|Home | About | Journals | Submit | Contact Us | Français|
When deprived of combined nitrogen, some filamentous cyanobacteria contain two cell types: vegetative cells that fix CO2 through oxygenic photosynthesis and heterocysts that are specialized in N2 fixation. In the diazotrophic filament, the vegetative cells provide the heterocysts with reduced carbon (mainly in the form of sucrose) and heterocysts provide the vegetative cells with combined nitrogen. Septal junctions traverse peptidoglycan through structures known as nanopores and appear to mediate intercellular molecular transfer that can be traced with fluorescent markers, including the sucrose analog esculin (a coumarin glucoside) that is incorporated into the cells. Uptake of esculin by the model heterocyst-forming cyanobacterium Anabaena sp. strain PCC 7120 was inhibited by the α-glucosides sucrose and maltose. Analysis of Anabaena mutants identified components of three glucoside transporters that move esculin into the cells: GlsC (Alr4781) and GlsP (All0261) are an ATP-binding subunit and a permease subunit of two different ABC transporters, respectively, and HepP (All1711) is a major facilitator superfamily (MFS) protein that was shown previously to be involved in formation of the heterocyst envelope. Transfer of fluorescent markers (especially calcein) between vegetative cells of Anabaena was impaired by mutation of glucoside transporter genes. GlsP and HepP interact in bacterial two-hybrid assays with the septal junction-related protein SepJ, and GlsC was found to be necessary for the formation of a normal number of septal peptidoglycan nanopores and for normal subcellular localization of SepJ. Therefore, beyond their possible role in nutrient uptake in Anabaena, glucoside transporters influence the structure and function of septal junctions.
IMPORTANCE Heterocyst-forming cyanobacteria have the ability to perform oxygenic photosynthesis and to assimilate atmospheric CO2 and N2. These organisms grow as filaments that fix these gases specifically in vegetative cells and heterocysts, respectively. For the filaments to grow, these types of cells exchange nutrients, including sucrose, which serves as a source of reducing power and of carbon skeletons for the heterocysts. Movement of sucrose between cells in the filament takes place through septal junctions and has been traced with a fluorescent sucrose analog, esculin, that can be taken up by the cells. Here, we identified α-glucoside transporters of Anabaena that mediate uptake of esculin and, notably, influence septal structure and the function of septal junctions.
Filamentous cyanobacteria of the orders Nostocales and Stigonematales fix atmospheric nitrogen in specialized cells called heterocysts (1). Heterocysts are formed from vegetative cells when the filaments of those cyanobacteria lack a source of combined nitrogen (2). The heterocysts provide the vegetative cells with fixed nitrogen, and the vegetative cells, which fix carbon dioxide through oxygenic photosynthesis, provide the heterocysts with reduced carbon (3). Substances exchanged between the two cell types include regulators, such as PatS- and HetN-derived peptides, and nutrients, including amino acids and sugars (4). In the model heterocyst-forming cyanobacterium Anabaena sp. strain PCC 7120 (here called Anabaena) grown in the absence of combined nitrogen, heterocysts constitute about 10% of the cells and are distributed with a semiregular pattern along the filament (2). This implies that one heterocyst feeds more than one vegetative cell with fixed nitrogen. Two routes have been considered for intercellular molecular transfer, the continuous periplasm of the filament (5, 6) and cell-cell joining structures (7), now termed septal junctions (8,–10). The latter would represent a kind of symplasmic route (11) implying intercellular transfer between vegetative cells as well as between heterocysts and vegetative cells.
Proteins SepJ, FraC, and FraD, which are located at the cell poles in the intercellular septa of the filaments of Anabaena, are integral membrane proteins (12, 13). SepJ and FraD have predicted extramembrane domains that appear to reside in the periplasm (10, 14,–16). Intercellular molecular exchange in the cyanobacterial filament can be traced with fluorescent markers, including calcein, 5-carboxyfluorescein (5-CF), and esculin, and transfer has been found to be impaired in inactivated mutants of sepJ, fraC, and fraD (7, 14, 17, 18). Additionally, perforations (termed nanopores) that have been observed in septal peptidoglycan disks from heterocyst-forming cyanobacteria (19) are present at decreased numbers in those mutants (18). Structures observed by electron tomography of Anabaena that have been termed channels (20) likely correspond to the nanopores. SepJ, FraC, and FraD appear to contribute to the formation of cell-cell joining structures (septal junctions) that traverse the septal peptidoglycan through the nanopores. Differential impairment in the transfer of calcein and 5-CF in the sepJ and fraC-fraD mutants has suggested that two types of septal junction complexes exist, one related to SepJ and another related to FraCD (14).
Sucrose appears to be a quantitatively important metabolite transferred from vegetative cells to heterocysts (21,–25). Intercellular transfer of sucrose has been probed in Anabaena using esculin (6,7-dihydroxycoumarin β-d-glucoside), a fluorescent analog of this sugar (18). Esculin is taken up into the cells by a mechanism that can be inhibited by the presence of sucrose. Once inside the cells, esculin can be transferred cell to cell in the filament by diffusion through the septal junctions (18, 26). Thus, septal junctions are functionally analogous to the gap junctions of metazoans (18, 26).
In this work, we addressed the transporters that are involved in esculin uptake in Anabaena and their role, if any, in intercellular molecular transfer. The genome of Anabaena contains several open reading frames (ORFs) predicted to encode components of sugar transporters (27). We have identified three genes that are involved in uptake of esculin, two that encode components of two different ABC uptake transporters and one that encodes a major facilitator superfamily (MFS) transporter. We also found that the three identified glucoside transporters influence intercellular molecular exchange in Anabaena. One of the ABC transporter components, an ATP-binding subunit, is needed for the correct subcellular localization of SepJ, and the two other transporters appear to affect SepJ function.
We have previously shown that esculin can be taken up by Anabaena filaments grown in BG11 medium (containing nitrate as the nitrogen source) or grown in BG11 medium and incubated for 18 h in BG110 medium (lacking any source of combined nitrogen), and that uptake is linear for at least 70 min and takes place at higher levels in the filaments incubated in BG110 medium (18). To understand better the process of esculin uptake, we determined the dependence of esculin uptake on esculin concentration. Esculin uptake was faster in cells that had been incubated in the absence of nitrate compared to nitrate-grown cultures, with Vmax values of about 0.31 and 0.57 nmol (mg chlorophyll a [Chl])−1 min−1 for BG11-grown filaments and filaments incubated in BG110 medium, respectively (Fig. 1). Esculin concentrations giving half-maximal uptake rates (Ks) were 150 and 119 µM in BG11 and BG110, respectively. Because a concentration somewhat lower than the Ks would permit observation of effects such as competitive or noncompetitive inhibition, we have used 100 µM esculin as a standard concentration in our uptake assays.
