University of Groningen

Engineering of sugar metabolism in Lactococcus lactis Pool, Weia Arianne

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

Publication date: 2008

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Pool, W. A. (2008). Engineering of sugar metabolism in Lactococcus lactis. s.n.

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

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

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

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

Download date: 06-10-2021

CHAPTER 3

FUNCTIONAL CHARACTERIZATION OF THREE DIFFERENT UPTAKE ROUTES IN LACTOCOCCUS LACTIS

Wietske A. Pool, Ana R. Neves, Helena Santos, Oscar P. Kuipers, Jan Kok.

Chapter 3

52 Glucose transport and metabolism

SUMMARY

Molecular and physiological characteristics of lactococcal strains harbouring mutations in three different glucose uptake or dissimilation routes were determined, to gain insight into the specific roles of the two glucose-transporting phosphoenolpyruvate-dependent phosphotransfer systems (PEP:PTS) (EIIglc/man and EIIcel/glc) and the non-PTS glucose transport route(s) in Lactococcus lactis. Disruption of any of the pathways resulted in a considerable change in glucose transport. EIIcel/glc and the non-PTS route(s), had a preference for transport of the β-anomer of glucose. The main route for glucose transport was via EIIglc/man. Furthermore, a regulatory role for glucokinase (Glk) is proposed. Disruption of either glk, EIIglc/man, EIIcel/glc or both PTS systems had significant effects on overall glucose metabolism: changes in the pools of intracellular metabolites (fructose-1,6-bisphosphate (FBP), 3-phosphoglycerate (3PGA) and phosphoenolpyruvate (PEP)), in end-products formed (PEP, 3PGA, lactate, , , 2,3-butanediol), and in key activities were observed. The lower (LDH) and pyruvate kinase (PK) activities, which might be caused by a lower FBP pool, together with transcriptional regulation, indicated by the induction of the gene encoding pyruvate formate lyase (pfl) in the glucose-PTS negative strain, resulted in the shifted pattern. This study demonstrates and specifies the roles that the three sugar uptake systems play in directing glucose uptake and subsequent metabolic processes in L. lactis.

53 Chapter 3

INTRODUCTION

Lactococcus lactis, a bacterium used in the dairy industry for the production of fermented milk products, has been studied intensely over the last decades. Its relatively simple carbon metabolism and the availability of a number of molecular cloning tools make L. lactis an attractive organism for in depth studies on sugar utilization (84, 96). In L. lactis, glucose is taken up either via a phosphoenolpyruvate-dependent phosphotransferase system (PEP:PTS), or it is imported via a non-PTS permease, and phophorylated by glucokinase (Fig. 1). The PTS is considered to be the main route for glucose metabolism (36, 148, 191). PEP:PTSs are group translocators, which import and phosphorylate sugars via a phosphoryl-transfer process. The sugar-specific enzyme II (EII) imports the sugar, while the non-sugar-specific phosphocarriers enzyme I (EI) and histidine protein (HPr) are involved in the phosphate cascade in which the phosphate moiety of PEP is transferred to the incoming sugar (42, 148, 205). Recently, we have shown that L. lactis can import glucose via two different PTS-systems, the mannose/glucose-PTS encoded by ptnABCD, and the cellobiose/glucose-PTS specified by ptcBAC (141). PTSman/glc consists of a cytosolic EIIAB protein and an integral membrane EIICD protein, the PTScel/glc is made up of a cytosolic EIIAB protein and an integral membrane EIIC protein. The glucokinase of L. lactis is an ATP-dependent protein of 33.8 kD (based on the nucleotide sequence of glk), which is able to phosphorylate intracellular glucose (195). The protein has three domains: an ATP-binding site at the N-terminus, a NagC-domain, specifying a possible transcriptional regulatory site, spanning almost the complete protein and a ROK-motif also present in certain sugar kinases and regulator proteins, in the middle of the protein. The main role of glucokinase in L. lactis is so far thought to be the prevention of accumulation of unphosphorylated glucose from the hydrolysis of disaccharides, such as lactose (195). Therefore, a deletion of glucokinase is expected to have no or at most a limited effect on the metabolism of the monosaccharide glucose in L. lactis.. The main end-product of lactococcal sugar fermentation is lactic acid (homolactic fermentation). The control of the flux through glycolysis has been studied extensively over the last decades: glycolytic intermediates have been determined using different conditions, and several glycolytic and their effectors have been described (122).

54 Glucose transport and metabolism

Figure 1: Glucose Schematic Out overview of Non- glucose Membrane EIIman/glc EIIcel/glc PTS metabolism in L. In PEP lactis NZ9000. ATP Pyruvate α- And β-glucose GLK can be imported via Glucose Glucose-6P PTSman/glc (EII man/glc; ATP cel/glc ADP + Pi ptnABCD), PTS ADP + Pi (EII cel/glc; ptcBAC), or FBP by (an) as yet unidentified non-PTS glucose DHAP GAP transporter(s). The + NAD points of formation ADP + Pi NADH ATP or expenditure of GAPDH ATP and NADH are 3-PGA depicted as well as some of the enzymes involved PEP (bold). ADP + Pi PK Abbreviations: ATP NADH NAD+ glucose-6P, glucose- 6-phosphate; FBP, fructose-1,6- α-Acetolactate Pyruvate LDH Lactate ALS bisphosphate; O2 NAD+ CO DHAP, CO 2 PFL PDH NADH 2 Formate dihydroxyacetone phosphate; GAP, Diacetyl CO Acetyl-CoA 2 glyceraldehyde-3- NADH ADP + Pi 2NADH NAD+ Acetoin phosphate; 3-PGA, ATP 2NAD+ NADH 3-phosphoglycerate; NAD+ Acetate Ethanol PEP, phosphoenolpyruvat 2,3-Butanediol e; CO2, carbon dioxide; O2, oxygen; ATP, ; ADP, adenosine diphosphate; Pi, inorganic phosphate; NAD+, nicotinamide adenine nucleotide; NADH, dihydronicotinamide adenine dinucleotide; GLK, glucokinase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PK, pyruvate kinase; LDH, lactate dehydrogenase; PFL, pyruvate formate lyase; PDH, ; ALS, α-acetolactate synthase.

55 Chapter 3

Although L. lactis is generally homofermentative, under conditions of limited glucose availability or when it has to metabolize less favourable sugars, it can perform a mixed-acid fermentation in which besides lactate also acetate, ethanol, 2,3- butanediol and formate are produced (30, 67, 112).Regulation takes place at several points in the glycolytic pathway. A high level of fructose-1,6-bisphosphate (FBP) is known to activate lactate dehydrogenase (LDH) and pyruvate kinase (PK), while high levels of inorganic phosphate (Pi) together with low FBP-concentrations (when the cells are starved) inhibit PK (109, 153, 186, 191, 199). Inhibition of PK leads to accumulation of phosphoenolpyruvate (PEP) and 3-phosphoglycerate (3PGA) (152). Pyruvate metabolism is also influenced by concentrations of the glycolytic intermediates FBP, glyceraldehyde-3-phosphate (GAP), and dihydroxyacetone- phosphate (DHAP). Under anaerobic conditions LDH and pyruvate formate lyase (PFL) compete for pyruvate. During fast growth, high FBP-levels are present that activate LDH, while the high levels of DHAP and GAP under these conditions inhibit the activity of PFL (50). When a less favourable sugar is used or growth rates are lowered, less FBP, DHAP, and GAP accumulate, relieving PFL inactivation, enabling the enzyme to metabolize some of the pyruvate. Regulation of glycolysis by - and energy-state modulation has also been described (50). The intracellular NADH / NAD+ ratio has been suggested to control the flux through glycolysis mainly via the activity of GAP-dehydrogenase, producing NADH (50), while other studies measuring metabolites in vivo suggest a more important role for the ATP/ADP/Pi content of the cells (127, 132). In addition to regulation by glycolytic intermediates, sugar metabolism in L. lactis is also regulated at the transcriptional level via the transcriptional regulator CcpA (carbon catabolite protein A), which is involved in global metabolic control (105). Recently, L. lactis CcpA was found to intertwine regulation of carbon and nitrogen metabolism by regulating pepQ (225), a gene located tail-to-tail to ccpA (55). Besides being part of the PTS sugar phosphorylation cascade, which ultimately leads to the transfer of a phosphate-group from PEP to the incoming sugar, HPr also plays a role in regulation, since HPr phosphorylated at serine 46 acts as a cofactor for DNA binding of CcpA in Gram-positive bacteria (42, 43). Although the glycolytic pathway for the metabolism of glucose in L. lactis has been studied widely (75, 122, 152, 192), a detailed understanding of glucose transport and the initial steps in glucose degradation is missing. In this paper we analyze the first steps in glucose metabolism of L. lactis and show that the mechanism and efficiency of transport are crucial for the characteristics of fermentation. Mutant strains using different glucose uptake routes (PTSglc/man, PTScel/glc or non-

