich, St. Louis, USA)

Ethidium bromide (Sigma-Aldrich, St. Louis, USA)

6.1.2 Growth media and solutions

For all solutions, destilated water was used.

LB medium: 20 g of LB in 800 ml of water, autoclaved at 120oC for 20 min and stored at 4oC.

nutrient-agar: 1.5% (w/v) agar in LB medium, autoclaved at 120oC for 20 min and stored at 4oC.

M9 5x stock solution: mixture of the compounds described on table 6.1. After the adjustment of the pH to 7.0 using 4M NaOH, the medium was autoclaved for 20 min at 120oC and stored at 4oC.

Table 6.1: M9 stock solution (5x) composition (per liter)

Compound Amount(g) Na2HPO4.2H2O 42.5 KH2PO4 15 NH4Cl 5.0 NaCl 2.5

Trace elements 100x stock solution: mixture of the compounds described on table 6.2.

The solution was ﬁlter-sterilized using a 0.2 µm pore size ﬁlter (Filtropur S, Sarstedt AG, Nuremberg, Germany) and stored in the dark at 4oC.

97 Table 6.2: Compounds and the amount (per liter) used in trace elements stock solution (100x) preparation

Compound Amount (mg) MnCl2.4H2O 100 ZnCl2 170 CoCl2.6H2O 60 Na2MoO4.2H2O 60

CaCl2 100 mM solution: 1.47 g of calcium chloride was dissolved in 100 mL of water. The solution was autoclaved at 120oC for 20 min and stored at 4oC.

MgSO4 1 M solution: 24.6 g of magnesium sulfate was dissolved in 100 ml of water. The solution was autoclaved at 120oC for 20 min and stored at 4oC.

FeCl3 50 mM in 100 mM citrate solution: 1.35 g of iron chloride was dissolved in 100 ml of water containing 2.10 g of citrate. The solution was ﬁlter-sterilized using a 0.2

µm pore size ﬁlter (Filtropur S, Sarstedt AG, Nuremberg, Germany) and stored in the dark at 4oC.

Glucose (50% w/v) solution: 500 g of glucose was dissolved in 1 l of water. Warm water bath was used to speed up dissolution. The solution was ﬁlter-sterilized using a o 0.2 µm pore size ﬁlter and stored in the dark at 4 C.

Malate (50% w/v) solution: 50 g of malate was dissolved in 1 l of water. Warm water bath was used to speed up dissolution. The solution was ﬁlter-sterilized using a 0.2 o µm pore size ﬁlter and stored in the dark at 4 C.

Pyruvate solution: 5.5 g of sodium pyruvate was dissolved in 50 ml of water. Warm water bath was used to speed up dissolution. The solution was ﬁlter-sterilized using a o 0.2 µm pore size ﬁlter and stored in the dark at 4 C.

M9 medium: Prepared as described by Harward and Cutting with small changes [192]. To prevent precipitation problems, magnesium sulfate and calcium chloride were added as separate solutions. Iron chloride solution was prepared with citric acid and was also added as a separate solution. For 1 L of medium, the solutions listed on table 6.3 were added in the order described.

M9GlcMal medium: M9 medium containing 5.54 mM of glucose and 7.4 mM of malate.

M9GlcMalGlut medium: M9 medium containing 5.54 mM of glucose, 7.4 mM of malate and 3 mM of glutamate.

98 Table 6.3: M9 medium composition

Solution Volume (ml) H2O adjusted to 1 L M9 (5x) stock 200 Trace Elements (100x) 10 C source adjusted to the required concentration MgSO4 1.0 CaCl2 1.0 FeCl33 in citrate 1.0

M9Glc20Mal5 medium: M9 medium containing 20 mM of glucose and 5 mM of malate.

M9Glc medium: M9 medium containing 40 mM of glucose.

M9GlcPyr medium: M9 medium containing 10 mM of glucose and 60 mM of pyruvate.

M9Pyr medium: M9 medium containing 80 mM of pyruvate.

Spizizen minimal medium (SMM): mixture of the compounds described on table 6.4. The pH was adjusted to 7.5. Stored at 4oC.

Table 6.4: SMM composition

Compound concentration (%)) (NH4)2SO4 0.2 K2HPO4 1.4 KH2PO4 0.6 C6H5O7Na3.2H2O 0.1

Competence medium: mixture of the compounds described on table 6.5. Stored at 4oC.

Table 6.5: Competence medium composition

Compound Volume (ml) SMM medium 10 D-glucose (40%) 0.125 Tryptophan solution 2 mg.ml−1 0.1 MgSO4 (1 M) 0.06 Casamino acid (20%) 0.01 Fe-NH4-citrate (0.22%) 0.05

99 Starvation medium: mixture of the compounds described on table 6.6. Stored at 4oC.

Table 6.6: Starvation medium composition

Compound Volume (ml) SMM medium 10 D-glucose (40%) 0.125 MgSO4 (1 M) 0.06

Tris-acetate-EDTA (TAE) stock solution (50x): mixture of the compounds described on table 6.7. The pH is not adjusted and should be about 8.5. Stored at room temperature.

Table 6.7: TAE buﬀer solution

Compound Amount Tris free base 242 g Glacial acetic acid 50 mM 57 ml Disodiumn EDTA 18.61 g H2O 943 ml

TAE solution (1x): 20 ml of TAE stock solution (50x) in 980 mL. Stored at room temperature.

TES buﬀer solution: mixture of Tris/HCl 0.2 M, EDTA 5 mM, and NaCl 100 mM. The pH was adjusted to 7.5. Stored at room temperature.

6.1.3 Strains collection

The B. subtilis strains used in these studies are listed in table 6.8. Except for JS01-JS04, strains were constructed and provided by collaborators [119]. Gene knockouts were achieved by replacing the gene coding sequence with an antibiotic resistance cassette

100 Table 6.8: B. subtilis strains used in the studies

Strain Genotype Source 168 wild type Newcastle University tryptophan auxotroph (trpC2−) 168∆pgcA trpC2 ∆pgcA::tet Newcastle University 168∆gtaB trpC2 ∆gtaB::erm Newcastle University 168∆ugtP trpC2 ∆ugtP::kan Newcastle University BSB1 wild type Newcastle University tryptophan prototroph (trpC2+) BSB1∆pgcA trpC2+ ∆pgcA::tet Newcastle University BSB1∆gtaB trpC2+ ∆gtaB::erm Newcastle University BSB1∆ugtP trpC2+ ∆ugtP::kan Newcastle University BSB1Goet wild type G¨ottingenUniversity trpC2+ BSB1∆pyk trpC2+ ∆pyk::kan G¨ottingenUniversity JSL01 trpC2+ ∆dacA::erm mcherry::spec Newcastle University JSL02 trpC2+ ∆ugtP::kan∆dacA::erm Newcastle University JSL03 trpC2+ ∆dacA::erm Newcastle University JSL04 trpC2+ ∆ugtP::kan mcherry::spec∆dacA::erm Newcastle University

6.2 Methods

6.2.1 Strains maintenance

All strains were streaked in nutrient agar plates from glycerol stock and incubated overnight at 37oC. −1 For genetic selections, speciﬁc antibiotics were added: 10 µg.ml of tetracycline for BSB1∆pgcA; −1 −1 0.5 µg.ml of erythromycin for BSB1∆gtaB, JSL0, JSL04, JSL03, and JSL04; 3 µg.ml of −1 kanamycin for BSB1∆ugtP, BSB1∆pyk; and 50 µg.ml of spectynomycin for JSL01, and JSL04. New glycerol stock of B. subtilis were prepared by cultivating cells from nutrient agar plates o in 500 µL of LB medium for 4h at 300 rpm and 37 C. Subsequently, 500 µL of sterile glycerol was added and kept at -80oC form long-term storage. (-80oC).

6.2.2 Isolation of chromosomal DNA

LB medium was inoculated with cells from an overnight culture and incubated for 5 h at o 37 C. Cells were pelleted (3000 g for 5 min) and resuspended with 100 µl EDTA (50 mM). −1 Lysozyme (3 µl of a 10 mg.ml stock prepared in TES buﬀer) and RNase (3 µl of a 10 mg.ml−1 stock in TES buﬀer) were added to the suspension and samples were incubated o for 1 h at 37 C. Nuclei lysis solution (500 µl) was added and cells were incubated for 5 min at 80oC then cooled down to room temperature (25oC). Protein precipitation solution (200

101 µl) was added and the mixture and vortexed for 20 s at high speed. Cells were incubated afterwards on ice for 10 min then centrifuged for 10 min at 13000 g. The supernatant was transferred to a clean microfuge tube containing 600 µl of isopropanol, mixed gently then o centrifuged (13000 g for 10 min at 4 C). The supernatant was discarded and 600 µl of 70% ethanol was added to the pellet. Samples were centrifuged as before, the supernatant was discarded carefully and the tubes were left to dry. DNA was resuspended in distilled water o (100 µl), incubated at 65 C for 15 min.

6.2.3 Transformation

The transformation was fulfilled according to the method developed by Anagnostopoulos and Spizizen with modifications [193, 194]. Cells were incubated overnight in 5 ml of competence medium at 30oC with agitation. Fresh competence medium (10 ml) was inoculated with 600 µl of the overnight culture and incubated for 4 h at 37oC with shaking. Pre-warmed starvation medium (10 ml) was added to the previous culture and cells were incubated at 37oC for another 2 h with shacking (competent cells). −1 o DNA (1-3 ng.ml ) was added to competent cells (900 µl) and incubated at 37 C for 15 min. Cells were plated on nutrient agar plates with selective antibiotics and incubated overnight at 37oC. Successful transformants grew on plates with antibiotics and a further polymerase chain reaction check was performed for the confirmation of mutants.

6.2.4 Polymerase chain reaction and restriction endonuclease digestion

Gotaq Flexi DNA Polymerase (Promega, Wisconsin, USA) was used for confirming gene deletions. Reactions conditions were prepared according to the manufacturer’s instructions. The PCR amplification steps consisted of an initial denaturation (98oC, 2 min), 30 amplification cycles and a final oligonucleotides extension step (72oC for 4 min). Each amplification cycle consisted of denaturation (98oC, 10 s), annealing (55oC, 30 s) and oligonucleotide extension (72oC, 6 min ). Restriction enzyme BgLII was used in the strains with the deletion of dacA gene, according to the instruction of the manufacturer. Purification of PCR and restriction digestion products QIAquick PCR Purification Kit (QI- AGEN) was used to clean up DNA in PCR reaction according to the manufacture’s protocol.

102 6.2.5 Agarose gel electrophoresis

A 1.0 % of agarose gel was melted in TAE buffer. The solidified gel was submerged in TAE buffer in an electrophoresis tank. The DNA sample was mixed with DNA loading buffer and loaded into the agarose gel wells. The electrophoresis was run at 100 V for about 30-40 −1 min. The gel was soaked in 1.25 µg.ml of ethidium bromide in TAE buffer for 15 min. The DNA bands were visualised by UV transilluminator.

6.2.6 Cultivation conditions

6.2.6.1 Growth cultivation for metabolic studies of the cell wall precursors mutants

The pre-cultures were incubated (New Brunswick Innova 42, Eppendorf, Germany) from isolated colonies for 4h in 5 ml of LB medium at 37oC and 300 rpm. These cultivations were used to prepare overnight cultures in LB medium in several dilutions (1:75 000, 1:150 000 000, 1:225 000 000 and 1:300 000 000). After 15h of incubation at 37oC and 260 rpm, main cultures were inoculated with an initialOD 600nm of 0.1 in LB medium from exponentially o growing overnight cultures (OD600nm between 0.8-1.2) at 37 C and 280 rpm. Cultivation in chemical deﬁned media was carried out in a similar way, whereas overnight and main cultures were incubated with the same M9 media (M9 medium with 5.5 mM of glucose and 7.4 mM of malate or M9 medium with 5.5 mM of glucose, 7.4 mM of malate and 3 mM of glutamate). The overnight cultures were cultivated in higher dilutions (1:25 000 000, 1.50 000 000, 1:75 000 000 and 1:150 000 000) and the cultures selected were in exponential growth phase between 0.5 and 0.8OD 600nm. The main cultures were incubated with an initialOD 600nm of 0.05.

6.2.6.2 Growth cultivation for metabolic studies of the pyruvate kinase mutant

The growth conditions for pyruvate kinase experiments were carried out as described above with some optimization. −1 B. subtilis BSB1Goet and BSB1∆pyk were streaked in nutrient agar plate (3 µg.ml of kanamycin for ∆pyk). The pre-cultures were incubated for 4h in 5 ml of LB medium at 300 rpm and 37oC. After incubation overnight at 37oC and 260 rpm in M9 medium with 20 mM of glucose and 5 mM of malate in several dilutions (1:25 000 000, 1:50 000 000, 1:75 000 000 for wt and 1:6

000, 1:8 000, 1:10 000 and 1:12 000 for ∆pyk), the cultures in exponential growth (OD600nm between 0.5-0.8) were centrifuged at 6 000 rpm and 4oC for 3 minutes (Heraus Multifuge X1R, Thermo Scientiﬁc, Massachusetts, USA). The pellet obtained was transferred to the main culture consisted in M9 and diﬀerent carbon sources (LB broth, 40 mM of glucose, 10 mM of glucose and 60 mM of pyruvate or 80 mM of pyruvate).

103 6.2.7 Sampling of extracellular metabolites

For the collection of extracelullar metabolites samples, the optical density was monitored and 2 ml cell suspension was sampled. Two ml of bacterial culture was sterile ﬁltered using a 0.45 µm pore size ﬁlter (Filtropur S, Sarstedt AG, Nuremberg, Germany) every 60 minutes and stored at -20oC prior to measurement.

6.2.8 Sampling of intracellular metabolites

At every sampling time point, 20OD units of cell culture were harvested via vacuum- dependent fast-ﬁltration system as described by Meyer et al. (ﬁgure 6.1)[141]. The main culture was transferred into a falcon tube and cooled 10 times with liquid nitrogen

extracellular analytical studies metabolites extraction

inoculation pre-culture overnight main culture culture

N2 pre-cooling

analytical studies intracellular metabolites N2 quenching extraction and lyophilization vaccum filter system

Figure 6.1: Experimental workﬂow

(1 s each time). During this in/out of liquid nitrogen cycle, the sample was carefully shaken to avoid freezing and metabolite leakage caused by cell lysis. Subsequently, the cooled cell culture was ﬁltered using a regenerated cellulose membrane TM ﬁlter with a 0.45 µm pore size and 100 mm of diameter (RC55, Whatman , GE Healthcare Life Sciences, Pittsburgh, USA) and washed 2 times with isotonic sodium chloride solution

104 at 4oC (0.8 or 0.9% when cultivated in chemical defined or complex medium, respectively). The filter was immediately transferred to a falcon tube containing 5 ml of ice-cold extraction solution (60% w/v) and ISTD constituted of 2.5 nmol of camphorsulphonic acid for LC-MS or 20 nmol each of ribitol, norvaline, N,N-dimethyl-phenylalanine, and p-chlorophenylalanine-hydroxide for GC-MS analysis). The metabolites were quenched by frozen the sample immediately in liquid nitrogen. The falcon tube was stored at -80oC until extraction. For cell disruption and metabolites extraction, a freeze/thaw cycle was performed by alter- nately thawing on ice, vortexing and shaking the sample for 10 times. Subsequently, the sample was centrifuged for 5 min at 4oC and 13000 rpm. The supernatant was collected to a new falcon tube and left on ice. The pellet formed was extracted once again with 5 ml distillated water. The new supernatant was mixed with the previously aqueous extraction. To get a final organic solution concentration of 10%, destilated water was added (30˜ ml) and stored at -80oC. The sample was lyophilized with a Christbeta Alpha 1-4 LSC lyophilized at -52°C and 0.25 mbar before further analysis. To avoid disturbances in HPLC columns associated with macromolecules accumulation (i.e. proteins), a third extraction with chloroform was added to the experimental procedure of pyruvate kinase analyses. These samples were lyophilized once again.

