MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains and plasmids used in this study are listed in Table 1.

Media and growth conditions. Escherichia coli strains were grown on Luria-Bertani (LB) medium at 37^oC (Sambrook et al., 1989). X. flavus strains were grown in minimal media supplemented with gluconate (10 mM), succinate (10 mM) or methanol (0.5%, v/v) at 30^oC as described previously (Meijer et al., 1990) X. flavus was grown on a mixture of gluconate (5 mM) and formate (20 mM) in a 3 L batch fermentor with automatic titration with formic acid (25%, v/v) to maintain a constant pH (Meijer et al., 1991). When appropriate, the following supplements were added: ampicillin, 100 µg ml^-1; X-Gal, 20 µg ml^-1; isopropyl-ß-D-thiogalactoside (IPTG) 0.1 mM; tetracycline, 12.5 µg ml^-1 (E. coli) or 7 µg ml^-1 (X. flavus);

kanamycin, 5 µg ml^-1. Agar was added for solid media (1.5%, w/v).

Mobilization of plasmids. Mobilization of plasmids to X. flavus using E. coli S17-1 containing the appropriate plasmids was performed as described previously (Simon et al., 1983).

DNA manipulations. Plasmid DNA was isolated via the alkaline lysis method of Birnboim &

Doly (Birnboim & Doly, 1979). DNA modifying enzymes were obtained from Boehringer Mannheim and were used according tot the manufacturer’s instructions. DNA fragments were isolated from agarose gels by using the Geneclean DNA purification kit from BIO 101.

Other DNA manipulations were done in accordance with standard protocols (Sambrook et al., 1989). Oligonucleotides were obtained from Eurogentec. PCR was carried out by using PWO DNA polymerase as recommended by the manufacturer (Boehringer).

Nucleotide sequencing and phylogenetic tree construction. Nucleotide sequencing was done using dye-primers in the cycle sequencing method with the thermosequenase kit RPN 2538 from Amersham Pharmacia Biotech AB. The samples were run on the A.L.F.-Express sequencing robot. The nucleotide sequence data were compiled and analyzed using the programs supplied in the PC/GENE software package (IntelliGenetics, Mountain View, Calif.). Amino acid sequences were aligned using ClustalW (Thompson et al., 1994).

Distance between sequences for phylogenetic inference with the neighbor-joining method (Saitou & Nei, 1987) was measured as numbers of amino acid substitutions per site correction for multiple substitutions with the Dayhoff matrix of the program Treecon for Windows version 1.3b (van de Peer & De Wachter, 1995). A bootstrap analysis was carried out to test the reliability of the tree (Felsenstein, 1985).

TABLE 1. Bacteria and plasmids used in this study Strain or

plasmid Genotype or characteristics Source or reference

E. coli

DH5α supE44 ∆lacU169 (φ80lacZ∆M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1

R22 cbbR::Km (van den Bergh et

al., 1993) Plasmids

pBluescriptKSII Ap^R, lacZ’, cloning vector Stratagene

pBC3 Tc^R, translational fusion, lacZ (Meijer et al., 1991)

pYS5 Tc^R incP1 mob tkt gap pgk, 5.8-kb HindIII-SacI fragment (Meijer et al., 1996) pYO2041 Ap^R, tpi, 1.1-kb ApaI fragment in pBluescript KSII (Meijer et al., 1997) pYSL2 Tc^R incP1 mob ‘gap φ(pgk’-lacZ) (Meijer et al., 1996) pBSfTPI Ap^R, tpi’, 300-bp BamHI-EcoRI fragment in pBluescript KSII This study

pBCfTPI Tc^R incP1 mob φ(tpi’-lacZ) This study

pBSfTkt416 Ap^R, tkt’, 325-bp BamHI-EcoRI fragment in pBluescript KSII This study