56 Glucose transport and metabolism

PTS/glucokinase) showed great differences in transport characteristics for the glucose anomers and in the dynamics of the intracellular metabolite pools.

57 Chapter 3

EXPERIMENTAL PROCEDURES

Microbial strains and growth conditions

Strains and plasmids used in this study are listed in Table 1. The strains were grown in M17 (Difco, Sparks, MD) with 0.5% galactose (w/v) or 0.5% glucose (w/v) at 30ºC or 37°C, or in chemically defined medium (CDM (144)), with 1% glucose (w/v). When appropriate, erythromycin or chloramphenicol was used at a final concentration of 5 μg/ml. For growth in a 2 L Biostat® MD fermentor (B. Braun Biotech, Inc., Allentown, PA), the medium was gassed with argon for 60 min prior to inoculation (4% inoculum from a culture grown overnight); the pH was kept at 6.5 by automated addition of 10 M NaOH, and an agitation rate of 70 rpm was used to keep the system homogeneous. Growth was monitored by measuring the optical density at 600 nm.

DNA techniques

General DNA techniques were performed essentially as described elsewhere (162). Plasmid DNA was isolated by the method of Birnboim and Doly (12). Restriction enzymes, T4 DNA ligase, Expand polymerase and Taq polymerase were obtained from Roche Applied Science (Mannheim, Germany) and used according to the supplier’s instructions. PCR was performed in an Eppendorf thermal cycler (Eppendorf, Hamburg, Germany).

Specific cloning procedures

Gene deletions were all performed in L. lactis NZ9000 and were constructed using a two-step homologous recombination method (96). This method does not leave antibiotic resistance markers in the chromosome, and multiple deletions in one strain can be easily realized. Primers used for cloning are listed in Table 1. Chromosomal DNA of L. lactis NZ9000 was used as a template in PCR amplifications. L. lactis NZ9000ΔptcBA and NZ9000ΔptnABCDΔptcBA, carrying only the first 36 bp of ptcB and the last 58 bp of ptcA, were made using the primerpairs ptc1/ptc2 and ptc3/ptc4 (L. lactis NZ9000 and NZ9000ΔptnABCD were used as parent strains).

58 Glucose transport and metabolism

TABLE 1: Lactococcal strains, plasmids and primers used in this study L. lactis strains Description Reference

NZ9000 MG1363 carrying pepN::nisRK (84) LL302 RepA+ MG1363, carrying a single copy of pWV01 repA in pepXP (95) LL108 RepA+ MG1363, Cmr, with multiple copies of pWV01 repA in the chromosome (95) NZ9000Δglk NZ9000 containing a 404-bp deletion in glk (141) NZ9000ΔptnABCD NZ9000 containing a 1736-bp deletion in ptnABCD (141) NZ9000ΔptcBA NZ9000 containing a 657-bp deletion in ptcBA This work NZ9000ΔglkΔptnABCD NZ9000Δglk with a 1736-bp deletion in ptnABCD (141) NZ9000ΔptnABCDΔptcBA NZ9000ΔptnABCD with a 657-bp deletion in ptcBA This work

Plasmids Description Reference pORI280 Emr, LacZ+, ori+ of pWV01, replicates only in strains providing RepA in trans (96) pORI280-ptcBA’ Emr, pORI280-derivative specific for integration in L. lactis ptcBA (141) pVE6007 Cmr, temperature-sensitive derivative of pWV01 (107)

Primer Sequence (5' to 3') Restriction site Location annealing part

Ptc1 GCTCTAGAGTCATCTCTGACCCCTTTC XbaI 560-541 bp up. ptcB TSS Ptc2 CGGGATCCTTAGGCTGCACATGCAAGTGC BamHI 19-36 bp down. ptcB TSS Ptc3 CGGGATCCCCTTGCAGTAGAAGTTGTTG BamHI 294-313 down. ptcA TSS Ptc4 CGGAATTCCGGATAAGTTACATCGCTAAATG EcoRI 509-531 bp down. ptcA SC

Abbreviations: TSS, (putative) Translational Start Site; SC, Stop-Codon; up., upstream; down., downstream.

In vivo NMR experiments

Cells were grown in CDM containing 1% glucose (w/v), harvested in the mid- logarithmic phase of growth, washed twice with 5 mM potassium phosphate (KPi) buffer (pH 6.5), and resuspended in 50 mM KPi (pH 6.5), to a protein concentration of approximately 15 mg protein/ml. In vivo NMR experiments were performed using the on-line system described earlier (124). Glucose specifically labeled with 13C on carbon one (20 mM) was added to the cell suspension at time-point zero. The time course of glucose consumption, product formation, and changes in the pools of intracellular metabolites were monitored in vivo. When the substrate was exhausted and no changes in the resonances of intracellular metabolites were observed, an NMR-sample extract was prepared as described previously (124, 127). Carbon-13 spectra were acquired at 125.77 MHz on a Bruker DRX500 spectrometer (Bruker BioSpin GmbH, Karlsruhe, Germany). All in vivo experiments were run using a

59 Chapter 3 quadruple nuclei probe head at 30oC, as described before (124). Lactate concentration was determined in the NMR-sample extract by 1H-NMR in a Bruker AMX300 (Bruker BioSpin GmbH). The concentrations of other metabolites were determined in fully relaxed 13C spectra of the NMR-sample extracts as described (127). Intracellular NAD+ and NADH were measured in vivo by 13C-NMR using a method described before (127).

Enzyme activity assays

Cell suspensions used in the in vivo NMR studies were diluted in 50 mM KPi-buffer pH 6.5, to a final OD600 of 25, and aliquoted in 500 μl quantities. The cells were disrupted with 0.5 g glass beads (∅ 50-105 µm), using a Mini-BeadBeater-8 (Biospec products Inc., Bartlesville, OK) with two 1 min pulses (homogenize), with a 1 min interval on ice. Cell debris was pelleted twice and the cell-free extracts obtained were immediately used for spectrophotometric determination of enzymatic activities using a microtiterplate-reader at 340 nm (GENios, Tecan Group Ltd., Maennedorf, Switzerland). Total protein was determined by the method of Bradford (19). Glucokinase: Glucokinase (Glk) (EC2.7.1.2) activity was assayed spectrophotometrically by the glucose-6-phosphate dehydrogenase (Glc-6P-DH) (EC1.1.1.49) : NADPH-coupled assay as described (147), with minor changes. The + assay mixture contained KPi-buffer (10 mM, pH 7.2), MgCl2 (5 mM), NADP (1 mM), ATP (1 mM), Glc-6P-DH (1 U). Glucose (20 mM) was used to start the reaction. Pyruvate Kinase: Pyruvate Kinase (PK) (EC2.7.1.40) activity was assayed spectrophotometrically by the lactate dehydrogenase (LDH) (EC1.1.1.27) : NADH- coupled reaction as described earlier (50). Lactate Dehydrogenase: Lactate Dehydrogenase (LDH) activity was assayed spectrophotometrically by NADH measurement, as described previously (50).