The pH dependence of uptake of esculin was investigated. As shown in Fig. 2, the rate of uptake was higher at pH 7 than at lower or higher pH values and was, in every case, higher in filaments incubated in BG110 medium than in filaments from BG11 medium. The difference in the rate of uptake between filaments from BG110 and BG11 media decreased as the pH of the assay buffer was increased, suggesting that a H+-dependent transporter is induced in filaments incubated in BG110 medium.
Esculin has been used to test the activity of some higher plant sucrose transporters (28) and, consistent with the possibility of uptake through a sucrose transporter(s), inhibition of uptake of esculin by sucrose has been observed in Anabaena (18). To characterize further the transporters involved, we tested whether uptake of esculin would be inhibited by various monosaccharides (glucose, fructose, and galactose) and disaccharides (sucrose, maltose, trehalose, and lactose). The results depicted in Fig. 3 show that, in BG11-grown filaments, uptake of esculin was inhibited mainly by sucrose and, to a lesser extent, by maltose. Other sugars tested appear to have stimulated uptake of esculin. In filaments that were incubated in BG110 medium, inhibition of uptake by sucrose and maltose again was evident. Our results suggest that although esculin is a β-glucoside, its uptake is inhibited mainly by some α-glucosides, sucrose (glucose 1α→2 fructose) and maltose (glucose 1α→4 glucose), whereas neither lactose (a β-galactoside; galactose 1β→4 glucose) nor trehalose (a different α-glucoside; glucose 1α→1α glucose) inhibits uptake of esculin.
Two genes, Ava_2050 and Ava_2748, that encode possible components of ABC uptake transporters for disaccharides or oligosaccharides, are induced in the heterocysts of Anabaena variabilis ATCC 29413 (29). BLAST analysis with the genomic sequence of Anabaena (30) identified Alr4781 and All0261, with 97% and 99% amino acid identity, respectively, as the products of the Anabaena orthologs of those A. variabilis genes. Among characterized proteins included in the Transporter Classification Database (TCDB; http://www.tcdb.org), Alr4781 is most similar (45.4% identity, 59.6% similarity; expect, 3.2 × 10−111) to MalK1, an ATP-binding subunit shared by the glucose/mannose (TCDB no. 3.A.1.1.24) and the trehalose/maltose/sucrose/palatinose (TCDB no. 3.A.1.1.25) transporters from Thermus thermophilus, and All0261 is most similar (36.4% identity, 58.9% similarity; expect, 3.3 × 10−51) to the AraQ permease component of the arabinosaccharide transporter AraNPQ-MsmX from Bacillus subtilis (TCDB no. 3.A.1.1.34). We denote alr4781 as glsC and all0261 as glsP (gls standing for glucoside). Neither glsC nor glsP is clustered with other ABC transporter-encoding genes in the Anabaena genome. To test whether the transporters encoded by these Anabaena genes can be involved in uptake of esculin, glsC was inactivated by insertion of gene cassette C.S3 (31), resulting in Anabaena strains that bear the DR3912a mutation (see Fig. S1 in the supplemental material), and glsP was inactivated by insertion of C.S3, resulting in Anabaena strains that bear the DR3915 mutation, or of C.CE1 (31), resulting in Anabaena strains that bear the DR3985a mutation (Fig. S1). BG11-grown filaments of the glsC and glsP mutants showed esculin uptake activities that were 49% and 59%, respectively, of the wild-type activity (Table 1), and filaments of the glsC and glsP mutants that had been incubated in BG110 medium showed 74% and 73%, respectively, of the wild-type activity. Thus, the products of both genes contribute to esculin uptake by Anabaena in medium containing nitrate (BG11) and after incubation in medium lacking combined nitrogen (BG110).
ABC uptake transporters typically comprise one periplasmic solute-binding protein, two integral membrane proteins (transmembrane domains or permeases), and two nucleotide-binding domains that hydrolyze ATP in the cytoplasm (32). If the GlsC ATP-binding subunit and the GlsP permease belong to the same ABC transporter, we would expect that mutation of the two genes would not increase the effect on the uptake of esculin over that of the single mutations. If, on the other hand, GlsC and GlsP belong to two different transporters, we would expect an additive effect of the mutations. A double glsC glsP mutant, i.e., an Anabaena strain bearing the DR3912a and DR3985a mutations (Fig. S1), showed 25% of the wild-type activity of esculin uptake in BG11-grown filaments and 50% in filaments incubated in BG110 medium (Table 1), percentages that represent decreased values compared to the effects of the single mutations (49% and 59% for BG11 and 74% and 73% for BG110). These results suggest that GlsC and GlsP are components of different ABC transporters that can mediate esculin uptake. Notably, significant uptake activity remains in the double mutant, especially in filaments that had been incubated in medium lacking combined nitrogen (BG110).
Genes all1711 (hepP) and alr3705 encode predicted MFS proteins that would facilitate movement of disaccharides or oligosaccharides across cell membranes. As shown by results with all1711::Tn5-1063 mutant strain FQ163 (33), HepP may be a glucoside transporter that is involved in production of the heterocyst-specific polysaccharide layer and may also mediate sucrose transport. According to BLAST analysis, Alr3705 is the predicted Anabaena genomic product most similar to higher plant sucrose transporters. alr3705 was mutated by insertion of C.S3-containing plasmid pCSRL49, producing strain CSRL15, and insertion of pCSRL49 was also combined with all1711::Tn5-1063 to produce a double mutant, strain CSMN3 (Fig. S2). None of the strains FQ163, CSRL15, or CSMN3, when grown in BG11 medium, was significantly affected in uptake of esculin (P values of 0.553 to 0.703 by Student's t test), and CSRL15 was not significantly affected when incubated in BG110 medium (Table 1). In contrast, filaments of mutants FQ163 and CSMN3 incubated in BG110 medium showed similarly decreased activities, 69% and 67% of the wild-type activity, respectively (Table 1). These results indicate that HepP, but not Alr3705, contributed to uptake of esculin in filaments deprived of combined nitrogen.
In conclusion, two ABC transporters, of which GlsC and GlsP are independent components, together are responsible for about 75% and 50% of uptake of esculin in BG11 and BG110 filaments, respectively, and HepP is responsible for about 30% of uptake of esculin in BG110 filaments when tested at pH 7. Other transporters therefore should contribute to uptake of esculin in both BG11 and BG110 media.
