Copper Response Element and Crr1-Dependent Ni²⁺-Responsive Promoter for Induced, Reversible Gene Expression in Chlamydomonas reinhardtii

Abstract

The Cpx1 and Cyc6 genes of Chlamydomonas reinhardtii are activated in copper-deficient cells via a signal transduction pathway that requires copper response elements (CuREs) and a copper response regulator defined by the CRR1 locus. The two genes can also be activated by provision of nickel or cobalt ions in the medium. The response to nickel ions requires at least one CuRE and also CRR1 function, suggesting that nickel interferes with a component in the nutritional copper signal transduction pathway. Nickel does not act by preventing copper uptake/utilization because (i) holoplastocyanin formation is unaffected in Ni²⁺-treated cells and (ii) provision of excess copper cannot reverse the Ni-dependent activation of the target genes. The CuRE is sufficient for conferring Ni-responsive expression to a reporter gene, which suggests that the system has practical application as a vehicle for inducible gene expression. The inducer can be removed either by replacing the medium or by chelating the inducer with excess EDTA, either of which treatments reverses the activation of the target genes.

Copper is an essential micronutrient for most organisms, including Chlamydomonas reinhardtii, because copper functions as a catalyst of redox reactions or reactions involving molecular oxygen. Well-characterized copper enzymes in Chlamydomonas include plastocyanin (50), cytochrome oxidase (2), and a ferroxidase (20, 32). Nevertheless, Chlamydomonas cells survive copper deficiency by activating adaptive mechanisms. Among these is the degradation of copper proteins such as plastocyanin to facilitate redistribution of copper to other, presumably more physiologically important, copper enzymes such as cytochrome oxidase (41). The cells retain photosynthetic metabolism in the face of plastocyanin loss by transcriptional activation of the Cyc6 gene encoding an alternate electron transfer catalyst, cytochrome c₆ (37). This response requires copper response elements (CuREs) associated with the Cyc6 gene and a wild-type CRR1 locus (45, 47). In addition to Cyc6, a number of other genes are regulated by copper nutritional state, including Cpx1, Crd1, and Cth1, whose gene products function in tetrapyrrole metabolism (23, 39, 40).

Previously, we noted that the response of the Cyc6 gene was highly specific for copper as the metal ion regulator. Other metal ions, such as Ag(I), Mn(II), Co(II), Zn(II), and Ni(II), cannot replace copper in deactivating Cyc6 expression (22). Mercuric ions can mimic copper as a repressor of Cyc6 expression, but only when added at a 20-fold-greater concentration than the minimally effective concentration of copper, and this was interpreted as an interaction with a critical thiol in a signal transduction component (22, 23). We noted recently that Cyc6, Cpx1, and Crd1 expression can be induced in fully copper-replete Chlamydomonas cells by the addition of nickel or cobalt ions. Besides the pharmacological utility of nickel or cobalt as a probe for the mechanistic dissection of the CuRE- and CRR1-dependent signal transduction pathway, the observation also suggested to us the possibility for using the Cyc6 promoter to design a system for inducible gene expression in Chlamydomonas.

Metal-responsive promoters, developed initially for mammalian cells (26, 56, 57), are favored tools for the design of inducible promoter systems and have been exploited in various model systems, including cyanobacteria, fungi, and plants (1, 5, 63). The extraordinarily tight transcriptional regulation of the Cyc6 gene immediately suggested its utility for such a purpose when the promoter was shown to be sufficient for conferring copper responsiveness to a heterologous gene (24, 47). Nevertheless, although the Cyc6 gene is turned off essentially instantaneously by provision of copper ions to copper-deficient cells, turning the gene on is not instantaneous because it requires significant dilution (by cell division) to attain intracellular copper deficiency.

We show here that (i) nickel and cobalt induce the synthesis of the copper-deficiency response genes Cyc6, Cpx1, and Crd1; (ii) nickel does not prevent copper uptake or utilization; (iii) the core of the CuREs associated with the Cyc6 and Cpx1 genes is essential also for nickel activation as is Crr1, a master regulator of the nutritional copper response in Chlamydomonas; (iv) the promoter regions of Cpx1 and Cyc6 containing the CuREs are sufficient for conferring nickel responsiveness to a reporter gene; and (v) the effect of nickel ions can be reversed either by chelation or by washing the cells and transferring them to Ni-free medium.

