INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In plants, hypoxia is relevant in several physiological contexts, but it has been studied most extensively in the context of root biology where several situations resulting in an oxygen-depleted environment have been described. At the cellular level, the responses to oxygen deprivation involve changes in energy and oxidative metabolism, which are effected through regulatory processes at all levels, including transcriptional activation of genes, post-transcriptional processes, translational control, and enzyme activation (for review, see Bailey-Serres and Dawe, 1996; Sachs et al., 1996; Germain et al., 1997; Rivoal et al., 1997). The alcohol dehydrogenase gene has been studied as a prototypical target for the root hypoxic response. Its activation in anaerobiosis requires cis-regulatory sequences that were identified through reporter gene analysis (Olive et al., 1991; Walker et al., 1997) and transcription factors of which one has been identified recently (Hoeren et al., 1998). Experiments involving differential screening have revealed additional genes that are induced in oxygen deficiency and, not surprisingly, most encode enzymes involved in sugar metabolism, including glycolysis and fermentation. Although these metabolic pathways have been substantially unraveled, the signal transduction mechanisms resulting in differential gene expression remain an open question.

Target genes are activated with different kinetics in response to oxygen deprivation and to different extents (e.g. Peschke and Sachs, 1993; Huq and Hodges, 1999), indicating the presence of multiple sensing and response pathways that may allow adaptation to be tuned to organ function and different ranges of oxygen tension. An important example of this is the response of photosynthetic organs to oxygen deprivation, which is distinct from the root response (Okimoto et al., 1980; Ellis et al., 1999), and is not well described. In particular, chloroplast metabolism is largely unexplored, except for adaptive processes in anaerobic algal cells, including Chlamydomonas reinhardtii. Metabolic, physiological, and molecular responses within the chloroplast to oxygen-deficient or anaerobic growth conditions have long been known for C. reinhardtii (Wood, 1978; Happe et al., 1994; Melis et al., 2000). One of these is the accumulation of chloroplast cytochrome c₆ in poorly aerated medium (Wood, 1978). Cytochrome c₆ functions in copper-deficient cells as a heme-containing substitute for a copper-containing protein, plastocyanin, in the photosynthetic apparatus (Merchant and Bogorad, 1986b; Merchant, 1998). Cytochrome c₆ is induced in copper-deficiency by transcriptional activation of the Cyc6 gene through copper-response elements (CuREs). C. reinhardtii also up-regulates Cpx1 (encoding coprogen oxidase, an enzyme of tetrapyrrole biosynthesis) and Crd1 (encoding a putative di-iron enzyme) during adaptation to copper deficiency (Hill and Merchant, 1995; Moseley et al., 2000; Quinn et al., 2000a). Because each copper-responsive gene is also expressed in hypoxic cells, a mechanistic and physiological connection between copper- and oxygen-deficiency-induced gene expression was proposed (Moseley et al., 2000; Quinn et al., 2000a). The availability of well-characterized target genes and the amenability of C. reinhardtii for molecular genetic analysis of signaling pathways indicated that oxygen deficiency-activated expression of Cyc6, Cpx1, and Crd1 might serve as another important model for hypoxic responses in photosynthetic eukaryotes, particularly in the context of chloroplast metabolism.

Here, we show that the target gene Crd1 is physiologically essential for normal chloroplast biogenesis under oxygen-deficient conditions, that the core of the CuRE associated with the Cpx1 gene is necessary but not sufficient for transcriptional activation under oxygen deficiency, which requires also a hypoxia-response element (HyRE), and that Crr1, a trans-acting master regulator of the copper-deficiency response, is also required for the oxygen deficiency response. Anoxic induction of a well-known anaerobically induced gene, Hyd1 (encoding an Fe-hydrogenase that is induced under anaerobic conditions), in the crr1 mutant speaks to multiple hypoxia/anoxia-sensing mechanisms in C. reinhardtii.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Hypoxic Induction of Copper Deficiency Responsive Genes

Cyc6, Cpx1, and Crd1 are three target genes of a copper deficiency signal transduction pathway. Each is activated during adaptation to copper deficiency. We noted that these three genes are also induced in copper-replete cells under oxygen-deficient or near anoxic conditions as might be experienced in their natural environment or during normal laboratory growth conditions (Moseley et al., 2000; Quinn et al., 2000a). For instance, when heterotrophic C. reinhardtii cultures are suspended by slow basal stirring (versus vigorous agitation on a shaker at 250 rpm) under normal room lighting (approximately 10-15 µmol m² s¹), they can become oxygen depleted within a day as a consequence of respiratory activity (Fig. 1A). Under these conditions, Cyc6 mRNA accumulates (Fig. 1A; Wood, 1978). As photosynthetic activity predominates after depletion of acetate in the culture, the oxygen content is restored and the mRNA falls in response (Fig. 1A). We established that the critical variable was oxygen (rather than CO₂ or pH; Quinn et al., 2000a), and that Cpx1 and Crd1 mRNAs also increased when the oxygen content of the culture was decreased. In all experiments, oxygen content was varied by mixing air with nitrogen. The CO₂ content was kept constant at 2%.