PBCfTkt416 Tc^R incP1 mob φ(tkt’-lacZ) This study

Construction of promoter fusion vectors. The promoter of tpi present on plasmid pYO2041 was amplified by PCR using the oligonucleotides TPIBAM CAAGGATCCGGCGCGGAAGCGGCGCATCATAGG-3') and TPIECO (5'-CGCGAATTCATCTTCCAGTTGCCTGCGATCAGG-3') containing respectively a BamHI and EcoRI restriction site. The resulting DNA fragment was digested with BamHI and EcoRI and ligated into pBluescript KS digested with the same enzymes. The nucleotide sequence of the resulting plasmid (pBSfTPI) was determined to verify that unwanted mutations were not introduced in the PCR. Plasmid pBSfTPI was digested with BamHI and EcoRI and ligated into pBC3 which was digested with the same enzymes. The resulting in-frame tpi-lacZ fusion plasmid pBCfTPI was mobilized to X. flavus strains H4-14 and R22 using E. coli S17-1.

The promoter of tkt present on plasmid pYS5 was amplified by PCR using the oligonucleotides TktBAM416 CGCGGATCCCTTATGTTGCGTCG-3') and TktECO (5'-CATGAATTCGTGACGCTGTCGCTGGTG-3') containing respectively a BamHI and EcoRI restriction site. The strategy as described above was applied to create a tkt-lacZ fusion plasmid resulting in plasmids pBSfTkt416 and pBCfTkt416. Plasmid pBCfTkt416 was mobilized to X. flavus strains H4-14 and R22 using E. coli S17-1.

Enzyme assays. Cell free extracts were prepared by using a French pressure cell as described previously (Meijer et al., 1991). RuBisCO (EC 4.1.1.39) activity was determined by measuring the incorporation of ¹⁴CO2 into acid-stable compounds (Gibson & Tabita, 1977). Phosphoglycerate kinase (PGK, EC 2.7.2.3), and triosephosphate isomerase (TPI, EC 5.3.1.1) activities were determined as described previously (Meijer et al., 1997; Meijer, 1994). Transketolase (EC 2.2.1.1) and transaldolase (EC 2.2.1.2) activities were determined as described by Levering et al (Levering et al., 1982). β-Galactosidase activity was determined according to Miller except that cell free extracts were used instead of cell suspensions (Miller, 1972). Protein was determined according to Bradford using bovine serum albumin as standard (Bradford, 1976).

Nucleotide sequence accession number. The nucleotide sequence presented in this paper has been assigned GenBank accession no. AF223218

RESULTS

Nucleotide sequence of the tkt gene

We previously reported the presence of an additional transketolase gene upstream of the gap-pgk operon (Meijer et al., 1996). The nucleotide sequence of the pYS5 insert was determined in order to characterize the complete tkt gene and its upstream region (Fig. 1).

One large open reading frame (ORF) of 2004 bp was identified (Fig. 2) which was preceded by a putative ribosome binding site and could encode a protein of 668 amino acids with a molecular mass of 70,202 Da. The deduced amino acid sequence of the ORF was compared with entries in GenBank at the National Center for Biotechnology Information (Bethesda, Md) using BLASTP. The putative protein encoded by the ORF was found to be highly similar (57%) to transketolase proteins . Features typical for transketolase (Schenk et al., 1998) were present in the deduced amino acid sequence. A thiamine pyrophosphate (TPP) binding motif was found between amino acids 160 and 191 and a transketolase motif representing a modified dinucleotide binding motif was found between amino acids 463 and 498 of the Tkt protein (Fig. 2). The region upstream of the tkt gene was also sequenced and compared with entries in Genbank by using the program BLASTN. However no significant similarities were found. A putative promoter is present directly upstream of the tkt gene which resemble E. coli σ⁷⁰ promoter-like structures.