6.2.9 Analytical methods

6.2.9.1 1H-NMR spectroscopy measurement and data analysis of extracellular metabolites

Supernatant culture sample was thawed at room temperature. A 400 µL sample was mixed with 200 µL of sodium hydrogen phosphate buffer (0.2 mM, pH 7.0) and TSP (1 mM) 1 made up with 50% D2O to provide a nuclear magnetic resonance H-NMR-lock signal. All 1H-NMR spectra were obtained at 600.27 MHz at 310 K using a Bruker Avance-II 600 1H-NMR spectrometer operated by TOPSPIN 3.2 software (Bruker Biospin GmbH, Rhein- stetten, Germany). A modified 1D-NOESY pulse sequence was used with presaturation on the residual HDO signal during both the relaxation delay and the mixing time. A total of 64 free induction decays (FID scans) were collected, using a spectral width of 30 ppm for a one dimensional spectrum. The identification and quantification analysis were done using AMIX v3.9.11 software (Bruker Biospin GmbH, Rheinstetten, Germany). The signal peak identification was based on spectra alignment of pure standard compounds (Sigma-Aldrich, St. Louis, USA). Quantification was done by integration and comparison of designated peaks to an external QuantRef signal at 15 ppm. Unidentified signals were relatively quantified due to their unknown quantity of protons.

105 6.2.9.2 GC-MS measurement and data analysis of intracellular metabolites

o The dried samples were derivatized first with 60 µL of MeOx for 90 min at 37 C and second o with 120 µL of MeOx for 30 min at 37 C. Samples were centrifuged for 2 min at room temperature and the supernatant was transferred into GC-vial for injection. GC-MS analysis was performed with an Agilent 6890N GC system with an autosampler G2614A model coupled to a mass selective detector 5973N model (Agilent Technologies, USA). A 2 µL sample was injected with a G2613A model SeriesInjector and split 1:10 at 250oC using helium as the carrier gas (split flow of 10 ml.min −1 and 8.8 Psi). The chromatographic run was performed using a 30 m DB 5-column (JW Scientific, Folsom, USA) with 0.25 mm inner diameter and 0.25 mm film thickness and a constant gas flow of 1 ml.min−1. The oven program started with an initial temperature hold at 70°C for 1 min and continued with a heating rate of 1oC.min−1 up to 76oC, 5oC.min−1 up to 220oC, and 20oC.min−1 up to 330oC, with a hold for 3 min followed by a 10 min isothermal cool-down to 70oC. The analytes were transferred to a quadrupole mass analyzer operated in theEI mode with an ionization energy of 70 eV. Data acquisition was done in 40 min runtime. A full scan mass spectra were acquired from 50 to 500 m/z at a rate of 2 scans.s−1 and with a 6 min solvent delay. The qualitative analysis of the detected compounds was performed using ChromaTOF software (LECO Corporation, Michigan, USA). Metabolite identification was carried out by comparison of retention time and fragmentation patterns peaks detected to the National Institute of Standards and Technology (NIST) mass spectral database 2.0 (Gaithersburg, USA) and an in-house database. Identification of peaks was carried out by comparison of mass spectra with a database with a similarity of 75%. For relative quantification, the area of the quantifier ion of each metabolite was integrated and normalized to the area of the quantifier ion of the ISTD. In all experiments, ribitol was the ISD chosen for normalization. For precision analysis control, daily quality control (dQC) samples were analyzed during the batches. The dQC consisted in 53 metabolites with 100 nmol concentration for each metabolite. Precision analysis was determined by assessing the measured dQC in calibration curves with concentrations ranging from 0.5 nmol to 500 nmol of each metabolite. The calibration curves fitting were performed with a polynomial of degree 2 and 1/x weighting based on minimum of 6 calibration points.

6.2.9.3 LC-MS measurement and data analysis of intracellular metabolites

For intracellular metabolites analysis in LC-MS, the lyophilized samples were dissolved in

100 µL of water HPLC grade and centrifuged for 2 min at room temperature. The supernatant was transferred into LC-vial for injection. LC-MS analysis was carry out using an Agilent 1100 HPLC system consisted of a degaser, a quaternary pump and a G1329A autosampler with controlled temperature coupled to Bruker microtime of ﬂight (TOF) mass spectrometer (Bruker Daltonics, Bremen, Germany).

106 Chromatography was performed on a SymmetryShield RP18 column (3.5 µm, 150 x 4.6 mm) (Waters, Milford, USA) with a SecurityGuard cartridge C18 pre-column (4 x 3.0 mm) (Phenomenex, Torrance, USA) using an ion-pairing reagent and a methanol gradient. In detail, the mobile phase consisted in eluent A: 95% water and 5% methanol, containing 10 mM of tributylamine as the ion-pairing reagent and 15 mM of acetic acid, pH 4.9; and eluent B: 100% methanol. Data acquisition was done in 42 min runtime with a flow rate of 0.4 ml.min-1. The gradient elution started with 100% A for 2 min, 0-31% B in 2 min and continued with 31 to 50% in 18 min. Followed by 50-60% B in 2 min, 60-100% B in 1 min and left 100% for 7 min. The eluent A returned to 100% in 1 min and was left for 10 min until the end of the run. The gradient elution started with 100% A for 2 min, 0-31% B in 2 min and continued with 31 to 50% in 18 min. Followed by 50-60% B in 2 min, 60-100% B in 1 min and left 100% for 7 min. The eluent A returned to 100% in 1 min and was left for 10 min until the end of the run. Mass spectrometry was operated in ESI and negative-ion mode using a mass scan range of 50 to 3000 m/z. Internal MS calibration was carry out in the beginning of each chromatographic run with 16 different masses from a sodium formate solution tune mix (49.4% water, 49.4% isopropanol, 0.2% formic acid, and 10 mM sodium hydroxide). Metabolite identification was carried out by comparison of retention time and m/z values of detected peaks ([M-H]1− or [M-2H]2−) with database alignment of the calculated exact mass. The quantitative analysis was done using QuantAnalysis (Bruker Daltonik, Bremen, Germany). The extracted ion peaks were integrated and normalized to the ISTD (CSA) area. The dQC samples consisted in 22 metabolites with 10 nmol concentration of each metabolite. Precision analysis was determined by assessing the measured dQC in calibration curves with concentrations ranging from 0.5 nmol to 500 nmol of each metabolite. The calibration curves fitting were performed with a polynomial of degree 2 and 1/x weighting based on minimum of 6 calibration points. For precision analysis control, dQC samples were analyzed during the time runs. The dQC consisted in 53 metabolites with 100 nmol concentration for each metabolite. Precision analysis was determined by assessing the measured dQC in calibration curves with concentrations ranging from 0.25 nmol to 100 nmol of each metabolite. The calibration curves fitting were performed with a polynomial of degree 2 and 1/x weighting based on minimum of 6 calibration points.

6.2.9.4 LC-MS/MS measurement and data analysis of intracellular metabolites

The amino acids proline, arginine, and citrulline of were analyzed without derivatization on an Intrada amino acid column (50 x 3 mm, 3 µm)(Imtakt Corporation) with acetonitrile/100 mM ammonium-formate (20/80, v/v) as eluent A and acetonitrile/THF/25 mM ammonium formate /formic acid (9/75/16/0.3, v/v/v/v) as eluent B. A 2 µL sample was injected in the column stabilized at 40oC. Data acquisition was done in 26 min runtime with a ﬂow rate

107 Table 6.9: Gradient elution method in LC system

Time (min) Eluent A (%) Eluent B (%) 0.0 100 0.0 2.0 69 31 20 50 50 23 40 60 24 100 0.0 31 100 0.0 32 0.0 100 42 0.0 100

of 0.4 ml.min-1. The gradient elution started with 100% B for 5 min, followed by 0-17% A in 4,45 min. Eluent A reached 100% in 1,15 min and was left in this condition for 6 min. Afterwards, eluent B returned to 100% in 50 sec and was left for 8 min until the end of the run (table 6.10). Analyses were performed on an HPLC system (1200) consisting of a degasser, a binary pump, a temperature-controlled autosampler, and a column oven coupled to a 6460 triple quadrupole mass spectrometer with ESI Jet stream source (all Agilent Technologies). The mass spectrometer was operated in positive mode (table 6.11) using multiple reaction monitoring (MRM). Proline - 13C, 15N and arginine - 13C, 15N were used as internal standard compounds for relative quantiﬁcation (Cell Free Amino Acid Mixture – 13C, 15N; Sigma Aldrich). Quantiﬁcation was done using Mass Hunter Quantitative Analysis (for QQQ) B 06.00 from Agilent.

Table 6.10: Gradient elution method in LC-MS/MS system

Time (min) Eluent A (%) Eluent B (%) 0.0 0.0 100 4.5 0.0 100 9.8 17 83 11 100 0.0 17 100 0.0 17.5 0.0 100 25.5 0.0 100

108 Table 6.11: MS source parameters used during amino acids measurements

Parameter Value Gas ﬂow 10 ml.min−1 Gas temperature 350oC Nebulizer pressure 40 psi Sheath gas ﬂow 12 l.min−1 Sheath gas temperature 350oC Capillary voltage 2000 V Nozzle voltage 0 V

6.3 Incorporation and release of labelled D-alanine (NADA) in the peptidoglycan and ﬂuorescent microscopy

The staining of PG structure with the fluorescent D-amino acid (NADA) was done according to a technique described previously by de Jong et al. [195]. Cells were cultivated in LB o medium with NADA (final concentration of 50 µM) at 37 C and 280 rpm. A time lapse-microscopy was performed. After fully labelling with NADA, live cells were collected and sampled. When the desired time was reached (0, 25, and 80 min), 500 µl of cells were pelleted at 10000 rpm for 1 min. The cells were washed with 1 ml of PBS, fixed with paraformaldehyde solution (5% in PBS) and immediately observed. The preparation of cell sample for time-lapse was performed as quickly as possible. Images acquisition were performed with brightfield illumination with 50 ms exposure time. In red and green wavelength emissions, images were collected with 50 ms and 3 sec of exposure time, respectively. Microscopic images were taken using Nikon TiE microscope coupled to a Hamamatsu C9100 EMCCD or a Sony CoolSnap HQ2 camera.

6.4 Transmission electron microscopy

o For TEM studies, cells were cultivated in LB medium at 37 C and 280 rpm. WhenOD 600nm 0.5 was reached, 15 ml of cultivation was centrifuged at 3 000 rpm for 3 min. The pellet formed was resuspended in 1 ml of fixation solution containing 1% of glutaraldehyde and 4% of paraformaldehyde, 50 mM of sodium azide and 0.2% picric acid in 5mM HEPES buffer. The samples were immediately stored at 4oC until further processing. Subsequently, the cells were treated with 0.5% glutaraldehyde and 1% osmium tetroxide in washing buffer (100 mM cacodylate buffer pH 7, 1 mM calcium chloride, 0.09 M sucrose) for 1 h at 4oC. Next, the cells were embedded in low-gelling agarose, postfixed in 2% osmium tetroxide in washing buffer for 1 h at room temperature and stained overnight with 0.5% uranyl acetate in 0.9%

109 sodium chloride at 4oC. After dehydration in graded series of ethanol (30%-100% ethanol), the material was transferred step wise into propylene oxide and ﬁnally embedded in AGAR- LV resin (Plano, Wetzlar, Germany). Sections were cut on an ultramicrotome (Reichert Ultracut, Leica UK Ltd, Milton Keynes, United kingdom), stained with 4% aqueous uranyl acetate for 5 min and subsequently by lead citrate for 1 min. The analyses were done in a Leo 906 transmission electron microscope (Zeiss, Oberkochen, Germany).

6.5 Statistical analysis and visualization

Microsoft Excel software 2007 was used for metabolites quantification and calculation of FC when comparing two sample groups. The visualization of the time-resolved extracellular metabolites changes was performed using Excel and VANTED software v2.01. The heat maps of metabolites were generated using MeV v4.8.1 with hierarchical clustering analysis using euclidean distance and average linkage method. PCA projections were created using Metabonanalyst v3.0 platform with hierarchical clustering analysis using and euclidean distance and Wards linkage method. Data were normalized in logarithmic and scaled using Pareto scaling. The ANOVA studies were also performed with Metabonanalyst using Tukey post-hoc method whereas p-values≤0.05 were considered as being statistically significant. Bar-charts, XY-plots and volcano plots were generated using GraphPad PRISM software v6.01. Unpaired t-tests were also done in GraphPad PRISM. The two-sided homoscedastic t-tests were used to calculate p-values, whereas p-values≤0.05 were considered as being statistically significant. For metabolome missing values were replaced with half the value of the minimum positive value in the original data, assuming to be the detection limit. Fluorescent and phase-contrast microscopic images were taken using inverted Nikon eclipse Ti microscope coupled to a Sony The images acquired during the fluorescence microscopy were handle with ImageJ v1.49 (National Institutes of Health, Bethesda, USA). Data were analyzed by XCMS with the default parameters: centwave, span: 0.4, bw: 60 and mzwid: 0.5.

110 Bibliography

[1] Oliver Fiehn. Combining genomics, metabolome analysis, and biochemical modelling to understand metabolic networks. Comparative and functional genomics, 2(3):155– 168, 2001.

[2] Mlawule R Mashego, Karl Rumbold, Marjan De Mey, Erick Vandamme, Wim Soetaert, and Joseph J Heijnen. Microbial metabolomics: past, present and future methodolo- gies. Biotechnology letters, 29(1):1–16, 2007.

[3] KD Nadella, Soma S Marla, and P Ananda Kumar. Metabolomics in agriculture. Omics: a journal of integrative biology, 16(4):149–159, 2012.

[4] Rasmus Madsen, Torbj¨ornLundstedt, and Johan Trygg. Chemometrics in metabolomics—a review in human disease diagnosis. Analytica chimica acta, 659(1-2):23–33, 2010.

[5] Aihua Zhang, Hui Sun, Ping Wang, Ying Han, and Xijun Wang. Modern analytical techniques in metabolomics analysis. Analyst, 137(2):293–300, 2012.

[6] Pubmed. https://www.ncbi.nlm.nih.gov/pubmed/?term=metabolomics. Accessed on 2018-09-20.

[7] Phillip D Whitﬁeld, Alexander J German, and Peter-John M Noble. Metabolomics: an emerging post-genomic tool for nutrition. British Journal of Nutrition, 92(4):549–555, 2004.

[8] Silas G Villas-Bˆoas,Sandrine Mas, Mats Akesson,˚ Jørn Smedsgaard, and Jens Nielsen. Mass spectrometry in metabolome analysis. Mass spectrometry reviews, 24(5):613–646, 2005.

[9] Nadia Koen, Ilse Du Preez, et al. Chapter three-metabolomics and personalized medicine. Advances in protein chemistry and structural biology, 102:53–78, 2016.

[10] Vladimir Shulaev. Metabolomics technology and bioinformatics. Brieﬁngs in bioinformatics, 7(2):128–139, 2006.