To understand better the role of the transporters identified in this work in the physiology of Anabaena, we investigated their subcellular localization. The localization of HepP in the cytoplasmic membrane of both vegetative cells and heterocysts has been described previously (33). To study the subcellular localization of GlsC and GlsP, strains producing GlsC-GFP and GlsP-GFP fusion proteins were constructed. As a putative nucleotide-binding domain of an ABC transporter, GlsC is expected to reside in the cytoplasmic face of the cytoplasmic membrane. GlsP is a predicted integral membrane protein that bears six putative transmembrane segments with both the N and C termini in the cytoplasm. Because GFP folds efficiently in the cytoplasm (34), the gfp-mut2 gene was added to the 3′ end of the glsC and glsP genes, and the corresponding constructs were transferred to Anabaena (Fig. S3). Visualization of filaments of the corresponding strains, CSMN13 (glsC-gfp) and CSMN15 (glsP-gfp), incubated in BG11 or BG110 medium showed a relatively low GFP signal that was spread through the periphery of the cells, including the septal regions, where the signal was increased (Fig. 4). Quantification of GFP fluorescence was performed as described in Materials and Methods and is summarized in Fig. S4. The data show that fluorescence was roughly 2-fold higher in the septa than in lateral areas for both GlsC-GFP and GlsP-GFP in cells grown in BG11 medium as well as in cells incubated in BG110 medium, indicating that the increased fluorescence from the septa corresponds to the combination of the fluorescence from the adjacent cytoplasmic membranes. Nonetheless, somewhat higher-level GFP fluorescence was observed in septal areas of cells grown in BG11 medium than of cells incubated in BG110 medium. In filaments incubated in BG110 medium, the GFP signal was present at similar levels in heterocysts and vegetative cells. These results indicate that GlsC and GlsP are located throughout the cytoplasmic membrane of both vegetative cells and heterocysts. Our results also indicate that levels of GlsC-GFP or GlsP-GFP are generally similar in cells incubated in BG11 and BG110 media (Fig. S4).
The Fox− phenotype denotes inability to grow fixing N2 under oxic conditions, and it is frequently associated with malformation of the heterocyst envelope, as in the case of the hepP mutant (33). The growth phenotype was investigated here for the glsC and glsP mutants. On solid medium, the glsC and glsP single mutants and the glsC glsP double mutant could grow using nitrate or N2 as the nitrogen source, but the glsP mutant showed poorer diazotrophic growth than the wild type, and the glsC and glsC glsP mutants showed poorer growth in both media (Fig. 5). To determine growth rate constants, growth tests were carried out in liquid medium. In the presence of nitrate (BG11 medium), the growth rate of the single mutants was identical to that of the wild type, whereas the growth rate of the double mutant was 75% of that of the wild type (Table 2). In the absence of combined nitrogen (BG110 medium), the growth rate of the three mutants was lower than that of the wild type, being especially low in the case of the double mutant (Table 2). Thus, the glsC, glsP, and glsC glsP mutants cannot grow normally fixing N2 under oxic conditions and therefore show, at best, a weak Fox+ phenotype. The phenotype of diminished growth of the single mutants could be complemented by introducing in the corresponding mutant a replicative plasmid bearing the wild-type gene, glsC or glsP; however, when tested on solid medium, complementation was incomplete (Fig. S5). To investigate whether incomplete complementation could result from insufficient expression of the genes in the complemented strains, reverse transcription-quantitative PCR (RT-qPCR) analysis was performed as described in Materials and Methods. Rather than low expression, this analysis indicated 6-fold and 11-fold higher expression of the glsC and glsP genes, respectively, in the complemented mutants than in the wild type. Therefore, it is possible that overexpression of these genes is deleterious for Anabaena.
Production of heterocysts and nitrogenase activity were determined in filaments grown in BG11 medium and incubated for 48 h in BG110 medium. The glsC, glsP, and glsC glsP mutants showed about 60%, 85%, and 24%, respectively, of the number of heterocysts observed in the wild type (Table 2). Under oxic conditions, nitrogenase activity was about 10% of the wild-type activity in the two single mutants and about 6.5% in the double mutant (Table 2). Thus, the heterocysts produced in the mutants exhibited low nitrogenase activity. Assay under anoxic conditions showed little or no increase in activity, in contrast to what is normally observed in mutants that bear a defect in the heterocyst envelope (see, for instance, reference 33). The heterocyst envelope-specific polysaccharide layer can be stained with alcian blue, a stain useful to detect bacterial polysaccharides (35). Microscopic inspection of filaments of the glsC glsP double mutant stained with alcian blue showed the presence of stained heterocysts, indicating the existence of a polysaccharide layer in the double mutant (Fig. 6). Microscopic inspection also showed that the filaments of the double mutant were very short (Fig. 6). Inspection of cultures of the three mutants showed the presence of short filaments in the glsC glsP double mutant in both BG11 and BG110 media, but filament fragmentation was strongest in BG110 medium (Fig. S6). Such short filaments were not observed in the glsC or glsP single mutants. Thus, the phenotypic alterations were stronger in the glsC glsP double mutant than in the glsC or glsP single mutants, which is consistent with independent action of the GlsC and GlsP proteins as concluded above from the esculin uptake data.
Because the glsC, glsP, and hepP mutants are impaired in glucoside transport and diazotrophic growth, the proteins encoded by these genes could influence intercellular transfer of sucrose. We therefore tested intercellular exchange of esculin in the glsC, glsP, and hepP mutants by means of FRAP (fluorescence recovery after photobleaching) analysis. The results of these tests were analyzed to determine the recovery constant (R) of fluorescence in the cells in which esculin had been bleached (see Materials and Methods and Text S1). To attain adequate labeling of esculin to carry out the FRAP analysis, filaments were incubated for 1 h with 150 μM esculin. Transfer of esculin between vegetative cells of BG11-grown filaments was decreased in a limited way (by about 22%) in the glsC mutant but not in the glsP mutant (Table 3). However, the effect was larger in the glsC glsP double mutant (about 33% inhibition). In the hepP mutant, esculin transfer was 43% lower than that in the wild type.
In filaments of the wild type that had been incubated for 48 h in BG110 medium, esculin transfer between vegetative cells was similar to transfer between BG11-grown vegetative cells, but transfer from vegetative cells to heterocysts was decreased to about 38% of the value between vegetative cells (Table 3). These results are consistent with previously reported data (18). In the mutants, esculin transfer was lower in the BG110-incubated than in the BG11-grown vegetative cells, and it was especially decreased in the glsC mutant (Table 3). In contrast, esculin transfer from vegetative cells to heterocysts was increased in all of the mutants compared to the wild type, and this increase was particularly significant in the hepP mutant. In summary, esculin transfer was impaired between vegetative cells of heterocyst-containing filaments but not from vegetative cells to heterocysts.
To assess how specific the effect on intercellular transfer could be, transfer of calcein and 5-CF between nitrate-grown vegetative cells was also tested in the mutants. Calcein transfer was significantly impaired in the three single mutants, and it was lowest (21% of the wild-type activity) in the glsC glsP double mutant (Table 4). Transfer of 5-CF was also significantly impaired in the glsC and glsP mutants, although the effect of the mutations was lower in this case than on calcein transfer, and it was not impaired in the hepP mutant. These studies showed that GlsC, GlsP, and HepP are required for normal intercellular molecular exchange in Anabaena, but this requirement is more evident when the exchange is tested with calcein than with 5-CF or, as shown above, esculin (compared to BG11-grown filaments).