Phylogeny of transketolase

The presence of two transketolase genes in a single organism is not unprecedented. For example, E. coli, Ralstonia eutropha and Saccharomyces cerevisiae all have two or more transketolase encoding genes (Schäferjohann et al., 1993; Sprenger, 1995; Schaaff-Gerstenschlager & Zimmermann, 1993). The multiple transketolase genes in these organisms are closely related and most likely resulted from gene duplication. To examine the phylogenetic relationship between the transketolase genes in X. flavus, the amino acid sequences of CbbT and Ttk were aligned with other transketolase sequences using Clustal W (Thompson et al., 1994). A distance matrix was subsequently calculated with the model of Dayhoff et al. (Hasegawa & Fujiwara, 1993; Dayhoff et al., 1978) and used to construct a phylogenetic tree based on the neighbor joining method (Saitou & Nei, 1987). The branching order of the tree is conventional and in good agreement with the extensive study on the molecular evolutionary analysis of transketolases performed by Schenk et al. (Schenk et al., 1998). The transketolase proteins from the proteobacteria including the enzymes from X.

flavus, the Gram positive bacteria, yeast, plants and mammals each form distinct clusters

Figure 2. (Other page) Nucleotide sequence of tkt and upstream region. A putative ribosome binding site is underlined. Putative E. coli σ⁷⁰ promoter-like sequences are indicated by the numbers –35 and – 10 and by italicized nucleotides. The amino acid sequence of tkt is represented by single-letter code. A TPP binding motif and a transketolase motif are indicated by bold-faced underlined amino acids.

Figure 3. Unrooted phylogenetic tree showing the relationships between transketolases based on a distance matrix calculated using Dayhoff’s model constructed via the neighbor-joining method using the program Treecon for Windows version 1.3b. Numbers refer to the percentage of bootstrap samples (100 replicates) that support each topological element. Bar represents 0.1 substitutions per site. Amino acid sequences were taken from Aquifex aeolicus (AE000755), Bacillus subtilis (Z99113), Corynebacterium glutamicum (AB023377), Escherichia coli Tkt1 (X68025), Escherichia coli Tkt2 (D12473), Haemophilus influenza TktA (U32783), Homo sapiens (L12711), Kluyveromyces lactis (U65983), Mus Musculus (U05809), Mycobacterium leprae (Z99125), Mycobacterium tuberculosis (Z95844), Ralstonia eutropha CBBTc (M68904), Ralstonia eutropha CBBTp (M68905), Rhodobacter capsulatus (L48803), Rhodobacter sphaeroides (M68914), Saccharomyces cerevisiae Tkt1 (X73224), Saccharomyces cerevisiae Tkt2 (X73532), Solanum tuberosum (CAA90427), Spinacia oleracea (L76554), Streptococcus pneumoniae (M31296), Streptomyces coelicolor (AL031107), Synechocystis sp (D90905), Xanthobacter flavus CBBT (U29134).

(Fig. 3). In contrast to the proteins from R. eutropha, S. cerevisiae and E. coli, the CbbT and Tkt proteins of X. flavus do not group together. This strongly suggest that these genes did not originate from gene duplication.

The expression of tkt is not subject to autotrophic regulation

Transketolase is encoded by two genes in X. flavus. One of these, cbbT, is located in the cbb operon which is regulated by the NADPH-sensor CbbR. Nothing is known regarding the regulation of tkt, the second transketolase gene of X. flavus. To determine whether induction of the Calvin cycle genes affects expression of tkt, enzyme activity levels were determined during a transition from heterotrophic to autotrophic growth. The Calvin cycle was induced in wild type X. flavus by the addition of formate to cells growing on gluconate-containing medium. Transketolase enzyme activity in the wild type strain was present at a low and constant level before autotrophic growth induction. Following induction of the Calvin cycle by the addition of formate, transketolase activity increased threefold in parallel with RuBisCO

