In vivo substitution of zinc by cobalt in carbonic anhydrase of a marine diatom

Transcription

1 Notes 573 Limnol. Oceanogr., 41(3), 1996, C 1996, by the American Society of Limnology and Oceanography, Inc. In vivo substitution of zinc by cobalt in carbonic anhydrase of a marine diatom Abstract We examined the mechanism by which Co additions partially alleviate the growth limitation of marine diatoms in cultures with low Zn concentrations. Cultures of the coastal diatom Thalassiosira weissflogii were grown under lowzn conditions and provided with low and high levels of Co. Enzyme assays and autoradiographs of electrophoresis gels of cellular proteins revealed that a large fraction of the Co is found in the main isoform of the enzyme carbonic anhydrase (CA), where most of the Zn is located. The affinity of CA for Co is lower than for Zn, and the Cosubstituted enzyme is less active than the native Zn form. Comparisons of growth under different partial pressures of CO, show that Co can ameliorate the growth of carbonlimited cultures. We conclude that Co alleviates Zn limitation via direct substitution in CA, which is important for carbon acquisition. It has been shown that cobalt can substitute partially for the zinc requirement of marine diatoms (Price and Morel 199; Sunda and Huntsman 1995). The addition of Co in Znlimited cultures of both coastal and oceanic diatoms results in partial restoration of growth. Recent work has demonstrated a link between Zn availability and carbon acquisition in Thalassiosira weissjlogii (Morel et al. 1994). This link is apparently a result of the function of carbonic anhydrase (CA), a major Zncontaining soluble intracellular protein. There is evidence that Co restores CA activity in Znlimited T. weissjlogii (Morel et al. 1994). Here, we demonstrate that the mechanism of Co replacement of Zn is by in vivo substitution for Zn in at least one CA of T. weissflogii. This substitution restores some of the CA activity lost by Zn limitation. The concurrent improvement in growth rate apparently results from an improvement in the ability of Znlimited cells to acquire carbon. Cultures of T. weissflogii were grown in synthetic culture media based on the standard recipe for Aquil (Price et al. 1989) that contained reduced concentrations of EDTA (1 PM total, to allow a low total amount of radioactive and stable Co to achieve the desired available trace metal activities) and modified Zn and Co levels: regular Zn and Co levels (16 and 22 pm inorganic metal, i.e. free metal and complexes with inorganic ligands, and 8 and 5 nm total metal, respectively), a tenth the Zn, a tenth the Co, or a tenth both the Zn and Co. Radioactive Co (NEN/DuPont, 18.5 x lo4 Bq 57Co liter culture, as CoCl,) was added before the medium was inoculated with cells. Batch cultures were maintained in polycarbonate bottles under continuous artificial illumination at an intensity of.9 x lo6 quanta sl cm2. Bottles were kept open to the atmosphere (loosely capped) and unmixed at 2 C. Cells were counted with a hemacytometer. Once cells grew to a density of 1 O* cells mll, they were collected on 3pmpore polycarbonate filters (Poretics) and stored in liquid nitrogen. Cells were then resuspended in deionized water with 2 PM leupeptin and.5 mm PMSF (Sigma) added as protease inhibitors. Resuspended cells were broken via sonication (Branson Sonifier 25) on ice at 5% power and a 5% duty cycle for two 3s sessions with a 15s pause for cooling. Unbroken cells and insoluble cell fragments were removed by centrifugation at 16, x g for 2 min. An equal volume of soluble cell homogenate (all the cultures were grown to about the same cell density and resuspended and lysed in the same volume) was loaded into each lane. Samples were separated by means of nondenaturing electrophoresis on 1% polyacrylamide gels (ph 8.8) with a 5% stack (ph 6.8) in a 25 mm Tris, 25 gm glycine buffer. Subsamples of the cell homogenate were also taken and purified on an affinity column for CA at 4 C (Yang et al. 1985). Each sample was loaded on the column and