111 [11] Christoph J Bolten, Patrick Kiefer, Fabien Letisse, Jean-Charles Portais, and Christoph Wittmann. Sampling for metabolome analysis of microorganisms. Ana- lytical chemistry, 79(10):3843–3849, 2007.

[12] Susanne K Wiedmer and Tuulia Hy¨otyl¨ainen.Selection of analytical methodology for metabolomics. Chromatographic methods in metabolomics, (18):1, 2013.

[13] William J Griﬃths, Therese Koal, Yuqin Wang, Matthias Kohl, David P Enot, and Hans-Peter Deigner. Targeted metabolomics for biomarker discovery. Angewandte Chemie International Edition, 49(32):5426–5445, 2010.

[14] Lee D Roberts, Amanda L Souza, Robert E Gerszten, and Clary B Clish. Targeted metabolomics. Current protocols in molecular biology, 98(1):30–2, 2012.

[15] Xiaoli Zhang, Xinyue Zhu, Caihong Wang, Haixia Zhang, and Zhiming Cai. Non- targeted and targeted metabolomics approaches to diagnosing lung cancer and pre- dicting patient prognosis. Oncotarget, 7(39):63437, 2016.

[16] Frederique Lisacek. Metabolome analysis: An introduction. Proteomics, 7(19):3634– 3634, 2007.

[17] Katja Dettmer, Pavel A Aronov, and Bruce D Hammock. Mass spectrometry-based metabolomics. Mass spectrometry reviews, 26(1):51–78, 2007.

[18] Guilhem Desbrosses, Dirk Steinhauser, Joachim Kopka, and Michael Udvardi. Meta- bolome analysis using gc-ms. In Lotus japonicus Handbook, pages 165–174. Springer, 2005.

[19] Zhentian Lei, David V Huhman, and Lloyd W Sumner. Mass spectrometry strategies in metabolomics. Journal of Biological Chemistry, 286(29):25435–25442, 2011.

[20] Rawi Ramautar, Oleg A Mayboroda, Govert W Somsen, and Gerhardus J de Jong. Ce-ms for metabolomics: Developments and applications in the period 2008–2010. Electrophoresis, 32(1):52–65, 2011.

[21] M Bauer, A Bertario, G Boccardi, X Fontaine, R Rao, and D Verrier. Reproducibility of 1h-nmr integrals: a collaborative study1. Journal of pharmaceutical and biomedical analysis, 17(3):419–425, 1998.

[22] Hector C Keun, Timothy MD Ebbels, Henrik Antti, Mary E Bollard, Olaf Beckonert, G¨otzSchlotterbeck, Hans Senn, Urs Niederhauser, Elaine Holmes, John C Lindon, et al. Analytical reproducibility in 1h nmr-based metabonomic urinalysis. Chemical research in toxicology, 15(11):1380–1386, 2002.

112 [23] John L Markley, Rafael Br¨uschweiler, Arthur S Edison, Hamid R Eghbalnia, Robert Powers, Daniel Raftery, and David S Wishart. The future of nmr-based metabolomics. Current opinion in biotechnology, 43:34–40, 2017.

[24] Niko Jukarainen. Nmr metabolomics techniques and mathematical tools as an aid in neurological diagnosis. Kuopio Univ Pub Nat and Environ Sci, 267:18, 2009.

[25] Ryan T McKay. How the 1d-noesy suppresses solvent signal in metabonomics nmr spectroscopy: An examination of the pulse sequence components and evolution. Con- cepts in Magnetic Resonance Part A, 38(5):197–220, 2011.

[26] Patrick Giraudeau, Virginie Silvestre, and Serge Akoka. Optimizing water suppression for quantitative nmr-based metabolomics: a tutorial review. Metabolomics, 11(5):1041–1055, 2015.

[27] Santosh Kumar Bharti and Raja Roy. Quantitative 1h nmr spectroscopy. TrAC Trends in Analytical Chemistry, 35:5–26, 2012.

[28] PäiviPöhöand Tuulia Hyötyläinen.Mass spectrometric detection for chromatography. Chromatographic Methods in Metabolomics; The Royal Society of Chemistry: London, UK, pages 43–63, 2013.

[29] Christian J Wachsmuth, Franziska C Vogl, Peter J Oefner, and Katja Dettmer. Gas chromatographic techniques in metabolomics. Chromatographic Methods in Metabolo- mics; Hyotylainen, T., Wiedmer, S., Eds, pages 87–105, 2013.

[30] Kalpesh N Patel, Jayvadan K Patel, Manish P Patel, Ganesh C Rajput, and Hitesh A Patel. Introduction to hyphenated techniques and their applications in pharmacy. Pharmaceutical methods, 1(1):2–13, 2010.

[31] Ambrose Furey, Merisa Moriarty, Vaishali Bane, Brian Kinsella, and Mary Lehane. Ion suppression; a critical review on causes, evaluation, prevention and applications. Talanta, 115:104–122, 2013.

[32] Georgios A Theodoridis, Filippos Michopoulos, Helen G Gika, Robert S Plumb, and Ian D Wilson. Liquid chromatographic techniques in metabolomics. Chromatographic Methods in Metabolomics, (18):64, 2013.

[33] Patrice J Larger, Massimo Breda, Daniela Fraier, Heather Hughes, and Christopher A James. Ion-suppression eﬀects in liquid chromatography–tandem mass spectrometry due to a formulation agent, a case study in drug discovery bioanalysis. Journal of pharmaceutical and biomedical analysis, 39(1-2):206–216, 2005.

[34] M Herderich, E Richling, R Roscher, C Schneider, W Schwab, H-U Humpf, and P Schreier. Application of atmospheric pressure ionization hplc-ms-ms for the analysis of natural products. Chromatographia, 45(1):127–132, 1997.

113 [35] G Zheng and MF Slavik. Isolation, partial puriﬁcation and characterization of a bacteriocin produced by a newly isolated Bacillus subtilis strain. Letters in applied microbiology, 28(5):363–367, 1999.

[36] F Kunst, N Ogasawara, I Moszer, AM Albertini, GO Alloni, V Azevedo, MG Bertero, P Bessieres, A Bolotin, Sea Borchert, et al. The complete genome sequence of the gram-positive bacterium Bacillus subtilis. Nature, 390(6657):249–256, 1997.

[37] Mario L Arrieta-Ortiz, Christoph Hafemeister, Ashley Rose Bate, Timothy Chu, Alex Greenﬁeld, Bentley Shuster, Samantha N Barry, Matthew Gallitto, Brian Liu, Thadeous Kacmarczyk, et al. An experimentally supported model of the Bacillus subtilis global transcriptional regulatory network. Molecular systems biology, 11(11):839, 2015.

[38] Valerie Barbe, Stephane Cruveiller, Frank Kunst, Patricia Lenoble, Guillaume Meurice, Agnieszka Sekowska, David Vallenet, Tingzhang Wang, Ivan Moszer, Clau- dine M´edigue,et al. From a consortium sequence to a uniﬁed sequence: the Bacillus subtilis 168 reference genome a decade later. Microbiology, 155(6):1758–1775, 2009.

[39] Kazuo Kobayashi, Stanislav Dusko Ehrlich, A Albertini, G Amati, KK Andersen, M Arnaud, K Asai, S Ashikaga, S Aymerich, P Bessieres, et al. Essential Bacillus subtilis genes. Proceedings of the National Academy of Sciences, 100(8):4678–4683, 2003.

[40] J-W Veening, WK Smits, LW Hamoen, and OP Kuipers. Single cell analysis of gene expression patterns of competence development and initiation of sporulation in Bacillus subtilis grown on chemically deﬁned media. Journal of applied microbiology, 101(3):531–541, 2006.

[41] Nicolas Mirouze and David Dubnau. Chance and necessity in Bacillus subtilis development. Microbiology spectrum, 1(1), 2013.

[42] Patrick J Piggot and David W Hilbert. Sporulation of Bacillus subtilis. Current opinion in microbiology, 7(6):579–586, 2004.

[43] Lynne S Cairns, Laura Hobley, and Nicola R Stanley-Wall. Bioﬁlm formation by Bacillus subtilis: new insights into regulatory strategies and assembly mechanisms. Molecular microbiology, 93(4):587–598, 2014.

[44] Hera Vlamakis, Yunrong Chai, Pascale Beauregard, Richard Losick, and Roberto Kolter. Sticking together: building a bioﬁlm the Bacillus subtilis way. Nature Re- views Microbiology, 11(3):157–168, 2013.

114 [45] Gaurav Pant, Anil Prakash, JVP Pavani, Sayantan Bera, GVNS Deviram, Ajay Ku- mar, Mitali Panchpuri, and Ravi Gyana Prasuna. Production, optimization and partial puriﬁcation of protease from Bacillus subtilis. Journal of Taibah University for Science, 9(1):50–55, 2015.

[46] Yun Xia Duan, Tao Chen, Xun Chen, and Xue Ming Zhao. Overexpression of glucose- 6-phosphate dehydrogenase enhances riboﬂavin production in Bacillus subtilis. Applied microbiology and biotechnology, 85(6):1907–1914, 2010.

[47] Simon T¨annler,Seraina Decasper, and Uwe Sauer. Maintenance metabolism and carbon ﬂuxes in bacillus species. Microbial cell factories, 7(1):19, 2008.

[48] Torsten Stein. Bacillus subtilis antibiotics: structures, syntheses and speciﬁc functions. Molecular microbiology, 56(4):845–857, 2005.

[49] Estibaliz Sansinenea and Aurelio Ortiz. Secondary metabolites of soil bacillus spp. Biotechnology letters, 33(8):1523–1538, 2011.

[50] Debarun Dhali, Fran¸coisCoutte, Anthony Arg¨uellesArias, Sandrine Auger, Vladimir Bidnenko, Gabrielle Chataign´e,Michael Lalk, Joachim Niehren, Joana Sousa, Cristian Versari, et al. Genetic engineering of the branched fatty acid metabolic pathway of Bacillus subtilis for the overproduction of surfactin c14 isoform. Biotechnology Journal, 2017.

[51] Long Liu, Yanfeng Liu, Hyun-dong Shin, Rachel R Chen, Nam Sun Wang, Jianghua Li, Guocheng Du, and Jian Chen. Developing bacillus spp. as a cell factory for production of microbial enzymes and industrially important biochemicals in the context of systems and synthetic biology. Applied microbiology and biotechnology, 97(14):6113–6127, 2013.

[52] Anne Goelzer, Fadia Bekkal Brikci, Isabelle Martin-Verstraete, Philippe Noirot, Philippe Bessieres, St´ephaneAymerich, and Vincent Fromion. Reconstruction and analysis of the genetic and metabolic regulatory networks of the central metabolism of Bacillus subtilis. BMC systems biology, 2(1):20, 2008.

[53] Daniel R Zeigler, ZoltánPrágai,Sabrina Rodriguez, Bastien Chevreux, Andrea Muf- fler, Thomas Albert, Renyuan Bai, Markus Wyss, and John B Perkins. The origins of 168, w23, and other Bacillus subtilis legacy strains. Journal of bacteriology, 190(21):6983–6995, 2008.

[54] Colin Harwood and Anil Wipat. Microbial Synthetic Biology, volume 40. Elsevier, 2013.

[55] Pierre Nicolas, Ulrike Mäder, Etienne Dervyn, Tatiana Rochat, Aurélie Leduc, Nathalie Pigeonneau, Elena Bidnenko, Elodie Marchadier, Mark Hoebeke, Stéphane

115 Aymerich, et al. Condition-dependent transcriptome reveals high-level regulatory ar- chitecture in Bacillus subtilis. Science, 335(6072):1103–1106, 2012.

[56] Bacillus genetic stock center. http://www.bgsc.org/. Accessed on 2018-09-20.

[57] Subtilist. http://genolist.pasteur.fr/SubtiList/. Accessed on 2018-09-20.

[58] Ivan Moszer, Louis M Jones, Sandrine Moreira, C´eciliaFabry, and Antoine Danchin. Subtilist: the reference database for the Bacillus subtilis genome. Nucleic acids research, 30(1):62–65, 2002.

[59] Genolist. http://genolist.pasteur.fr/. Accessed on 2018-09-20.

[60] Sporeweb. http://sporeweb.molgenrug.nl/. Accessed on 2018-09-20.

[61] Bsubcyc. https://bsubcyc.org/. Accessed on 2018-09-20.

[62] Dbtbs. http://dbtbs.hgc.jp/. Accessed on 2018-09-20.

[63] Raphael H Michna, Bingyao Zhu, Ulrike Mäder,and JörgStülke. Subti wiki 2.0—an integrated database for the model organism Bacillus subtilis. Nucleic acids research, 44(D1):D654–D662, 2015.

[64] Subtiwiki. http://subtiwiki.uni-goettingen.de/. Accessed on 2018-09-20.

[65] Leigh G Monahan, Isabella V Hajduk, Sinead P Blaber, Ian G Charles, and Eliza- beth J Harry. Coordinating bacterial cell division with nutrient availability: a role for glycolysis. MBio, 5(3):e00935–14, 2014.

[66] Richard B Weart, Amy H Lee, An-Chun Chien, Daniel P Haeusser, Norbert S Hill, and Petra Anne Levin. A metabolic sensor governing cell size in bacteria. Cell, 130(2):335–347, 2007.

[67] Stephen Vadia and Petra Anne Levin. Growth rate and cell size: a re-examination of the growth law. Current opinion in microbiology, 24:96–103, 2015.

[68] Anthony M Sperber and Jennifer K Herman. Metabolism shapes the cell. Journal of Bacteriology, 199(11):e00039–17, 2017.

[69] David William Adams and Jeﬀ Errington. Bacterial cell division: assembly, maintenance and disassembly of the z ring. Nature reviews. Microbiology, 7 9:642–53, 2009.

[70] Rut Carballido-L´opez and Alex Formstone. Shape determination in Bacillus subtilis. Current opinion in microbiology, 10(6):611–616, 2007.

[71] Kathrin Schirner, Jon Marles-Wright, Richard J Lewis, and Jeﬀ Errington. Distinct and essential morphogenic functions for wall-and lipo-teichoic acids in Bacillus subtilis. The EMBO journal, 28(7):830–842, 2009.

116 [72] Ling Juan Wu and Jeﬀ Errington. Coordination of cell division and chromosome segregation by a nucleoid occlusion protein in Bacillus subtilis. Cell, 117(7):915–925, 2004.

[73] Isabella V Hajduk, Christopher DA Rodrigues, and Elizabeth J Harry. Connecting the dots of the bacterial cell cycle: Coordinating chromosome replication and segregation with cell division. 53:2–9, 2016.

[74] Christopher DA Rodrigues and Elizabeth J Harry. The min system and nucleoid occlusion are not required for identifying the division site in Bacillus subtilis but ensure its eﬃcient utilization. PLoS genetics, 8(3):e1002561, 2012.

[75] Marc Bramkamp, Robyn Emmins, Louise Weston, Catriona Donovan, Richard A Daniel, and Jeﬀ Errington. A novel component of the division-site selection system of Bacillus subtilis and a new mode of action for the division inhibitor mincd. Molecular microbiology, 70(6):1556–1569, 2008.

[76] Herv´eJo¨elDefeu Soufo and Peter L Graumann. Actin-like proteins mreb and mbl from Bacillus subtilis are required for bipolar positioning of replication origins. Current biology, 13(21):1916–1920, 2003.

