SLC4 family transporters in a marine diatom directly pump bicarbonate from seawater

Candidate HCO₃⁻ Transporters in Diatoms.

Diatom genomes contain numerous genes that apparently share origins with the mammalian genome, and these genes are believed to be derived from the ancestral eukaryotic host cell (probably a biciliate) of secondary endosymbiosis (26–28). In the genome of P. tricornutum, there are at least 10 candidate HCO₃⁻ transporter genes. Of these genes, seven genes were found to be similar to members of the mammalian SLC4 family, and the remainder were similar to the SLC26 family (Fig. 1A). In mammals, SLC4 and SLC26 are unrelated multigene families of HCO₃⁻ transporters, which have evolved independently (29, 30). These proteins transport HCO₃⁻ across the mammalian plasma membrane and play a critical role in maintaining intracellular and extracellular pH within a narrow range (31), and mutations in these proteins are known to cause various genetic diseases in humans (30).

All candidate SLC4 and SLC26 in the P. tricornutum genome were transcriptionally active (Fig. 1A). Of these genes, ptSLC4-5, -6, and -7 were previously predicted to encode HCO₃⁻ transporters by Kroth et al. (32). Three SLC4-type genes (ptSLC4-1, ptSLC4-2, and ptSLC4-4) displayed transcriptional activation specifically under atmospheric CO₂ condition (low CO₂), whereas they were repressed in high CO₂ (5% CO₂) (Fig. 1 A and B), suggesting that these ptSLC4 products are functional under CO₂-limiting conditions. A phylogenetic relationship of SLC4 family genes was analyzed (Fig. 1C, Dataset S1). Together with ptSLC4-3 and ptSLC4-5, aforementioned CO₂-responsive genes formed a diatom-specific cluster that likely shares a common origin with Homo sapiens SLC4 genes (supported by a bootstrap value of 776) (Fig. 1C). Interestingly, three CO₂-responsive PtSLC4 genes (ptSLC4-1, -2, and -4) were most closely related (Fig. 1C). Three other SLC4 candidates in P. tricornutum (ptSLC4-6 and -7) and Thalassiosira pseudonana (tpSLC4-2) were clustered with heterokont genes, which were also related to the aforementioned human/diatom cluster (supported by a bootstrap value of 983) (Fig. 1C). Conversely, T. pseudonana tpSLC4-1 was not related to any clade with a reliable bootstrap value; hence, its phylogenetic position was not clear (Fig. 1C).

In silico translation of full-length cDNA sequences of the CO₂-responsive ptSLC4-1, -2, and -4 revealed them to be 76–82% similar (Fig. S1) and have significant similarity to human SLC4A1 (29%, 29%, and 27% sequence homology, respectively). Moreover, 12 membrane-spanning helices of human SLC4A1 were highly conserved (Fig. 2A and Fig. S1, TM2–TM13). With the exception of the first transmembrane domain, which was unique to PtSLC4-2, a comparison of membrane topology deduced from the translated amino acid sequences revealed a striking similarity between PtSLC4-2 and human SLC4A1 (Fig. 2A). The aforementioned PtSLC4 candidates possessed one or two of the N-glycosylation site N-X-S (Fig. S1, underline), a characteristic of human SLC4A1 and SLC4A4 with a proposed role in promoting protein folding (33, 34). PtSLC4-1 and PtSLC4-2 also possessed amino acid sequences homologous to human SLC4s, Asp-Gly-Asp-Asp (DGDD) and/or Ser-Asp-Asp-Val (SDDV), at their C termini (Fig. 2A and Fig. S1, bold). The C terminus of human SLC4A4 possesses two acidic motifs, Asp-Asn-Asp-Asp (DNDD) and Leu-Asp-Asp-Val (LDDV), which are speculated to form a transport metabolon with CAII (35). The structural properties imply that the transport mechanism of these CO₂-responsive PtSLC4s is similar to the mechanism of human SLC4s.

Localization of Exogenously Introduced PtSLC4-2:GFP Fusion.