washed with 25 mm Tris, 22 mm Na,SO, buffer at ph 8.2, washed with 25 mm Tris, 3 mm NaClO, buffer at ph 8.7, and then eluted from the column with a solution of 25 mm Tris, 5 mm NaClO, at ph 5.6. Eluted fractions were collected and concentrated by ultrafiltration through filters with a nominal molecular weight cutoff of 1, Da (Millipore Ultrafree MC). Purified samples were also separated by nondenaturing gel electrophoresis. A separate lane of bovine CA (Sigma) was run on each gel as a control. After electrophoresis, the presence of CA activity was assayed by the method of Patterson et al. (197 1). The ph change from the catalyzed dissolution of CO2 was observed by blowing gaseous CO, on the surface of a gel stained in a 25 mm Tris, 25 mm glycine buffer (ph 7.6) containing.1% (wt/vol) of the color ph indicator bromcrescol purple (Sigma). The gels were frozen on dry ice to stop the reaction, photographed under UV illumination, then dried on a heated vacuum gel dryer. Dried gels were exposed to Xray film (Kodak Xomat AR) to determine the location of material containing radioactive co. To examine whether Co is similar to Zn in alleviating growth limitation in cultures grown under low Pco2, we grew T. weissflogii in Aquil with lofold lower Zn and either IOfold lower or regular Aquil Co concentrations, and in Aquil with regular Zn with 1 Ofold lower Co. These cultures were aerated continuously with an artificial air source containing nitrogen and oxygen in atmospheric proportions, and 1, 3, or 1, ppm C2. Cultures were kept at 22 C and grown under artificial fluorescent light at an intensity of 1.1 x lo6 quanta sl cm2. Sub

2 a b Fig. I. [a.] PAGE for samples from cultures grown at different concentrations ofzn and Co assayed for CA activity using the color ph indicator bromcrescol purple, with low ph (CO, dissolution catalyzed by CA) indicated by light bands on the gel. [b.] Gel previously assayed with bromcresol purple for CA, dried, and exposed to Xray film. Samples are from cultures grown in media containing the following combinations of Zn and Co (relative to standard Aquil concentrations): 1 Zn/lO, Co/IO; 2Zn/lO. Co; 3Z, Co/lo; 4Z, Co. The same amount of Co was added to all cultures (I 8.5 x IO Bq literr ). All four lanes are from the same gel. Dividing lines are provided to aid in differentiating lanes where the edges of bands overlap. samples of the cultures were removed periodically and counted with a Coulter Counter. The final ph of the cultures varied depending on the level of CO, supplied; ph ranged from 8.5 in the 1 ppm cultures to 8. in the 3 ppm and 7.5 in the 1, ppm cultures. Among cultures with varying inorganic Zn and Co concentrations, those with the smaller amounts of both metals grew at the lowest rate (.4 d ). The addition of Zn alone to Zn = 16 pm (the regular Aquil concentration ofinorganic Zn) resulted in nearly the usual rate ofgrowth (1.3 dl); adding Co alone increased growth to a lesser extent (.8 d&l). Addition of both Zn and Co at regular concentrations (Co = 22 pm and Zn = 16 PM) resulted in the usual maximum growth rate (1.4 d ). Measurements of CA activity show that both Zn and Co concentrations affect CA activity. The first lane of the PAGE gel in Fig. la, a sample from a culture with low concentrations of both Zn and Co, has undetectable CA activity. A control of bovine CA was used to ensure that the assay was functioning in the event that none of the samples showed any activity. However the control can provide no additional information about protein size in a b Fig. 2. [a.] PAGE for samples from cultures grown at dlfferent concentrations of Zn and Co, purified and concentrated on a CA affinity column, then assayed for CA activity using the color ph indicator bromcrescol purple, with low ph (CO, dissolution catalyzed by CA) indicated by light bands on the gel. [b.] The same PAGE for samples purified on a CA affinity column. After a bromcresol purple assay for CA activity, the gel was dried and exposed to Xray film. Details as in Fig. I a nondenaturing gel and is therefore not