The fragmentation of filaments observed in the glsC glsP double mutant and the effect of the mutation of the glucoside transporters on calcein exchange described above are reminiscent of effects of inactivation of sepJ in Anabaena (12, 18). We therefore investigated the effect of the inactivation of glsC, glsP, and hepP on the subcellular localization of SepJ. For this investigation, plasmids bearing a sepJ-gfp fusion gene were transferred to mutants of those genes, producing strains CSMN9 (glsC sepJ-gfp), CSMN10 (glsP sepJ-gfp), and CSMN16 (hepP sepJ-gfp) (for PCR analysis of the genomic structure of each strain, see Fig. S7). Confocal microscopic inspection of filaments of strains producing SepJ-GFP showed that whereas the glsP and hepP mutations did not impair SepJ-GFP localization at the intercellular septa, the glsC mutation had a strong effect on localization (Fig. 7; for SepJ-GFP localization in four independent clones inspected by fluorescence microscopy, see Fig. S8). In the glsC sepJ-gfp strain, spots of GFP were only sporadically observed in the center of the septa, and the GFP signal was frequently found throughout the periphery of the cells, including the intercellular septa. Thus, GlsC, but not GlsP or HepP, appears necessary for proper subcellular localization of SepJ.
To corroborate delocalization of SepJ as a result of inactivation of glsC, immunolocalization of SepJ was performed using antibodies raised against its coiled-coil domain (17). The antibodies localized SepJ at the cell poles of Anabaena (Fig. 8), as previously described (15). In the glsC mutant, the signal was largely delocalized, being observed at the cell poles only sporadically. In the complemented glsC-C strain [DR3912a(pCSMN22)], SepJ was observed clearly at the cell poles (Fig. 8). These results are fully consistent with the observation that the SepJ-GFP fusion protein shows delocalization of SepJ as the result of inactivation of glsC (Fig. 7 and Fig. S8).
Because SepJ is necessary for Anabaena to make a normal number of septal peptidoglycan nanopores (18), the number of nanopores was counted in septal peptidoglycan disks observed in murein sacculi isolated from the wild type and the glsC, glsP, and hepP mutants (Fig. 9). Whereas the glsP and hepP mutants contained a number of nanopores per septum similar to that of the wild type, the septa of the glsC mutant contained about 48% of the nanopores found in the wild-type septa.
The results in the previous section, showing that GlsC is necessary for proper localization of SepJ and formation of septal peptidoglycan nanopores, provides a rationale for understanding the effect of inactivation of glsC on the intercellular transfer of calcein, but no effect of inactivation of glsP or hepP was found. We then studied possible protein-protein interactions involving the glucoside transporters and SepJ using the bacterial adenylate cyclase two-hybrid (BACTH) assay, in which adenylate cyclase activity is reconstituted from two fragments, T25 and T18, of an adenylate cyclase from Bordetella pertussis brought together by interacting proteins fused to each of those fragments (36). Reconstituted adenylate cyclase in Escherichia coli produces cyclic AMP (cAMP) that promotes induction of lacZ encoding β-galactosidase. We have previously shown that SepJ-T25 and SepJ-T18 fusions (where the order of protein names denotes N-terminal to C-terminal orientation) are functional in SepJ self-interactions that produce high β-galactosidase activity (15, 16). As is the case for GlsP, HepP is a predicted integral membrane protein with both the N and C termini in the cytoplasm (33). Because of possible copy number or steric hindrance problems (37, 38), here we tested possible interactions of SepJ-T25 and SepJ-T18 with both N-terminal and C-terminal fusions to T18 and T25, respectively, of each of the glucoside transporter components investigated in this work, GlsC, GlsP, and HepP. The negative control in this analysis was an E. coli strain carrying plasmids that produce nonfused T25 and T18 fragments, and additional negative controls producing nonfused T25 or T18 and some of the tested fusions were used. None of these controls produced β-galactosidase activity significantly different from that of the T25/T18 control (Table 5). Combinations of protein fusions involving SepJ that produced β-galactosidase activity significantly higher than the controls included SepJ-T25/SepJ-T18 (positive control), SepJ-T25/T18-HepP, SepJ-T18/T25-HepP, and SepJ-T18/T25-GlsP, but no fusion involving GlsC. These results suggest significant interactions between SepJ and GlsP and, more strongly, between SepJ and HepP. On the other hand, significant interactions were also observed between HepP and GlsP. Finally, HepP self-interactions and GlsC self-interactions were also observed, suggesting that HepP and GlsC can form homo-oligomers.
Esculin has been successfully used as a fluorescent analog of sucrose to study intercellular molecular exchange in the filaments of Anabaena by means of FRAP analysis (18). This analysis requires esculin to be taken up by the cells in the filament, and we have now identified three genes, glsC, glsP, and hepP, that encode components of transporters that mediate esculin uptake in Anabaena. The glsC (alr4781) gene encodes an ATP-binding subunit of an ABC transporter, and the glsP (all0261) gene encodes an integral membrane (permease) subunit of a different ABC transporter. These genes were investigated because they are the possible Anabaena orthologs of genes highly expressed in the heterocysts of a closely related cyanobacterium, A. variabilis (29). In Anabaena, because the effect of inactivating glsC and glsP is evident in filaments grown in the presence of nitrate (Table 1), GlsC and GlsP appear to be active in vegetative cells. Additionally, as observed with GFP fusions, GlsC and GlsP are present in heterocysts as well as in vegetative cells (Fig. 4; see also Fig. S4 in the supplemental material). In transcriptomic analysis of Anabaena, these genes appear to have low expression, and their expression is not affected by nitrogen deprivation (39). According to the results of inhibition of uptake of esculin by sugars (Fig. 3), the natural substrate of these transporters can be sucrose or an α-glucoside. Sucrose uptake by vegetative cells of Anabaena has previously been reported (40), and sucrose transporters that can also transport maltose are frequently found in plants (41). The Anabaena glucoside transporters could have a role in the recovery of glucosides from extracellular polysaccharides produced under certain physiological conditions, as has been shown to occur in cyanobacterial mats (42). Consistent with this, biomass of the glsC and glsP mutants in old BG11 plates is shiny (not shown), which may be indicative of exopolysaccharide accumulation (33). It is also of interest that although Anabaena has been considered an obligatory photoautotroph (43), recent data suggest that it can grow using fructose, although this sugar has to be provided at a high concentration unless Anabaena is engineered to express a fructose transporter (27, 44). On the other hand, trehalose, lactose, glucose, fructose, and galactose could stimulate esculin uptake (Fig. 3), suggesting that Anabaena can use these sugars to support physiological activities such as active transport. As noted earlier, the Anabaena genome bears several genes putatively encoding sugar transporters (27), some of which could be involved in the uptake of those sugars.
