The Plant-Like C2 Glycolate Cycle and the Bacterial-Like Glycerate Pathway Cooperate in Phosphoglycolate Metabolism in Cyanobacteria

Genes Encoding Enzymes Involved in 2-PG Metabolism in Synechocystis

Genes of Synechocystis homologous to those encoding enzymes involved in the C2 cycle of plants or the glycerate pathway of bacteria are shown in Table II . For PGP, HPR, and GCL, the Synechocystis genome appears to encode two proteins of identical enzymatic activity, which might catalyze the same reaction during 2-PG metabolism. It is also remarkable that the enzymes listed in Table II represent a mixture of plant- and bacterial-type enzymes. For example, Synechocystis harbors a bacterial-type glycerate kinase, which is distinct from the glycerate kinase present in plants, fungi, and some other cyanobacteria, such as Anabaena sp. strain PCC 7120 (Boldt et al., 2005). The same is true for PGP and Glc, which show no similarity to the corresponding enzymes from Arabidopsis. In contrast, Ala:glyoxylate aminotransferase (AGT), the four GDC subunits, SHMT, and HPR are clearly related to plant proteins (Table II).

From this in silico analysis, we conclude that genes encoding all the enzymes known to participate in glycolate metabolism in other organisms exist in Synechocystis and other cyanobacteria, too (results of corresponding searches in other cyanobacterial genomes are not shown). These findings allowed the construction of a hypothetical cyanobacterial glycolate cycle, comprising a complete C2 cycle, which overlaps with and is short-circuited by two enzymatic steps of the glycerate pathway (Fig. 1).

Synechocystis Exhibits Glc and SHMT Activities

To gain direct experimental support for this hypothesis, we first examined the biochemical functionality of two enzymes of this putative 2-PG-metabolizing pathway, encoded by glcD (sll0404) and shm (sll1931) identified in the Synechocystis genome. The putative GlcD was selected, since in plants the C2 cycle employs a glycolate oxidase and, despite high similarities to bacterial glycolate dehydrogenases, the ORF sll0404 is annotated as a subunit of glycolate oxidase in CyanoBase (http://www.kazusa.or.jp/cyanobase/Synechocystis/cgi-bin/orfinfo.cgi?title=Chr&name=sll0404&iden=1). The occurrence of SHMT in cyanobacteria is not clear at all (Norman and Colman, 1988). The respective genes, glcD (sll0404) and shm (sll1931), were heterologously expressed in Escherichia coli. Western-blot analyses, using a specific Strep-Tactin antibody, verified the expression and purification of the recombinant proteins, designated Syn-GlcD and Syn-SHMT, of 57 and 50 kD, respectively (Fig. 2 ). Specific activities were measured in crude extracts from E. coli cultures overexpressing these genes and with purified enzymes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Overexpression of glcD and shm from Synechocystis in E. coli to verify their biochemical activities. A, GlcD from Synechocystis with a molecular mass of 57 kD. B, SHMT from Synechocystis with a molecular mass of 50 kD. The relevant genes were cloned in IBA6, overexpressed, purified by the fused Strep-tag, and the activity tested (see “Materials and Methods”). M, Molecular mass standard (broad range; Bio-Rad); H, n.i., homogenate not induced; H, i., homogenate induced; Control, raw extract from not induced E. coli culture; Syn-GlcD, purified Sll0404; Syn-SHMT, purified Sll1931.

In agreement with the protein's primary structure, which indicates a flavin-binding site and shows highest similarities with glycolate dehydrogenase from E. coli (Table II), the GlcD protein showed NAD⁺-dependent glycolate dehydrogenase but no glycolate oxidase activity. The specific activity of affinity-purified Syn-GlcD was distinctly higher in comparison with controls prepared from E. coli cultures grown under identical conditions but without the cyanobacterial glcD (Fig. 2). These data correspond to earlier reports on the occurrence of Glc in cyanobacteria (Sallal and Codd, 1975; Grodzinski and Colman, 1976) and show very clearly that, in Synechocystis, conversion of glycolate to glyoxylate is catalyzed by a glycolate dehydrogenase. This differs from the situation in plants where photorespiratory glycolate is converted to glycolate by glycolate oxidase. However, it should be mentioned that the genome of another cyanobacterium, Anabaena sp. strain PCC 7120, in addition to GlcD, encodes a protein (All0170), which is very similar to the plant-type peroxisomal glycolate oxidase.