shown. Regular Co in the culture medium restores CA activity to a detectable level, and cultures grown at regular Zn (whether at low or regular Co) exhibit yet higher activity. Thus, Co addition promotes parallel effects on the restoration of CA activity and on growth rates in low Zn cultures. The restoration of CA activity by Co may be due either to the displacement of Zn in other proteins or to a direct functional substitution of Co for Zn in CA. The autoradiagram of a PAGE gel (Fig. lb) from samples of 5 Colabeled cultures shows the Co in bands where there is CA activity (Fig. la). The major radioactive band found in each of the lanes corresponds to the location of the majority of CA activity. Because of the broadness of the bands in the gel, coelution of CA and s7co does not positively demonstrate that Co is present in CA. To establish the presence of Co in CA, we purified the soluble material from 57Colabeled cultures on a CA affinity column, after which only one radioactive band remained in each lane (Fig. 2b). The location of these bands corresponds TV the location of CA activity (Fig. 2a), and the relative amounts of radioactive Co found in each lane (lane 1>2>3>4) appear similar to those observed in the main CA bands in the unpurified samples (Fig. la). Densitometty measurements of square areas around the radioactive bands confirms this simi

3 Notes 575 Table 1. Averaged gray density on autoradiograms of 57Co in region of main CA activity ( is white, 255 is black). Background Lane 1 (Zn/1, Co/lo) Lane 2 (Zn/l, Co) Lane 3 (Zn, Co/l ) Lane 4 (Zn, Co) Unpurified Purified (Fig. lb) (Fig. 2b) v E le+6 _ o I I low CO2.28 dl I I a q med CO2.34 d * high CO2.46 d larity (Table 1). Thus, Co is indeed substituting for Zn in CA. Because the radioactive cultures were all grown with the same concentration of 57Co (but not the same total Co or Zn), variations in the intensity and size of the 57Co bands from lane to lane, each containing material from approximately an equal number of cells (although this does not mean the quantities of protein in each lane are equal, it is equivalent to a percell comparison), provide insight into Co replacement of Zn in T. weissflogii. In the main CA band, the amount of radioactive Co is greatly reduced by the presence of (cold) Zn (cf. lanes 1 and 3 in Table 1). The addition of an equivalent concentration of cold Co rather than Zn results in much less displacement of the 57Co (lanes 2 vs. 3). Likewise, additional cold Co in a Znsufficient culture (lane 4) causes only a slight decrease in the 57Co activity of the main CA band. From the autoradiogram data, the CA of T. weissflogii seems to have a higher affinity for Zn than for Co. The assays of CA activity also show that the Zn form of the enzyme has more activity than the Co form. However, when Zn is insufficiently available, Co can clearly substitute for Zn in CA and provide partial activity. The autoradiograms show that under conditions of low Zn, significant amounts of Co appear in two unidentified protein bands that have no measurable CA activity and that are not retained by the CA affinity column (and therefore are likely not another CA isoform). However, the largest fraction of the soluble cellular Co is in CA, which parallels the distribution of Zn in T. weissflogii (Morel et al. 1994), where a large fraction of the soluble Zn protein is CA. If a major function of Co under lowzn conditions is as a cofactor for CA (as Zn normally is), we expect that like Zn, Co might be involved in carbon acquisition. We tested this assumption by varying Pco2 along with Zn and Co in T. weissflogii cultures. Cultures with low Zn and Co aerated under different partial pressures of CO, exhibited large differences in growth rates (Fig. 3a)an effect that disappeared at high Zn (Fig. 3~). For all Pco2, Fig. 3. Growth curves of Thalassiosira weissflogii cultures containing different concentrations of Zn and CO relative to standard Aquil concentrations and grown under different partial pressures of CO, (low 1; medium 3; high 1, ppm). 7 F. 2 i A Zn/l, Co/l le+4 I I I I I 1 e+6 le+5 1 e+4 1 e+6 I I I 3 CJ low CO2.44 d b med CO2.53 dl * high CO2.78 dl Zn/IO, Co I I I I le+3 I I I I I A Zn, Co/l Days