The third gene that encodes an esculin transporter is hepP (all1711), which encodes an MFS protein that is also necessary for production of the heterocyst-specific polysaccharide layer (33). We have previously shown that HepP is present at higher levels in developing heterocysts (proheterocysts) and heterocysts than in vegetative cells, and that HepP could mediate sucrose uptake specifically in (pro)heterocysts (33). Because the contribution of HepP to uptake of esculin is evident only in filaments that had been incubated in the absence of combined nitrogen, and because uptake of esculin in these filaments is inhibited by sucrose, HepP may be involved in uptake of sucrose/esculin by (pro)heterocysts. MFS proteins, including sucrose transporters, frequently act as secondary transporters that mediate symport with protons (45). Uptake of esculin that is associated with incubation in BG110 medium minus uptake in BG11 medium decreases with increasing pH beyond pH 6 (Fig. 2). This observation suggests that a H+-dependent transporter is induced in filaments incubated in BG110 medium. Because HepP contributes to esculin uptake associated with incubation in BG110 medium, our results are consistent with the idea that HepP is a sucrose-H+ or α-glucoside-H+ symporter.
Inactivation of hepP leads to a Fox− phenotype that has been described in detail (33). We have found that the glsC and glsP mutants exhibit a weak Fox+ phenotype: they grow slowly without a source of combined nitrogen under oxic conditions and express low levels of nitrogenase activity (Table 2). Combination in the same strain of the two mutations, glsC and glsP, resulted in a greater impairment of diazotrophic growth, very low nitrogenase activity, and a low percentage of heterocysts. Nonetheless, these heterocysts bore an envelope polysaccharide layer (Fig. 6) and their nitrogenase activity was not substantially increased in anoxic assays, suggesting that they do not have a cell envelope problem. To explore the possibility of a limited sucrose supply to the heterocysts, we investigated whether the glsC and glsP mutations affect intercellular molecular exchange tested with the fluorescent sucrose analog esculin. We have observed that the transfer of esculin in filaments of strains mutated in glsC, glsP, or hepP is impaired between vegetative cells but not from vegetative cells to heterocysts (Table 3). Impairment of sucrose transfer between vegetative cells might eventually limit sucrose supply to heterocysts, and a low supply of reductant would explain the low nitrogenase activities detected in glsC, glsP, and glsC glsP mutants. In the case of the glsC glsP double mutant, the small number of vegetative cells in heterocyst-containing filaments, which are short (Fig. 6 and Fig. S6), may further limit the supply of reductant for nitrogenase. On the other hand, esculin transfer to heterocysts was substantially increased in the hepP mutant (Table 3). At least some sucrose transporters of the MFS family can function bidirectionally (45), and this could be the case for HepP, which appears to export saccharides from the heterocysts (33). Therefore, the apparently increased transfer of esculin to heterocysts in the hepP mutant might reflect increased retention of esculin in the heterocysts of this strain.
Starting from the observation that the glsC, glsP, and hepP mutants characterized in this work are impaired in the transfer of esculin between vegetative cells, we found that intercellular transfer of fluorescent markers is in general affected in these mutants, with the highest effect being observed on the transfer of calcein. A greater effect on transfer of calcein than of 5-CF or esculin is reminiscent of the effect of inactivation of sepJ (14, 17, 18). Hence, these observations suggest a role of the glucoside transporters in proper function of the SepJ-related septal junctions. GFP fusions indicate that GlsC, GlsP, and HepP are located in the periphery of the cells, including the intercellular septa (Fig. 4 and Fig. S4) (33), where they could interact with the septal junction complexes. To investigate whether such interactions are feasible, BACTH analysis was carried out with the glucoside transporter proteins and SepJ. This analysis showed that GlsP and, most strongly, HepP can interact with SepJ, whereas no interaction was observed between SepJ and GlsC. Hence, GlsP and HepP may affect SepJ function by means of protein-protein interactions. A functional dependence between SepJ and an ABC transporter for polar amino acids has also been described (46). These observations suggest that proper operation of SepJ and, hence, of the SepJ-related septal junctions requires interaction with other cytoplasmic membrane proteins.
GlsC is instead required for proper location of SepJ and maturation of the intercellular septa, as illustrated by the presence of a lower number of nanopores in the glsC mutant than in the wild type. How GlsC influences SepJ localization and nanopore formation is unknown, but we note (i) that an N-acetylmuramoyl-l-alanine amidase, AmiC, is required for drilling the septal peptidoglycan nanopores (19), and (ii) that the presence of septal proteins, including SepJ, is needed for the amidase to make the nanopores (18). In other bacteria, the ABC transporter-like FtsEX complex, in which FtsE is an ATP-binding subunit, is required for activation of amidases that split the septal peptidoglycan during cell division (47, 48) and of endopeptidases that function in cell elongation (49, 50). An appealing hypothesis is that GlsC participates in an ABC transporter-like complex that regulates amidases involved in nanopore formation with an effect on localization of SepJ.
The different effects of inactivation of glsC, i.e., impairment of esculin uptake and alteration of septal structure, indicate that GlsC has multiple functions. Multitask ATP-binding subunits that serve different ABC transporters have been described, e.g., in Streptomyces lividans (51), Streptococcus mutans (52), Bacillus subtilis (53), and Corynebacterium alkanolyticum (54), as well as in Anabaena (55). As checked at the Integrated Microbial Genomes webpage (https://img.jgi.doe.gov/cgi-bin/m/main.cgi), the glsC gene is not clustered with any other gene encoding an ABC transporter component in any cyanobacterium whose genome sequence is available. Therefore, no preferential association of GlsC to any particular ABC transporter can be established based on genomic data. Nonetheless, in a few cases the neighboring genes are related to cell wall biosynthesis, including an N-acetylmuramoyl-l-alanine amidase-encoding gene in Spirulina major PCC 6313, consistent with the idea of a relationship between GlsC and cell wall maturation.
In summary, we have identified three genes encoding components of transporters that mediate α-glucoside uptake, including sucrose uptake, in Anabaena. These transporters appear to influence septal junction maturation in the case of glsC or function in the case of glsP and hepP. As a consequence, inactivation of these genes impairs molecular transfer between vegetative cells, negatively affecting diazotrophy. A major task for future research is to explore whether the interplay between these transporters and SepJ has a function regulating the activity of septal junctions.