[77] Federica Laddomada, Mayara M Miyachiro, and Andrea Dessen. Structural insights into protein-protein interactions involved in bacterial cell wall biogenesis. Antibiotics, 5(2):14, 2016.

[78] Zemer Gitai, Natalie Dye, and Lucy Shapiro. An actin-like gene can determine cell polarity in bacteria. Proceedings of the National Academy of Sciences, 101(23):8643– 8648, 2004.

[79] Andrew K Fenton and Kenn Gerdes. Direct interaction of ftsz and mreb is required for septum synthesis and cell division in escherichia coli. The EMBO journal, 32(13):1953– 1965, 2013.

[80] William Stubbings and Harald Labischinski. New antibiotics for antibiotic-resistant bacteria. F1000 biology reports, 1, 2009.

[81] Jennifer Campbell, Atul K Singh, Jonathan G Swoboda, Michael S Gilmore, Brian J Wilkinson, and Suzanne Walker. An antibiotic that inhibits a late step in wall teichoic acid biosynthesis induces the cell wall stress stimulon in staphylococcus aureus. Antimicrobial agents and chemotherapy, 56(4):1810–1820, 2012.

[82] Marie-Pierre Chapot-Chartier and Saulius Kulakauskas. Cell wall structure and function in lactic acid bacteria. Microbial Cell Factories, 13(1):S9, Aug 2014.

117 [83] Jean van Heijenoort. Lipid intermediates in the biosynthesis of bacterial peptidoglycan. Microbiol Mol Biol Rev, 71, 2007.

[84] Athanasios Typas, Manuel Banzhaf, Carol A Gross, and Waldemar Vollmer. From the regulation of peptidoglycan synthesis to bacterial growth and morphology. Nature Reviews Microbiology, 10(2):123–136, 2012.

[85] Dirk-Jan Scheﬀers and Jeﬀery Errington. Pbp1 is a component of the Bacillus subtilis cell division machinery. Journal of bacteriology, 186(15):5153–5156, 2004.

[86] Jean van Heijenoort. Formation of the glycan chains in the synthesis of bacterial peptidoglycan. Glycobiology, 11(3):25R–36R, 2001.

[87] Matthew G Percy and Angelika Gr¨undling. Lipoteichoic acid synthesis and function in gram-positive bacteria. Annual review of microbiology, 68:81–100, 2014.

[88] Nathalie T Reichmann and Angelika Gr¨undling. Location, synthesis and function of glycolipids and polyglycerolphosphate lipoteichoic acid in gram-positive bacteria of the phylum ﬁrmicutes. FEMS microbiology letters, 319(2):97–105, 2011.

[89] Guoqing Xia, Thomas Kohler, and Andreas Peschel. The wall teichoic acid and lipoteichoic acid polymers of staphylococcus aureus. International Journal of Medical Mi- crobiology, 300(2):148–154, 2010.

[90] Francis C Neuhaus and James Baddiley. A continuum of anionic charge: structures and functions of d-alanyl-teichoic acids in gram-positive bacteria. Microbiology and Molecular Biology Reviews, 67(4):686–723, 2003.

[91] Nathalie T Reichmann, Carolina Picarra Cassona, and Angelika Gr¨undling.Revised mechanism of d-alanine incorporation into cell wall polymers in gram-positive bacteria. Microbiology, 159(9):1868–1877, 2013.

[92] Olaf Schneewind and Dominique Missiakas. Lipoteichoic acids, phosphate-containing polymers in the envelope of gram-positive bacteria. Journal of bacteriology, 196(6):1133–1142, 2014.

[93] Marie Brown, Warwick B Dunn, P Dobson, Y Patel, CL Winder, S Francis-McIntyre, P Begley, K Carroll, D Broadhurst, A Tseng, et al. Mass spectrometry tools and metabolite-speciﬁc databases for molecular identiﬁcation in metabolomics. Analyst, 134(7):1322–1332, 2009.

[94] Edward WC Sewell and Eric D Brown. Taking aim at wall teichoic acid synthesis: new biology and new leads for antibiotics. The Journal of antibiotics, 67(1):43–51, 2014.

118 [95] Michael A D’Elia, Kathryn E Millar, Terry J Beveridge, and Eric D Brown. Wall teichoic acid polymers are dispensable for cell viability in Bacillus subtilis. Journal of bacteriology, 188(23):8313–8316, 2006.

[96] Blazenka Soldo, Vladimir Lazarevic, Marco Pagni, and Dimitri Karamata. Teichuronic acid operon of Bacillus subtilis 168. Molecular microbiology, 31(3):795–805, 1999.

[97] Amit P Bhavsar, Laura K Erdman, Jeﬀrey W Schertzer, and Eric D Brown. Tei- choic acid is an essential polymer in Bacillus subtilis that is functionally distinct from teichuronic acid. Journal of bacteriology, 186(23):7865–7873, 2004.

[98] J Wright and John E Heckels. The teichuronic acid of cell walls of Bacillus subtilis w23 grown in a chemostat under phosphate limitation. Biochemical Journal, 147(1):187– 189, 1975.

[99] Michael A D’Elia, Kathryn E Millar, Amit P Bhavsar, Ana M Tomljenovic, Bernd Hutter, Christoph Schaab, Gabriel Moreno-Hagelsieb, and Eric D Brown. Probing teichoic acid genetics with bioactive molecules reveals new interactions among diverse processes in bacterial cell wall biogenesis. Chemistry & biology, 16(5):548–556, 2009.

[100] Alex Formstone, Rut Carballido-López, Philippe Noirot, Jeffery Errington, and Dirk- Jan Scheffers. Localization and interactions of teichoic acid synthetic enzymes in Bacillus subtilis. Journal of bacteriology, 190(5):1812–1821, 2008.

[101] Jonathan G Swoboda, Jennifer Campbell, Timothy C Meredith, and Suzanne Walker. Wall teichoic acid function, biosynthesis, and inhibition. Chembiochem, 11(1):35–45, 2010.

[102] Nathalie T Reichmann, Carolina Pi¸carraCassona, JoãoM Monteiro, Amy L Bottom- ley, Rebecca M Corrigan, Simon J Foster, Mariana G Pinho, and Angelika Gründling. Differential localization of lta synthesis proteins and their interaction with the cell division machinery in staphylococcus aureus. Molecular microbiology, 92(2):273–286, 2014.

[103] Andreas Peschel, Michael Otto, Ralph W Jack, Hubert Kalbacher, G¨unther Jung, and Friedrich G¨otz.Inactivation of the dlt operon instaphylococcus aureus confers sensitivity to defensins, protegrins, and other antimicrobial peptides. Journal of Biological Chemistry, 274(13):8405–8410, 1999.

[104] Vladimir Lazarevic, Blazenka Soldo, NoëlMédico,Harold Pooley, Sierd Bron, and Dimitri Karamata. Bacillus subtilis α-phosphoglucomutase is required for normal cell morphology and biofilm formation. Applied and environmental microbiology, 71(1):39– 45, 2005.

119 [105] Christopher Weidenmaier and Andreas Peschel. Teichoic acids and related cell-wall glycopolymers in gram-positive physiology and host interactions. Nature Reviews Mi- crobiology, 6(4):276–287, 2008.

[106] Leigh G Monahan and Elizabeth J Harry. You are what you eat: Metabolic control of bacterial division. Trends in microbiology, 24(3):181–189, 2016.

[107] Daniel P Haeusser and William Margolin. Splitsville: structural and functional insights into the dynamic bacterial z ring. Nature Reviews Microbiology, 14(5):305–319, 2016.

[108] Zhiwei Pan, Tao Zhu, Nathan Domagalski, Saleem Khan, Richard R Koepsel, Michael M Domach, and Mohammad M Ataai. Regulating expression of pyruvate kinase in Bacillus subtilis for control of growth rate and formation of acidic byprod- ucts. Biotechnology progress, 22(5):1451–1455, 2006.

[109] Teddy Charbonnier, Dominique Le Coq, Stephen McGovern, Magali Calabre, Olivier Delumeau, Stephane Aymerich, and Matthieu Jules. Molecular and physiological logics of the pyruvate-induced response of a novel transporter in Bacillus subtilis. MBio, 8(5):e00976–17, 2017.

[110] Reinhold Br¨uckner and Fritz Titgemeyer. Carbon catabolite repression in bacteria: choice of the carbon source and autoregulatory limitation of sugar utilization. FEMS microbiology letters, 209(2):141–148, 2002.

[111] Boris Görke and JörgStülke. Carbon catabolite repression in bacteria: many ways to make the most out of nutrients. Nature Reviews Microbiology, 6(8):613–624, 2008.

[112] Roelco J Kleijn, Joerg M Buescher, Ludovic Le Chat, Matthieu Jules, Stephane Aymerich, and Uwe Sauer. Metabolic ﬂuxes during strong carbon catabolite repression by malate in Bacillus subtilis. Journal of Biological Chemistry, 285(3):1587–1596, 2010.

[113] Pavel Kotrba, Masayuki Inui, and Hideaki Yukawa. Bacterial phosphotransferase system (pts) in carbohydrate uptake and control of carbon metabolism. Journal of Bioscience and Bioengineering, 92(6):502–517, 2001.

[114] Josef Deutscher, Christof Francke, and Pieter W. Postma. How phosphotransferase system-related protein phosphorylation regulates carbohydrate metabolism in bacteria. Microbiology and Molecular Biology Reviews, 70(4):939–1031, 2006.

[115] Yasutaro Fujita. Carbon catabolite control of the metabolic network in Bacillus subtilis. Bioscience, biotechnology, and biochemistry, 73(2):245–259, 2009.

[116] Marielle H. van den Esker, Akos´ T. Kov´acs,and Oscar P. Kuipers. From cell death to metabolism: Holin-antiholin homologues with new functions. MBio, 8(6):e01963–17, 2017.

120 [117] Marielle H. van den Esker, Akos´ T. Kov´acs,and Oscar P. Kuipers. Ysba and lytst are essential for pyruvate utilization in Bacillus subtilis. Environmental microbiology, 19(1):83–94, 2017.

[118] Matthieu Jules. The logics of metabolic regulation in bacteria challenges biosensor- based metabolic engineering. Microbial Cell, 5(1):56–56, 2018.

[119] Jad Sassine. Cell wall synthesis in Bacillus subtilis. 2017.

[120] Iris Fedtke, Diana Mader, Thomas Kohler, Hermann Moll, Graeme Nicholson, Raja Biswas, Katja Henseler, Friedrich G¨otz,Ulrich Z¨ahringer,and Andreas Peschel. A staphylococcus aureus ypfp mutant with strongly reduced lipoteichoic acid (lta) content: Lta governs bacterial surface properties and autolysin activity. Molecular microbiology, 65(4):1078–1091, 2007.

[121] AH Hughes, IC Hancock, and J Baddiley. The function of teichoic acids in cation control in bacterial membranes. Biochemical Journal, 132(1):83–93, 1973.

[122] Obaidur Rahman, Lynn G Dover, and Iain C Sutcliﬀe. Lipoteichoic acid biosynthesis: two steps forwards, one step sideways? Trends in microbiology, 17(6):219–225, 2009.

[123] Natividad Cabrera-Valladares, Luz M Mart’ınez, Noem’ı Flores, Georgina Hernández- Chávez, Alfredo Mart’ınez, Francisco Bol’ıvar, and Guillermo Gosset. Physiologic consequences of glucose transport and phosphoenolpyruvate node modifications in Bacillus subtilis 168. Journal of molecular microbiology and biotechnology, 22(3):177– 197, 2012.

[124] B Fry, T Zhu, MM Domach, RR Koepsel, C Phalakornkule, and MM Ataai. Charac- terization of growth and acid formation in a Bacillus subtilis pyruvate kinase mutant. Applied and environmental microbiology, 66(9):4045–4049, 2000.

[125] Jefrey Lodge and Gary R Jacobson. Starvation-induced stimulation of sugar uptake in streptococcus mutans is due to an eﬀect on the activities of preexisting proteins of the phosphotransferase system. Infection and immunity, 56(10):2594–2600, 1988.

[126] PW Postma, JW Lengeler, and GR Jacobson. Phosphoenolpyruvate: carbohydrate phosphotransferase systems of bacteria. Microbiological reviews, 57(3):543–594, 1993.

[127] Silvia Zorrilla, Denis Chaix, Alvaro Ortega, Carlos Alfonso, Thierry Doan, Emmanuel Margeat, Germ’ n Rivas, Stephan Aymerich, Nathalie Declerck, and Catherine A Royer. Fructose-1, 6-bisphosphate acts both as an inducer and as a structural cofactor of the central glycolytic genes repressor (cggr). Biochemistry, 46(51):14996–15008, 2007.

121 [128] Holger Ludwig, Georg Homuth, Matthias Schmalisch, Frank M. Dyka, Michael Hecker, and J¨orgSt¨ulke. Transcription of glycolytic genes and operons in Bacillus subtilis: evidence for the presence of multiple levels of control of the gapa operon. Molecular Microbiology, 41(2):409–422, 2001.

[129] Thierry Doan and St’phane Aymerich. Regulation of the central glycolytic genes in Bacillus subtilis: binding of the repressor cggr to its single dna target sequence is modulated by fructose-1, 6-bisphosphate. Molecular Microbiology, 47(6):1709–1721, 2003.

[130] Michael Dauner, Tazio Storni, and Uwe Sauer. Bacillus subtilis metabolism and ener- getics in carbon-limited and excess-carbon chemostat culture. Journal of bacteriology, 183(24):7308–7317, 2001.

[131] Marcel Emmerling, Michael Dauner, Aaron Ponti, Jocelyne Fiaux, Michel Hochuli, Thomas Szyperski, Kurt W¨uthrich, J. E. Bailey, and Uwe Sauer. Metabolic ﬂux responses to pyruvate kinase knockout in escherichia coli. Journal of Bacteriology, 184(1):152–164, 2002.

[132] Eugenio Meza, Judith Becker, Francisco Bolivar, Guillermo Gosset, and Christoph Wittmann. Consequences of phosphoenolpyruvate: sugar phosphotranferase system and pyruvate kinase isozymes inactivation in central carbon metabolism ﬂux distribution in escherichia coli. Microbial Cell Factories, 11(1):127–127, 2012.

[133] Yoshihiro Toya, Nobuyoshi Ishii, Kenji Nakahigashi, Takashi Hirasawa, Tomoyoshi Soga, Masaru Tomita, and Kazuyuki Shimizu. 13c-metabolic ﬂux analysis for batch culture of escherichia coli and its pyk and pgi gene knockout mutants based on mass iso- topomer distribution of intracellular metabolites. Biotechnology Progress, 26(4):975– 992, 2010.

[134] Oliver Schilling, Oliver Frick, Christina Herzberg, Armin Ehrenreich, Elmar Heinzle, Christoph Wittmann, and JörgStülke. Transcriptional and metabolic responses of Bacillus subtilis to the availability of organic acids: transcription regulation is important but not sufficient to account for metabolic adaptation. Applied and environmental microbiology, 73(2):499–507, 2007.

[135] Zhaopeng Li, Manfred Nimtz, and Ursula Rinas. The metabolic potential of escherichia coli bl21 in deﬁned and rich medium. Microbial cell factories, 13(1):45, 2014.

[136] Abraham L Sonenshein. Control of key metabolic intersections in Bacillus subtilis. Nature Reviews Microbiology, 5(12):917–927, 2007.