4 576 Notes & 6 z 9 1 I North Atlantic I Northeast Pacific n North Pacific lon I I I I I 1 II a Dissolved CO, (FM) Fig. 4. Comparison of surface dissolved inorganic (DI) Co concentrations and surfacesaturated CO, concentration from surfacewater temperature, using the equation of Edmond and Gieskes (197) for the Henry s law constant. Surface Co and temperature measurements from Martin et al. ( ) and Martin and Gordon (1988). growth rates were higher in the cultures provided with high Co concentrations. Thus, in those cultures that are clearly C limited (by combination of low Zn and low partial pressures of CO,), addition of Co partially restores growth rates (Fig. 3b), in parallel with its effect on CA activity. The failure of Co to eliminate the sensitivity of T. weissflogii to Pco,! at low Zn parallels its inability to fully restore growth in Znlimited cultures at atmospheric CO2 levels. Our culture data show that growth limitation induced by lowered Zn concentrations in cultures of T. weissflogii is partially offset by adding Co. This result confirms previous data of Price and Morel (199) in the same organism and is similar to results for two other Thalassiosira species, T. pseudonana and T. oceanica (Sunda and Huntsman 1995). Although all of our culture media, including those containing low inorganic Co (Co = 2.2 pm, Cotot =.5 nm), contain.4 nm vitamin B12, this organic Co source is unable to satisfy the Zn and Co requirements (normally Co = 22 pm, Co,,,= 5 nm) ofznlimited algae. Sunda and Huntsman (1995) noted that the amount of Co in vitamin B12 does not satisfy the Co requirement in Znlimited T. pseudonana or Emiliania huxleyi. Clearly, Co has a role in some phytoplankton independent of its function as a metal cofactor in vitamin B12. Our data from PAGE gels, which show the effect of Zn and Co concentrations on CA activity, extend previous results that demonstrate a partial restoration of CA activity by Co in lowzn cultures (Morel et al. 1994) and a An loss of CA activity in the same band from which Zn is lost upon Zn limitation. The appearance of Co in the CA band (as shown by radiolabeling) under lowzn conditions indicates that the positive effect of Co on CA activity is not merely the freeing of Zn from other functions, but rather a direct substitution of Co for Zn in this enzyme in vivo. The apparent affinity for Zn in T. weissflogii CA is much higher than for Co, and the activity of the Cosubstituted CA is lower than in the native Zncontaining enzyme. These differences may explain why adding Co improves but does not eliminate the growth limitation caused by low Zn concentration. This partial activity of the Cosubstituted CA of T. weissflogii is similar to results for the bovine erythrocyte CA substituted with Co in vitro (Tu and Silverman 1985). We note that this similarity in the efficacy of the Cosubstituted enzyme is seen despite the absence of any sequence homology between bovine and T. weissflogii CAs (Roberts 1995) and also despite differences in the methods by which the substituted metal is introduced into the enzyme. Under lowzn conditions, a large fraction of the Co in soluble cellular proteins is found in the largest CA band (only one ZnCA band appears in the figures presented here because the concentration of the cell extract was not sufficient to visualize the minor bands previously seen), where we previously found much of the Zn under ordinary culture conditions (Morel et al. 1994). In contrast, when Cd is added to lowzn cultures, it is found primarily in a different CA band (in the figures presented here, the CdCA does not appear because no Cd was added to any of the cultures) that migrates much less in a nondenaturing PAGE (Lee et al. 1995). The appearance of Co and Cd in different bands indicate that they substitute in different CA isoforms, and this might explain the differences in the growth response of cells: unlike Co, Cd is