SHMT activity was not recognized in earlier studies using crude cyanobacterial extracts (Codd and Stewart, 1973) or the SHMT inhibitor isonicotinyl hydrazide (Cheng et al., 1972; Ingle and Colman, 1976; Norman and Colman, 1988). However, it must be noted that the estimation of SHMT activity is difficult since the methylene acceptor molecule tetrahydrofolate (THF) is not very stable in aqueous O₂-containing solution. In our experiments, using recombinant cyanobacterial SHMT, the specific activity of overexpressed Syn-SHMT was 7 times higher in comparison with control cell extracts from E. coli cells without the cyanobacterial shm (Fig. 2B). This finding provides clear evidence for the presence of a functional SHMT in Synechocystis, which is encoded by sll1931 as annotated in CyanoBase.

Mutational Approach to Study the Involvement of Specific Proteins in 2-PG Metabolism

To clarify the participation and importance of proteins possibly involved in 2-PG metabolism, we next generated mutants in almost all enzymatic steps of the hypothetical 2-PG-metabolizing pathway depicted in Figure 1. In the first reaction, the 2-PG produced by the oxygenase activity of Rubisco is converted to glycolate by the activity of PGP. Schäferjohann et al. (1993) expressed the cbbZ gene from the nonphotosynthesizing bacterium Alcaligenes eutrophus and showed that the gene product exhibits PGP activity. Synechocystis possesses two ORFs, sll1349 and slr0458, which encode proteins with homology to cbbZ (http://www.kazusa.or.jp/cyanobase/Synechocystis/index.html). Inactivation of each of these genes resulted in completely segregated mutants, whose growth characteristics under various conditions, including high or low CO₂ levels, did not differ from that of the wild type (data not shown). A double mutant where both genes were inactivated grew very slowly on agar plates maintained at 3% CO₂ but not under air level of CO₂. The colonies obtained were yellowish and we could not grow the cells in liquid cultures to examine PGP activity. One colony that possibly emerged following a suppression mutation grew well on an agar plate and in liquid culture. It showed normal appearance and normal PGP activity (data not shown). While the data described here do not provide conclusive evidence, they lend support to the notion that PGP activity is essential even under high CO₂ conditions since it should prevent 2-PG inhibition of triosephosphate isomerase (Norman and Colman, 1991).

We then generated mutants impaired in GlcD (sll0404, glcD), the aminotransferase (sll1559), the T protein of GDC (sll0171, gcvT), SHMT (slr1931; shm), the two putative GCLs (sll1981 or slr2088, gcl), and TSR (slr0229, tsr; compare Table I). After selection of antibiotic-resistant Synechocystis colonies and progressive segregation by increasing drug concentrations, the genotypes were examined using PCR. Analysis of the DNA from single and double mutants, where glcD, gcvT, gcl, and tsr or gcvT/gcl and gcvT/tsr were inactivated, respectively, showed only the fragments corresponding to the inactivated gene but not fragments derived from the wild-type genes. Representative results for mutants ΔglcD and ΔgcvT are shown in Figure 3A . These data indicated complete segregation of the chromosomal copies and suggested that under the growth conditions used here, the Glc, GDC, GCL, and TSR were dispensable for cyanobacterial metabolism. These results also indicate that the hypothetical C2 cycle as well as the glycerate pathway are not essential for viability under ambient CO₂ conditions at laboratory conditions, which, at first sight, may support the view that oxygenase activity of Rubisco is nearly zero in cyanobacteria.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Segregation of the Synechocystis mutants ΔglcD and ΔgcvT (A) and Δsll1559, Δshm, and Δslr2088 (B) was checked by PCR using total DNA of the mentioned strains and the gene-specific primers given in Table I. Abbreviations in A: M, λ-DNA EcoRI/HindIII; WT, wild-type 1.5 kb or 2.0 kb for glcD and gcvT, respectively. The modified (due to the insertions) glcD is 2.7 kb and gcvT is 3.0 kb, as expected. Abbreviations in B: M, λ-DNA EcoRI/HindIII; WT, wild-type 1.2 kb, 1.3 kb, or 2.0 kb; Δsll1559, 3.2 kb; Δshm, 2.5 kb; Δslr2088, 3.0 kb.

