In marine diatoms, the metals cadmium, cobalt and zinc can functionally substitute for one another to maintain optimal growth rates. This effect is at least partly due to metal replacement in the metal-binding site of the enzyme carbonic anhydrase4, which is involved in the acquisition of inorganic carbon for photosynthesis. In addition to a zinc carbonic anhydrase that can substitute cobalt in its active site in vivo5, T. weissflogii has a putative cadmium carbonic anhydrase that is also involved in acquiring inorganic carbon6.

We purified a protein, CDCA1, from this organism that has carbonic anhydrase activity and contains cadmium (Fig. 1a, b; Genbank accession number AY772014). Determination of its sequence was complicated by the presence of a triple repeat (see supplementary information): the three amino-acid sequences are virtually identical (about 85% identity), but there is more variation in their encoding DNA (about 78% identity). Contrary to an earlier estimate6, we have determined the relative molecular mass of CDCA1 as about 69K (for methods, see supplementary information). CDCA1 has probably not been sequenced before, as there are no hits in the NCBI database. It is significantly different from any of the known major classes of carbonic anhydrases7,8, and therefore represents the first member of a new class of carbonic anhydrases, the ζ class.

Figure 1: Characterization of the carbonic anhydrase CDCA1 from a marine diatom.
figure 1

a, High-pressure liquid chromatography (HPLC) of the CDCA1 enzyme, showing co-elution of purified CDCA1 (black trace, absorbance at 280 nm) and of 109Cd-radiolabelled enzyme (red, c.p.m., 330 c.p.m at peak) at 24 min. Blue line, NaCl gradient. b, Non-denaturing gel electrophoresis of HPLC fractions containing the main peak of protein and radiolabel shows an intense Coomassie-blue-stained protein band (left) that co-migrates with the 109Cd radiolabel (centre) and carbonic anhydrase activity (right). c, Cadmium K-edge X-ray absorption spectra from purified CDCA1 (top) compared with two tetrahedral, thiolate-coordinated species, [Cd(SPh)4](Me4N)2 (green solid line), cadmium phytochelatin (green broken line), and with two octahedral species [Cd(H2O)6]2+ (purple solid line) and [Cd(imidazole)6](NO3)2 (purple broken line). The similarity of the CDCA1 spectrum to that of the tetrahedral species suggests that the metal site has tetrahedral symmetry and involves cysteinyl ligands to the metal. (For methods, see supplementary information.)

The genome of T. pseudonana contains a single sequence that is highly homologous to the three repeats in T. weissflogii, corresponding to a protein of 25.5K (see supplementary information). The presence of expressed-tag sequences shows that this putative cadmium-containing carbonic anhydrase is expressed, indicating that only a single repeat of CDCA1 may be necessary for activity. The amino-terminal sequence of the T. pseudonana enzyme (not available for CDCA1) has only 15 amino acids preceding the homologous sequence from T. weissflogii.

X-ray absorption near-edge spectroscopy of the purified protein confirms the presence of the cadmium-binding site (Fig. 1c). Comparison with standards9 indicates that the site probably has a roughly tetrahedral geometry and that the cadmium ion is bound by two or more thiolates — as in the zinc-containing β class of carbonic anhydrases in higher plants, which have two cysteines and a histidine at the metal-binding site10. The spectra are also consistent with an active site containing an activated water molecule, as in other carbonic anhydrases.

Gene-expression analysis shows that there is an increase in the abundance of cdca1 transcripts within 1 hour of increasing the concentration of cadmium or of decreasing the partial pressure of carbon dioxide in the medium (results not shown). Expression of CDCA1 may therefore be controlled in part by the availability of cadmium and carbon dioxide in sea water.

High-throughput sequencing of a seawater sample has revealed that the marine environment may contain unique genes1. We have identified and partially characterized one such gene — for a carbonic anhydrase from a marine diatom that, to our knowledge, is the first native enzyme so far discovered to contain cadmium. Because of the extraordinarily low concentrations of many essential trace metals in sea water, it is likely that there are other metalloenzymes in marine organisms that use unusual metals for activity and contribute to trace-metal geochemical cycling in the oceans.