Anabaena sp. strain PCC 7120 and derivative strains (described in Table S1 in the supplemental material) were grown in BG11 medium modified to contain ferric citrate instead of ferric ammonium citrate (43) or BG110 medium (BG11 further modified by omission of NaNO3) at 30°C in the light (ca. 25 to 30 μmol photons m−2 s−1) in shaken (100 rpm) liquid cultures. For tests on solid medium, medium BG11 or BG110 was solidified with 1% (wt/vol) Difco Bacto agar. For isolation of the glsC (alr4781), glsP (all0261), and glsC glsP mutants, Anabaena was grown, with shaking, in flask cultures of AA/8 liquid medium with nitrate (56) or in medium AA with nitrate solidified with 1.2% (wt/vol) purified (Difco) Bacto agar (56) at 30°C and illuminated as described above. When appropriate, antibiotics were added to the cyanobacterial cultures at the following concentrations: in liquid cultures, streptomycin sulfate (Sm), 2 to 5 μg ml−1; spectinomycin dihydrochloride pentahydrate (Sp), 2 to 5 μg ml−1; erythromycin (Em), 5 μg ml−1; and neomycin sulfate (Nm), 5 to 25 μg ml−1; in solid media, Sm, 5 to 10 μg ml−1; Sp, 5 to 10 μg ml−1; Em, 5 to 10 μg ml−1; and Nm, 30 to 40 μg ml−1. Chlorophyll a (Chl) content of cultures was determined by the method of Mackinney (57).
E. coli strains were grown in LB medium, supplemented when appropriate with antibiotics at standard concentrations (58). E. coli strain DH5α or DH5αMCR was used for plasmid constructions. E. coli strain DH5α or ED8654, bearing a conjugative plasmid, and strain HB101 or DH5αMCR, bearing a methylase-encoding helper plasmid and the cargo plasmid, were used for conjugation with Anabaena, unless stated otherwise (59).
The alr4781 (glsC) mutant, DR3912a, was generated by a diparental mating between Anabaena and DH5αMCR carrying pRL443, pRL3857a, and pRL3912a (plasmids described in Table S1). The single recombinant was selected on Em, tested for sucrose sensitivity, and then went through a sucrose selection cycle, as described by Cai and Wolk (60), for selection of the double recombinant (Fig. S1). Similarly, an all0261 (glsP) double recombinant deletion mutant, DR3915 (Fig. S1), was generated by mating between Anabaena and DH5αMCR carrying pRL443, pRL3857a, and pRL3915. Because DR3912a and DR3915 carry the same antibiotic resistance marker (Smr Spr), a new plasmid, pRL3985a, was constructed for creation of the glsC glsP double mutant (Fig. S1). In this case, pRL3985a was introduced into DR3912a by conjugation, and the mutant was selected as described above.
For complementation of the glsC mutant (DR3912a), a fragment containing ORF alr4781 and 202 bp of upstream and 49 bp of downstream DNA was amplified using Anabaena DNA as the template and primers alr4781-3 and alr4781-4 (oligodeoxynucleotide primers are described in Table S1). The PCR product was cloned into vector pSpark I, producing pCSMN21. This construct was verified by sequencing and transferred as a BamHI fragment to pRL25C (61) digested with the same enzyme, producing pCSMN22. This plasmid was transferred to DR3912a by conjugation. Clones resistant to Sm, Sp, and Nm were isolated and their genetic structure verified by PCR with primers alr4781-3 and alr4781-4 (Fig. S5). This strain was named CSMN11. For complementation of the glsP mutant (DR3915), a fragment containing ORF all0261 and 103 bp of upstream and 40 bp of downstream DNA was amplified using Anabaena DNA as the template and primers all0261-3 and all0261-4. The PCR product was cloned into pSpark I, producing pCSMN19, which was confirmed by sequencing and transferred as a BamHI fragment to pRL25C digested with BamHI, producing pCSMN20. This plasmid was transferred to DR3915 by conjugation. Clones resistant to Sm, Sp, and Nm were insolated and their genetic structure was verified by PCR with primers all0261-3 and all0261-4 (Fig. S5). This strain was named CSMN12.
For inactivation of alr3705, an internal fragment of 560 bp was amplified by PCR using Anabaena DNA as the template and primers alr3705-1 (bearing a BamHI site in its 5′ end) and alr3705-2. The amplified fragment was cloned into pMBL-T (http://www.molbiolab.es/uploads/phpgSgmue.pdf; Dominion MBL, Spain) and transferred as a BamHI-ended fragment (the second BamHI site is from the vector multiple-cloning site) to BamHI-digested pCSV3 (62), producing pCSRL49. This plasmid was transformed into E. coli HB101 carrying pRL623 and transferred to Anabaena and to hepP (all1711) mutant strain FQ163 (33) by conjugation with selection for Smr and Spr (because FQ163 is itself Nmr, Smr, and bleomycin resistant, in this case effective selection is only for Spr). Clones that had incorporated pCSRL49 by single recombination were selected for further study and named strain CSRL15 (wild-type background) and CSMN3 (hepP background) (Fig. S2).
To prepare an Anabaena strain producing a fusion of GFP to GlsC, a 950-bp DNA fragment from the 3′ region of glsC (alr4781) was amplified using Anabaena DNA as the template and primers alr4781-5 and alr4781-6. The 950-bp PCR product was cloned into pSpark I, producing pCSMN23. This construct was validated by sequencing and transferred to SacI-XhoI-digested pRL277 (63) as a SacI-NheI fragment together with NheI-SalI-digested gfp-mut2 (64), producing pCSMN24, in which the gfp-mut2 gene is fused to glsC. pCSMN24 was transferred to Anabaena by conjugation. Clones resistant to Sm and Sp were selected and their genetic structure was verified by PCR with primer pairs alr4781-3/gfp-5 and alr4781-3/alr4781-4. This strain was named CSMN13 (Fig. S3). To prepare an Anabaena strain producing a fusion of GFP to GlsP, a 482-bp DNA fragment from the 3′ region of glsP (all0261) was amplified using Anabaena DNA as the template and primers all0261-6 and all0261-5. The PCR product, a SacI-NheI fragment, was inserted together with NheI-SalI-digested gfp-mut2 into SacI-XhoI-digested pRL277 (63), producing pCSMN25, which bears a fusion of the all0261 coding sequence to the gfp-mut2 gene. This construct was verified by sequencing and transferred to Anabaena by conjugation. Clones resistant to Sm and Sp were isolated, and integration of the glsP-gfp construct was verified by PCR using primer pairs all0261-4/gfp-5 and all0261-3/all0261-4. This strain was named CSMN15 (Fig. S3).
To study the effect of inactivation of transporter genes on the localization of SepJ-GFP, Nmr plasmid pCSVT22, bearing sepJ-gfp (13), was transferred to strains DR3912a (alr4781::C.S3) and DR3915 (all0261::C.S3) by conjugation. Similarly, the Smr Spr plasmid pCSAM137, bearing sepJ-gfp (12), was transferred to FQ163 (hepP::Tn5-1063) (33). The genetic structure of selected clones bearing sepJ-gfp fusions was studied by PCR with DNA from those clones and primer pair alr2338-3/gfp-5 to test recombination in the correct genomic location (sepJ). We also verified the mutant background in the exconjugants using the following primer pairs: for alr4781, alr4781-3/alr4781-4; for all0261, all0261-3/all0261-4; and for hepP, all1711-3/all1711-4 (Fig. S7). Clones bearing the sepJ-gfp fusion were named strain CSMN9 (alr4781 background), CSMN10 (all0261 background), and CSMN16 (hepP background).