MATERIALS AND METHODS

Chlamydomonas strains and culture conditions.

C. reinhardtii wild-type strains CC125, 2137, and CC425; mutant strain crr1-1; and transformants of strain CC425 were cultured in copper-supplemented or copper-deficient TAP medium (46). Where indicated, cultures were supplemented with CuCl₂, NiCl₂, or CoCl₂ from stock solutions.

Nucleic acid analysis.

Total RNA was prepared and analyzed by hybridization as described by Quinn and Merchant (47) for Ars2 or as described by Hill et al. (22) for all other transcripts. Probes for Cpx1, Cyc6, Ars2, and RbcS2 (encoding the small subunit of Rubisco) were prepared as described previously (48). For Cβlp, a 915-bp EcoRI fragment from the cDNA insert in plasmid pcf8-13 was used (55). The RbcS2 or Cβlp hybridization signal is used for normalization between samples (not shown in every figure). For detection of Crd1 transcripts an insert from pCRD1-5 (39) was used.

Immunoblot analysis.

Total soluble protein was prepared (34) and separated on a 15% anionic gel for immunoblot analysis (22, 37). Blots were incubated overnight with a 1:1,000 dilution of anti-plastocyanin as the primary antibody and a 1:2,000 dilution of alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G (Southern Biotechnology Associates) as the secondary antibody. Bound antibody was detected by using the alkaline phosphatase color reaction (54).

Chimeric constructs.

The chimeric constructs used to test the requirement for the CuRE in Ni-induced gene expression were described previously (44, 47). For the Cyc6-derived constructs, construct A is the full-length Cyc6 promoter fused to a promoterless Ars2 reporter gene (8) called pCU1 (44, 47). Constructs B to F represent parts of the promoter region fused to the Ars2 reporter gene driven by the minimal promoter from the β-tubulin gene in plasmid pJD100 (7) and correspond to constructs 41, 40, 43, 39, and 42, respectively, shown in that study (44). The CuREs activate gene expression above the basal level of expression from pJD100 alone. The level of basal expression varies in individual transformants, depending presumably on the site of integration. The Cpx1-derived constructs G to J, correspond to fusions of the indicated part of the Cpx1 promoter to βtub-Ars2 reporter. They are the same constructs as A, C, E, and F shown in Fig. 3 in Quinn et al. (44). The promoter region of the Cyc6 gene containing two CuREs (from positions −129 to −7 relative to the 5′ end of the mRNA) is available from the Chlamydomonas Culture Collection as plasmid P620. A longer upstream region (from positions −852 to −7) is available as plasmid P619.

FIG. 3.

Ni²⁺ ions do not prevent the utilization of copper ions from the medium. CuCl₂ (+Cu), NiCl₂ (+Ni), or both together (Cu+Ni) were added as indicated to replicate copper-deficient cultures of CC125 at t = 0. Soluble protein was prepared from an aliquot of each culture at the indicated times and analyzed by native gel electrophoresis, followed by immune detection with anti-plastocyanin serum at a 1/1,000 dilution. The positions of migration of holoplastocyanin and cytochrome c₆ are indicated.

Reversible Ni-induced gene expression.

A large starter culture was divided into smaller cultures for individual treatments. All manipulations of the cultures, including division, removal of aliquots for analysis, changing of medium, or addition of salts and/or chelators were performed under sterile conditions, and any solutions were filter sterilized prior to addition. To transfer cells to fresh medium, the cells were collected by centrifugation (5 min at 2,100 × g) in sterile screw-cap centrifuge bottles, and the pellet was resuspended by gentle swirling in 10 ml of fresh medium before readjustment to the original volume. This process was repeated. The washed cells were maintained for analysis or tested for their ability to respond to Ni²⁺ by readdition of Ni²⁺ as indicated.

RESULTS

The copper deficiency target genes are activated by nickel and cobalt salts.