[137] Goutham N. Vemuri, Elliot Altman, Dipen Sangurdekar Sangurdekar, Arkady B. Kho- dursky, and Mark A. Eiteman. Overﬂow metabolism in escherichia coli during steady-

122 state growth: transcriptional regulation and eﬀect of the redox ratio. Applied and environmental microbiology, 72(5):3653–3661, 2006.

[138] Mariola Szenk, Ken A. Dill, and Adam M. R. de Graﬀ. Why do fast-growing bacteria enter overﬂow metabolism? testing the membrane real estate hypothesis. Cell systems, 5(2):95–104, 2017.

[139] Alan J Wolfe. The acetate switch. Microbiology and molecular biology reviews, 69(1):12–50, 2005.

[140] Frank J Grundy, Debra A Waters, SH Allen, and TM Henkin. Regulation of the Bacillus subtilis acetate kinase gene by ccpa. Journal of bacteriology, 175(22):7348– 7355, 1993.

[141] Hanna Meyer, Hendrikje Weidmann, Ulrike Mäder,Michael Hecker, Uwe Völker, and Michael Lalk. A time resolved metabolomics study: the influence of different carbon sources during growth and starvation of Bacillus subtilis. Molecular BioSystems, 10(7):1812–1823, 2014.

[142] Zijun Xiao and Ping Xu. Acetoin metabolism in bacteria. Critical reviews in microbiology, 33(2):127–140, 2007.

[143] Hugo Cruz Ramos, Tamara Hoﬀmann, Marco Marino, Hafed Nedjari, Elena Presecan- Siedel, Oliver Dreesen, Philippe Glaser, and Dieter Jahn. Fermentative metabolism of Bacillus subtilis: physiology and regulation of gene expression. Journal of Bacteriol- ogy, 182(11):3072–3080, 2000.

[144] Andrew J Turinsky, Frank J Grundy, Jeong-Ho Kim, Glenn H Chambliss, and Tina M Henkin. Transcriptional activation of the Bacillus subtilis ackA gene requires sequences upstream of the promoter. Journal of bacteriology, 180(22):5961–5967, 1998.

[145] Elisabeth H¨artigand Dieter Jahn. Regulation of the anaerobic metabolism in Bacillus subtilis, volume 61, pages 195–216. Elsevier, 2012.

[146] Kirsten Dörriesand Michael Lalk. Metabolic footprint analysis uncovers strain specific overflow metabolism and d-isoleucine production of staphylococcus aureus col and hg001. PLoS One, 8(12):e81500, 2013.

[147] Kenji Manabe, Yasushi Kageyama, Takuya Morimoto, Tadahiro Ozawa, Kazuhisa Sawada, Keiji Endo, Masatoshi Tohata, Katsutoshi Ara, Katsuya Ozaki, and Naotake Ogasawara. Combined eﬀect of improved cell yield and increased speciﬁc productivity enhances recombinant enzyme production in genome-reduced Bacillus subtilis strain mgb874. Applied and environmental microbiology, 77(23):8370–8381, 2011.

123 [148] Boris R. Belitsky, Hyun-Jin Kim, and Abraham L. Sonenshein. Ccpa-dependent regulation of Bacillus subtilis glutamate dehydrogenase gene expression. Journal of Bac- teriology, 186(11):3392–3398, 2004.

[149] Cuauhtemoc Licona-Cassani, Alvaro R Lara, Natividad Cabrera-Valladares, Adelfo Escalante, Georgina Hernandez-Chavez, Alfredo Martinez, Francisco Bol’ıvar, and Guillermo Gosset. Inactivation of pyruvate kinase or the phosphoenolpyruvate: sugar phosphotransferase system increases shikimic and dehydroshikimic acid yields from glucose in Bacillus subtilis. Journal of molecular microbiology and biotechnology, 24(1):37–45, 2014.

[150] Dong-Feng Liu, Guo-Min Ai, Qing-Xiang Zheng, Chang Liu, Cheng-Ying Jiang, Li- Xia Liu, Bo Zhang, Yi-Ming Liu, Chen Yang, and Shuang-Jiang Liu. Metabolic ﬂux responses to genetic modiﬁcation for shikimic acid production by Bacillus subtilis strains. Microbial Cell Factories, 13(1):40–40, 2014.

[151] Ming Jiang and Haoran Zhang. Engineering the shikimate pathway for biosynthesis of molecules with pharmaceutical activities in e. coli. Current opinion in biotechnology, 42:1–6, 2016.

[152] Astrid G Chapman, Lana Fall, and Daniel E Atkinson. Adenylate energy charge in escherichia coli during growth and starvation. Journal of Bacteriology, 108(3):1072– 1086, 1971.

[153] Hanna Meyer, Hendrikje Weidmann, and Michael Lalk. Methodological approaches to help unravel the intracellular metabolome of Bacillus subtilis. Microbial cell factories, 12(1):69, 2013.

[154] Guennadi Sezonov, Dani`eleJoseleau-Petit, and Richard D’Ari. Escherichia coli physiology in luria-bertani broth. Journal of bacteriology, 189(23):8746–8749, 2007.

[155] Jonathan Solomon, Laura Su, Stanley Shyn, and Alan D Grossman. Isolation and characterization of mutants of the Bacillus subtilis oligopeptide permease with altered speciﬁcity of oligopeptide transport. Journal of bacteriology, 185(21):6425–6433, 2003.

[156] Carlos Guijas, J Rafael Montenegro-Burke, Xavier Domingo-Almenara, Amelia Palermo, Benedikt Warth, Gerrit Hermann, Gunda Koellensperger, Tao Huan, Winnie Uritboonthai, Aries E Aisporna, et al. Metlin: a technology platform for identifying knowns and unknowns. Analytical chemistry, 90(5):3156–3164, 2018.

[157] Metlin. https://metlin.scripps.edu/landing_page.php?pgcontent=mainPage. Accessed on 2018-05-17.

[158] Michael K Dahl. Enzyme iiglc contributes to trehalose metabolism in Bacillus subtilis. FEMS microbiology letters, 148(2):233–238, 1997.

124 [159] L B¨urklen,F Sch¨ock, and MK Dahl. Molecular analysis of the interaction between the Bacillus subtilis trehalose repressor trer and the tre operator. Molecular and General Genetics MGG, 260(1):48–55, 1998.

[160] Kazunori Sawada, Masaru Wada, Takuya Hagiwara, Susumu Zen-in, Keita Imai, and Atsushi Yokota. Eﬀect of pyruvate kinase gene deletion on the physiology of corynebacterium glutamicum atcc13032 under biotin-suﬃcient non-glutamate-producing conditions: enhanced biomass production. Metabolic Engineering Communications, 2:67–75, 2015.

[161] TjaˇsaDanevˇciˇc,Leif Rilfors, Janez Strancar,ˇ G¨oranLindblom, and David Stopar. Ef- fects of lipid composition on the membrane activity and lipid phase behaviour of vibrio sp. dsm14379 cells grown at various nacl concentrations. Biochimica et Biophysica Acta (BBA)-Biomembranes, 1712(1):1–8, 2005.

[162] Sara E Diomandé, Christophe Nguyen-The, Marie-HélèneGuinebretière,Véronique Broussolle, and Julien Brillard. Role of fatty acids in bacillus environmental adaptation. Frontiers in microbiology, 6, 2015.

[163] Ayako Nishibori, Jin Kusaka, Hiroshi Hara, Masato Umeda, and Kouji Matsumoto. Phosphatidylethanolamine domains and localization of phospholipid synthases in Bacillus subtilis membranes. Journal of bacteriology, 187(6):2163–2174, 2005.

[164] Yiling Yang, Abiola M. Pollard, Carolin Höfler,Gernot Poschet, Markus Wirtz, Rüdi- ger Hell, and Victor Sourjik. Relation between chemotaxis and consumption of amino acids in bacteria. Molecular Microbiology, 96(6):1272–1282, 2015.

[165] Fabian M Commichau, Fabian M Rothe, Christina Herzberg, Eva Wagner, Daniel Hellwig, Martin Lehnik-Habrink, Elke Hammer, Uwe Völker, and JörgStülke. Novel activities of glycolytic enzymes in Bacillus subtilis interactions with essential proteins involved in mrna processing. Molecular & Cellular Proteomics, 8(6):1350–1360, 2009.

[166] Uwe Sauer and Bernhard J. Eikmanns. The pep—pyruvate—oxaloacetate node as the switch point for carbon ﬂux distribution in bacteria. FEMS microbiology reviews, 29(4):765–794, 2005.

[167] Yasutaro Fujita, Hiroshi Matsuoka, and Kazutake Hirooka. Regulation of fatty acid metabolism in bacteria. Molecular microbiology, 66(4):829–839, 2007.

[168] John E Cronan and Grover L Waldrop. Multi-subunit acetyl-coa carboxylases. Progress in lipid research, 41(5):407–435, 2002.

[169] Gustavo E Schujman, Marcelo Guerin, Alejandro Buschiazzo, Francis Schaeﬀer, Leti- cia I Llarrull, Georgina Reh, Alejandro J Vila, Pedro M Alzari, and Diego De Mendoza.

125 Structural basis of lipid biosynthesis regulation in gram-positive bacteria. The EMBO journal, 25(17):4074–4083, 2006.

[170] Joshua B Parsons and Charles O Rock. Bacterial lipids: metabolism and membrane homeostasis. Progress in lipid research, 52(3):249–276, 2013.

[171] Daniela Albanesi and Diego De Mendoza. Fapr: From control of membrane lipid homeostasis to a biotechnological tool. Frontiers in molecular biosciences, 3, 2016.

[172] Gustavo E Schujman, Luciana Paoletti, Alan D Grossman, and Diego de Mendoza. Fapr, a bacterial transcription factor involved in global regulation of membrane lipid biosynthesis. Developmental cell, 4(5):663–672, 2003.

[173] Luc’ıa Porrini, Larisa E Cybulski, Silvia G Altabe, Mar’ıa C Mansilla, and Diego Mendoza. Cerulenin inhibits unsaturated fatty acids synthesis in Bacillus subtilis by modifying the input signal of desk thermosensor. Microbiologyopen, 3(2):213–224, 2014.

[174] Maryam Lahooti and Colin R Harwood. Transcriptional analysis of the Bacillus subtilis teichuronic acid operon. Microbiology, 145(12):3409–3417, 1999.

[175] Dina Petranovic, Ole Michelsen, Ksenija Zahradka, Catarina Silva, Mirjana Petra- novic, Peter Ruhdal Jensen, and Ivan Mijakovic. Bacillus subtilis strain deﬁcient for the protein-tyrosine kinase ptka exhibits impaired dna replication. Molecular microbiology, 63(6):1797–1805, 2007.

[176] Michael Y Kiriukhin, Dmitri V Debabov, Dean L Shinabarger, and Francis C Neuhaus. Biosynthesis of the glycolipid anchor in lipoteichoic acid of staphylococcus aureus rn4220: role of ypfp, the diglucosyldiacylglycerol synthase. Journal of bacteriology, 183(11):3506–3514, 2001.

[177] Jun Kasahara, Yuuka Kiriyama, Mari Miyashita, Takuma Kondo, Takeshi Yamada, Kazuya Yazawa, Ritsuko Yoshikawa, and Hiroki Yamamoto. Teichoic acid polymers aﬀect expression and localization of dl-endopeptidase lyte required for lateral cell wall hydrolysis in Bacillus subtilis. Journal of bacteriology, 198(11):1585–1594, 2016.

[178] Warawan Eiamphungporn and John D Helmann. The Bacillus subtilis σm regulon and its contribution to cell envelope stress responses. Molecular microbiology, 67(4):830– 848, 2008.

[179] Bryson D Bennett, Elizabeth H Kimball, Melissa Gao, Robin Osterhout, Stephen J Van Dien, and Joshua D Rabinowitz. Absolute metabolite concentrations and implied enzyme active site occupancy in escherichia coli. Nature chemical biology, 5(8):593, 2009.

126 [180] Katrin Gunka and Fabian M Commichau. Control of glutamate homeostasis in Bacillus subtilis: a complex interplay between ammonium assimilation, glutamate biosynthesis and degradation. Molecular microbiology, 85(2):213–224, 2012.

[181] Fabian M Commichau, Christina Herzberg, Philipp Tripal, Oliver Valerius, and J¨org St¨ulke. A regulatory protein–protein interaction governs glutamate biosynthesis in Bacillus subtilis: the glutamate dehydrogenase rocg moonlights in controlling the transcription factor gltc. Molecular microbiology, 65(3):642–654, 2007.

[182] Boris R Belitsky and Abraham L Sonenshein. Role and regulation of Bacillus subtilis glutamate dehydrogenase genes. Journal of bacteriology, 180(23):6298–6305, 1998.

[183] Longzhu Cui, Hiroko Murakami, Kyoko Kuwahara-Arai, Hideaki Hanaki, and Kei- ichi Hiramatsu. Contribution of a thickened cell wall and its glutamine nonamidated component to the vancomycin resistance expressed by staphylococcus aureus mu50. Antimicrobial agents and chemotherapy, 44(9):2276–2285, 2000.

[184] Takenori Satomura, Daisuke Shimura, Kei Asai, Yoshito Sadaie, Kazutake Hirooka, and Yasutaro Fujita. Enhancement of glutamine utilization in Bacillus subtilis through the glnk-glnl two-component regulatory system. Journal of bacteriology, 187(14):4813– 4821, 2005.

[185] Jill M Zalieckas, Lewis V Wray, and Susan H Fisher. Expression of the Bacillus subtilis acsA gene: Position and sequence context aﬀect cre-mediated carbon catabolite repression. Journal of bacteriology, 180(24):6649–6654, 1998.

[186] Robert A Bender. Regulation of the histidine utilization (hut) system in bacteria. Microbiology and Molecular Biology Reviews, 76(3):565–584, 2012.

[187] Susan H Fisher, Kellie Rohrer, and Amy E Ferson. Role of cody in regulation of the Bacillus subtilis hut operon. Journal of bacteriology, 178(13):3779–3784, 1996.

[188] Yasuhiko Miwa, Mitsuyuki Saikawa, and Yasutaro Fujita. Possible function and some properties of the ccpa protein of bacillus subtilis. Microbiology, 140(10):2567–2575, 1994.

[189] Daniel E Atkinson. Energy charge of the adenylate pool as a regulatory parameter. interaction with feedback modiﬁers. Biochemistry, 7(11):4030–4034, 1968.

[190] Yusuke Oku, Kenji Kurokawa, Miki Matsuo, Sakuo Yamada, Bok-Luel Lee, and Kazuhisa Sekimizu. Pleiotropic roles of polyglycerolphosphate synthase of lipoteichoic acid in growth of staphylococcus aureus cells. Journal of bacteriology, 191(1):141–151, 2009.

127 [191] Stefan G Richter, Derek Elli, Hwan Keun Kim, Antoni PA Hendrickx, Joseph A Sorg, Olaf Schneewind, and Dominique Missiakas. Small molecule inhibitor of lipoteichoic acid synthesis is an antibiotic for gram-positive bacteria. Proceedings of the National Academy of Sciences, 110(9):3531–3536, 2013.

[192] Colin R Harwood and Simon M Cutting. Molecular biological methods for Bacillus. Wiley, 1990.

[193] C Anagnostopoulos and John Spizizen. Requirements for transformation in Bacillus subtilis. Journal of Bacteriology, 81(5):741, 1961.