Figure 4. Activities of (A) RuBisCO (■) and PGK (▼); (B) β-galactosidase (●) and transketolase (▲) and (C) transaldolase (●) in X. flavus wild type (closed symbols) and X. flavus R22 (open symbols) following the addition of 20 mM formate to a culture growing on 5 mM gluconate at t=0 h. The pH is kept constant by automatic titration with formic acid (25% v/v). Graphs representing typical data from two independent experiments are shown. Each point in the graph represents the mean of two measurements. The values differed from the mean by < 10%. Enzyme activities are in nmol min^-1 (mg protein)^-1.

and PGK activities. In contrast, the activity of transaldolase, an enzyme of the pentose phosphate cycle, remained at a constant level throughout the experiment (Fig. 4). To assess whether the transcriptional regulator CbbR was required for the super-induction of transketolase activity, the experiment was repeated in a cbbR mutant strain, X. flavus R22.

As expected, induction of RuBisCO, and super-induction of PGK did not occur following addition of formate. The levels of transketolase and transaldolase activities also remained constant throughout autotrophic growth induction.

The results from these experiments show that CbbR is required for the observed increase in transketolase activity. Since cbbT is not expressed in the CbbR mutant strain, the observed transketolase activity in X. flavus R22 must be due to constitutive expression of the tkt gene.

However, this does not rule out that CbbR super-induces tkt expression as was seen for the gap-pgk operon or that CbbR represses tkt expression resulting in replacement of Tkt by CbbT. In order to analyze expression of the tkt gene in more detail, a fusion between tkt and lacZ was constructed (pBCfTkt416) and mobilized to X. flavus. The Calvin cycle was induced in the transconjugant as before, by adding formate to cells growing on gluconate-containing medium. The activity of the tkt promoter, as indicated by ß-galactosidase activity, remained constant following induction of the Calvin cycle (Fig. 4). Similar results were

Time (h) after induction

obtained using the X. flavus R22, indicating that CbbR is not involved in regulation of tkt expression.

The tpi gene is not part of the CbbR regulon

X. flavus has a single tpi gene, which is required for both heterotrophic and autotrophic growth (Meijer et al., 1997). The activity levels of GapDH and PGK rapidly increase following induction of the Calvin cycle due to super-induction of the gap-pgk operon (Fig. 4). To determine whether the tpi gene is also subject to super-induction by CbbR, a fusion between tpi and lacZ was constructed (pBCfTPI) and mobilised to X. flavus. X. flavus harbouring pBCfTPI was grown on gluconate and the Calvin cycle was induced by the addition of formate as before. In contrast to the rapid increase in RuBisCO, GapDH, PGK activity following addition of formate (data not shown), the activity of TPI and ß-galactosidase, indicative for tpi transcription, remained constant throughout the experiment (Fig. 5). Similar activities were observed when the experiment was repeated in the cbbR mutant X. flavus R22. These data show that the tpi gene is constitutively expressed and is not part of the CbbR regulon.

Figure 5. Activities of TPI (♦) and β-galactosidase (●) in X. flavus wild type (closed symbols) and X.

flavus R22 (open symbols) following the addition of 20 mM formate to a culture growing on 5 mM gluconate at t=0 h. The pH is kept constant by automatic titration with formic acid (25% v/v). Graphs representing typical data from two independent experiments are shown. Each point in the graph represents the mean of two measurements. The values differed from the mean by < 10%. Enzyme activities are in nmol min^-1 (mg protein)^-1.

gap and pgk constitute an operon under autotrophic growth conditions

We previously showed that expression of pgk during heterotrophic growth is dependent on transcription from a promoter upstream from the gap gene (Meijer et al., 1996). However, inspection of the intergenic region between gap and pgk revealed extensive similarities between the gap-pgk intergenic region and the cbb operon promoter (Fig. 1). To determine whether the CbbR mediated super-induction of pgk expression is due to activation of a promoter in between gap and pgk, a fusion (pYSL2) between pgk and lacZ was introduced in X. flavus. The Calvin cycle was subsequently induced in this strain by adding formate to the gluconate-grown X. flavus. ß-galactosidase activity was not detectable before or after induction of autotrophic growth (data not shown). This shows that the sequences upstream of pgk do not contain promoter activity, despite the similarities with the promoter of the cbb operon.