Each copper deficiency target is also activated by hypoxia (40, 44, 45). A classic characteristic of mammalian hypoxia-induced factor target genes is their activation by cobalt or nickel ions (14). Therefore, we tested copper-supplemented Chlamydomonas cultures for Ni- and Co-dependent activation of Cpx1 and Cyc6. Wild-type cells were treated with NiCl₂ or CoCl₂ at various concentrations for 5 h prior to extraction and analysis of total RNA. We found that both Ni²⁺ and Co²⁺ can activate the expression of Cpx1 and Cyc6 (Fig. 1A) although, in contrast to the situation for mammalian cells, Ni²⁺ is effective at a lower concentration (≥25 μM) than is Co²⁺ (a minimum concentration of 75 μM is required). The maximum induction of Cpx1 and Cyc6 by Ni²⁺ (75 μM) is ca. 2.5 times greater than the maximum induction observed after Co²⁺ addition (at 300 μM) (Fig. 1B). We conclude that Ni²⁺ is a more effective inducer than Co²⁺ in this system, and we focused the experimental work on Ni²⁺ as the inducer in subsequent experiments. Crd1, another target of Cu deficiency, is also induced by Ni²⁺ (Fig. 2). The greater effectiveness of Ni²⁺ versus Co²⁺ may reflect either a bias for uptake or the more effective action of Ni²⁺ at the intracellular target site. It is worth noting that the effective concentrations of Ni²⁺ and Co²⁺ are much higher than would be physiologically meaningful for a nutrient. We therefore assume that the compounds are interfering with a signal transduction component. Consistent with this expectation, NiCl₂-EDTA and CoCl₂-EDTA are ineffective in inducing the expression of Cyc6, Cpx1, or Crd1 at concentrations up to 250 and 500 μM, respectively (data not shown). When NiCl₂ was added to the medium at high concentrations, we noted that the cells became visibly chlorotic after 24 h of growth, which is a typical metal toxicity response. To assess toxicity, we measured chlorophyll accumulation in cultures grown for 2 days in medium containing Ni²⁺ in a range from 10 to 100 μM. For the cultures containing 10 or 25 μM NiCl₂, there was no effect on chlorophyll accumulation. The culture containing 50 μM NiCl₂ showed only a modest effect that increased at higher Ni²⁺ concentrations (data not shown). Therefore, we chose 25 μM NiCl₂ as the inducing concentration for all subsequent experiments since this concentration results in strong induction of the target genes with no obvious toxicity to the cells.

FIG. 1.

Nickel and cobalt ions can activate Cyc6 and Cpx1 expression. (A) CC125 cells were sampled 5 h after the addition of the indicated concentrations of NiCl₂ or CoCl₂, and total RNA was isolated and analyzed by blot hybridization. (B) The RNA blots were exposed to a PhosphorImager screen, and the relative band intensities were quantified and normalized to the respective RbcS2 hybridization signals. Bars: □, Cpx1; ▪, Cyc6.

FIG. 2.

Comparison of Cu²⁺, Ni²⁺, and O₂ signals. Relative accumulation of Cpx1, Cyc6, and Crd1 transcripts in strain CC125 in response to each signal: Cu²⁺ deficiency, O₂ deficiency (0% air, 98% nitrogen, 2% CO₂), or Ni²⁺ (25 μM) addition as described in the legend to Fig. 1.

In a previous study we showed that, whereas the copper deficiency targets were each induced by hypoxia, each gene showed a distinct pattern of induction with respect to hypoxia and copper deficiency, suggesting some differences in the signal transduction components for each response (45). We therefore compared the relative induction of target genes Cyc6, Cpx1, and Crd1 under copper-deficient conditions versus nickel addition versus hypoxia and noted that the response to Ni²⁺ was more similar to the copper deficiency response than it was to the hypoxic response with respect to the magnitude of induction of the target genes (Fig. 2). This suggests that the site of action of Ni²⁺ might be a component of the copper signal transduction pathway.

Nickel does not inhibit copper uptake.