14C-Glucose transport assays

Mid-exponential phase cells were washed twice in KPi-buffer (5 mM, pH 6.5) and resuspended in the assay-buffer KPi (50 mM, pH 6.5). Glucose uptake was analyzed at 30ºC in cell suspensions with an appropriate cell density, dependent on the strain tested. The transport reaction (3 ml) was initiated by adding D-[U-14C]- Glucose (100 nM), with a specific activity of 317 mCi/mmol; glucose concentrations were increased up to 10 mM by using cold D-glucose (Merck, Darmstadt, Germany).

60 Glucose transport and metabolism

Time-point samples (0.5 ml cell suspension) were taken from the reaction and filtered using a vacuum pump and nitrocellulose filters with a pore size of 0.45 μm (Millipore, Billerica, MA). The reaction was immediately stopped by washing the cells on the filter with 5 ml KPi-buffer (50 mM, pH 6.5). The filters were placed in a vial, covered with scintillation fluid (Ultima Gold LSC Cocktail, Packard BioScience, Groningen, the Netherlands) and the 14C-activity was determined using a Packard TriCarb 2000 CA liquid scintillation analyzer (Packard Instrument, Meriden, CT). Data from at least two independent experiments were analyzed. The calculations were compared to a Michaelis-Menten model. The glucose uptake data resembled Michaelis-Menten uptake kinetics.

Estimation of kinetic properties for α- and β-glucose consumption

A mathematical model was developed that accounted for two different Michaelian uptake kinetics of α- and β-glucose anomers and for the first-order kinetics of the anomers’ interconversion. The first-order rate constants of glucose anomerization were determined by performing 13C-NMR time series of solutions of α-[1-13C]- glucose (20 mM) in KPi (50 mM, pH 6.5), which were allowed to anomerize to equilibrium at 30ºC; at this temperature the relative percentages of α-glucose and β- glucose in equilibrium were 38 and 62%, respectively. The first-order rate constants determined were: for the conversion of α-glucose into β-glucose 0.108±0.001 min-1 and for the conversion of β-glucose into α-glucose 0.063±0.001 min-1. With these constants, the α- and β-glucose consumption kinetics were fitted to the model and the Vmax and Km determined. The model was developed in Matlab v7.3.0 (The MathWorks, Inc., Natick, MA) as a set of ordinary differential equations. The ODE23s function was used to solve the differential equations and the least squares method was used to find the set of parameters that best fitted the experimental data.

⎧∂[]α Glc V α ⋅[α Glc] ⎪ = k ⋅[]β Glc − k ⋅ []α Glc − max ⋅ϕ ∂t β −−>α α −−>β K α + α Glc α ⎪ m [] ⎨ β ⎪∂[]β Glc V ⋅[]β Glc = k ⋅ α Glc − k ⋅ β Glc − max ⋅ϕ ⎪ α −−>β []β −−>α []β β ⎩ ∂t Km + []β Glc

Where φx is a linear function defined as:

61 Chapter 3

⎧ 1 − a x ⎪a x + bx ⋅ t ⇐ a x + bx ⋅ t <1 ⇔ t < ⎪ bx ϕ = ⎨ x 1 − a ⎪1 ⇐ t ≥ x ⎪ ⎩ bx

This function was required in order to obtain a good fit to the initial time-points of the glucose consumption kinetics. Indeed, in some strains the maximal glucose consumption rate was only reached after a lag time of a few minutes. To model this phenomenon, we assumed that initially (at t=0), the transport rate is only a fraction

(φx=ax and 0

Transcriptome analysis

Levels of mRNA in L. lactis strains NZ9000 and NZ9000ΔptnABCDΔptcBAC were compared by transcriptome analysis using full-genome amplicon-based L. lactis IL1403 DNA-microarrays (83). Cells grown in M17 with 0.5% glucose were harvested at the mid-exponential phase of growth. The strains were grown and analyzed independently in triplicate, with Cy3 and Cy5 dye-swaps of each repetition. This resulted in a total of 6 hybridized slides, all containing duplicate spots of each amplicon. Thus, maximally 12 measurements were obtained for each gene. The experiments were performed essentially as described by van Hijum et al, 2005 (207), with the modifications introduced by Pool et al., 2006 (141).

62 Glucose transport and metabolism

RESULTS

Glucose fermentation patterns differ greatly between L. lactis transport mutants

Various isogenic L. lactis glucose transport mutants were made that use different glucose uptake systems. The mutant strains have a deletion in either glk (encoding glucokinase), in ptnABCD (encoding EIIman/glc), or in ptcBA (encoding EIIBAcel/glc) or in both ptnABCD and ptcBA. The glk mutation was made to disable the further metabolism of glucose imported by (a) yet unknown non-PTS glucose permease system(s). The chromosomal organization of the genes and the position and extent of the deletions are depicted in Fig 2A.

A Δglk glk yvaB

ΔptnABCD ptnAB ptnC ptnD

ΔptcBA ptcB ptcA yecA ptcC B 3.0

2.5 -1 ) Strain μmax (h )

600 2.0 NZ9000 0.72

1.5 NZ9000Δglk 0.51 NZ9000ΔptnABCD 0.59 1.0 NZ9000ΔptcBA 0.55 Growth (OD 0.5 NZ9000ΔptnABCDΔptcBA 0.61 0.0 024681012 Time (h)

Figure 2: Genetic structure of L. lactis deletion strains and their growth characteristics. A) Schematic overview of the genes and the deletions studied in this work. Hooked arrow, putative promoter; lollipop, terminator structure; dashed areas, deleted sequence. B) L. lactis strains were grown at 30°C without pH control (standing bottles), under anaerobic conditions in CDM + 1% glucose. The maximum growth rates (μmax) for each strain is given at the right.

63 Chapter 3

The growth curves for each deletion strain studied, grown anaerobically in CDM with

1% glucose, are shown in Figure 2B, together with the maximum growth rates (μmax) for each strain. All the deletion strains show a lower maximum growth rate (between 0.51 and 0.61 h-1) than the parental strain L. lactis NZ9000 (0.72 h-1). Glucose consumption and the dynamics of appearance of glycolytic intermediates in the mutants were compared mutually and with L. lactis NZ9000. The fate of [1-13C]- glucose during metabolism in all of the strains was followed in time using in vivo NMR. (Table 2, Fig. 3). L. lactis NZ9000 displayed a homolactic fermentation with a maximum glucose consumption rate of 0.44 μmol·min-1·mg protein-1. FBP reached a maximum concentration of about 50 mM just prior to glucose depletion, upon which around 8 mM 3-PGA was formed (Fig. 3A). L. lactis strain NZ9000ΔptnABCD showed a heavily disrupted glucose metabolism, which was expected since the EIIman/glc encoded by ptnABCD has been proposed to be the main sugar transport system in L. lactis. The maximal consumption rate is lowered to 0.24 and the overall glucose consumption rate is lowered to only 0.06 μmol·min-1·mg protein-1 and the strain shifted to a mixed acid fermentation (Fig. 3C). Besides that, high concentrations of 3-PGA and PEP accumulated. Based on these results EIIman/glc appears indeed to be the main glucose transport system. Interestingly, the consumption of glucose appeared to consist of two stages: a first phase with a relatively fast rate followed by a much lower rate. To see if these two phases in glucose consumption might be a consequence of anomeric specificity of the different transport systems, the glucose consumption is split into the consumption of α-glucose and β-glucose in Fig. 3. This showed that the non-PTS transport route and EIIcel/glc probably have a preference for the β-anomer of glucose, which we will later quantify (see below). The strain in which both glucose-transporting PTS systems were disrupted, L. lactis strain NZ9000ΔptnABCDΔptcBA, showed a glucose fermentation pattern comparable to that of NZ9000ΔptnABCD, although the maximum concentration of FBP accumulating during fermentation and the maximal glucose consumption rate was lower.