[194] Leendert W Hamoen, Wiep Klaas Smits, Anne de Jong, Siger Holsappel, and Oscar P Kuipers. Improving the predictive value of the competence transcription factor (comk) binding site in Bacillus subtilis using a genomic approach. Nucleic acids research, 30(24):5517–5528, 2002.

[195] Imke G de Jong, Katrin Beilharz, Oscar P Kuipers, and Jan-Willem Veening. Live cell imaging of Bacillus subtilis and Streptococcus pneumoniae using automated time-lapse microscopy. Journal of visualized experiments: JoVE, (53), 2011.

128 Chapter 7

Supplementary information

Table 7.1: OD determined for wt and ∆pyk in M9GlcPyr, M9Pyr, and M9Glc in four biological replicates.

129 M9Pyr time (min) wt Δpyk 0 0.109 0.097 0.131 0.070 0.094 0.097 0.102 0.101 60 0.090 0.089 0.124 0.066 0.098 0.095 0.109 0.105 120 0.079 0.086 0.127 0.067 0.112 0.112 0.125 0.118 180 0.083 0.108 0.164 0.084 0.145 0.144 0.154 0.146 240 0.109 0.145 0.220 0.114 0.178 0.178 0.191 0.181 300 0.140 0.186 0.300 0.146 0.220 0.233 0.228 0.215 360 0.180 0.227 0.380 0.170 0.260 0.260 0.280 0.270 420 0.240 0.290 0.460 0.230 0.320 0.330 0.340 0.300 480 0.310 0.400 0.590 0.280 0.400 0.410 0.490 0.360 540 0.390 0.470 0.710 0.340 0.500 0.500 0.510 0.440 600 0.490 0.590 0.900 0.420 0.660 0.640 0.650 0.540

M9Glc time (min) wt Δpyk 0 0.096 0.108 0.082 0.103 0.094 0.097 0.101 0.103 60 0.147 0.156 0.127 0.163 0.104 0.105 0.109 0.112 120 0.242 0.262 0.234 0.296 0.127 0.123 0.129 0.132 180 0.430 0.450 0.460 0.570 0.143 0.143 0.151 0.154 240 0.680 0.790 0.840 1.100 0.164 0.164 0.174 0.174 300 1.160 1.330 1.620 2.200 0.188 0.188 0.199 0.204 360 1.940 2.280 2.480 3.240 0.213 0.214 0.228 0.233 420 3.200 3.950 4.120 4.280 0.239 0.244 0.258 0.265 480 5.010 5.430 4.810 4.480 0.290 0.300 0.294 0.302 540 4.580 4.320 4.210 3.890 0.330 0.340 0.330 0.350 600 3.320 3.090 3.700 3.270 0.380 0.370 0.390 0.390 660 2.160 2.130 2.900 2.480 0.400 0.400 0.420 0.440 720 1.590 1.700 2.130 1.940 0.460 0.500 0.470 0.490

130 Figure 7.1: Time-resolved extracellular metabolites concentrations under M9GlcPyr medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis ∆pyk. Data are shown as mean concentrations ± SD (shaded) of three biological replicates. 131 Figure 7.2: Time-resolved extracellular metabolite concentrations under M9Pyr medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis wt (black) and ∆pyk (purple) are displayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

132 Figure 7.3: Time-resolved extracellular metabolite concentrations under M9Glc medium cultivation. Absolute concentrations of consumed133 and secreted metabolites by B. subtilis wt. Data are shown as mean concentrations ± SD (shaded) of three biological replicates. Figure 7.4: Time-resolved extracellular metabolite concentrations under M9Glc medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis ∆pyk. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

134 (…)

Figure 7.5: Relative amount of intracellular PG precursors in wt and ∆pyk in minimal media. Data are presented as relative amount ± SD of three biological replicates. Statistical diﬀerences between control and mutants were considered signiﬁcant (∗) for p-values≤0.05.

135 Table 7.2: FC and statistical significance of metabolites of wt and ∆pyk in chemically defined media. FC and p-values were calculated using the mean values of four biological replicates. Statistical differences were determined using unpaired t-tests for each metabolite.

Fold Change (Δpyk/wt) p-value Metabolite M9GlcPyr M9Pyr M9Glc M9GlcPyr M9Pyr M9Glc 2-oxoglutarate 1.024 0.702 0.583 0.942 0.278 0.141 2-phosphoglycerate 3.185 1.560 21.946 0.057 0.546 0.001 3-hydroxybutyrate 1.275 3.121 21.185 0.303 0.093 0.356 3-phosphoglycerate 3.072 1.056 15.634 0.048 0.949 0.000 4-hydroxy-L-proline 2.984 1.312 0.978 0.502 0.818 0.986 5-methyluridine 0.466 1.184 2.981 0.058 0.278 0.049 5-oxo-proline 0.644 0.862 1.745 0.203 0.853 0.212 Acetyl adenylate 2.108 1.235 2.751 0.089 0.480 0.058 Acetyl-CoA 1.470 2.077 0.331 0.552 0.108 0.050 Adenylsuccinate 1.832 0.519 1.028 0.261 0.161 0.925 ADP 1.004 0.855 0.861 0.986 0.226 0.417 Alanine 0.705 0.771 1.407 0.267 0.588 0.418 AMP 2.144 0.728 0.858 0.005 0.013 0.429 Arginine 0.093 0.852 2.015 0.357 0.299 0.012 Asparagine 1.161 0.756 2.471 0.826 0.451 0.093 Aspartate 0.890 0.833 2.831 0.765 0.848 0.075 ATP 0.578 0.990 0.534 0.028 0.966 0.020 CDP 1.196 0.854 0.795 0.493 0.261 0.196 CDP-glycerol 0.760 0.849 0.709 0.387 0.589 0.266 Citrate 0.948 0.774 1.014 0.589 0.509 0.859 Citrulline 1.466 0.624 3.957 0.367 0.224 0.001 CMP 1.242 0.686 0.931 0.434 0.022 0.755 CTP 0.375 1.151 0.509 0.016 0.564 0.010 Cysteine 1.784 0.913 1.127 0.240 0.137 0.842 Cytidine 1.002 0.622 0.794 0.998 0.336 0.181 dAMP 1.066 0.691 0.391 0.826 0.040 0.004 dATP 0.509 1.029 0.330 0.083 0.893 0.001 dCDP 0.713 0.692 0.827 0.095 0.051 0.721 dCTP 0.268 0.991 0.220 0.004 0.968 <0.001 Deoxycytidine 1.209 0.616 3.931 0.726 0.300 0.011 Dihydroxyacetone P 2.299 1.034 7.061 0.075 0.909 0.001 dTDP 0.750 0.691 0.561 0.145 0.033 0.005 dTMP 0.469 0.611 0.403 0.015 0.030 0.008 dUMP 1.421 0.894 1.331 0.092 0.629 0.224 Erythrose 4-P 1.883 1.229 1.124 0.135 0.343 0.568 FAD 1.264 0.991 0.722 0.185 0.958 0.042 FAICAR 2.516 1.169 0.019 0.214

136 Fold Change (Δpyk/wt) p-value Metabolite M9GlcPyr M9Pyr M9Glc M9GlcPyr M9Pyr M9Glc Fructose 1,6bis-P 1.846 0.999 5.268 0.211 0.998 0.003 Fructose 6-P 4.571 1.467 11.685 0.025 0.596 0.001 Fumarate 1.443 1.018 0.841 0.472 0.966 0.497 GDP 1.441 0.997 0.935 0.044 0.980 0.710 Gluconate 6-P 1.427 1.380 0.783 0.233 0.204 0.043 Glucono 1,5lactone 6-P 0.650 1.037 2.228 0.236 0.944 0.513 Glucose 1.308 0.548 Glucose 6-P 5.017 1.975 8.324 0.019 0.511 0.001 Glutamate 0.628 0.919 1.715 0.132 0.907 0.160 Glutamine 0.641 0.746 2.884 0.335 0.764 0.109 Glycine 0.556 0.761 0.884 0.273 0.616 0.711 GMP 1.833 0.714 1.133 0.026 0.028 0.511 GTP 0.652 1.111 0.506 0.144 0.615 0.049 Histidine 0.876 1.682 1.043 0.427 0.356 0.921 IMP 0.395 0.754 0.206 0.001 0.233 0.005 Isoleucine 0.412 0.577 0.395 0.127 0.373 0.143 ITP 0.571 0.984 0.522 0.034 0.945 0.022 Lactate 0.754 1.186 0.731 0.436 0.476 0.530 Leucine 0.601 0.814 4.431 0.104 0.695 0.389 Lysine 0.390 0.591 1.673 0.186 0.397 0.118 Malate 2.721 1.062 0.872 0.361 0.873 0.215 Malonyl-CoA 0.448 2.422 0.597 0.389 0.328 0.220 Methionine 0.591 0.690 1.509 0.112 0.584 0.295 myo-inositole 0.496 1.122 0.720 0.368 0.824 0.787 Ornithine 1.138 0.766 5.932 0.821 0.721 0.011 Phenylalanine 0.730 0.897 1.272 0.447 0.809 0.591 Phenylpyruvate 0.901 1.384 2.735 0.653 0.422 0.005 Phosphoenolpyruvate 10.035 1.900 89.998 0.031 0.442 0.001 Proline 0.679 1.646 8.036 0.578 0.363 0.055 PRPP 0.820 0.821 0.633 0.416 0.291 0.093 Pyruvate 0.478 0.004 Ribose/Ribulose-P 1.184 0.944 0.997 0.346 0.775 0.984 SAICAR 0.916 0.749 0.150 0.812 0.223 0.003 Sedoheptulose 7-P 3.210 1.198 4.260 0.020 0.526 0.001 Serine 1.433 1.441 4.337 0.600 0.745 0.067 Shikimate 3-P 0.166 0.500 0.035 0.004 0.030 0.005 Succinate 0.836 0.783 0.863 0.465 0.563 0.563 Threonine 0.291 0.652 0.755 0.082 0.632 0.680 Tryptophan 0.751 1.586 1.515 0.406 0.452 0.260 Tyrosine 0.694 0.842 2.229 0.266 0.772 0.048 UDP 1.202 0.780 0.638 0.445 0.105 0.036

137 Fold Change (Δpyk/wt) p-value Metabolite M9GlcPyr M9Pyr M9Glc M9GlcPyr M9Pyr M9Glc UDP-GlucNAc 0.877 0.901 0.505 0.672 0.690 0.034 UDP-GlucNAcenolpyruvate 1.306 0.846 0.544 0.036 0.057 0.005 UDP-Glucose 1.005 0.891 1.148 0.986 0.604 0.517 UDP-Glucuronate 2.185 1.118 2.358 0.100 0.681 0.017 UDP-MurNAc 1.373 0.888 1.635 0.207 0.532 0.025 UDP-MurNAc-ala 1.340 1.266 0.635 0.451 0.379 0.028 UDP-MurNAc-ala-glu 1.341 1.295 0.372 0.212 0.037 <0.001 UDP-MurNAc-ala- 0.898 1.109 0.394 0.637 0.199 0.002 glu-pm-ala-ala UMP 1.051 0.701 0.571 0.788 0.047 0.003 Urea 0.861 1.093 1.08 0.566 0.807 0.737 UTP 0.400 0.976 0.295 0.022 0.924 0.008 Valine 0.532 0.564 1.297 0.077 0.511 0.450 XMP 0.888 0.739 0.622 0.519 0.199 0.099 XTP 0.635 1.004 0.569 0.107 0.982 0.051

138 0.001 M9GlcPyr M9Pyr 0.01

0.01 e u e l u a l a v 0.1 - v - p p 0.1

1 1 -4 -2 0 2 4 -2 -1 0 1 2

Log2 Fold Change Log2 Fold Change

0.000001 M9Glc

0.00001

0.0001 e u l a 0.001 v - p 0.01

0.1

1 -5 0 5 10

Log2 Fold Change

Figure 7.6: Volcano-plots of intracellular metabolome data of B. subtilis wt and ∆pyk in diﬀerent media. In volcano-plots are presented the FC given by the relative amount of wt and ∆pyk intracellular metabolites (in logarithmic scale) as function of unpaired t-tests (p- value results) under M9GlcPyr, M9Pyr, and M9Glc. Metabolites with signiﬁcant changes (p-value≤0.05) and -1>log2FC<1 are displayed in the upper left and right region of the plot (black dots).

139 Table 7.3: OD determined for wt and ∆pyk in LB medium in four biological replicates.

140 Table 7.4: FC and statistical signiﬁcance of metabolites of wt and ∆pyk in LB medium. FC and p-values were calculated using the mean values of four biological replicates. Statistical diﬀerences were determined using unpaired t-tests for each metabolite.

Metabolite Fold Change (Δpyk/wt) p-value 2-oxoglutarate 0.301 0.000 2-phosphoglycerate 3.103 0.013 3-hydroxybutyrate 0.610 0.042 3-phosphoglycerate 6.098 0.002 4-hydroxy-L-proline 0.307 0.058 5-methyluridine 0.755 0.011 5-oxo-proline 0.557 0.012 Acetyl adenylate 0.736 0.016 Adenylsuccinate 0.861 0.662 ADP 1.148 0.269 Alanine 0.544 0.008 AMP 1.319 0.377 Asparagine 0.067 0.000 Aspartate 0.505 0.003 ATP 0.771 0.375 beta-Alanine 0.351 0.129 CAIR 0.812 0.013 CDP 1.592 0.085 CDP-glycerol 4.695 0.017 Chorismate 0.728 0.037 Citrate 0.730 0.340 CMP 1.125 0.550 CoA 0.853 0.266 CTP 0.798 0.526 cUMP 0.613 0.006 Cysteine 0.349 0.025 Cytidine 0.808 0.023 dADP 1.384 0.037 dATP 0.877 0.709 dCDP 0.845 0.223 dCMP 0.579 0.001 dCTP 0.444 0.028 Deoxycytidine 0.671 0.012 Deoxyuridine 0.759 0.035 dGDP 1.150 0.239 Dihydroxyacetone P 1.635 0.010 dTDP 1.152 0.556

141 Metabolite Fold Change (Δpyk/wt) p-value dTMP 1.200 0.440 dTTP 0.536 0.128 dUMP 0.885 0.771 Erythrose 4-P 0.794 0.210 FAD 0.754 0.045 FAICAR 1.144 0.536 Fructose 6-P 2.655 0.060 Frutose 1,6-bis-P 2.865 0.129 Fumarate 0.504 0.001 GDP 0.809 0.194 Gluconate 0.657 0.011 Gluconate 6-P 5.783 0.014 Glucono 1,5-lactone 6-P 0.680 0.007 Glucose 2.607 0.335 Glucose 6-P 2.853 0.061 Glutamate 0.415 0.002 Glutamine 0.338 0.012 Glycerate 0.695 0.021 Glycine 0.577 0.018 GMP 1.199 0.414 GTP 0.417 0.070 Histidine 0.483 0.032 IMP 0.836 0.603 Isoleucine 0.561 0.011 ITP 0.647 0.094 Lactate 0.563 0.006 Leucine 0.588 0.006 Lysine 0.597 0.015 Malate 0.801 0.013 Methionine 0.553 0.014 Myo-inositole 0.603 0.017 NAD 1.221 0.307 Ornithine 0.937 0.698 Phenylalanine 0.575 0.012 Phenylpyruvate 0.772 0.003 Phosphoenolpyruvate 7.412 0.004 PRA 0.795 0.018 Proline 0.567 0.014 PRPP 0.540 0.222 Pyruvate 0.132 0.001 Ribulose 5-P 0.738 0.047 SAICAR 0.764 0.006

142 Metabolite Fold Change (Δpyk/wt) p-value Sedoheptulose 7-P 4.249 0.006 Serine 0.461 0.004 Shikimate 3-P 0.262 0.015 Succinate 0.567 0.002 Threonine 0.577 0.014 Tryptophan 0.642 0.118 Tyrosine 0.572 0.016 UDP 1.024 0.921 UDP-GlcA 0.648 0.184 UDP-GlucNAc/UDP-ManNAc 0.212 0.038 UDP-GlucNAcenolpyruvate 0.278 0.003 UDP-Glucose 1.121 0.672 UDP-MurNAc 1.116 0.663 UDP-MurNAc-ala 10.640 0.028 UDP-MurNAc-ala-glu 12.046 0.051 UDP-MurNAc-ala- 0.879 0.351 glu-pm-ala-ala UMP 1.031 0.880 Urea 0.785 0.041 UTP 0.422 0.178 Valine 0.555 0.011 Xanthosine 1.548 0.329 XMP 0.948 0.807 XTP 0.465 0.117

143 Table 7.5: OD determined for 168 and BSB1 wt and respective mutants in LB medium for three biological replicates.

144 Figure 7.7: Time-resolved extracellular metabolite concentrations of asparagine, aspartate and alanine under LB medium cultivation. Absolute concentrations of consumed metabolites by B. subtilis 168 (A) and BSB1 (B) in LB medium. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

145 Figure 7.8: Time-resolved extracellular metabolite concentrations of 2-methylbutyrate and isovalerate under LB medium cultivation. Absolute concentrations of secreted metabolites by B. subtilis 168 (A) and BSB1 (B) in LB medium. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

146 Table 7.6: Consumption of tryptophan, phenylalanine, lysine, and tyrosine during sampling of B. subtilis 168 and BSB1 in LB medium. Data are shown as mean concentrations ± SD of three biological replicates.