One possible way in which nickel could induce copper deficiency-responsive genes is by inhibiting copper uptake in copper-replete medium, thus resulting in intracellular copper deficiency. To test this hypothesis, we monitored the copper-dependent biosynthesis of plastocyanin in the presence versus the absence of Ni²⁺ ions to assess intracellular copper availability (36). In copper-deficient cells, apoplastocyanin continues to be synthesized, but it is rapidly degraded by thermodynamic destabilization and induced proteolysis (33). This process can be blocked by the provision of copper, which allows holoprotein formation and also blocks the activity of the protease. The response is highly selective for copper ions. The accumulation of holoplastocyanin is therefore a measure of intracellular copper availability. When copper is added to copper-deficient cells, holoplastocyanin accumulates to comparable abundance even if Ni²⁺ is added simultaneously (Fig. 3, lanes 4 to 6, −Ni²⁺, compared to lanes 7 to 9, +Ni²⁺). RNA was analyzed in parallel to assess the effect of copper and nickel on the activation of target gene expression. Cyc6 and Cpx1 transcript levels dropped transiently in response to copper and then increased again in response to nickel (data not shown). Note that holocytochrome c₆ is a very stable protein. Therefore, even when transcription of the Cyc6 gene is turned off, the protein remains (see, for example, reference 36). The experiment shown in Fig. 3 was repeated with concentrations of copper as low as 400 nM, with the same result (data not shown). Nickel ions cannot support holoplastocyanin accumulation (Fig. 3, lanes 10 to 12). We conclude that Ni²⁺ does not interfere with copper uptake into Chlamydomonas cells.

The CuRE is required for Ni²⁺-responsive activation from the Cyc6 and Cpx1 promoters.

If the response to Ni²⁺ occurs via the nutritional copper signal transduction pathway, we would predict that expression is controlled at the transcriptional level and, furthermore, that the same DNA sequences are required for the response to both copper and nickel. Reporter gene constructs, Cyc6-Ars2 or Cpx1-Ars2, introduced in strain CC425 were tested for Ni²⁺-responsive expression (Fig. 4 and 5). The strains were grown in copper-sufficient medium (no expression of Cyc6 and low expression of Cpx1) and analyzed for Ars2 (reporter gene) and endogenous Cyc6 and Cpx1 gene expression after the addition of NiCl₂ to 25 μM. The full-length Cyc6 (Fig. 4, construct A) or Cpx1 (Fig. 5, construct G) promoter sequences were each capable of conferring nickel-responsive expression on the Ars2 reporter gene, indicating that Ni²⁺ acts at the level of transcription.

FIG. 4.

Analysis of Ni-responsive expression of Cyc6-Ars2 reporter gene constructs. Strains containing the indicated Cyc6 promoter sequences fused to the Ars2 reporter gene were grown to mid to late log phase (5 × 10⁶ to 10 × 10⁶) in copper-supplemented TAP medium. Nickel chloride was added to 25 μM, and samples of the cultures were harvested at the indicated times for preparation and analysis of total RNA. X, mutation of GTAC core of a CuRE; 0, mutation of GTAC that is not part of a CuRE. The series of panels labeled Ars2 were probed with the Ars2 cDNA to analyze expression of the reporter gene. Expression of the endogenous Cyc6 was probed as a positive control for nickel induction.

FIG. 5.

Analysis of Cpx1-Ars2 reporter gene constructs. An analysis of nickel-responsive expression was carried out as follows. Strains containing the indicated Cpx1 5′ upstream sequences fused to the Ars2 reporter gene were grown to 4 × 10⁶ to 9 × 10⁶ cells/ml in copper-supplemented TAP medium. NiCl₂ was added to 25 μM, and samples of the cultures harvested at the indicated times for preparation and analysis of total RNA. X, mutation of GTAC core of CuRE(s); 0, mutation of GTAC that is not part of a CuRE. Endogenous Cpx1 was probed as a positive control for the efficacy of the nickel treatment.