Time (h) after induction

-2 0 2 4 6

TPI, β-galactosidase

0 250 500 750 1000 1250 1500 1750 2000

DISCUSSION

Induction of the Calvin cycle to allow autotrophic growth is characterized by the appearance of RuBisCO, PRK and SBPase activities. In principle, acquisition of these enzyme activities, in conjunction with existing enzymes of the pentose phosphate cycle, glycolysis and gluconeogenesis, should allow operation of the Calvin cycle and hence CO2 assimilation.

However, CO2 assimilation to support autotrophic growth requires a high flux of carbon and a different type of regulation of the non-unique Calvin cycle enzymes compared to heterotrophic growth. The induction of RuBisCO and PRK activities is therefore accompanied by extensive rearrangement of the central metabolic pathways. This manifests itself by an increase in activity of enzymes other than RuBisCO and PRK and synthesis of isoenzymes with different catalytic or regulatory properties.

This is in part achieved by induction of the cbb operon, which in addition to RuBisCO, PRK and SBPase, encodes a number of other Calvin cycle enzymes. Some of the latter have different properties than their isoenzyme counterparts which are active during heterotrophic growth. These are for example cbbF and cbbA encoding respectively FBPase/SBPase and a class II fructosebisphosphate aldolase (van den Bergh et al., 1996; van den Bergh et al., 1995). The former enzyme has both different regulatory and catalytic properties than its heterotrophic counterpart. The latter enzyme employs a different catalytic mechanism than the class I fructosebisphosphate aldolase found during heterotrophic growth.

Transketolase is a Calvin cycle enzyme of X. flavus which is also encoded both within and outside the cbb operon. The phylogeny of transketolase proteins of X. flavus presented in the present paper indicates that cbbT and tkt did not originate from a gene duplication event, as is clearly the case for the two cbbT genes of R. eutropha. The same is true for the fructosebisphosphate aldolase enzymes of X. flavus, which are completely unrelated. In fact, class II fructosebisphosphate aldolase encoded by cbbA is more closely related to enzymes of Gram positive bacteria than to those found in proteobacteria to which X. flavus belongs (van den Bergh et al., 1996). Considering that cbb operons are frequently present on self transmissible plasmids in both proteobacteria and Gram positives (Kusian & Bowien, 1997), it is plausible that the acquisition of the Calvin cycle in X. flavus is due to a lateral gene transfer of the cbb operon.

Acquisition of Calvin cycle genes by itself is not sufficient to allow autotrophic growth; the new pathway needs to be integrated into overall metabolism. In X. flavus, the cbb operon does not encode GapDH and PGK, yet high activity levels of these enzymes are necessary to sustain a high rate of CO2 fixation required for autotrophic growth. This problem is solved by placing gap-pgk under CbbR control. Consequently, the gap-pgk operon of X. flavus is constitutively transcribed during heterotrophic growth and is super-induced by CbbR during autotrophic growth (Meijer et al., 1996). The data presented in this paper show that the CbbR regulon does not comprise tkt and tpi. These genes are expressed at a constitutive level and, in contrast to gap-pgk, do not respond to signals emanating from autotrophic metabolism. However, operation of the Calvin cycle requires high transketolase and triosephosphate isomerase activities. This is achieved by induction of the cbb operon, which results in expression of cbbT and hence increased transketolase activities (van den Bergh et al., 1996; Meijer et al., 1991). The transketolase activity profile therefore mimics that of GapDH and PGK, even though the mechanism to achieve it is completely different. The constitutive high level expression of tpi is apparently sufficient for both heterotrophic and autotrophic growth.