In contrast, only partial segregation could be achieved for the genomes of those mutants where we attempted to inactivate sll1559, shm, or slr2088. In all these latter cases, a dominant fragment of the expected wild-type size was observed in addition to a faint band corresponding to the mutated gene copies (Fig. 3B). Raising the CO₂ concentration during cultivation hardly affected the degree of segregation, suggesting essential functions for the putative aminotransferase Sll1559, SHMT, and the hypothetical GCL Slr2088. A similar result was reported for mutations in the L-protein subunit of GDC, which also could not be completely segregated (Engels and Pistorius, 1997). Apparently, the respective gene products fulfill essential tasks in the cyanobacterial cell. For the L protein, such a task is probably its additional function as a subunit of the pyruvate dehydrogenase complex. SHMT, as in other organisms, might be essential for one-carbon metabolism. The Sll1559 and Slr2088 proteins are most probably involved in basic amino acid metabolism and therefore not dispensable (Kouhen and Joset, 2002).

CCM Activity and Expression of Low-CO₂-Dependent Genes in a GDC Mutant

It was possible that the lack of a clear phenotype in some of the Synechocystis mutants impaired in either the C2 cycle or the glycerate pathway is due to enhanced CCM activity or faster acclimation of the mutants to low levels of CO₂ in comparison with the wild type. To test this possibility, we used a membrane-inlet mass spectrometer to measure inorganic carbon uptake (Tchernov et al., 1997, 2003) and reverse transcription (RT)-PCR to assess the level of genes whose expression is strongly affected by the concentration of CO₂ (Shibata et al., 2002; McGinn et al., 2004; Wang et al., 2004). Detailed analyses did not show marked differences between the rate of CO₂ uptake in high-CO₂-grown wild type and mutants impaired in GlcD or GDC. Displacements of CO₂ concentration from equilibrium in the light, consequent of CO₂ uptake and inorganic carbon cycling (Tchernov et al., 2003), were essentially similar in the wild type and the mutants. Furthermore, acclimation of the cells to air level of CO₂ enhanced the rate of CO₂ uptake in both the wild type and the mutant to the same extent (data not shown).

Analysis of the abundance of transcripts originating from ndhF3, cmpA, and sbtA, which are known to be up-regulated in Synechocystis cells exposed to a low level of CO₂ (Shibata et al., 2002), were analyzed in wild type and selected mutant cells. In Figure 4 , we show only representative data from the wild type and a mutant impaired in gcvT grown under high CO₂ and transferred to air level of CO₂ for 1 h. In principal, the mutant affected in the T-protein subunit of GDC showed similar CO₂-dependent regulation as wild-type cells. Under high CO₂, we did not find transcripts; however, a shift to air level of CO₂ resulted in the induction of these genes, even though a weaker transcription of these genes was detected in the mutant (Fig. 4). Taken together, the data from the membrane-inlet mass spectrometer and the analysis of the abundance of transcripts did not support the possibility that ability to withstand these mutations (albeit with some decline in the growth rates in mutants ΔglcD and ΔgcvT; see below) rested on enhanced CCM activity and thereby reduced metabolic flux in the glycolate pathway.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Influence of carbon limitation on the expression of certain genes in the wild type (WT) and a ΔgcvT mutant of Synechocystis. Total RNA was obtained from cells grown at 5% CO₂ (HC) or air level of CO₂ (LC). The transcript abundance was assessed by RT-PCR using the gene-specific primers given in the “Materials and Methods” section.

Growth Experiments

To find out whether there exist more subtle effects on the metabolic performance, we examined the growth rates under 5% CO₂ or air level of CO₂ for those mutants where the genes encoding GlcD (glcD), T protein of GDC (gcvT), or TSR (tsr) were individually inactivated. To distinguish between the effects on 2-PG turnover brought about by either the C2 cycle or the glycerate pathway or both (see Fig. 1), we also included the double mutant ΔgcvT/Δtsr in these analyses.