In previous work, we had shown that all of the copper-responsive activity associated with the Cyc6 gene was localized to a small ∼1.2 × 10²-bp fragment, while for the Cpx1 gene, an approximately 4 × 10²-bp fragment was sufficient for copper responsiveness (47, 48). These regions were tested also in the context of the reporter gene for Ni-responsive expression (Fig. 4, constructs B to E, and Fig. 5, constructs H to J). We found that the copper-responsive regulatory regions containing CuREs, corresponding to the region from −129 to −7 for the Cyc6 promoter and the region from −197 to + 207 for the Cpx1 promoter, also contained the Ni²⁺-responsive elements. The region from −129 to −7 in the Cyc6 promoter has two CuREs, each of which is independently functional for transcriptional activation during copper deficiency (44). To confirm that the CuREs were responsible for the Ni²⁺ response, we tested constructs in which the core GTAC of each CuRE was mutated, either singly (constructs B and C) or together (construct D). In construct C, the proximal CuRE is mutated (indicated by an X) in construct B, the distal CuRE is mutated (indicated by an X), and construct D carries both mutations (indicated by the two X's). The single mutations did not affect Ni²⁺-reponsive gene expression while the construct carrying both mutations did not show Ni²⁺-responsive expression, indicating that the CuREs are required for the Ni²⁺ response but, just as for the nutritional copper response, they are functionally redundant. The −129 to −7 fragment also contains another GTAC sequence, but this is not part of a CuRE: when this sequence is mutated, either alone (Fig. 4, construct E) or in combination with a CuRE (Fig. 4, construct F), there is no effect on Ni²⁺-responsive expression. For Cpx1, we had previously identified one CuRE, again with an absolutely essential GTAC core (48). When the CuRE is mutated, the constructs now also lose Ni²⁺-responsive expression (Fig. 5, compare constructs I to H). On the other hand, mutation of a GTAC in the promoter that is not part of a CuRE has no effect on Ni²⁺-responsive expression (construct J). On this basis we conclude that the Ni²⁺ response occurs through the CuREs.

Nickel-responsive expression of Cyc6 and Cpx1 requires the trans-acting factor Crr1.

Another component in the nutritional copper response pathway is the genetically defined CRR1 locus. In crr1 mutants, all copper deficiency responses are blocked, including transcriptional activation of Cyc6 and Cpx1 (M. Eriksson, J. Moseley, J. del Campo, S. Tottey, and S. Merchant, unpublished data). If the Ni²⁺ response acts through the copper deficiency pathway, we would predict that crr1 mutants do not respond to Ni²⁺, and indeed this is the case (Fig. 6). On the other hand, the crr1 mutant responds normally to sulfate or iron deficiency independently of copper availability, indicating that the phenotype is specific for the copper deficiency target genes (45).

FIG. 6.

CRR1 is required for Ni-responsive expression of Cpx1 and Cyc6. Total RNA was isolated from the mutant crr1 and a wild-type strain (CRR1 = CC425) 5 h after addition of NiCl₂ (to 25 μM) and analyzed for Cpx1 and Cyc6 expression.

Nickel can be used as a reversible inducer of gene expression.

The tight regulation of the Cyc6 promoter by Ni²⁺ (Fig. 1A) and the fact that Ni²⁺ responsiveness can be transferred to a reporter gene via short (100- to 400-bp) promoter fragments from the Cyc6 or Cpx1 genes (Fig. 4) suggests that this system could be used to design an inducible promoter in Chlamydomonas. This raised the question of whether gene expression could be turned off by removal of the inducer. Chlamydomonas cells were treated with NiCl₂ and sampled 5, 10, or 24 h later to monitor Ni²⁺ activation of Cyc6 and Cpx1 (Fig. 7, lanes 1 through 4). If EDTA is added 5 h after Ni²⁺ addition, activation is completely blocked in cells sampled either 5 or 19 h later (Fig. 7, lanes 5 and 6). EDTA alone has no effect on the target genes (Fig. 7, lanes 7 and 8). The action of Ni²⁺ can also be reversed by washing the Ni²⁺-treated cells in Ni²⁺-free medium and replacing it with standard TAP medium (Fig. 7, lanes 9 and 10). To assess whether centrifugation and washing is deleterious with respect to gene expression, Ni²⁺ was added back to the washed cells, and the activation of Cyc6 and Cpx1 expression was monitored. The expression of both genes is activated with the same kinetics as noted previously (Fig. 7, lanes 11 and 12).

FIG. 7.