For RT-qPCR, RNA was isolated as described previously (33) from 50 to 100 ml of shaken Anabaena cultures. RNA was treated with Ambion TURBO DNA-free DNase according to the manufacturer's protocol. Three independent RNA samples were analyzed from each strain (the wild type and the complemented glsC and glsP strains), and three technical replicates were carried out for each sample. RNA (200 ng) was reverse transcribed using a QuantiTect reverse transcription kit (Qiagen) with random primers as indicated by standard protocols of the manufacturer. Quantitative real-time PCR was performed on an iCycler iQ real-time PCR detection system equipped with the iCycler iQ v 3.0 software from Bio-Rad. PCR amplification was performed in a 20-μl reaction mix according to standard protocols of the SensiFAST SYBR and fluorescein kit (Bioline). qPCR conditions were the following: 1 cycle at 95°C for 2 min, followed by 30 cycles of 95°C for 15 s, 67.5°C for 20 s, and 72°C for 30 s. PCR products were checked by a single-peak melting curve. The threshold cycle (CT) of each gene was determined and normalized to those of reference genes ispD (all5167) and dxs (alr0599) to obtain ΔCT values from each sample. Relative gene expression was calculated using the 2−ΔΔCT method (65), and the data presented correspond to the average of data obtained with each reference gene. The following primer pairs were used: all0261-11/all0261-12, alr4781-9/alr4781-10, all5167-1/all5167-2, and alr0599-1/alr0599-2 (Table S1).
Anabaena strains grown in BG11 medium, with antibiotics for the mutants, were harvested by centrifugation, washed three times with BG11 or BG110 medium without antibiotics, and incubated for 18 h in the same medium under culture conditions. Cells were harvested, washed, and resuspended in the corresponding growth medium supplemented with 10 mM HEPES-NaOH buffer (pH 7, unless indicated otherwise), and 1 mM the indicated sugar in the experiment depicted in Fig. 3. Assays of uptake were started by addition of esculin hydrate (Sigma-Aldrich) at 100 μM, and suspensions were incubated at 30°C in the light (~170 μmol photons m−2 s−1) for up to 70 min. One-ml samples were withdrawn and filtered. Cells on the filters were washed with 10 mM HEPES-NaOH buffer of the same pH used in the assay and were resuspended in 2 ml of 10 mM HEPES-NaOH buffer (pH 7). Fluorescence of the resulting cell suspension was measured in a Varian Cary eclipse fluorescence spectrophotometer (excitation, 360 ± 10 nm; emission, 462 ± 10 nm). Esculin solutions in the same buffer (pH 7) were used as standards. Significance in the differences of uptake between strains (as well as in other parameters investigated in this work) was assessed by unpaired Student's t tests, assuming a normal distribution of the data. Data sets with P values of <0.05 are considered significant.
The growth rate constant (μ = [ln2]/td, where td is the doubling time) was calculated from the increase in the optical density at 750 nm (OD750) of shaken liquid cultures. Cultures were inoculated with an amount of cells giving an OD750 of about 0.05 (light path, 1 cm) and grew logarithmically until reaching an OD750 of about 0.8 to 0.9. The suspensions of filaments were carefully homogenized with a pipette before taking the samples.
For determination of nitrogenase activity, filaments grown in BG11 medium were harvested, washed with BG110 medium, and resuspended in BG110 medium. After 48 h of incubation under growth conditions, the filaments were used in acetylene reduction assays performed under oxic or anoxic conditions at 30°C in the light (ca. 150 μmol photons m−2 s−1). For these assays, the cell suspensions (2 ml; ca. 10 μg Chl ml−1) were placed in flasks sealed with rubber stoppers (total volume, 12 to 14 ml). For the anoxic assays, the cell suspensions were supplemented with 10 μM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), bubbled thoroughly with argon for 3 min, and incubated for 60 min under assay conditions before starting the reaction. Production of ethylene, determined by gas chromatography in 1-ml samples from the gas phase, was monitored for up to 3 h after starting the reaction by addition of acetylene (2 ml).
Cultures were routinely observed by light microscopy. To stain the polysaccharide layer of heterocysts, cell suspensions were mixed (1:2) with a filtered 1% (wt/vol) alcian blue (Sigma) solution.
For visualization by confocal microscopy of filaments of strains producing genetic fusions to GFP, small blocks of agar-solidified BG11 or BG110 medium bearing the filaments were excised and placed in a sample holder with a glass coverslip on top. GFP fluorescence was visualized using a Leica HCX Plan Apo 63× 1.4-numeric-aperture (NA) oil immersion objective attached to a Leica TCS SP2 confocal laser-scanning microscope. GFP was excited using 488-nm irradiation from an argon ion laser. Fluorescent emission was monitored by collection across windows of 498 to 541 nm (GFP imaging) and 630 to 700 nm (cyanobacterial autofluorescence). GFP fluorescence intensity was analyzed using ImageJ 1.45s software. To determine the relative fluorescence intensity in different cell zones, integrated density was recorded in squares of 0.2 to 0.8 μm2. About 80 to 190 measurements were made for each of the lateral walls and septal areas of vegetative cells from BG11 or BG110 medium, and 50 to 60 measurements were made for lateral walls of heterocysts. We could not accurately quantify GFP fluorescence from heterocyst-vegetative cell septa, which are thinner than the septa between vegetative cells. Because fluorescence did not follow a normal distribution, data are presented as median and interquartile ranges (66).
For fluorescence microscopy, filaments of cells were imaged using a Leica DM6000B fluorescence microscope and an ORCA-ER camera (Hamamatsu). GFP fluorescence was monitored using a fluorescein isothiocyanate (FITC) L5 filter (excitation band pass [BP], 480/40 nm; emission BP, 527/30 nm), and red autofluorescence was monitored using a Texas red TX2 filter (excitation BP, 560/40 nm; emission BP, 645/75 nm).
Cells from 1.5 ml of liquid cultures were collected by centrifugation, placed atop a poly-l-lysine-precoated microscope slide, and covered with a 45-μm-pore-size Millipore filter. The filter was removed and the slide was left to dry at room temperature, immersed in 70% ethanol at −20°C for 30 min, and dried for 15 min at room temperature. The cells were washed twice (2 min each time, room temperature) by covering the slide with PBS-T (PBS supplemented with 0.05% Tween 20). Subsequently, the slides were treated with a blocking buffer (5% milk powder in PBS-T) for 15 min. Cells on the slides were then incubated for 90 min with anti-SepJ-CC antibodies (17) diluted in blocking buffer (1:250), washed three times with PBS-T, incubated for 45 min in the dark with anti-rabbit antibody conjugated to FITC (1:500 dilution in PBS-T; Sigma), and washed three times with PBS-T. After being dried, several drops of FluorSave (Calbiochem) were added atop, covered with a coverslip, and sealed with nail lacquer. Fluorescence was monitored as described above, and images were analyzed with ImageJ software (http://imagej.nih.gov/ij).