64 Glucose transport and metabolism

TABLE 2: Glucose fermentation pattern and glycolytic intermediates formed by L. lactis NZ9000 and its isogenic deletion strains analyzed by in vivo NMR.

Strain Functional FT a) β-glc Glycolytic Intermediates (mM) c) b) transport routes pref. FBPmax PEP 3-PGA

NZ9000 EIIman/glc, EIIcel/glc, non-PTS H +/- 49.5 ± 0.5 - 8.3 ± 2.5 Δglk EIIman/glc, EIIcel/glc MA ++ 44.2 ± 0.9 13.2 ± 2.4 31.8 ± 2.7 Δptn d) EIIcel/glc, non-PTS MA ++ 43.5 ± 1.0 13.2 ± 2.4 35.2 ± 2.8 Δptc e) EIIman/glc, non-PTS MA +/- 20.2 ± 2.1 - - ΔptnΔptc d,e) non-PTS MA ++ 33.2 ± 0.7 13.6 ± 2.1 33.6 ± 2.5

a) FT, fermentation type; H, homolactic fermentation; MA, mixed-acid fermentation b) β-glc pref., qualitative preference for β-glucose shown: +/-, slight preference; ++, preference. c) FBPmax, maximum fructose-1,6-bisphosphate concentration; PEP, phosphoenolpyruvate; 3-PGA, 3-phosphoglycerate d) Δptn, ΔptnABCD e) Δptc, ΔptcBA

Surprisingly, disruption of the non-PTS-route for glucose metabolism in L. lactis strain NZ9000Δglk also led to quite large differences with respect to glucose utilization compared to its parent L. lactis NZ9000 (Table 2, Fig. 3B). This was not expected since the non-PTS uptake system(s) were suggested to play no or only a minor role in glucose consumption (195). The maximal glucose consumption rate was lowered about 4-fold to 0.20 μmol·min-1·mg protein-1 in L. lactis NZ9000Δglk, with a 4-fold increase of the 3-PGA concentration. Also considerable amounts of PEP (13.2 mM) accumulated, which was not detectable in L. lactis NZ9000. FBP accumulated to a slightly lower level in L. lactis NZ9000Δglk compared to L. lactis NZ9000. Furthermore, deletion of glk resulted in a shift to a mixed acid fermentation. In fact, the glucose fermentation pattern of L. lactis NZ9000Δglk resembled that of L. lactis strains NZ9000ΔptnABCD and NZ9000ΔptnABCDΔptcBA. L. lactis NZ9000ΔptcBA displayed a glucose fermentation pattern that differed completely from that of L. lactis NZ9000 and all the other deletion strains studied here (Fig. 3D). L. lactis NZ9000ΔptcBA showed a mixed-acid fermentation, but no PEP nor 3-PGA were formed as final fermentation products. The maximum level of accumulated FBP was less than half of that in L. lactis NZ9000. Furthermore, the consumption of glucose in two stages, as shown in L. lactis strains NZ9000ΔptnABCDΔptcBA, NZ9000ΔptnABCD, and less pronounced in NZ9000Δglk, was not observed in L. lactis NZ9000ΔptcBA.

65 Chapter 3

A 50 B 50 NZ9000 GCR: 0.44 Δglk GCR: 0.20 40 40

30 30

20 20

10 10 Concentration (mM) Concentration (mM) 0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 Time (min) Time (min)

C 50 D 50 ΔptnABCD GCR: 0.24 ΔptcBA GCR: 0.13 40 40

30 30

20 20

10 10 Concentration (mM) Concentration (mM) 0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 Time (min) Time (min) E F 50 40 ΔptnABCD ΔptcBA GCR: 0.15 40 30 30 20 20 10 10 Concentration (mM) Concentration (mM) 0 0 0 5 10 15 20 25 30 35 NZ9000 Δglk Δptn ΔptcBA ΔptnABCD ABCD ΔptcBA Time (min)

Figure 3: Kinetics of 20 mM [1-13C]-glucose metabolism, shown by in vivo NMR. A-E) [1-13C]-glucose consumption, product formation and pools of intracellular metabolites of L. lactis under anaerobic conditions at 30°C with pH controlled at 6.5. Results are shown for L. lactis strains NZ9000 (A), NZ9000Δglk (B), NZ9000ΔptnABCD (C), NZ9000ΔptcBA (D) and -1 NZ9000ΔptnABCDΔptcBA (E). Maximal glucose consumption rates (GCR, in μmol·min ·mg protein-1) are boxed in the upper right corners of the graphs. Symbols used: yellow diamond, α-glucose; black diamond, β-glucose; red triangle, fructose-1,6-bisphosphate; blue circle, 3- phosphoglycerate; green diamond, phosphoenolpyruvate. The lines drawn in the graph are interpolations. F) End-products formed by the indicated strains. Symbols used: White, lactate; black, acetate; stripes, ethanol; grey, 2,3-butanediol.

66 Glucose transport and metabolism

Mixed-acid fermentation is accompanied by a high NADH/NAD+ ratio

Since the genetic deletions in the L. lactis strains studied here led to a shift from a homolactic to a mixed-acid fermentation, the redox balance of one of the deletion strains (L. lactis NZ9000ΔptnABCDΔptcBA) was compared to that of the wildtype, L. lactis NZ9000. Besides the regulation of carbox flux by the FBP concentration, it has been suggested before that glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which is influenced by a high ratio of NADH/NAD+, controls the flux (50). The kinetics of glucose consumption, accumulation and breakdown of FBP and the levels of nicotinamide adenine nucleotide (NAD+) and dihydronicotinamide adenine dinucleotide (NADH) were measured in vivo in order to investigate the redox-state of the cells of both strains (Fig. 4).

A B 80 6 80 6 70 70 5 5 60 60 4 4 50 50 40 3 40 3 / NADH (mM) NADH / 30 30 (mM) NADH / + 2 2 + 20 20

1 NAD 1 10 10 NAD Glc (mM) FBP/ Products / Glc(mM) Products / FBP/ 0 0 0 0 -10 0 10 20 30 40 50 60 70 80 -10 0 10 20 30 40 50 60 70 80 Time (min) Time (min)

Figure 4: Evolution of glucose, FBP, lactate, acetate, NAD+ and NADH in non-growing cells of L. lactis strains NZ9000 (A) and NZ9000ΔptnABCDΔptcBA (B). The cells were grown in CDM with 5 mg/L of [5-13C]-nicotinic acid, before and after a 40 mM [1-13C]-glucose pulse. Glucose was added at time 0 min; dark-grey shaded area is before glucose addition. Light-grey shaded area is the fast phase of glucose consumption in L. lactis NZ9000ΔptnABCDΔptcBA, Conditions were anaerobic, 30°C and pH was controlled at 6.5. Symbols: dark-blue diamond, glucose; red triangle, FBP; green square, lactate; purple triangle, acetate; gold square, NAD+; light-blue circle, NADH. Fitted lines are simple interpolations.