Figure 7.9: 1H-NMR spectrum of a sample in LB medium

147 Figure 7.10: Growth curves of wt and ∆pgcA M9GlcMalGlut medium cultivation. Data are shown as mean values ± SD of three biological replicates.

148 Table 7.7: FC and statistical signiﬁcance of metabolites of cell wall mutants in LB medium. For the calculation of FC in undetected metabolites, it was considered half of the lowest metabolite intensity detected.

Fold Change (Δ/wt) p-value Metabolite ΔpgcA ΔgtaB ΔugtP ΔpgcA ΔgtaB ΔugtP 2-oxoglutarate 1.169 1.180 1.116 0.370 0.141 0.354 2-phosphoglycerate 6.561 2.205 2.721 0.023 0.464 0.360 3-phosphoglycerate 2.806 1.221 1.466 0.019 0.458 0.194 5-methyluridine 1.224 1.132 1.097 0.032 0.531 0.231 5-oxoproline 1.079 1.024 1.093 0.308 0.857 0.521 Acetyl adenylate 1.212 1.130 1.045 0.079 0.309 0.505 Acetyl-CoA 1.663 1.453 1.321 0.016 0.071 0.506 Adenine 1.748 2.224 1.423 0.001 0.004 0.316 Adenosine 0.928 1.114 1.159 0.621 0.349 0.227 Adenylsuccinate 0.997 0.679 0.520 0.981 0.048 0.011 ADP 0.688 0.809 0.966 0.089 0.048 0.853 AICAR 1.267 1.293 1.109 0.354 0.169 0.537 AMP 0.575 0.740 1.405 0.220 0.417 0.182 Asparagine 0.909 1.227 1.500 0.337 0.349 0.016 Aspartate 1.410 1.413 1.192 0.005 0.053 0.319 ATP 1.566 1.265 1.112 0.001 0.163 0.679 CAIR 1.203 1.273 1.111 0.533 0.257 0.578 cCMP 1.124 1.064 1.017 0.394 0.746 0.895 cdiAMP 1.238 1.116 1.008 0.018 0.489 0.923 CDP 0.549 0.707 1.156 0.003 0.092 0.459 CDP-glucose 0.108 0.449 0.658 0.004 0.043 0.547 CDP-glycerol 9.189 1.707 3.924 0.002 0.112 0.151 Citrate 1.119 1.250 1.175 0.386 0.003 0.026 CMP 0.724 0.690 0.640 0.361 0.334 0.252 CoA 0.854 0.940 0.892 0.519 0.735 0.377 CTP 1.151 1.188 1.890 0.340 0.345 0.145 Cysteine 1.343 1.134 1.225 0.051 0.442 0.298 Cytidine 1.436 1.288 1.163 0.246 0.522 0.611 dADP 0.530 0.687 0.947 0.026 0.041 0.805 D-Ala-D-Ala 1.984 1.991 0.727 0.012 0.165 0.391 D-Alanine 1.353 1.387 1.327 0.541 0.511 0.594 dATP 1.257 1.028 1.047 0.097 0.858 0.902 dCDP 0.565 0.596 0.623 0.006 0.071 0.011 dCMP 1.187 1.106 1.042 0.210 0.555 0.714 dCTP 1.441 1.514 1.499 0.012 0.016 0.130 Deoxycytidine 1.122 0.931 0.922 0.261 0.709 0.376 Deoxythymidine 1.069 1.077 1.124 0.161 0.391 0.053 Deoxyuridine 1.113 1.085 1.117 0.345 0.666 0.346 dGDP 0.698 0.815 0.966 0.099 0.051 0.852 dGMP 0.569 0.737 1.433 0.219 0.411 0.169 Dihydroxyacetone Phosphate 11.038 1.128 2.213 0.008 0.935 0.639

149 Fold Change (Δ/wt) p-value Metabolite ΔpgcA ΔgtaB ΔugtP ΔpgcA ΔgtaB ΔugtP dTDP 0.579 0.614 1.118 0.059 0.046 0.360 dTTP 1.598 1.094 1.289 0.004 0.307 0.158 Erythrose 4-Phosphate 1.015 0.981 0.879 0.951 0.943 0.623 FAD 0.647 0.928 0.912 0.012 0.523 0.546 FGAR 1.000 1.315 1.516 0.999 0.431 0.200 Fructose 1,6-bis-Phopshate 2.936 1.348 1.211 0.006 0.544 0.516 Fructose 6-Phosphate 1795 1679 665 0.111 0.266 0.249 Fumarate 1.835 1.601 1.403 0.000 0.035 0.058 GAR 0.775 0.958 0.833 0.021 0.671 0.160 GDP 0.510 0.748 0.717 0.058 0.146 0.163 Gluconate 1.013 0.931 1.168 0.888 0.610 0.576 GluconateP 1.087 1.030 1.086 0.259 0.778 0.142 Glucose 0.842 0.952 1.004 0.031 0.424 0.903 Glucose 6-Phosphate 796 404 229 0.015 0.248 0.221 Glucuronate 6.870 0.037 0.037 0.042 0.374 0.374 Glutamate 1.106 1.844 1.424 0.833 0.040 0.029 Glutamine 22.77 18.16 8.65 0.001 0.037 0.011 Glycerate 1.263 1.199 1.130 0.051 0.065 0.020 Glycine 1.098 1.066 1.040 0.293 0.564 0.850 GMP 1.200 1.112 1.105 0.035 0.531 0.136 GTP 1.456 1.243 0.930 0.012 0.209 0.823 Guanine 1.344 0.980 0.882 0.013 0.917 0.323 Guanosine 1.161 0.974 1.508 0.593 0.938 0.259 Histidine 8.699 4.509 4.911 0.001 0.077 0.137 Hypoxanthine 1.038 0.974 1.072 0.828 0.814 0.644 IMP 0.571 0.712 1.331 0.182 0.315 0.204 Isoleucine 0.905 0.903 0.919 0.526 0.588 0.686 ITP 1.584 1.293 1.159 0.001 0.133 0.554 Lactate 1.258 1.192 1.211 0.019 0.074 0.068 Leucine 1.055 1.020 0.957 0.750 0.926 0.877 Lysine 1.002 1.046 1.089 0.982 0.708 0.657 Malate 1.578 1.482 0.957 0.010 0.009 0.630 Malonyl-CoA 4.674 2.969 2.139 0.001 0.019 0.364 Methionine 1.126 1.024 1.057 0.194 0.845 0.710 Myo-inositol 1.396 1.203 1.196 0.038 0.325 0.546 NAcGlucosamine-Phopshate 1.304 2.095 1.763 0.129 0.016 0.013 NAD 1.191 1.234 1.191 0.227 0.146 0.388 NADH 1.089 1.204 1.199 0.541 0.261 0.390 NADP 1.274 1.221 1.079 0.231 0.129 0.802 NADPH 1.156 1.238 1.201 0.417 0.103 0.588 Ornithine 1.262 1.120 1.170 0.022 0.393 0.391

150 Fold Change (Δ/wt) p-value Metabolite ΔpgcA ΔgtaB ΔugtP ΔpgcA ΔgtaB ΔugtP Phosphoenolpyruvate 4.085 1.731 2.021 0.029 0.267 0.165 PRA 1.096 1.158 1.197 0.473 0.412 0.168 Proline 0.913 0.899 0.864 0.605 0.626 0.599 PRPP 1.165 0.974 1.308 0.366 0.895 0.400 Pseudouridine 1.126 1.152 1.026 0.176 0.507 0.748 Pyruvate 0.570 0.673 0.722 0.035 0.070 0.086 Ribose 1,5bis-Phopshate 0.787 1.279 1.029 0.327 0.386 0.750 SAICAR 1.247 1.133 0.976 0.050 0.270 0.773 Sedoheptulose17bisP 1.863 1.247 0.956 0.046 0.666 0.854 Sedoheptulose7P 1.032 1.294 1.417 0.908 0.394 0.099 Serine 1.104 1.072 1.057 0.229 0.584 0.725 Succinate 1.066 1.153 1.254 0.475 0.230 0.063 SuccinylCoA 1.131 1.284 0.605 0.776 0.547 0.185 Threonine 1.096 1.028 1.052 0.293 0.807 0.742 Trehalose 6-Phosphate 0.716 1.197 1.065 0.370 0.663 0.877 Tyrosine 1.078 1.009 1.082 0.380 0.943 0.611 UDP 0.586 0.646 1.047 0.042 0.013 0.729 UDP-GlcA 0.001 0.001 0.485 0.000 0.000 0.011 UDP-GlcNAc 1.460 1.051 1.536 0.004 0.478 0.009 UDP-GlcNAc-enolpyruvate 0.854 0.789 1.026 0.453 0.149 0.845 UDP-Glucose 0.000 0.000 0.994 0.000 0.000 0.985 UDP-MurNAc 1.606 1.465 1.790 0.005 0.019 0.002 UDP-MurNAc-5aa 1.562 1.535 1.721 0.013 0.012 0.009 UDP-MurNAc-ala 2.500 3.129 5.588 0.001 0.000 0.018 UDP-MurNAc-ala-glu 1.554 1.818 2.734 0.001 0.000 0.030 UDP-MurNAc-ala-glu-pm 1.078 1.078 1.472 0.780 0.774 0.296 UMP 0.686 0.740 1.341 0.042 0.100 0.033 Uracil 0.976 1.080 1.020 0.730 0.270 0.921 Uridine 1.044 1.089 1.039 0.647 0.696 0.621 UTP 1.814 1.286 1.508 0.001 0.021 0.158 Valine 1.339 1.280 1.158 0.004 0.199 0.641 XMP 0.641 0.771 0.828 0.000 0.011 0.017 XTP 1.445 1.301 0.941 0.023 0.143 0.848

151 Table 7.8: OD determined for wt and ∆pgcA in M9GlcMalGlut in four biological replicates.

152 Figure 7.11: Time-resolved extracellular metabolite concentrations under M9GlcMalGlut medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis wt (black) and ∆pgcA (green) are displayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

153 Figure 7.12: Extracellular concentration proﬁles of BCAA metabolism of wt and ∆pgcA under M9GlcMalGlut medium cultivation. Absolute concentrations of consumed and secreted metabolites by B. subtilis wt (black) and ∆pgcA (green) are displayed. Data are shown as mean concentrations ± SD (shaded) of three biological replicates.

154

-3.0 0 3.0

OD OD

FC ΔpgcA

1.0 0.5

1.2 D D log2

O OD O FC wt

A) 0

1.2 0.5

B) G)

Figure 7.13: Heat-map of intracellular metabolite levels of intermediates of purine and pyrimidine metabolism (A), cell wall precursors (B), amino acids (C), glycolysis (D), PPP (E), TCA cycle (F), and cofactors (G) in wt and ∆pgcA under M9GlcMalGlut. Color-code represent the log2 FC ratio between mutant155 and control cells, whereas increased levels are indicated in red and lower levels in green. Data are shown as the mean of four biological replicates. Table 7.9: FC and statistical signiﬁcance of metabolites of wt and ∆pgcA in M9GlcMalGlut. FC and p-values were calculated using the mean values of three biological replicates. Sta- tistical diﬀerences were determined using unpaired t-tests for each metabolite.

Fold Change (Δ/wt) p-value Metabolite 0.5 OD 1.0 OD 1.2 OD 0.5 OD 1.0 OD 1.2 OD 2-oxoglutarate 0.622 0.419 0.612 0.364 0.054 0.278 3-phosphoglycerate 0.473 0.848 0.706 0.379 0.487 0.648 5-methyluridine 2.601 2.364 1.501 0.214 0.186 0.443 5-oxoproline 0.643 1.355 1.023 0.604 0.501 0.955 Acetyl adenylate 2.742 3.745 4.275 0.001 0.001 0.010 Acetyl-CoA 0.117 0.442 0.072 0.377 0.326 0.334 Adenosine 1.403 1.550 1.542 0.047 0.021 0.006 Adenylsuccinate 1.443 2.622 0.989 0.826 0.089 0.979 ADP 1.082 1.204 1.521 0.223 0.308 0.043 AIR 1.707 2.275 1.786 0.488 0.103 0.097 Alanine 1.288 0.638 1.001 0.856 0.409 0.998 AMP 1.909 1.474 1.641 0.053 0.061 0.339 Aspartate 2.848 1.719 1.335 0.240 0.403 0.797 ATP 0.814 1.107 1.333 0.689 0.679 0.178 CAIR 2.359 2.073 1.368 0.152 0.006 0.051 cCMP 1.703 1.300 1.108 0.283 0.579 0.828 CDP 1.152 1.518 2.035 0.196 0.221 0.085 CDP-glucose 0.190 0.364 0.430 0.110 0.281 0.191 CDP-glycerol 0.970 1.604 1.107 0.425 0.117 0.718 Citrate 0.870 0.897 0.964 0.877 0.893 0.969 CMP 1.687 3.119 1.968 0.336 0.258 0.361 CoA 0.085 0.396 0.056 0.409 0.393 0.370 CTP 0.942 1.238 1.825 0.626 0.671 0.019 Cytidine 3.053 3.306 5.049 0.025 0.058 0.004 dADP 1.128 1.272 1.507 0.562 0.321 0.108 D-Ala-D-Ala 0.861 0.957 1.183 0.453 0.544 0.054 dAMP 1.170 1.301 1.592 0.264 0.246 0.105 dATP 0.760 1.048 1.284 0.311 0.898 0.299 dCDP 1.227 1.801 1.911 0.082 0.098 0.031 dCMP 1.566 2.076 2.356 0.217 0.502 0.196 dCTP 0.968 1.189 1.382 0.227 0.634 0.208 Deoxyguanosine 0.540 3.167 0.937 0.842 0.290 0.929 Dihydroxyacetone P 1.161 1.468 1.154 0.156 0.101 0.882 dTDP 1.305 1.523 1.661 0.051 0.072 0.041 dTMP 0.811 1.206 1.125 0.933 0.496 0.583 dTTP 0.888 1.210 1.297 0.587 0.573 0.343 dUMP 1.044 1.600 1.676 0.738 0.179 0.061 dUTP 1.047 1.196 1.354 0.200 0.592 0.200 Erythrose 4-P 3.023 1.674 3.003 0.008 0.123 0.072 FAD 1.052 0.864 1.593 0.403 0.564 0.094 FADH2 1.097 0.775 1.615 0.360 0.507 0.076 FGAR 0.783 1.758 1.326 0.484 0.021 0.715 Frutose 1,6-bisP 0.639 1.376 1.213 0.349 0.164 0.753 Frutose 6-P 2.583 2.536 2.753 0.010 0.006 0.098 Fumarate 0.764 0.452 0.669 0.191 0.032 0.226 GAR 0.952 1.177 1.407 0.730 0.308 0.013 GDP 1.219 1.169 2.138 0.322 0.304 0.086 GDP-glucose 1.115 1.718 1.885 0.011 0.052 0.047