The murein sacculi (which are made of peptidoglycan) were isolated from filaments grown in BG11 medium and analyzed as described previously (18, 19). The purified sacculi were deposited on Formvar/carbon film-coated copper grids and stained with 1% (wt/vol) uranyl acetate. All of the samples were examined with a Zeiss Libra 120 Plus electron microscope at 120 kV.
For assays of intercellular transfer of esculin, filaments were harvested, resuspended in 500 μl of fresh growth medium, mixed with 15 μl of saturated (~5 mM) aqueous esculin hydrate solution, incubated for 1 h in the dark with gentle shaking at 30°C, and then washed three times with growth medium, followed by incubation in the dark for 15 min in 1 ml medium at 30°C with gentle shaking. Cells were then washed and spotted onto a BG11 or BG110 agar plate (1%, wt/vol), and excess medium was removed. Small blocks of agar with cells adsorbed on the surface were placed in a custom-built temperature-controlled sample holder under a glass coverslip at 30°C. Cells were visualized with a laser-scanning confocal microscope (Leica TCS SP5) using a Leica HCX Plan Apo 63× 1.4-NA oil-immersion objective. Fluorescence was excited at 355 nm, with detection of esculin at 443 to 490 nm and detection of Chl at 670 to 720 nm. High-resolution imaging used a 6× line average with an optical section of ~0.7 μm. FRAP measurements were without line averaging and with a wide pinhole, giving an optical section of ~4 μm. After capturing a prebleach image, the fluorescence of a defined region of interest was bleached out by scanning this region at ~6× higher laser intensity, and recovery was then recorded in a sequence of full-frame images.
For calcein and 5-CF transfer assays, calcein and 5-CF staining and FRAP analysis were performed as previously reported (7, 14). Cell suspensions were spotted onto agar and placed in a custom-built temperature-controlled sample holder with a glass coverslip on top. All measurements were carried out at 30°C. For both calcein and 5-CF, cells were imaged with a Leica HCX Plan Apo 63×, 1.4-NA oil immersion objective attached to a Leica TCS SP5 confocal laser-scanning microscope as previously described for calcein (7), with a 488-nm line argon laser as the excitation source. Fluorescent emission was monitored by collection across windows of 500 to 520 nm or 500 to 527 nm in different experiments and a 150-μm pinhole. After an initial image was recorded, the bleach was carried out by an automated FRAP routine which switched the microscope to X scanning mode, increased the laser intensity by a factor of 10, and scanned a line across one cell for 0.137 s before reducing the laser intensity, switching back to XY imaging mode, and recording a sequence of images typically at 1-s intervals.
For FRAP data analysis, we quantified kinetics of transfer of the fluorescent tracer to either (i) a terminal cell (with one cell junction) or (ii) a cell somewhere in the middle of a filament (i.e., with two cell junctions). For the first option, the recovery rate constant, R, was calculated from the formula CB = C0 + CR (1 − e−Rt), where CB is fluorescence in the bleached cell, C0 is fluorescence immediately after the bleach and tending toward (C0 + CR) after fluorescence recovery, t is time, and R is the recovery rate constant due to transfer of the tracer from one neighbor cell (14). For the second option, the formula CB = C0 + CR (1 − e−2Rt) was used. Development of equations for FRAP analysis is described in the supplemental material (Text S1).
The possible interaction of the different glucoside transporters with SepJ was tested using bacterial adenylate cyclase two-hybrid (BACTH) analysis. For this analysis, all tested genes were amplified using Anabaena DNA as the template. The following primers were used: alr4781-7 and alr4781-8 to amplify glsC, all0261-7 and all0261-8 to amplify glsP, and all1711-9 and all1711-10 to amplify hepP. The PCR products were cloned in vector pSpark I, transformed into E. coli DH5α, and sequenced. Inserts with the correct sequence were transferred as XbaI- and KpnI-digested fragments to pUT18, pUT18C, pKNT25, and pKT25 (37), producing fusions to the 5′ and 3′ ends of the genes encoding the adenylate cyclase T18 and T25 fragments, respectively. The resulting plasmids were transformed into E. coli XL1-Blue to amplify the plasmids. Fusions of the sepJ gene to the 5′ end of T18 or T25 were as previously described (15). Isolated plasmids were cotransformed into E. coli BTH101 (cya-99). Transformants were plated onto LB medium containing selective antibiotics and 1% glucose. Efficiencies of interactions between different hybrid proteins were quantified by measurement of β-galactosidase activity in cells from liquid cultures.
To determine β-galactosidase activity, bacteria were grown in LB medium in the presence of 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and appropriate antibiotics at 30°C for 16 h. Before the assays, cultures were diluted 1:5 into buffer Z (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, and 1 mM MgSO4). To permeabilize cells, 30 μl of toluene and 35 μl of a 0.1% SDS solution were added to 2.5 ml of bacterial suspension. The tubes were vortexed for 10 s and incubated with shaking at 37°C for 30 min for evaporation of toluene. For the enzymatic reaction, 875 μl of permeabilized cells was added to buffer Z supplemented with β-mercaptoethanol (25 mM final concentration) to a final volume of 3.375 ml. The tubes were incubated at 30°C in a water bath for at least 5 min. The reaction was started by adding 875 μl of 0.4 mg ml−1 o-nitrophenol-β-galactoside (ONPG) in buffer Z. Samples of 1 ml, taken at different times (up to 12 min), were added to 0.5 ml of 1 M Na2CO3 to stop the reaction. A420 was recorded, and the amount of o-nitrophenol produced was calculated using an extinction coefficient, ε420, of 4.5 mM−1 cm−1 and referred to the amount of total protein, determined by a modified Lowry procedure.
We thank Antonia Herrero (Seville, Spain) for useful discussions, Alexandra Johnson (East Lansing, MI) for her cloning work related to generation of DR3912a and DR3915, and Sergio Camargo (Seville, Spain) for help with the immunofluorescence and RT-qPCR analyses.
M.N.-M. was the recipient of an FPU (Formación de Profesorado Universitario) fellowship/contract from the Spanish government. Work in Seville was supported by grant no. BFU2014-56757-P from Plan Nacional de Investigación, Spain, cofinanced by the European Regional Development Fund. Work in East Lansing was supported by the Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (grant DOE FG02-91ER20021).
Supplemental material for this article may be found at https://doi.org/10.1128/JB.00876-16.