Before addition of glucose, ~ 5 mM of NAD+ was present in the cells of both strains while NADH could not be detected. The addition of 40 mM [1-13C]-glucose at time- point zero resulted in an immediate accumulation of FBP in both strains due to glucose metabolism. When FBP levels decrease at the onset of glucose depletion, NADH was formed at the expense of NAD+. During the formation of the end-

67 Chapter 3 products, NAD+ is recovered slowly to 5 mM in both strains. The two stages in glucose consumption in L. lactis NZ9000ΔptnABCDΔptcBA as shown by Fig. 3E can also be seen in Fig. 4B. During the first (fast) part of the glucose consumption, in this case during the first 10 min after glucose addition, mainly lactate was formed (data can be obtained from Fig. 3, although in that case only half the amount of glucose (20 mM) was added, and, thus, the fast phase only lasted for about 5 min). This fast phase of glucose consumption led to a mainly homolactic fermentation. During the second (slow) phase of glucose consumption, L. lactis NZ9000ΔptnABCDΔptcBA shifted to a mixed-acid fermentation (see Fig. 3E). The onset of this mixed-acid fermentation is accompanied by a high NADH/NAD+ ratio (above 0.2). L. lactis NZ9000ΔptnABCDΔptcBA recovers the original concentration of NAD+ (~ 5 mM) at the moment the glycolysis stops.

PK and LDH activities are lowered in all mutants, while Glk activity varies

Since LDH and PK are crucial and highly regulated enzymes in glycolysis (186, 191, 199), the activity of these enzymes was analyzed in L. lactis NZ9000 and the mutant strains. Furthermore, the activity of glucokinase was determined, as a glk-deletion had a considerable effect on glucose metabolism. In all 4 mutant strains the PK and LDH activities were lower than those in the wildtype (Table 3). PK and LDH activities were decreased about 1.5 times and 2 times respectively, which is in agreement with the mixed-acid fermentation seen in all mutants. The decrease in LDH-activity, apparently, does not cause a problem in NAD+ recovery, since Fig. 4 shows that NAD+ is formed quickly from NADH after FBP depletion in L. lactis NZ9000ΔptnABCDΔptcBA. The activity of Glk in L. lactis NZ9000 is only 0.14 U·mg protein-1. As expected, L. lactis NZ9000Δglk does not exhibit glucokinase activity. Interestingly, mutations in PTSman/glc (ptnABCD) and PTScel/glc (ptcBA) had opposite effects on the activity of Glk. Deletion of ptnABCD decreased the Glk-activity about 2.5 to 3 times, while deletion of ptcBA increased the Glk-activity about 3.5 to 4 times. When both PTSs are disrupted (NZ9000ΔptnABCDΔptcBA), the Glk-activity was comparable to that of the PTSman/glc single deletion strain.

68 Glucose transport and metabolism

TABLE 3: Enzyme activity in L. lactis of glucokinase (Glk), pyruvate kinase (PK) and lactate dehydrogenase (LDH) a).

L. lactis strain Glk PK LDH

NZ9000 0.14 (0.02) 1.90 (0.06) 30.44 (1.65) NZ9000Δglk 0.00 (0.00) 1.15 (0.05) 14.29 (2.06) NZ9000ΔptnABCD 0.05 (0.02) 1.14 (0.02) 14.02 (0.31) NZ9000ΔptcBA 0.52 (0.01) 1.37 (0.04) 15.54 (1.36) NZ9000ΔptnABCDΔptcBA 0.07 (0.02) 1.26 (0.07) 11.70 (0.50)

a) In all cases enzyme activities are given (in U·mg protein-1) and standard deviations determined on at least 3 independent experiments are given between brackets.

The two-phase kinetics of glucose consumption is explained by the preferential utilization of glucose anomers

The ability of NMR to distinguish between α- and β-anomers of glucose allowed investigation of the kinetics of glucose utilization in the L. lactis deletion strains in more detail (Fig. 5). A mathematical model accounting for the Michaelian uptake kinetics of each anomer as well as the first-order kinetics of the α- to β-anomer interconversion was developed (for details see Experimental Procedures). The Vmax for the consumption of the α- and β-anomers and the relative fluxes for the uptake of each anomer as well as for the conversion of α- into β-anomer were estimated (Fig. 5). In all strains examined, β-glucose was consumed at a maximal rate higher than that of α-glucose. In L. lactis strains NZ9000 and NZ9000ΔptcBA, exhibiting a single phase of glucose consumption, both anomers were transported as evidenced by their relative uptake fluxes (Fig. 5A and 5D). The lower maximal glucose consumption rate in L. lactis NZ9000ΔptcBA results from a reduction in the uptake of both α- and β-glucose. This assumption is further supported by the ratio between the anomers, similar to that found in L. lactis NZ9000 (Fig. 5E). In L. lactis NZ9000ΔptnABCD, the rate of glucose consumption was moderately high only when β-glucose was available. Upon depletion of β-glucose, glucose consumption is limited by the flux of α- to β-anomer conversion (Fig. 5B). The data strongly suggests that the glucose transporters present in this strain have high specificity for the β-anomer.

69 Chapter 3

AB 20 20 NZ9000 0.3 NZ9000ΔptnABCD 0.3 15 15

0.2 0.2 10 10 mol/min/(mg protein) mol/min/(mg

0.1 mol/min/(mg protein) 0.1 μ

5 μ 5 Concentration (mM) Concentration (mM) Flux ( Flux 0 0.0 ( Flux 0 0.0 02468 0 10203040 C Time (min) D Time (min) 20 20 NZ9000Δglk NZ9000ΔptcBA 0.3 0.3 15 15

0.2 0.2 10 10 mol/min/(mg protein) 0.1 0.1 protein) mol/min/(mg μ 5 5 μ Concentration (mM) Concentration (mM) Flux ( Flux 0 0.0 0 0.0 Flux ( 0246810121416 0246810121416 Time (min) Time (min) E

Vmax L. lactis strain μmol·min-1·mg protein-1 Total α-Glc β-Glc β-Glc/α-Glc NZ9000 0.44 0.13 0.36 3 NZ9000ΔptnABCD 0.24 0.02 0.23 8 NZ9000Δglk 0.20 0.02 0.19 8 NZ9000ΔptcBA 0.13 0.03 0.11 3

Figure 5: Kinetics of consumption of the α- and β-anomers of glucose. Time course for the consumption of α- and β-glucose (experimental data and simulated line) and estimated fluxes (µmol·min-1·mg protein-1) for the conversion of α- into β-glucose and for uptake of α- and β-glucose in L. lactis strains NZ9000 (A), NZ9000ΔptnABCD (B), NZ9000Δglk (C), NZ9000ΔptcBA (D). Experimental data was obtained during the metabolism of 20 mM glucose (see Fig. 3). The mathematical model assumed Michaelis-Menten uptake kinetics of α- and β-glucose and first-order kinetics of the anomer interconversion (see

Experimental Procedures). Vmax for total glucose uptake as well as for the uptake of α- and β- anomers was also estimated (E). Model simulation of the kinetics of α- and β-glucose consumption (full lines) and model-predicted fluxes (dashed lines). Symbols used: open diamond, glucose; black diamond, β-glucose; red diamond, α-glucose; black dotted line, β- glucose; red dotted line, α-glucose; blue dotted line, conversion of α-glucose into β-glucose. Note the different time scales.