Reversibility of Ni-induced gene expression. Expression of target genes in strain CC125 at 3 × 10⁶ cells/ml is shown in lane 1 (t = 0). The culture was treated with either 25 μM NiCl₂ for 5 h (lane 2) or 50 μM EDTA for 5 or 19 h (lanes 7 and 8) The reversibility of the 5 h Ni²⁺ treatment was ascertained either by addition of excess EDTA (50 μM) to chelate the Ni²⁺ ions or by transfer to fresh medium and then sampled 5 or 19 h later to assess gene expression (lanes 5 and 6 and lanes 9 and 10, respectively). Ni-treated cultures were maintained in parallel for another 5 or 19 h as a control for the EDTA treatment (lanes 3 and 4, respectively), or in the case of the washed cells, NiCl₂ was added back, and the culture sampled 5 or 19 h later (lanes 11 and 12). The total time of exposure to Ni²⁺ was, therefore, 10 or 24 h. Total RNA was isolated and analyzed by RNA blot hybridization for Cyc6 and Cpx1 expression. Cβlp expression was used as a loading control. The circular arrow indicates that the cells were transferred to fresh TAP medium as described in Materials and Methods.

Strain- and culture stage-dependent variation in the Ni²⁺ response.

To test the applicability of this system to Chlamydomonas strains that are popularly used in various laboratories, we tested three standard strains—CC125, CC425, and 2137—for Ni-dependent activation of Cyc6 and Cpx1. We noted some strain-dependent variation in the minimally effective concentration of Ni²⁺ and the time course of the response (Fig. 8A). Nevertheless, all three strains do respond. We noted also no change in the expression of a marker gene encoding a G-protein subunit (55), and this is consistent with a lack of toxicity of Ni²⁺ at these concentrations. There is also no indication of chlorosis during the course of these experiments. The response of strain CC125 was also tested for the effect of culture stage from early to late exponential and stationary phase. We noted distinct variation, with the denser culture showing a much stronger response (Fig. 8B). This is consistent with Ni²⁺ entry on a nutrient transporter that is more active in denser cultures with an increased nutrient demand.

FIG. 8.

Strain- and culture density-dependent optimization of Ni-responsive Cyc6 and Cpx1 expression. (A) Wild-type strains (CC125, CC425, and 2137) were treated with nickel concentrations between 25 and 50 μM. Samples were taken at the indicated times over a 24-h period and analyzed by RNA blot hybridization for Cyc6 and Cpx1 gene expression. Cβlp expression was monitored as a loading control. (B) The indicated concentration of NiCl₂ was added to cultures of CC125 at various cell densities. The cultures were sampled 5 h later for isolation and analysis of total RNA.

DISCUSSION

Site of nickel action.

We show that provision of nickel salts to Chlamydomonas can turn on three target genes, Cyc6, Cpx1, and Crd1, that are known to be coordinately expressed in response to copper deficiency (Fig. 1 and 2). There are two formal possibilities for the observed effect of nickel on gene expression: (i) there is a physiological connection between the function of the target genes and nickel metabolism or (ii) the action of nickel represents a nonphysiological interaction with the nutritional copper signal transduction pathway, perhaps through binding at a copper site. Precedence suggests that we consider both models. Nickel is an essential trace micronutrient for organisms that use Ni-containing enzymes such as hydrogenase, urease, or glyoxylase (13, 19) and, in some cases, it is well established that Ni²⁺ regulates the expression of genes involved in the biogenesis of these enzymes (9, 30). There are also some examples in Streptomyces spp. of nickel-containing alternative enzymes for other metalloenzymes, where nickel represses the biosynthesis of that metalloenzyme (6, 29). Nevertheless, nickel is required at trace levels only, and the changes in gene expression in these systems occur at nickel concentrations that are 2 to 3 orders of magnitude lower than those used in these experiments. The high concentrations required in this case probably reflect a threshold for entry of nickel into cells or a threshold for intracellular competition at the metal-binding site with the natural ligand. Furthermore, there is no known connection between the nickel targets noted here and nickel metabolism. Therefore, we consider the first option less likely, leaving us with the second as a working model, where nickel is proposed to interact at a metal-binding site of a component of copper homeostasis.