Surprisingly, transaldolase is also constitutively expressed, which was also observed for R.

eutropha (Bowien & Schlegel, 1972). This enzyme functions in the pentose phosphate cycle, and catalyses the same reaction as the combined activities of SBPase and sedoheptulosebisphosphate aldolase. However, all autotrophs examined to date employ an SBPase variant of the Calvin cycle, which, in contrast to the transaldolase variant, is irreversible. This may allow the cell to obtain high intracellular levels of ribulose-5-phosphate, the substrate of PRK. In addition, it apparently introduces another control point

in the Calvin cycle since the activity of SBPase of X. flavus is stimulated by ATP (van den Bergh et al., 1995). The simultaneous use of SBP aldolase/SBPase and transaldolase as is the case in X. flavus could therefore create a futile cycle. A possible explanation for this paradox could be that the enzymes involved in the Calvin cycle and pentose phosphate cycle are organised in enzyme complexes, allowing channeling of substrates through a particular set of enzymes. These complexes have been observed for both Calvin cycle and pentose phosphate cycle enzymes, e.g., (Joint et al., 1972; Suss et al., 1993; Gontero et al., 1988; Wood et al., 1985). Alternatively, the constitutive expression of transaldolase and Tkt during autotrophic growth could indicate that X. flavus only recently acquired the cbb operon and that mechanisms to repress tkt and the transaldolase encoding gene have ot yet evolved. Future experiments will therefore assess the role of the enzymes in autotrophic metabolism.

Reference list

Birnboim, H. C. & Doly, J. (1979). A rapid alkaline extraction procedure for screening of recombinant plasmid DNA. Nucleic Acids Res 7, 1513-1523.

Bowien, B. & Schlegel, H. G. (1972). Der Biosyntheseweg der RNS-Ribose in Hydrogenomonas eutropha Stamm H16 und Pseudomonas facilis. Arch Microbiol 85, 95-112.

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anal Biochem 73, 248-254.

Dayhoff, M. O., Schwartz, R. M., & Orcutt, B. C. (1978). A model for evolutionary change in proteins.

Dayhoff, M. O. Atlas of protein sequence and structure. pp 345-352. Washington, D.C.: National Biomedical Research Foundation.

Felsenstein, J. (1985). Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783-791.

Gibson, J. L. & Tabita, F. R. (1977). Different molecular forms of D-ribulose-1,5-bisphosphate carboxylase from Rhodopseudomonas sphaeroides. J Biol Chem 252, 943-949.

Gontero, B., Cardenas, M. L., & Ricard, J. (1988). A functional five-enzyme complex of chloroplasts involved in the Calvin cycle. Eur J Biochem 173, 437-443 .

Hasegawa, M. & Fujiwara, M. (1993). Relative efficiencies of the maximum likelihood, maximum parsimony, and neighbor-joining methods for estimating protein phylogeny. Mol Phyl Evol 2, 1-5 . Joint, I. R., Morris, I., & Fuller, R. C. (1972). Purification of a complex of alkaline fructose 1,6-bisphosphatase and phosphoribulokinase from Rhodospirillum rubrum. J Biol Chem 247, 4833-4838.

Kusian, B. & Bowien, B. (1997). Organization and regulation of cbb CO2 assimilation genes in autotrophic bacteria. FEMS Microbiol Rev 21, 135-155.

Levering, P. R., Dijkhuizen, L., & Harder, W. (1982). Enzymatic evidence for the operation of the FBP aldolase cleavage and TK/TA re-arrangement variant of the RuMP cycle in Arthrobacter P1. FEMS Microbiol Lett 14, 257-261.

Lidstrom-O'Conner, M. E., Fulton, G. L., & Wopat, A. E. (1983). 'Methylobacterium methanolicum': a

Lidstrom-O'Conner, M. E., Fulton, G. L., & Wopat, A. E. (1983). 'Methylobacterium methanolicum': a

In document University of Groningen Regulation of carbon dioxide fixation in the chemoautotroph Xanthobacter flavus Keulen, Geertje van (pagina 59-69)