70 Glucose transport and metabolism

Similar results were obtained for L. lactis strains NZ9000ΔptnABCDΔptcBA and NZ9000ΔglkΔptnABCD (data not shown). In L. lactis NZ9000Δglk, the maximal consumption rates of the glucose anomers were lower, but the decrease was more pronounced for β-glucose, as in the two strains lacking ptnABCD. However, during the slower phase of glucose utilization the flux of α-glucose uptake was greater than the conversion of α-glucose to β-glucose, indicating that the α-anomer is taken up in the glk mutant. Our data show that L. lactis strains with an intact EIIglc/man system transport both anomers of glucose, whereas strains in which ptnABCD is inactivated have a clear preference for β-glucose.

The affinity of glucose transport is highly decreased when EIIman/glc is deleted

To elucidate which of the transport systems has the highest apparent affinity for glucose, glucose transport was analyzed in cell suspensions of L. lactis NZ9000 and its isogenic mutants using 14C-labelled glucose (Table 4). For each strain the assay conditions were optimized seperately, since the Km for glucose transport differed greatly between the strains. L. lactis NZ9000 had a very high affinity for glucose transport (Km = 5.2 μM), but also low affinity transport could be detected using higher glucose concentrations in the assay. The exact kinetic properties of this low affinity transporter could not be determined. Deletion of EIIcel/glc (NZ9000ΔptcBA) glc/man decreased both the Vmax and the Km 2-fold. When EII was removed (NZ9000ΔptnABCD), a drastic change in glucose transport affinity was observed. man/glc The Km increased to around 5.2 mM, showing that EII has the highest affinity for the uptake of glucose of all the glucose uptake systems. L. lactis

NZ9000ΔptnABCDΔptcBA showed a similar Km for glucose (8.0 mM) as L. lactis

NZ9000ΔptnABCD, but the Vmax was slightly lower. Although the non-PTS permease system(s) was/were expected to be of minor importance in glucose transport, L. lactis NZ9000Δglk showed a 4-fold decreased glucose transport affinity (Km = 20

μM), together with an 18-fold reduction of Vmax for one specific transporter. Deletion of glk seems to lower the capacity of EIIman/glc about 20 times, while the affinity for glucose seems to decrease about 4-fold. Besides this, a low-affinity transporter was detected with a Km of 4.8 mM (Table 4).

71 Chapter 3

TABLE 4. Kinetic parameters obtained for the transport of glucose in whole cells of L. lactis NZ9000 and its glucose transport mutants.

a) App a) L. lactis strain Functional Vmax Km transport routes (nmol·min-1·mg protein-1) (uM)

NZ9000 EIIman/glc, EIIcel/glc, non-PTS 188 ± 11 5.2 ± 0.2 > 160 NDb) NZ9000Δglk c) EIIman/glc, EIIcel/glc, non-PTS 10 ± 1 20 ± 2.3 224 ± 30 4809 ± 245 NZ9000ΔptnABCD EIIcel/glc, non-PTS 171 ± 8 5166 ± 372 NZ9000ΔptcBA EIIman/glc, non-PTS 92 ± 6 3.5 ± 0.8 NZ9000ΔptnABCDΔptcBA non-PTS 120 ± 1 7986 ± 674

a) Values of two independent experiments were averaged and are reported ± SD of the two App measurements. Vmax and Km (apparent Km) were determined using the following glucose concentrations: NZ9000, 0.1 to 500 µM and 500 µM to 25 mM; NZ9000Δglk, 0.1 µM to 10 mM; NZ9000ΔptnABCD and NZ9000ΔptnABCDΔptcBA, 0.5 µM to 10 mM; NZ9000ΔptcBA, 0.1 to 500 µM. b) Not exactly determined, above 5 mM. c) Experimental data obtained with whole cells of L. lactis NZ9000Δglk were fitted using non-linear least squares regression analysis to the sum of two independent Michaelis-Menten equations.

EIIman/glc is a high-capacity / high-affinity transporter; EIIcel/glc is also a high-capacity transporter but it has a low affinity for glucose; the non-PTS(s) is/are (a) low- capacity and low-affinity transporter(s). Different types of transport kinetics (different transporters) could be measured using the appropriate glucose concentration ranges.

Transcriptome analysis shows regulation at the genetic level in L. lactis NZ9000ΔptnABCDΔptcBA

To analyze the transcriptional effect of deletion of both glucose transporting PTSs, the transcriptomes of L. lactis strains NZ9000 and NZ9000ΔptnABCDΔptcBA were compared using DNA-microarrays of L. lactis. Any possible non-PTS transport system(s) might be overexpressed in NZ9000ΔptnABCDΔptcBA grown on glucose and may, thus, be identified. Transcriptome analysis of cells grown to the mid- exponential phase of growth in GM17 showed that several genes were differentially expressed in L. lactis NZ9000ΔptnABCDΔptcBA relative to L. lactis NZ9000, and these are categorized in Table 5. Additionally, the results from the transcriptome comparison of L. lactis NZ9000ΔptnABCDΔglk with L. lactis NZ9000 are shown in Table 5 (data obtained from previous results (141)). Although we show that glucokinase enzyme activity in L. lactis NZ9000ΔptnABCDΔptcBA was lowered 2-

72 Glucose transport and metabolism fold compared to L. lactis NZ9000 (Table 3), glk expression was not significantly reduced in the mutant strain. Also the enzymatic activities of PK and LDH were lowered resp. 1.5- and 2.6-fold in L. lactis NZ9000ΔptnABCDΔptcBA compared to L. lactis NZ9000 (Table 3), while the expression of the genes of the las-operon (pfk, pyk and ldh) was not significantly different. The expression of the gal-operon genes (galPMKTE) was not significantly different between both strains (data not shown), showing that the growth on galactose during construction of L. lactis NZ9000ΔptnABCDΔptcBA did not have a permanent effect on the gene expression (data not shown). Gene pfl (encoding pyruvate formate lyase) was expressed to a higher level in L. lactis NZ9000ΔptnABCDΔptcBA, which was expected since the protein encoded by this gene is involved in pyruvate metabolism and is responsible for the increased mixed-acid fermentation observed in this mutant strain. Also pdhA and pdhB (encoding part of the pyruvate dehydrogenase complex) were expressed to a higher level in L. lactis NZ9000ΔptnABCDΔptcBA. The cells used for these DNA-microarray studies were grown semi-anaerobically (the Pdh proteins might function slightly). Since this accounts for all strains tested, the overexpression of pdhA and pdhB in L. lactis NZ9000ΔptnABCDΔptcBA is caused by the genotype. The expression pattern of the genes involved in glycolysis and pyruvate metabolism of L. lactis NZ9000ΔptnABCDΔglk resembles that of L. lactis NZ9000ΔptnABCDΔptcBA. Gene ptsH encoding histidine protein (HPr) was not differentially expressed in L. lactis NZ9000ΔptnABCDΔptcBA, but was overexpressed 1.6 times (p-value 0.005) in L. lactis NZ9000ΔptnABCDΔglk, relative to L. lactis NZ9000. The deletion of glk seems to have an effect on the expression of this regulatory gene. The expression of ccpA, encoding carbon catabolite protein A is differentially overexpressed in both L. lactis NZ9000ΔptnABCDΔptcBA (1.5 times, p-value 0.007) and NZ9000ΔptnABCDΔglk (2.3 times). The increased expression of the noxABE genes, encoding NADH dehydrogenases (noxA and noxB) and an NADH-oxidase (noxE) in both deletion strains tested is in agreement with the overexpression of the pdh genes: it is probably required to keep a proper NAD+/NADH redox balance.