Comparison of the nickel and copper deficiency responses indicates that nickel action exactly parallels the copper response (Fig. 2), and this is consistent with the model. Also consistent with the model is the requirement of CuREs and Crr1 (Fig. 4 to 6). One obvious site for nickel action that would readily explain the pattern of gene expression is a copper transporter. If nickel inhibited copper uptake by competing for a copper binding site, then the cells would be internally copper deficient despite an adequate extracellular supply of nickel. We have ruled out this simple possibility by showing that copper is available within the cell to support holoplastocyanin synthesis (Fig. 3). Another candidate site for Ni²⁺ action is Crr1. Preliminary analysis of Crr1 indicates that it contains a histidine-rich domain, which is compatible with Ni²⁺ binding (J. Kropat, unpublished data). Whether the domain represents the copper-sensing site or is a DNA-binding domain is not known.

The same target genes are also turned on by hypoxia, and we considered the possibility that Ni²⁺ and/or Co²⁺ interact with the hypoxic pathway as they do in the mammalian hypoxic response (27). Moreover, many nickel enzymes are involved in anaerobic metabolism, and a role for nickel in gene regulation at low oxygen tensions could be physiologically relevant. However, comparison of the nickel response to the copper and hypoxia responses indicates that nickel action exactly parallels the copper response but not the hypoxia response (Fig. 2). Specifically, (i) Cpx1 is induced more strongly than Cyc6 under hypoxic conditions, whereas Cyc6 is more strongly activated by nickel, and (ii) the hypoxic response of Cpx1 requires a distinct HyRE in addition to CuRE (45).

A reversible, inducible gene expression system for Chlamydomonas.

The Cyc6 promoter has for some time generated interest as a candidate for the development of a regulatable switch for functional studies of transgenes in Chlamydomonas. Nevertheless, while it is easy to turn off the Cyc6 promoter by the addition of copper to the medium, turning on the promoter by copper deficiency is much more difficult because Cyc6 expression is sensitive to nanomolar levels of copper in the medium, even in the presence of copper chelators (46). CO₂- and O₂-responsive promoters are also available, but these conditions alter metabolism drastically and are, therefore, not ideal (31, 45). Metal-regulated promoters, derived in most cases from metal stress response genes, have proven highly useful in other experimental systems (see the introduction). We therefore tested whether Ni-dependent activation of Cyc6 and Cpx1 expression could be exploited for this purpose for Chlamydomonas.

An important feature of a regulatable switch is that it needs to be reversible. For nickel, this can be achieved either by addition of a chelator or by replacing the medium with Ni-free medium (Fig. 7). The persistence of the mRNA varies as a function of its half-life, so this will be a consideration in experimental design. Another important feature is utility in a range of strains. We tested three common laboratory strains, and each responds to nickel (Fig. 8A). Nevertheless, the effective concentration varies within a twofold range, which means that the most useful concentration will be have to be determined by the user. We also noted an effect of cell density on the minimally effective concentration and the time course of expression (Fig. 8B). We attribute this feature to the fact that cells in late log phase have induced nutrient uptake pathways that probably facilitate nickel entry. Perhaps the most important consideration is the potential toxicity of the inducing signal. At the concentration ranges used in these experiments, we saw no evidence of chlorosis nor any decreases in gene expression for nontarget genes (Fig. 8). There are reports in the literature of an effect of Ni²⁺ on swimming speed but the concentrations used in those studies were very much higher (4, 49).

Chlamydomonas is a powerful experimental system for the study of ciliary motility and flagellar function, chloroplast metabolism (especially photosynthesis), energy metabolism, gametogenesis, photoperception, sexual reproduction, and the mechanisms of organelle inheritance (10, 15, 17, 18, 21, 35, 43, 51-53, 59, 61). A number of molecular tools have been developed in the last decade, including tools for chloroplast and nuclear transformation and for the expression of proteins in these compartments (3, 11, 42, 60). Green fluorescent protein, and luciferase reporters, organelle proteome maps, molecular map of the nuclear genome, an extensive collection of ESTs, and a draft genome sequence are also available (12, 16, 25, 28, 38, 58, 62). A reversible, inducible promoter system adds one more tool to the repertoire of approaches in this experimental system and is especially timely given the current interest in functional genomics.