able to restore full growth in Znlimited cells. Nonetheless, as expected from cofactors in CA isoforms, both metals are clearly involved in carbon acquisition. The difference in function of CoCA and CdCA simply may arise from differences in the locations of the enzymes (e.g. internal and external to the cell). The use of Co in phytoplankton CA when Zn is limiting should be reflected in the distribution of Co concentrations in the surface oceans. Concentration of Co in seawater typically ranges from 5 to 5 pmol kgl (Martin and Gordon 1988; Martin et al. 199) and is fairly invariant at depth, dropping noticeably only very near the surface. Sunda and Huntsman (1995) noted that this Co depletion in surface seawater coincides with regions where Zn concentrations are very low (~.5 nmol kgl). The depletion of Co in regions of low Zn is expected if Co is a replacement for Zn in phytoplankton. According to our data, Co utilization by phytoplankton, and hence its depletion in surface seawater, ought to be linked also to CO, availability. Unlike Co concentrations at depth, Co concentrations in the top 2 m from the North Atlantic and North Pacific (where Zn concentrations are low) are quite variable. We compared these surface Co concentrations

5 Notes 577 to the corresponding (saturated) CO, concentrations (which depend on temperature) in the water. Consistent with our expectations, the results (Fig. 4) demonstrate a greater depletion in Co where CO, is lowest. Even though the situation may be complicated by the different metal requirements of the dominant phytoplankton species in different areas of the ocean, depletion of Co from surface seawater seems to be chiefly determined by both low Zn and CO* concentrations. Conversely, biological CO* uptake and export in the open ocean is likely dependent on the availability of some key trace metals, including Co. Donald Yee R. M. Parsons Laboratory Department of Civil and Environmental Engineering Massachusetts Institute of Technology Cambridge Department of Geology Princeton University Princeton, New Jersey 8544 References Fraqois M. M. Morel EDMOND, J. M., AND T. M. GIESKES On the calculation of the degree of saturation of sea water with respect to calcium carbonate under in situ conditions. Geochim. Cosmochim. Acta 34: LEE, J. G., S. B. ROBERTS, AND F. M. M. MOREL Cadmium: A nutrient for the marine diatom Thalassiosira weiss$ ogii. Limnol. Oceanogr. 4: 156l 63. l Present address: Dept. of Geology, Princeton University, Princeton, New Jersey MARTIN, J.H.,S.E. FITZWATER, R.M. GORDON,C.N.HUNTER, AND S. J. TANNER Iron, primary production and carbonnitrogen flux studies during the JGOFS North Atlantic bloom experiment. DeepSea Res. 4: 115l 34., AND R. M. GORDON Northeast Pacific iron distributions in relation to phytoplankton productivity. DeepSea Res. 35: ~ AND S. E. FITZWATER Iron in Antarctic waters. N&ure 345: MOREL, F. M. M., AND OTHERS Zinc and carbon colimitation of marine phytoplankton. Nature 369: PATTERSON, B. D., C. A. ATKINS, D. GRAHAM, AND R. B. H. WILLS Carbonic anhydrase: A new method of detection on polyacrylamide gels using lowtemperature fluorescence. Anal. Biochem. 44: PRICE, N. M., AND OTHERS Preparation and chemistry of the artificial algal culture medium Aquil. Biol. Oceanogr. 6: ,AND F. M. M. MOREL Cadmium and cobalt substitution for zinc in a marine diatom. Nature 344: ROBERTS, S. B Carbonic anhydrase in the marine diatom Thalassiosira weissflogii. MS. thesis, Mass. Inst. Technol. SUNDA, W. G., AND S. A. HUNTSMAN Cobalt and zinc interreplacement in marine phytoplankton: Biological and geochemical implications. Limnol. Oceanogr. 4: Tu, C. K., AND D. N. SILVERMAN Catalysis by cobalt(i1) substituted carbonic anhydrase of the exchange of oxygen 18 between CO, and H,O. Biochemistry 24: YANG, S. Y., M. TSUZUKI, AND S. MIYACHI Carbonic anhydrase of Chlamydomonas: Purification and studies on its induction using antiserum against Chlamydomonas carbonic anhydrase. Plant Cell Physiol. 26: Submitted: 15 May 1995 Accepted: 19 October I995 Amended: 21 November 1995