73 Chapter 3

TABLE 5: Gene expression in L. lactis NZ9000ΔptnABCDΔptcBA or NZ9000ΔptnABCDΔglk, both compared to L. lactis NZ9000

Category Gene Ratio a) Ratio b) Product function ΔptnΔptc ΔptnΔglk

Glycolysis glk 0.9 del Glucokinase and pgi 1.0 1.6 Phosphoglucose isomerase pyruvate pfk 0.7 s 0.5 s 6-Phosphofructokinase metabolism fba 0.8 0.7 s Fructose-bisphosphate aldolase tpiA 0.7 1.0 Triose-phosphate isomerase gapA 1.3 1.4 Glyceraldehyde-3P-dehydrogenase pgk 0.9 0.8 s Phosphoglycerate kinase pmg 0.7 s 0.8 Phosphoglycerate mutase enoA 1.1 1.0 Enolase pyk 1.1 0.7 Pyruvate kinase ldh 0.9 0.8 Lactate dehydrogenase (LDH) pfl 5.1 s 4.2 s Pyruvate-formate lyase (PFL) pflA 1.6 s 1.4 s PFL activating enzyme pdhA 3.6 s 1.4 Pyruvate dehydrogenase E1 comp. α pdhB 2.1 s 2.0 s Pyruvate dehydrogenase E1 comp. β-subunit pdhD 1.4 1.8 s Pyruvate dehydrogenase complex E3 comp. als 1.0 0.9 α-Acetolactate synthase aldB 1.0 1.2 Acetolactate decarboxylase butA 1.2 2.1 s Acetoin reductase butB 1.3 2.0 s 2,3-Butanediol dehydrogenase ackA1 1.5 1.6 s Acetate kinase ackA2 1.4 2.5 s Acetate kinase

Regulation ptsH 0.9 1.6 Histidine protein ccpA 1.5 2.3 s Carbon catabolite protein A

Energy atpB 0.8 0.6 ATP synthase F0, A subunit production atpD 0.8 0.6 s ATP synthase F1, beta subunit and atpG 1.0 0.7 s ATP synthase F1, gamma subunit conversion atpH 0.9 0.7 s ATP synthase F1, delta subunit noxA 1.4 1.5 s NADH dehydrogenase noxB 1.6 2.4 s NADH dehydrogenase noxE 2.0 1.7 s NADH oxidase

Other ytgA 11.3 s 3.0 s Unknown, 2 TMS, pI 9.2 genes c) ytgB 2.9 1.2 Unknown, 2 TMS, pI 12 ytgH 6.6 5.3 s Unknown, part of the ytgBAH operon ymgH 3.2 s 1.6 Unknown, 2 TMS, pI 10 chiA 8.0 s 1.9 s Chitinase yucG 3.1 1.9 Putative chitin binding protein

Abbreviations: del, deleted; s, significant differential expression (p-value < 0.001) a) L. lactis NZ9000ΔptnABCDΔptcBA over L. lactis NZ9000 b) L. lactis NZ9000ΔptnABCDΔglk over L. lactis NZ9000; taken from previous data (141) c) Highly overexpressed genes in L. lactis NZ9000ΔptnABCDΔptcBA, coding for putative membrane proteins or putatively involved in carbohydrate metabolism. Criteria for selection were the presence of putative transmembrane segments (TMS) and the predicted isoelectric point (pI) (169). TMS prediction was performed using the HMMTOP tool (204). Clone Manager Suite (Sci Ed Software, USA) was used to calculate the pI.

74 Glucose transport and metabolism

The gene with the highest fold overexpression in L. lactis NZ9000ΔptnABCDΔptcBA compared to L. lactis NZ9000 is ytgA. This gene is part of the putative operon ytgBAH. All three genes of this operon are overexpressed in L. lactis NZ9000ΔptnABCDΔptcBA. ytgBAH encodes three small proteins of 115, 156, and 91 amino-acid residues, respectively, with as yet unknown functions. Both ytgB and ytgA have 2 putative transmembrane segments and both have a very high pI, which make them promising candidates to be membrane proteins potentially involved in glucose transport. For similar reasons, another candidate gene that could be involved in glucose transport is ymgH. Other interesting differentially expressed genes are chiA and yucG, also located in a putative operon, which are overexpressed resp. 8.0 and 3.1 times in L. lactis NZ9000ΔptnABCDΔptcBA relative to L. lactis NZ9000. These genes are (putatively) involved in the breakdown of chitin, which is a polysaccharide of β-1,4-linked N-acetylglucosamine units. In L. lactis NZ9000ΔptnABCDΔglk, ytgAH, ymgH, chiA and yucG are also overexpressed, but to lesser extent than in L. lactis NZ9000ΔptnABCDΔptcBA.

75 Chapter 3

DISCUSSION

L. lactis NZ9000 has two types of glucose transport systems, the PEP-PTSs represented by PTSman/glc and PTScel/glc, and an unknown number of non-PTS transporters. Glucose fermentation through the latter requires glucokinase for phosphorylation of glucose after its internalization. In L. lactis NZ9000Δglk no glucose phosphorylating activity was detected, suggesting Glk is the only enzyme in L. lactis NZ9000 able to phosphorylate glucose intracellularly. Disruption of glk leaves only the PTSs to transport and phosphorylate glucose. The PTSs are considered to be the main glucose transporters in L. lactis (36, 148, 191). PTScel/glc appears to have the smaller contribution to the overall glucose uptake by the two glucose-transporting PTSs, since deletion of PTScel/glc only decreased man/glc cel/glc Vmax and Km about 2-fold. Deletion of PTS alone or together with PTS , resulted in an approximately 1000-fold decreased affinity for glucose, showing that PTSman/glc is the major glucose transporter in L. lactis.

Removal of the glk gene led to several changes in the uptake and further metabolism of glucose by the cell, such as a lower glucose consumption rate and a shift to a mixed-acid fermentation. The kinetic parameters of glucose uptake were also affected by deletion of glk. Especially at low glucose concentrations, when PTSman/glc operates, the maximal capacity was lowered almost 20 times, while the affinity decreased about 4 times. Furthermore, deletion of glk together with ptnABCD resulted in transcriptional activation of ptsH and ccpA, which did not occur (ptsH) or was less pronounced (ccpA) in a ptnABCDptcBA deletion strain. All these data suggest a direct or indirect regulatory role for Glk in glucose metabolism of L. lactis. The presence of a ROK-motif in Glk motivates this suggestion. Proteins belonging to the ROK-family known so far are bacterial sugar kinases, transcriptional repressors, or have as yet uncharacterized functions (196). Interestingly, glucokinases from the Gram-positive bacteria Streptomyces coelicolor, Staphylococcus xylosus, Bacillus megaterium and Corynebacterium glutamicum, which also contain a ROK-motif, contribute to carbon catabolite repression (CCR), although their precise roles remain unknown (86, 134, 177, 214). Proteins having a NagC-domain usually also have a ROK-domain. The abbreviation NagC is derived from the regulation of the use of N-acetylglucosamine. E. coli NagC and Mlc are two sugar-specific regulatory proteins (140, 166) that belong to the ROK-family of proteins. They have additional almost identical DNA-binding motifs, although NagC coordinates the metabolism of aminosugars and Mlc regulates genes involved in

76 Glucose transport and metabolism sugar uptake together with the cAMP/CAP complex (139). Glk proteins without a ROK-motif have so far not been shown to fulfil regulatory functions. Although the main role of glucokinase in L. lactis so far was thought to be a catalytic function in the metabolism of glucose, the results obtained here suggest also a possible regulatory role for L. lactis Glk. Besides this, the activity of Glk itself was affected by the disruption of PTSman/glc or PTScel/glc. Interestingly, deletion of either of the two PTSs had opposite effects on Glk. Glucokinase activity was slightly decreased when PTSglc/man was deleted, but deletion of PTScel/glc resulted in a five-fold increased activity. Also, Glk activity was lower in L. lactis NZ9000ΔptnABCDΔptcBA than in L. lactis NZ9000, although glk transcription was not downregulated in the double mutant, indicating that Glk is regulated by means other than transcriptional control, e.g. translational or post- translational regulation. At this point, a full explanation for these results cannot be put forward.