View larger version (41K): [in this window] [in a new window]   Figure 1.   A, Cyc6 expression in copper-supplemented laboratory cultures of wild-type C. reinhardtii as a consequence of oxygen depletion. Cells were grown in Erlenmeyer flasks fitted with cotton plugs and were suspended by low basal stirring using a magnetic stir bar and stir plate under normal room lighting (approximately 1-15 µmol m2 s1 illumination). The oxygen content of the culture was measured with a standardized oxygen electrode each day (). The culture was sampled at the same time for preparation of RNA and was analyzed by hybridization (). B, Hypoxia-induced gene expression in C. reinhardtii. Wild-type strain CC125 was grown in copper-supplemented (6 µM) Tris-acetate-phosphate medium under normal aeration to a concentration of 1 × 106 cells mL1, and it was then bubbled with a gas mixture containing the indicated amount of air plus 2% CO2 with the balance as N2. Total RNA was prepared after 24 h and was analyzed by hybridization.

For the Crd1 and Cpx1 genes, even a small change in oxygen concentration, corresponding to 24-h growth at 50% air (2% CO₂ and balance N₂), results in increased transcript accumulation (Fig. 1B). As the amount of oxygen in the culture is decreased, each of the genes is induced more strongly. In contrast to Crd1 and Cpx1, Cyc6 expression is not as sensitive to oxygen depletion. Activation of Cyc6 is not detected until the concentration of oxygen is decreased to the amount in 5% air (about 1% O₂). The time course of response of Cpx1 and Crd1 versus Cyc6 is also different. The hypoxic response of the Cpx1 and Crd1 genes is noticeable within 30 min, but the hypoxic response of the Cyc6 gene is barely detectable even when the cells have been oxygen deficient for at least 1 h (Quinn et al., 2000a). In accordance with this, the Cpx1 response is saturated quickly (within 4 h of transfer to 0% air, 2% CO₂, and balance N₂), whereas the Cyc6 response is still increasing for up to 24 h (Fig. 2A). The response time course and threshold of sensitivity of the Crd1 gene to oxygen deprivation parallels Cpx1 exactly (not shown).

View larger version (50K): [in this window] [in a new window]   Figure 2.   Time course of the Cyc6 and Cpx1 responses. A, Cultures were bubbled with 98% N2/2% CO2, sampled in duplicate at the indicated times, and analyzed for RNA abundance by blot hybridization. B, In a comparable time course experiment, cultures were sampled and analyzed for protein accumulation by immunoblotting. The lanes marked Cu represent the accumulation of coprogen oxidase and cytochrome c6 in copper-deficient, fully aerated cultures.

The production of cytochrome c₆ and coprogen oxidase in copper-deficient cells follows the abundance of the Cyc6 and Cpx1 transcripts, except that the proteins are more stable than the corresponding mRNAs and hence persist even when the mRNAs are completely degraded following copper supplementation (Hill et al., 1991). The same is true in oxygen-deficient cells. Protein abundance parallels mRNA abundance with a slight lag (compare Fig. 2, A with B). For example, the increase in coprogen oxidase is not evident until 24 h after oxygen deprivation, whereas the corresponding Cpx1 mRNA increase peaks by 4 h. Upon re-aeration of cells, the proteins persist long after the mRNA has decayed to undetectable levels as a result of cessation of transcription (not shown).

Why are three targets of the copper deficiency signal transduction pathway also turned on by oxygen deficiency? A simple explanation for the complete overlap of the set of target genes is that low oxygen results in copper deficiency, which activates the copper deficiency response. For instance, oxygen deprivation could inhibit copper uptake, resulting in intracellular copper deficiency even in copper-replete medium. Therefore, intracellular copper availability in oxygen-depleted versus aerated cells was assessed. One measure of intracellular copper availability is the abundance of the copper protein plastocyanin in the thylakoid lumen. Plastocyanin does not accumulate unless copper is present in the lumen for holoprotein formation (Merchant and Bogorad, 1986a; Li and Merchant, 1995; Tottey et al., 2001). Other metals cannot substitute for copper, which makes this assay highly selective for intracellular copper (Hill et al., 1991).