When aerated with 5% CO₂ in air, the wild type and most of the mutants showed similar growth rates. The only exception was mutant ΔglcD (Fig. 5 ), where both the C2 cycle and the glycerate pathway (Fig. 1) were inactivated. This mutant strain grew significantly poorer and reached only about 85% of the wild-type growth rate. This can be interpreted as the first hint that 2-PG metabolism occurs in Synechocystis and is important even under high CO₂ conditions. When the wild-type and mutant cultures were transferred to air level of CO₂, their growth was significantly reduced. The least affected strains were wild type (growth rate of wild-type cells at 5% CO₂: 0.072 ± 0.007 h;⁻¹ growth rate of wild-type cells at air level of CO₂: 0.021 ± 0.005 h⁻¹) and the single mutant Δtsr whose growth rate was reduced to one-third as compared to growth of high CO₂ cells. The growth rate of the other strains was only 40% to 70% that of the wild type under low CO₂ level. Here, mutants ΔglcD and ΔgcvT showed only about 70% of the growth rate of wild-type cells. Growth of the double mutant ΔgcvT/Δtsr, where both branches of the hypothetical 2-PG metabolism are impaired, was even more reduced under air conditions (40% compared to wild type), unlike the corresponding single mutants ΔgcvT and particularly Δtsr (Fig. 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Influence of carbon limitation on growth of wild type (WT) and cells bearing mutations in the indicated genes of Synechocystis. Growth was measured as increase in OD₇₅₀ over time. Means and sds from three independent experiments are shown. Growth rates of the wild-type cells grown under 5% CO₂ or air were set to 100% (0.072 ± 0.007 h⁻¹ and 0.021 ± 0.005 h⁻¹, respectively; *, statistically significant growth difference between wild-type and mutant cells, P ≤ 0.05).

Apparently, low CO₂ conditions trigger the differences between the wild type and several mutants impaired in glycolate metabolization. This shows that, under these conditions, the presence of a fully functional 2-PG metabolism becomes more important and provides strong support to the suggestion that the two branches of 2-PG metabolism, as shown in Figure 1, are indeed metabolically active in Synechocystis.

Levels of Glycolate, Gly, Ser, and Lys

We next analyzed whether the observed growth retardation of several strains is accompanied by a corresponding accumulation of intermediates of 2-PG metabolism. To this end, we examined the levels of glycolate and selected amino acids in wild-type and mutant cells grown under 5% CO₂ as well as in cells transferred for a few hours from high to low CO₂ conditions (Figs. 6 and 7 ). As reported for other cyanobacterial strains (Colman, 1989), extracts of Synechocystis cells grown at standard conditions with 5% CO₂ were essentially free of glycolate. In contrast to wild type and the other mutant strains, intracellular glycolate accumulation was observed only in the case of mutant ΔglcD (where the conversion of glycolate to glyoxylate should be inhibited). Glycolate accumulation in this mutant occurred even at 5% CO₂ (Fig. 6). Glycolate was at undetectable levels not only in samples from wild type and mutants ΔgcvT and Δtsr, where either the C2 cycle or the glycerate pathway is inhibited, but also in the double mutant ΔgcvT/Δtsr, where both routes are inhibited (data not shown). The significant glycolate accumulation in mutant ΔglcD even after 24 h in air, i.e. long after the CCM induction should have been completed (Woodger et al., 2005), suggests that the CO₂ concentration in close proximity to Rubisco within the carboxysomes may not be high enough to completely abolish the enzyme's oxygenase activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Glycolate content in the cells of mutant ΔglcD of Synechocystis. Samples were taken 3 or 24 h after transfer of high-CO₂-grown cells into air or after 24 h growth at 5% CO₂, respectively. Low-molecular-mass compounds were isolated from cell pellets, and glycolate was detected and quantified by HPLC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

The intracellular content of Gly (A), Ser (B), and Lys (C) in cells of the wild type and mutants of Synechocystis grown under high (5%, gray columns) or ambient CO₂ (white columns) concentration. Low-molecular-mass compounds were isolated from cell pellets, and amino acids were detected and quantified by HPLC.