It has been described more than a decade ago that uptake of glucose in L. lactis takes place at (at least) two different sites with anomeric specificity (11). Using the transport system mutants in combination with in vivo NMR allowed elucidating the anomeric preferences of the transporters. The glucose transport system(s) operative in L. lactis NZ9000ΔptnABCD displayed clear specificity for β-glucose. As PTScel/glc is the main glucose transporter in this mutant, since the Vmax and the Km are slightly higher than in L. lactis NZ9000ΔptnABCDΔptcBA, we propose that PTScel/glc has a preference for β-glucose. To unequivocally prove this, the anomeric specificity should be tested in an L. lactis strain deleted for ptnABCD together with the gene(s) encoding the as yet unknown non-PTS transporter(s). The preference for β-glucose of L. lactis NZ9000ΔptnABCDΔptcBA, lacking both glucose-transporting PTS systems, indicates that the non-PTS transporter(s) also has/have a preference for β- glucose. PTSglc/man displays no clear anomeric specificity as L. lactis NZ9000ΔptcBA does not have an absolute preference for either anomer; only a slight preference for the β-anomer is seen. Possessing multiple systems for glucose transport, each with its own glucose anomer specificity, may be an evolutionary advantage for L. lactis.

Although transcriptional downregulation did not occur significantly for ldh and pyk in L. lactis NZ9000ΔptnABCDΔptcBA compared to L. lactis NZ9000, all deletion strains tested had lower PK and LDH enzyme activities than their parent, L. lactis NZ9000. The in vivo activities of PK and LDH in the mutants during glucose metabolism may be even lower than the activity measured in cell-free extracts, since the concentration of FBP, the positive effector of PK and LDH (191), is lower in the

77 Chapter 3 deletion strains than in L. lactis NZ9000 (Table 2). Furthermore, PK and LDH are negatively controlled by high levels of inorganic phosphate (Pi) (191). The concentration of Pi is much higher in L. lactis NZ9000ΔptnABCD (and probably similar to that in L. lactis NZ9000ΔptnABCDΔptcBA) than in L. lactis NZ9000 (data not shown). The concentrations of FBP and Pi counterbalance each other as far as regulation of PK- and LDH-activity is concerned. The lower LDH-activity is in accordance with the mixed-acid fermentation taking place in the deletion strains: the flux towards lactate is decreased and other products can be formed by PFL from pyruvate. Indeed, the pfl gene was shown to be transcriptionally overexpressed (five-fold) in L. lactis NZ9000ΔptnABCDΔptcBA relative to L. lactis NZ9000. The levels of the intracellular metabolites FBP, 3-PGA and PEP were closely related to the glucose transporter used. When the main glucose transporter PTSman/glc is removed alone or together with PTScel/glc, high accumulation of PEP and 3-PGA occured. This can be explained by the fact that without the PTSs no PEP is used for glucose uptake and concomitant phospho-transfer reactions. However, when PTSman/glc is the only PTS used, as in L. lactis NZ9000ΔptcBA, PEP and 3-PGA do not accumulate, and FBP accumulates to a lower maximum level. Deletion of PTScel/glc resulted in an approximately 4 times higher Glk activity compared to the wildtype strain. The higher activity of Glk may activate PTSman/glc, which would result in less PEP accumulation. This could also explain why high levels of PEP and 3PGA are formed in the glk deletion strain. In that case no activating Glk is present and PTSman/glc is less active, as shown by the glucose transport assays.The differences in the concentrations of PEP and 3PGA could also be caused by other regulatory effects, like those imposed by the concentrations of FBP and Pi. Furthermore, the pyruvate pool is controlled by the activity of PK, which catalyzes the reaction from PEP to pyruvate. PK was shown to be less active in all deletion strains tested.

In L. lactis, sugar consumption rate has been associated with the shift to mixed-acid fermentation (50). Furthermore, modulation of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) by the NADH/NAD+ ratio was put forward as the main factor controlling the glycolytic flux (50). In line with this view, homolactic metabolism is a consequence of high NADH/NAD+ ratios (0.05 or above), which inhibit GAPDH. Previously, we had shown that L. lactis glyceraldehyde 3-phosphate is able to support a high glycolytic flux despite a high NADH/NAD+ ratio (127). In this study, the glucose transport mutants are characterized by moderately low glucose consumption rates and they display a clear switch to mixed-acid metabolism (Fig. 2). However, production of acetate and ethanol during the slow phase of glucose consumption was associated with a high NADH/NAD+ ratio, whereas the faster

78 Glucose transport and metabolism conversion of glucose to lactate is accompanied by a low ratio of NADH/NAD+ (Fig. 3). These observations reinforce the view that the NADH/NAD+ ratio is not the main factor controlling the glycolytic flux (127). Our data support the hypothesis that glycolytic flux is limited by the transport step.

The transcriptome comparison of L. lactis NZ9000ΔptnABCDΔptcBA and L. lactis NZ9000 revealed a higher expression of pfl, pdhA and pdhB in the former strain. This is in accordance with the shift to a mixed-acid fermentation, as shown by the in vivo NMR experiments. Both the anaerobic (pfl) and the aerobic fermentation route (pdh) are overexpressed, since the cells used for the transcriptome analysis were grown under microaerobic conditions. Surprisingly, also chiA and the downstream gene yucG (both putatively involved in the degradation of chitin) were overexpressed, respectively 8 and 3 times, in L. lactis NZ9000ΔptnABCDΔptcBA. The overexpression of these genes cannot be fully explained, but one could speculate that overexpression of chiA and yucG is a stress response of the cell to prepare for metabolism of polysaccharides. Possible new candidates that could be directly or indirectly involved in glucose transport or (the regulation of) glucose metabolism and, thus, are interesting subjects for further investigation are ytgBAH. The function of these genes is yet unknown, but they are highly upregulated in L. lactis NZ9000ΔptnABCDΔptcBA. ytgA and ytgB are potential membrane proteins and could be involved in non-PTS glucose transport. Also ymgH is a potential candidate glucose transporter, since it is also a putative membrane protein.

Summarizing, we show that L. lactis strain NZ9000 is able to transport glucose via three different transport systems, via one of the two PTSs (PTSglc/man or PTScel/glc) or via non-PTS transport. Deletion of one or two of these systems gives rise to major changes in glucose metabolism. The main route for glucose transport was shown to be via PTSglc/man, and a regulatory role for Glk in glucose transport was proposed. It remains to be investigated which targets are regulated by Glk, and if regulation is direct or perhaps occurs via the global carbon metabolism regulator CcpA. Each transport system has its own characteristics, and a clear preference for the β- anomer of glucose was displayed by PTScel/glc and the non-PTS transporter(s).

79 Chapter 3

ACKNOWLEDGEMENTS

This work was sponsored by the European Commission through contract QLK1-CT- 2000-01376, within the research programme "Quality of Life and Management of Living Resources" under the Key Action "Food, Nutrition & Health" acronym: Nutra Cells. We would like to thank Thijs Kouwen for expert technical assistance and Luis

Fonseca for the Vmax and Km calculations of the β- and α-glucose anomers.

80