The enhanced GFP gene (egfp) was fused immediately downstream of the full-length ptSLC4-2 cDNA and transformed into P. tricornutum cells under the control of the fucoxanthin chlorophyll (Chl) a/c binding protein gene promoter, which functions regardless of [CO₂] in the medium (36). Quantitative RT-PCR (qRT-PCR) and Western blotting showed the constitutive expression of the GFP fusion product in the transformants carrying the ptSLC4-2::egfp fusion (SLC4G cells) (Fig. S2 A and B). The expressed GFP fusion was equally localized in the peripheral area of SLC4G cells (Fig. 2 B–E), whereas no GFP signal was observed in WT cells (Fig. 2 F–H). Cross-sections at the distal tip area of fusiform cells showed a distinct ring-shaped pattern of the GFP signal (Fig. 2 I and J), indicating that membrane-spanning helices of PtSLC4-2 target the plasmalemma. We had confirmed the occurrence of GFP signal before all of the following experiments (Fig. S2C). As shown in Fig. 2 D and E, large dots of GFP signal, which were not overlapped with either Hoechst or Mitotacker signals (Fig. S2D), were occasionally observed in SLC4G cells. We assume that this body is a part of the Golgi apparatus with exocytotic vesicles carrying PtSLC4-2:GFP fusion proteins, although details are, so far, not clear.

Influence of Exogenously Introduced PtSLC4-2:GFP on Photosynthetic Parameters.

Because of a down-regulation of the CCM (2), photosynthetic affinity for HCO₃⁻ was decreased significantly for WT cells grown in high CO₂ with a K_0.5[HCO₃⁻] of 515 µM (Table 1, Fig. S3A). Conversely, high-CO₂–grown SLC4G cells revealed 2.8-fold lower K_0.5[HCO₃⁻] value relative to high-CO₂–grown WT cells (Table 1). The [DIC] at the CO₂ compensation point was consistent with these changes in affinity, whereas P_max values were stable in WT and SLC4G cells (Table 1), indicating a significant improvement in photosynthetic affinity caused by constitutive expression of PtSLC4-2 under high-CO₂ conditions. Little difference in the photosynthetic parameters was observed between WT and SCL4G cells grown in low CO₂ (Table 1), probably because the effect of the boosted PtSLC4-2 expression was veiled by the fully operational CCM components in the low-CO₂–acclimated cells.

DIC Fluxes Across the Plasma Membrane and Gross HCO₃⁻ Uptake by PtSLC4-2.

A DIC depletion and evolution time course by P. tricornutum cells was determined by a filtration–centrifugation technique combined with a gas chromatographic (GC) detection of [DIC] (37) (methodological validation is in Fig. S3 B–D). These parameters were measured at an initial external DIC of 100 µM, a concentration well below what is required to saturate photosynthesis at the seawater level pH (thus, it is suitable to compare photosynthetic affinity). WT cells grown in high CO₂ showed little net DIC uptake rate (N_DIC), which was calculated from the initial rate of DIC depletion from the bulk medium (Fig. 3A, open circles). In contrast, N_DIC by low-CO₂–grown WT and SLC4G cells was fairly quick (3.54 and 2.38 µmol mg⁻¹ Chl min⁻¹, respectively), with ∼60% DIC taken up from the surrounding seawater in the first 30 s (Fig. 3A, closed symbols). Similarly, high-CO₂–grown SLC4G cells revealed fast N_DIC (1.27 µmol mg⁻¹ Chl min⁻¹) (Fig. 3A, open triangles). The CO₂ efflux rate (E_c) was determined from DIC evolution time course immediately after turning off the light in an assumption that the efflux of HCO₃⁻ is negligible relative to the efflux of CO₂ (20). High-CO₂–grown WT cells showed E_c of 1.15 µmol mg⁻¹ Chl min⁻¹ after turning off the light in a condition where external DIC was quickly replaced by DIC-free medium (Fig. 3 A, open circles, and B). Similar E_c values were obtained with high-CO₂–grown SLC4G cells and low-CO₂–grown WT cells (Fig. 3 A, open triangles and closed circles, and B) when light was turned off at the stationary phase of each DIC uptake. The Chl a content of SLC4G cells was marginally different from WT cells, irrespective of growth conditions (Fig. 3B). The net rates of CO₂ fixation in each cell were determined by the O₂ evolution rate at 100 µM DIC in the light (Fig. 3B). Together with the Chl a content and the net fixation rate, aforementioned parameters E_c and N_DIC are summarized in Fig. 3B.