In some filamentous cyanobacterial strains, glycolate excretion was observed under selected conditions (Norman and Colman, 1988; Renström and Bergman, 1989), which, in principle, could also explain the undetectable low intracellular amounts of glycolate in Synechocystis and most of the analyzed mutants. To consider this possibility, too, we analyzed the glycolate contents of medium from low-CO₂-grown ΔglcD and wild-type cells and did not find any sign of glycolate accumulation (data not shown). We conclude that Synechocystis does not excrete significant amounts of glycolate corresponding to the observed intracellular glycolate accumulation in mutant ΔglcD. Accordingly, the complete lack of a significant glycolate accumulation in wild-type cells under low and high CO₂ levels can therefore only be explained by the presence of an active glycolate metabolism in Synechocystis. The lack of glycolate accumulation in other mutants, particularly in the double mutant ΔgcvT/Δtsr, may be attributed to accumulation of other intermediates like glyoxylate or tartronic semialdehyde, which we could not analyze. It may also indicate the existence of additional routes for the metabolization of glycolate and/or glyoxylate. For example, complete decarboxylation of glycolate via formate was proposed in previous studies (Norman and Colman, 1992).

The plant-type C2 cycle, by aminotransferase reactions and by the Gly-to-Ser interconversion via GDC and SHMT, is closely related to amino acid metabolism. The low level of glycolate in the ΔgcvT/Δtsr mutant therefore prompted us to examine the possibility of changes in the level of selected amino acids (Fig. 7). When grown under 5% CO₂ in air, the intracellular levels of Gly, Ser, and Lys were not very strongly affected by the various mutations examined here. In contrast, when exposed to air level of CO₂, several of the mutants accumulated large quantities of Gly and Lys, but not Ser. The relatively small effect of CO₂ concentration on the accumulation of Ser is consistent with the data obtained with higher plants where the ratio of Gly to Ser is strongly affected by photorespiratory conditions, mainly due to altered levels of Gly (Snyder and Tolbert, 1974; Heineke et al., 2001; Novitskaya et al., 2002). Under air conditions, the intracellular contents of Gly and Lys were strikingly changed in the single ΔgcvT and in the double ΔgcvT/Δtsr mutant, where the T-protein subunit of GDC was inactivated. These mutants accumulated much more Gly and Lys than did wild-type cells exposed to low CO₂. This may explain the slower growth of these mutants under low CO₂ (see Fig. 5) since elevated Gly concentrations appear to be toxic for Synechocystis (Hagemann et al., 2005). On the other hand, the normal growth phenotype of air-grown mutant Δtsr is consistent with the lower levels of Gly and Lys observed in this mutant, relative to ΔgcvT (Fig. 7). The close correlation between the levels of Gly and Lys in the mutants points to the presence of a metabolic link between these two amino acids. Theoretically, Gly can be converted to Lys via the homoserine or, alternatively, via the malate-oxalacetate-Asp route. Most of the necessary enzymatic steps are obviously encoded in the Synechocystis genome (Kyoto Encyclopedia of Genes and Genomes; http://www.genome.ad.jp/dbget-bin/get_pathway?org_name=syn&mapno=00300).

Strains and Culture Conditions

The Glc-tolerant Synechocystis sp. strain PCC 6803 was obtained from Prof. N. Murata (National Institute for Basic Biology, Okazaki, Japan) and served as the wild type. Axenic cultures were grown on agar plates at 30°C under constant illumination using agar-solidified BG11 medium (Rippka et al., 1979) buffered with 20 mm TES-KOH or 20 mm HEPES-NaOH to pH 8.0. Transformants were initially selected on media containing 10 μg mL⁻¹ kanamycin (Km), 4 μg mL⁻¹ spectinomycin (Sp), or 5 μg mL⁻¹ chloramphenicol (Cm; Sigma), respectively, whereas the segregation of clones and cultivation of mutants were performed at 50 μg mL⁻¹ Km, 20 μg mL⁻¹ Sp, or 15 μg mL⁻¹ Cm. For the physiological characterization, axenic cultures (about 10⁸ cells mL⁻¹) were grown photoautotrophically in batch cultures at 29°C under continuous illumination of 130 μmol photons m⁻² s⁻¹ (warm light; Osram L58 W32/3) bubbling with air enriched with CO₂ (5% CO₂ in air) in the BG11 medium at pH 8.0. Cultivation at ambient CO₂ concentration (0.035% CO₂ in air) was performed by bubbling with air in BG11 medium at pH 7.0. Growth was monitored by measurements of the optical density at 750 nm (OD₇₅₀). Contamination by heterotrophic bacteria was checked by spreading of 0.2 mL of culture on LB plates.