156 Fold Change (Δ/wt) p-value Metabolite 0.5 OD 1.0 OD 1.2 OD 0.5 OD 1.0 OD 1.2 OD Gluconate 0.558 0.574 0.757 0.311 0.246 0.552 Gluconate 6-P 1.566 2.093 3.412 0.016 0.060 0.219 Glucose 1.244 0.762 2.430 0.706 0.545 0.510 Glucose 6-P 4.881 3.575 3.529 0.036 0.023 0.200 Glucuronic acid 0.794 0.753 0.779 0.731 0.556 0.587 Glutamate 0.915 1.113 1.077 0.911 0.752 0.831 Glutamine 0.589 1.928 0.679 0.846 0.486 0.673 Glycine 0.560 0.484 0.656 0.411 0.253 0.198 GMP 1.436 1.241 2.694 0.135 0.509 0.098 GTP 0.789 0.961 1.346 0.138 0.910 0.588 IMP 1.344 1.239 1.358 0.164 0.424 0.664 Inosine 1.277 1.090 1.840 0.898 0.859 0.526 Isoleucine 1.079 2.601 2.277 0.686 0.229 0.199 Lactate 0.650 0.663 0.745 0.801 0.676 0.548 Lysine 0.419 0.744 0.702 0.250 0.551 0.539 Malate 1.886 0.898 0.970 0.300 0.868 0.953 Malonyl-CoA 0.372 0.536 0.221 0.153 0.381 0.163 Methionine 0.542 0.722 0.655 0.493 0.581 0.347 MurNAc-Ala 0.624 0.587 0.833 0.320 0.192 0.789 NAcGlucosamine 1.728 6.206 1.356 0.179 0.040 0.688 NAD 1.099 1.244 1.683 0.116 0.232 0.006 NADH 1.099 1.375 1.623 0.260 0.302 0.005 NADP 1.056 0.800 1.712 0.572 0.485 0.079 NADPH 1.044 1.097 1.670 0.568 0.410 0.046 Ornithine 0.452 0.755 0.405 0.454 0.652 0.235 Phosphenolpyruvate 0.330 0.596 0.466 0.251 0.343 0.431 PRA 0.619 0.731 0.674 0.525 0.490 0.271 Proline 0.594 0.536 0.592 0.595 0.306 0.399 PRPP 0.980 1.289 1.481 0.427 0.306 0.545 Pseudouridine 1.356 2.728 1.455 0.297 0.170 0.277 Pyruvate 0.761 0.461 0.816 0.779 0.189 0.442 Ribose/Ribulose 5-P 0.810 1.394 1.440 0.897 0.218 0.430 SAICAR 1.073 2.463 2.180 0.530 0.004 0.359 Sedoheptulose 1,7-bisP 0.671 0.992 1.436 0.696 0.964 0.332 Sedoheptulose 7-P 0.880 1.398 1.618 0.939 0.232 0.047 Serine 0.473 0.438 0.609 0.303 0.108 0.439 Succinate 0.948 0.506 0.705 0.988 0.072 0.287 Succinyl-CoA 0.592 0.251 0.079 0.672 0.350 0.258 Threonine 1.912 8.366 5.896 0.397 0.109 0.073 Trehalose 6-P 0.244 0.359 0.220 0.374 0.374 0.374 Tryptophan 0.564 0.637 0.914 0.529 0.466 0.750 Tyrosine 0.465 0.532 0.511 0.374 0.397 0.202 UDP 1.099 1.484 1.575 0.606 0.113 0.030 UDP-GlcA 0.000 0.000 0.000 0.000 0.027 0.004 UDP-GlucNAc 1.313 1.852 1.991 0.062 0.077 0.015 UDP-GlucNAc-enolpyruvate 0.691 0.730 1.019 0.142 0.335 0.856 UDP-Glucose 0.004 0.004 0.005 0.001 0.026 0.079 UDP-MurNAc 1.099 1.586 2.564 0.265 0.163 0.052

157 Fold Change (Δ/wt) p-value Metabolite 0.5 OD 1.0 OD 1.2 OD 0.5 OD 1.0 OD 1.2 OD UDP-MurNAc-5aa 2.700 2.369 2.853 0.002 0.003 0.002 UDP-MurNAc-ala 3.636 3.284 3.587 0.000 0.005 0.011 UDP-MurNAc-ala-glu 2.187 2.267 2.927 0.001 0.013 0.004 UDPMurNAc-ala-glu-pm 6.207 3.755 8.017 0.000 0.114 0.006 UMP 0.911 0.984 0.806 0.580 0.951 0.728 Urea 0.881 1.264 0.976 0.984 0.655 0.959 Uridine 1.389 1.873 2.668 0.341 0.409 0.260 UTP 0.722 1.050 1.050 0.202 0.904 0.901 Valine 0.734 0.178 0.311 0.698 0.299 0.296 Xanthosine 0.342 0.713 0.825 0.760 0.665 0.655 XMP 0.305 0.627 0.636 0.023 0.082 0.539 XTP 0.851 1.042 1.393 0.511 0.893 0.511

158 159 160 Joana Dias de Sousa

Education 2014–Present Ph.D candidate in Biochemistry and Microbiology Institute for Biochemistry, Ernst-Moritz- Arndt-Universität Greifswald, Greifswald, Germany 2008–2010 Master in Environmental and Human Biology Faculty of Sciences, University of Lisbon, Lisbon, Portugal 2004–2008 Bachelor in Biochemistry Faculty of Sciences, University of Lisbon, Lisbon, Portugal Master thesis title Exposure to heavy metals in the workplace: establishing biomarkers of exposure to pollutants supervisors Dr. Teresa Pinheiro, Dr. Marta Almeida, and Dr. Ana Crespo description Study of Human exposure to heavy metals in a batteries workplace manufacture. Evaluation of the exhaled breath condensate, a noninvasive method, as a potential biomarker of occupational exposure to metals in the workplace. Experience acquired in mass spec. ICP-MS and nuclear activation techniques INAA and PIXE. Collaboration in two scientiﬁc articles. Working Experience 2014-Present Researcher in Microbiology and Biochemistry under the Marie Skłodowska-Curie Euro- pean Fellowship Institute for Biochemistry, Ernst-Moritz-Arndt-Universität Greifswald, Greifswald, Germany ITN network "AMBER-Advance Multidisciplinary training in Molecular Bacteriology". Investigation of metabolic pathways variations, mainly the central carbon metabolism, in Bacillus subtilis. Supervisor: Prof. Dr. Michael Lalk

10.2013- Junior Researcher - Trainee European Commission - Joint Research Centre, Ispra, Italy 12.2013 Study of the metabonomic eﬀects in carcinoma cells (CaCo2 cell line) exposed to gold nanoparticles. Structure identiﬁcation of chemical compounds provided by European Union Customs. Supervisor: Dr. Claude Guillou

10.2011- Researcher in Photochemistry Faculty of Science and Technology, New University of Lisbon, 10.2012 Lisbon, Portugal Synthesis and functionalization of organic compounds for binding onto metallic nanoparticles. Application of developed nanomaterials in the control of the polymerization of nucleic acids. Supervisor: Prof. Dr. João Lima

07.2010- Trainee Directorate General of Health - Directorate for the Prevention and Disease Control, Lisbon, 06.2011 Portugal Support in the national epidemiological surveillance and vaccination programs that prevent and control communicable diseases. Supervisor: Doctor Ana Leça

07.2009- Trainee Centro Interdisciplinario de Investigación para el Desarrollo Integral Regional, Durango, 09.2009 Mexico Scholarship awarded by the Exchange of Students for Technical Experience (IASTE). Studies for the development of waste treatments and landﬁlls. Supervisor: Dr. Ignacio Villanueva

Rua Venâncio 81, Nogueiró – 4715-324 Braga – Portugal Æ +49 (0) 15 757 226 035 Ó +351 253 041 931 Q [email protected] • • ¯ joana-dias-de-sousa-5b923025 * 13th October 1986 • Languages Portuguese Native English Fluent Spanish Fluent German Basic Publications

+ Dhali, D., Coutte, F., Arias, A.A., Auger, S., Bidnenko, V., Chataigné, G., Lalk, M., Niehren, J., Sousa, J., Versari, C. and Jacques, P., 2017. Genetic engineering of the branched fatty acid metabolic pathway of Bacillus subtilis for the overproduction of surfactin C14 isoform. Biotechnology journal, 12(7), p.1600574. + Félix, P.M., Almeida, S.M., Pinheiro, T., Sousa, J., Franco, C. and Wolterbeek, H.T., 2013. Assessment of exposure to metals in lead processing industries. International journal of hygiene and environmental health, 216(1), pp.17-24. + Pinheiro, T., Barreiros, M.A., Alves, L.C., Félix, P.M., Franco, C., Sousa, J. and Almeida, S.M., 2011. Particulate matter in exhaled breath condensate: a promising indicator of environmental conditions. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 269(20), pp.2404-2408. Conferences and training schools

+ LC-MS Data Processing and Statistics in Metabolomics workshop – May, 2016. West Coast Metabolomics Center, UC Davis - Sacramento (USA) + Fluorescence Microscopy, Scientiﬁc writing, and Presentation skills workshops – September, 2015. Newcastle University - Newcastle upon Tyne (England) th + 8 International Conference on Gram-Positive Microorganisms – June, 2015. Montecatini, (Italy) – Poster: How do glycolytic mutants impact the metabolome in Bacillus subtilis? + Metabolomics summer school - organiser – September, 2014. Institute for Biochemistry, Ernst-Moritz-Arndt-Universität Greifswald (Ger- many) + Metabolomics Workshop and Symposium – May, 2014. University of Florida, Gainesville (USA) – Poster: Unravelling the impact of glycolytic mutants in metabolome of Bacillus subtilis

Rua Venâncio 81, Nogueiró – 4715-324 Braga – Portugal Æ +49 (0) 15 757 226 035 Ó +351 253 041 931 Q [email protected] • • ¯ joana-dias-de-sousa-5b923025 * 13th October 1986 • Publications

I Dhali, D., Coutte, F., Arias, A.A., Auger, S., Bidnenko, V., Chataign´e,G., Lalk, M., Niehren, J., Sousa, J., Versari, C. and Jacques, P., 2017. Genetic engineering of the branched fatty acid metabolic pathway of Bacillus subtilis for the overproduction of surfactin C14 isoform. Biotechnology journal, 12(7), p.1600574.

II Bohorquez, L. C., Sousa, J., Garcia-Garcia T., Jonker M. J., Noirot M. F., Lalk M. and Hamoen L. W., The conserved WhiA protein inﬂuences the fatty acid composition of the Bacillus subtilis cell membrane. in preparation.

III Sousa, J., Westhoﬀ, P., Methling, K. and Lalk, M., Pyruvate kinase mutation in Bacillus subtilis leads to the relieve of carbon catabolite repression of pyruvate uptake. in preparation.

Others:

IV F´elix,P.M., Almeida, S.M., Pinheiro, T., Sousa, J., Franco, C. and Wolterbeek, H.T., 2013. Assessment of exposure to metals in lead processing industries. International journal of hygiene and environmental health, 216(1), pp.17-24.

V Pinheiro, T., Barreiros, M.A., Alves, L.C., F´elix,P.M., Franco, C., Sousa, J. and Almeida, S.M., 2011. Particulate matter in exhaled breath condensate: a promising indicator of environmental conditions. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 269(20), pp.2404- 2408.

163 164 Acknowledgements

Firstly, I would like to express my sincere gratitude to Prof. Dr. Michael Lalk for the opportunity to work in this project and under his supervision. I also want to acknowledge Marie Sk lodowska-Curie actions and the European Com- mission for the 3 years of funding. My special thanks go to all colleagues that made this working experience enjoyable. A big thank to Hanna, who helped me with all her knowledge in cultivation of Bacillus subtilis and in the optimization of the sampling protocol during my ﬁrst weeks. Also to Kirsten, for always having time to answer most my random questions, for her wisdom in keeping our analytical precious properly working, and for her contagious laugh. I am particularly grateful to Philipp for his genuine interest on my work and for the valuable discussions. For always having time to tell me inspiring words and to give me strength in diﬃcult times. Also, for the morning updates on german news and for the company during the evenings’ workouts. I also would like to thank Karen for the help every time LC didn’t collaborate and Dr. Sabine Witt for sharing the laboratory facilities of the Baltic Analytics GmbH. My appreciation also goes to Simone for all the lab assistance given and for making the labs running. I wish to acknowledge the help provided by Dr. Rabea Schlueter in the TEM experiments. Besides all, I cannot forget to thank Hanna, Kirsten, Linda, and Philipp for the help in my integration since day one. It might have been bigger than you realized it. I am also grateful to Prof. Dr. Richard Daniel and Prof. Dr. Waldemar Vollmer who provided me the opportunity to join their working groups for a few months in the Centre for Bacterial Cell Biology in Newcastle. I would like to thank the guidance and the constructive discussions. I couldn’t forget to thank Jad for taking his time to share his knowledge in microscopy and for making my geordi experience a little bit more fun. I’m also thankful to be part of the nice ITN network AMBER group. Thank you for the discussions during meeting presentations but also while drinking some beers. In particular to Laura. It was great to meet you all!

I also want to thank my friends that I luckily have in all places I’ve been. In particular, to the support of my friends from Greifswald Regina, Maria, Myriam, Olga, Ane, Kasia, Musli, Ani, Marcella, Ingrid, Ewa, Regina, Cynthia, Alberto, Carlos, and Steve. Thank you very much for the nice time we’ve spent together. I was far away

165 to imagine the great people I would encounter when I arrived in this city. A special thanks to my friends Vanessa, Soﬁa, Vasco, Joana L., Inˆes,Lara, Joana V., and Diana for giving me their continuous support abroad. Thank you to Filipa for the silly to the deepest and encourage talks. Don’t have words to describe how meaningful they were. Foremost, I would like to extend my sincere gratitude to Patr´ıcia, Joana, Andreia, and Kika, for their friendship and support throughout these years and also for making things much easier when returning home. A special thank you to Carlos, for all the encouragement and for always caring about me.

Finally, I would like to thank my family for the constant love and unﬂinching support even if so far away. Obrigada!

166