A DIC flux model across the plasma membrane was constructed according to Badger et al. (38) (Fig. S4), and the model was solved theoretically with the parameters described above (Fig. 3B and SI Results). As a result, a gross DIC uptake (G_DIC) in high-CO₂–grown SLC4G cells was about 2.4 times and 57% of G_DIC of high- and low-CO₂–grown WT cells, respectively (Fig. 3B). The contribution of PtSLC4-2 to the gross HCO₃⁻ uptake (G_SLC4-2) in high-CO₂–grown SLC4G cells was, thus, calculated to be 1.44 µmol mg⁻¹ Chl min⁻¹ (Fig. 3B). The DIC accumulation in high-CO₂–grown SLC4G cells was about 2.2 times higher than the DIC accumulation of high-CO₂–grown WT cells (Fig. S5A), reflecting the difference in G_DIC between these cells. The difference in E_c values in Fig. 3 was small, and thus, we used the N_DIC value for evaluations of the PtSLC4-2 activity in assays hereafter.

Inhibition of PtSLC4s.

4,4′-Diisothiocyanostilbene-2,2′-disulfonic acid (DIDS), an inhibitor commonly used for mammalian SLC-type HCO₃⁻ transporters (39), did not significantly inhibit photosynthetic efficiency in high-CO₂–grown WT cells (Fig. 4A). In sharp contrast, DIDS treatment of low-CO₂–grown WT cells and SLC4G cells (grown under both high- and low-CO₂ conditions) significantly increased the K_0.5[HCO₃⁻] values in a dose-dependent manner, whereas P_max values were unaffected by DIDS (Fig. 4A). The effect of DIDS on K_0.5[HCO₃⁻] of low-CO₂–grown cells was significant (Fig. 4A), indicating a major role for the DIDS-sensitive SLC-type transporters supporting high-affinity photosynthesis in low-CO₂–grown P. tricornutum. The same tendency was apparent in the effect of saturating (3 mM) DIDS on N_DIC. There was very little N_DIC in high-CO₂–grown WT cells, and DIDS exerted little inhibition; however, in low-CO₂–grown WT cells, 3 mM DIDS reduced N_DIC by 65% (Fig. 4B). By clear contrast, high-CO₂–grown SLC4G cells showed the fast N_DIC comparable with low-CO₂–grown cells, and 67% of this uptake was inhibited by DIDS (Fig. 4B). These results account for the function of the plasma membrane transporter, PtSLC4-2, and perhaps, PtSLC4-1 and -4, which support the majority of the DIC influx into diatoms. Our previous study of the physiological estimate and the molecular localization negated the occurrence of CA_ext in the strain used in this study (40, 41). Furthermore, treatment by a CA_ext inhibitor, N-(5-sulfamoyl-1,3,4-thiadiazol-2-yl) acetamide, showed little effect on N_DIC (Fig. S5B). The CA_ext was, thus, excluded from the CCM model in this study.

pH and Substrate Dependency.

There were clear differences in pH dependency of N_DIC between high-CO₂–grown WT and SLC4G cells. High-CO₂–grown WT cells displayed little N_DIC in the presence of initial HCO₃⁻ of 100 μM in the media with pH 8.2 and 9.5, where there was virtually no CO₂ available, whereas N_DIC was evident (∼1.0 μmol DIC mg⁻¹ Chl min⁻¹) at pH 7.5 (Fig. 4C, white bars), which is presumably supported by a rapid production of CO₂ at this pH. In sharp contrast, high-CO₂–grown SLC4G cells showed significant N_DIC at pH 9.5 (0.21 µmol DIC mg⁻¹ Chl min⁻¹), which reached a maximum level at pH below 8.2 (1.27–1.32 µmol DIC mg⁻¹ Chl min⁻¹), where HCO₃⁻ represented 94% of DIC (Fig. 4C, gray bars), strongly suggesting that HCO₃⁻ is preferentially taken up by PtSLC4-2 and functions optimally at the pHs typical of seawater. Together, from this evidence, it is clear that DIC influxes into low-CO₂–grown WT cells are primarily supported by direct DIDS-sensitive HCO₃⁻ uptake, for which PtSLC4-2 is one of critical components.

Na⁺ Dependency of PtSLC4-2.