The Escherichia coli strain TG1 (Sambrook et al., 1989), cultured in LB medium at 37°C, was used for routine DNA manipulations. E. coli strain BL21 Gold was used for heterologous expression of selected genes.

DNA Manipulations

Total DNA from Synechocystis was isolated according to Hagemann et al. (1997). All other DNA techniques, such as plasmid isolation, transformation of E. coli, ligations, and restriction analysis (restriction enzymes were obtained from Life Technologies and New England Biolabs), were standard methods (Sambrook et al., 1989). PCR with gene-specific oligonucleotides (see Table I) was carried out using the Taq-PCR Master Mix (Qiagen). PCR products were cloned into pGEM-T (Promega). Sequence comparisons were made using the PSI- and PHI-BLAST software packages (Altschul et al., 1997).

Isolation of RNA and RT-PCR Analyses

Cultures were grown to a cell density of OD₇₅₀ = 0.5 under 5% CO₂ and transferred to air level of CO₂ for various durations. Analyses of transcript abundances of ndhF3, cmpA, sbtA, and 16S rRNA genes were performed using RT-PCR on RNA isolated from the wild type and various mutants. The cells were lysed and total RNA was prepared as done by Katoh et al. (2001). One hundred nanograms of total RNA was used in the reverse transcriptase reaction using Superscript II (Invitrogen) and random hexamer primers to synthesize cDNA, followed by PCR amplification using gene-specific primers [cmpA (slr0040), Fw GAG CGG ATT GGG TAG ATA A, Rev GGT CAA TAT CGG TGT CAG GA; sbtA (slr1512), Fw CCT GGC TAA ACT GCC TAA C, Rev GGC TTG GTA AAT ATG CCC AG; ndhF3 (sll1732), Fw TTC CCG CCA CTA TTC TAA GG, Rev GGT AGG GCA ATA AGG GAT GC; rnpB, served as internal standard, Fw AGC GTC CGT AGG TGG TTA T, Rev GTG TCA GTT TCA GCC CAG T] for 12 to 25 cycles depending on the abundance of the transcript.

Generation of Mutants

To generate mutations in the selected ORFs, by insertion or deletion, the Km, Sp, or Cm resistance cartridges were integrated into the coding sequences at selected restriction sites (see Table I). The products were checked by restriction analysis. Plasmid DNA of these constructs was isolated from E. coli using the GFX micro plasmid prep kit (Amersham). About 1 μg of DNA was used for transformation of Synechocystis, and Km-, Sp-, and Cm-resistant clones, respectively, were selected (Hagemann et al., 1997).

Purification of Recombinant Syn-GlcD and Syn-SHMT

The genes encoding Syn-GlcD and Syn-SHMT were amplified by PCR using chromosomal Synechocystis wild-type DNA, gene-specific primers (Syn-GlcD: Fw 5′-CTC GAG ATG GCC ATT TTC TCC-3′ and Rev 5′-GAA TTC TCA ATA AAT TTC CTC-3′; Syn-SHMT: Fw 5′-CTC GAG GTG AAT CAA ACC AAC-3′ and Rev 5′-CAG CTG TCA GGC GAT CAC CGC-3′), and the proofreading Pfu polymerase. Each forward primer contained an extension with an additional XhoI site (underlined). PCR fragments were cloned into pGEM-T. The generated plasmids pG-GlcD and pG-SHMT were cleaved with XhoI/NcoI and XhoI/PstI, respectively. After gel elution the fragments were ligated into the plasmid pASK-IBA6 (IBA) between the XhoI and NcoI or PstI sites, respectively. The in-frame insertions led, after induction by anhydrotetracycline, to the expression of fusion proteins with an N-terminal Strep-tag, which facilitated protein purification using Strep-Tactin sepharose (IBA). E. coli BL 21 Gold strains containing IBA6-Syn-GlcD and IBA6-Syn-SHMT constructs, respectively, were grown in LB medium to an OD₆₀₀ of 0.6. Expression of the fusion proteins was induced by addition of anhydrotetracycline (200 ng mL⁻¹). The cultures were allowed to grow for 16 h at 20°C. The cells were harvested by centrifugation and pellets were resuspended in buffer W (100 mm Tris-HCl, pH 8.0, 150 mm NaCl, 1 mm EDTA, IBA). Under ice-cooling, the suspensions were sonicated (6 × 10 s, 90 W) and the homogenates were centrifuged at 13,000 rpm for 15 min at 4°C. The supernatant was used for affinity chromatography following the protocol for Strep-tag protein purification under native conditions (IBA). The fractions of expression and purification were checked by western blotting using Strep-Tactin alkaline phosphatase conjugate (IBA) for detection.