Many mammalian SLC4 transporters are Na⁺-dependent (39, 42, 43). DIC uptake in high-CO₂–grown SLC4G cells was severely inhibited after complete depletion of Na⁺ ions from F/2ASW in experiments where osmotic pressure was maintained constant with choline-chloride (Fig. 5A). A linear increase in the N_DIC of SLC4G cells was associated with increasing [Na⁺] and became saturated at levels above 100 mM Na⁺ (Fig. 5A). Oxygen evolution in high-CO₂–grown SLC4G cells was also Na⁺-dependent, with saturation over 100 mM (Fig. 5B). N_DIC by PtSLC4-2 was stimulated by Na⁺, and neither K⁺ nor Li⁺ provided functional substitution (Fig. 5C). The same ion selectivity was displayed for photosynthetic parameters. Photosynthetic affinity, K_0.5[HCO₃⁻], of high-CO₂–grown WT cells was unaffected by [Na⁺], whereas a marked decrease in K_0.5[HCO₃⁻] was evident in both low-CO₂–grown WT and high-CO₂–grown SLC4G cells in the presence of 100 mM NaCl or 50 mM Na₂SO₄ (Fig. 5D). This Na⁺-dependent stimulation of affinity did not occur in the presence of an equivalent concentration of K⁺ or Li⁺, and the maximum photosynthetic activity, P_max, was fairly stable at around 140 μmol mg⁻¹ Chl h⁻¹ in all experiments (Fig. 5D). These data indicate that Na⁺ is required for the DIC acquisition system and PtSLC4-2 activity.

Cells and Culture Conditions.

The marine diatom P. tricornutum Bolin (UTEX 642) was obtained from the University of Texas Culture Collection and cultured in artificial seawater, which was supplemented with half-strength Guillard’s f solution (F/2ASW) under continuous illumination (50–75 µmol m⁻² s⁻¹) at 20 °C under constant aeration with 5% CO₂ or ambient air (0.039% CO₂).

Semiquantitative RT-PCR and qRT-PCR.

Total RNA was isolated from high- and low-CO₂–grown WT cells using the RNeasy Plant Mini Kit (Qiagen). After the single-strand cDNAs were synthesized, semiquantitative RT-PCR and qRT-PCR were carried out by a set of primers shown in Table S1.

Construct of a Transformation Vector Containing the ptSLC4-2::egfp Fusion.

Inverse PCR was carried out on pFcpApGFP vector (46) as a template with high-fidelity PrimeSTAR HS DNA polymerase (TaKaRa) and a primer pair: egfp Fw: 5′-ATGGTGAGCAAGGGCGAGGAGCTGTTC-3′; fcpAp Rv: 5′-TCGAAACGGCAGACAAATTTGTG-3′. A ptSLC4-2 fragment was amplified from single-strand cDNA synthesized by the SMART RACE cDNA Amplification Kit, phosphorylated, and ligated at the upstream region of egfp in the inverse PCR product described above.

Transformation of WT Cells and Screening of Transformants.

Transformation was carried out as previously described (47), and transformants were screened on agar plates containing 100 µg mL⁻¹ Zeocin. GFP positive clones were further screened from the Zeocin-resistant clones.

Confocal Laser Microscopy.

Cells at midlogarithmic growth phase were harvested and resuspended in a small volume of F/2ASW. The nucleus in the SLC4G cells was stained by 5 μM Hoechst 33342 at room temperature for 30 min. The stained cells were washed three times with F/2ASW. Specimens were observed using a Nikon A-1 confocal microscope.

Determination of Photosynthetic Parameters.

After high- and low-CO₂–grown cells at the midlogarithmic growth phase were harvested by centrifugation, cells were washed with DIC-free F/2ASW. Cells were then suspended in the same buffer at a Chl a concentration of 10 µg mL⁻¹ (determined as described in SI Materials and Methods). The rate of photosynthetic O₂ evolution at various concentrations of DIC was measured with a Clark-type oxygen electrode (Hanzatech). [DIC] at CO₂ compensation point was measured by GC. K_0.5 and P_max values were determined by the least-squares method.

Determination of the DIC Flux Parameters.

Cells were washed and resuspended as described above, except that final Chl a concentration was set at 40 µg mL⁻¹. N_DIC and E_c values were determined by a filtration–centrifugation technique using an Eppendorf tube equipped with a filter cartridge. Cell culture maintained at the CO₂ compensation point was mixed in the filtrater cartridge with NaHCO₃ at a final concentration of 100 µM. The bulk medium was filtered by centrifugation after various reaction times in the light, and [DIC] was determined by GC. DIC depletion rate was calculated as N_DIC. E_c was determined as a CO₂ evolution rate in the dark. E_c of high-CO₂–grown WT cells was determined in the dark immediately after the removal of the external DIC. G_DIC and G_SLC4-2 values were calculated as described in SI Materials and Methods.

Additional details of materials and methods are provided in SI Materials and Methods.