Estimation of Enzyme Activities

Glycolate dehydrogenase (GlcD) activity was assayed by measuring the reduction of NAD⁺ as an increase of absorption at λ = 340 nm. Cell extracts or elution fractions of 100 μL (1 μg protein μL⁻¹) were added to a mixture comprising 50 μL of 3 mg mL⁻¹ NAD⁺ and 840 μL of 40 mm potassium phosphate buffer, pH 7.0. The reactions were initiated by adding 10 μL of 0.8 m glycolate adjusted to pH 6.0.

SHMT activity was assayed according to Taylor and Weissbach (1965). The rate of methylene-THF production from ¹⁴C-labeled Ser by SHMT was monitored. The produced methylene-THF was extracted after conversion to a dimedon complex with toluene and used for liquid scintillation counting, while the remaining substrate Ser and the second product Gly remained in the water phase. The complete assay mixture contained the following in a total volume of 1 mL: 10 μL 3-[¹⁴C]-l-Ser (2.11 GBq mmol⁻¹, Amersham), 40 μL of 100 mm l-Ser, 50 μL of 0.5 mm pyridoxal phosphate, 50 μL of 1% (w/v) THF, 50 μL of 50 mm KCl, and 200 μL of protein (2.5 μg/μL) in 0.4 m potassium phosphate buffer of pH 7.4. All components except radioactive Ser and THF were filled in a degassed tube and incubated for 5 min at room temperature. After addition of ¹⁴C-l-Ser and removal of 0.1-mL reference sample, the reaction was initiated by adding the substrate THF. After 5, 10, and 15 min, samples of 0.1 mL were taken and mixed with 0.1 mL of 0.1 m formaldehyde, 0.15 mL of 1 m sodium acetate, pH 4.5, and 0.15 mL of 0.4 m dimedon (in 50% ethanol). The mixtures were shaken and heated for 10 min at 100°C and stored on ice for 5 min. The dimedon complex was then extracted by vigorous shaking with 3 mL of toluene. Phase separation was reached by centrifugation for 5 min at 5,000 rpm. Two milliliters of the supernatant were prepared with 2 mL of scintillation mix in scintillation vials for counting.

Quantification of Internal Amino Acids and Glycolate Concentrations

Free amino acids were extracted from frozen cyanobacterial cell pellets of 2 mL of culture with 80% ethanol at 65°C for 3 h. After centrifugation, the supernatants were dried by lyophilization and redissolved in 8 mm Na₂HPO₄, pH 6.8, containing 2.5% tetrahydrofurane. Individual amino acids were assayed after derivatization with o-phthaldialdehyde (Gollan et al., 1992) as described by Geigenberger et al. (1996) using a Hypersil 120 ODS column (4.6 × 150 mm; Knauer) connected to a Class-Vp HPLC system (Shimadzu) with a fluorescence detector (RF-10AXL; Shimadzu).

Glycolate was extracted from frozen cyanobacterial cell pellets of 50 mL of culture with 80% ethanol at 65°C for 3 h. After centrifugation, the supernatants were dried by lyophilization and redissolved in 350 μL of water. The content of glycolate was determined by HPLC in ion exclusion mode. The operating parameters used were as follows: HPLC column, Aminex HPX-87H 300 × 7.8 mm (Bio-Rad); mobile phase, sulfuric acid 0.006 m; temperature, 65°C; and detection by refractive index. The measurements were carried out using an HPLC equipment of Knauer GmbH. Possible glycolate excretion was checked by the detection of glycolate in the medium. Fifty milliliters of suspension was taken from the cultures and cells were harvested by centrifugation (10 min, 4,000 rpm, 4°C). The supernatant was dried by lyophilized and redissolved in 800 μL of water before HPLC analysis.

All experiments were repeated at least three times using independent cell cultivations. Pair-wise t test was applied for the statistical comparison of mean values.