Wild-type cells were depleted of holoplastocyanin simply by growth under copper-deficiency (Merchant and Bogorad, 1986a), and were maintained in air or shifted to oxygen-depleted conditions (2% air) prior to initiation of holoplastocyanin formation by copper addition (Fig. 3). Holoplastocyanin is synthesized and accumulates to similar levels within hours in aerated and oxygen-deprived cultures (Fig. 3, lanes 5, 6, 9, and 10), indicating that copper is available inside the cell for de novo holoplastocyanin formation regardless of oxygen supply. RNA was isolated from the same cultures to confirm that oxygen removal was effective in activating the target genes, and hybridization analysis confirmed that the abundance of the Cyc6 and Cpx1 transcripts displayed the expected pattern of expression. That is, the transcripts decreased transiently in response to copper supplementation and then increased again in response to activation by oxygen deficiency (data not shown). From these results, which show that copper can access intracellular compartments, we conclude that the oxygen deficiency response is not manifested merely through an effect on intracellular copper metabolism/mobilization. We conclude that the oxygen deficiency response of Cyc6, Cpx1, and Crd1 is a separate physiological response from the nutritional copper response.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Intracellular copper availability in hypoxic-treated cells. Duplicate copper-deficient cultures of wild-type cells were kept in air or transferred to 2% air (CO2 was kept constant at 2%, balance N2) for 1 h prior to addition of copper to 6 µM final concentration (t = 0). Soluble protein was prepared from an aliquot of each culture at the indicated times after addition of copper and was analyzed by native gel electrophoresis and immunoblotting using anti-plastocyanin antiserum at 1:1,000 dilution. The antiserum, generated against plastocyanin, cross-reacts with cytochrome c6 and, therefore, both proteins are detected (lanes 1 and 2).

The Oxygen Deficiency Response Is Physiologically Relevant

That an oxygen deficiency response might be biologically relevant for species of C. reinhardtii is evident upon consideration of their natural habitats, which include damp soil, bogs, and sewage lagoons. Even laboratory cultures can become oxygen depleted quite rapidly as a consequence of respiration (Fig. 1A). The availability of strains carrying loss-of-function mutations in one of the hypoxic target genes, namely Crd1, gave us the opportunity to assess whether the oxygen deficiency response under study in this project was physiologically relevant and significant for acclimation to low oxygen.

Crd1 encodes a candidate redox enzyme required for normal accumulation of photosystem I and its associated light-harvesting complexes in copper-deficient cells. Therefore, copper-deficient crd1 strains are chlorotic (Moseley et al., 2000). If expression of Crd1 in hypoxic cells represents a requirement for its gene product in hypoxic cells, crd1 mutants should display chlorosis under hypoxic growth conditions. When strain crd1-1 is grown in 2% air, it becomes chlorotic, accumulating approximately 5-fold less chlorophyll per cell compared with wild-type hypoxic cells or to crd1-1 or wild-type cells grown with normal aeration (Fig. 4). We conclude that Crd1 function is critical in oxygen-deficient cells for normal development of the photosynthetic apparatus. Hence, the activation of Crd1 expression under oxygen-deficient growth conditions is biologically meaningful.

View larger version (26K): [in this window] [in a new window]   Figure 4.   Effect of oxygen-deficient growth on a crd1 mutant strain. Strain crd1 and wild-type (WT) cells were grown under normal aeration to 1 × 106 cells mL1. Each culture was then bubbled with a mixture of 2% air (constant 2% CO2, balance N2) or with 100% air, and was grown continuously under these conditions for about two generations.

Oxygen-Responsive Expression Requires CuREs and a HyRE

To test whether the two signals, i.e. copper and oxygen deficiency, might feed into the same pathway, we compared the expression of each target gene in copper or oxygen deficiency versus copper and oxygen deficiency (Fig. 5). The Cyc6 gene is induced maximally by copper deficiency, and oxygen deprivation does not further activate it (Fig. 5, compare lanes 3 and 4). Likewise, Cpx1 is maximally induced in oxygen-deficient cells, and copper deficiency does not further activate it (Fig. 5, compare lanes 2 and 4). This is consistent with a model in which the oxygen and copper deficiency responses share signal transduction components.

View larger version (32K): [in this window] [in a new window]   Figure 5.   Effect of copper and oxygen deficiency on Cpx1 and Cyc6 expression. Total RNA isolated from cultures grown under the indicated conditions was analyzed. O2, 1% air (constant 2% CO2, balance N2).

In previous work, we have shown that the copper deficiency response requires CuREs associated with the target genes (Quinn et al., 2000a). A functionally essential GTAC sequence forms the core of the CuREs. One CuRE was identified in the Cpx1 promoter. Constructs a and b (Fig. 6) consist of the indicated portions of Cpx1 fused to a reporter gene (Ars2 encoding arylsulfatase). These constructs are expressed in Cu-deficient cells coordinately with the endogenous Cpx1 gene. The same is true when these constructs are tested for oxygen deficiency expression, which indicates that all the information required for induction of Cpx1 in oxygen deficiency is within the region from 197 to +207. When a GTAC sequence corresponding to the core of the previously identified CuRE is mutated (Fig. 6, compare constructs c and b), the oxygen deficiency response of the reporter gene is lost (Quinn et al., 2000a). Construct c does not induce Ars2 at any concentration of air. Construct d, in which a downstream GTAC sequence that is not part of a CuRE is mutated, also fails to respond to oxygen deficiency. This suggests that induction of Cpx1 in oxygen deficiency requires a HyRE beyond the CuRE. The sequence GTAC is also a critical component of the HyRE. The GTAC cores of both elements must be intact to confer oxygen deficiency responsive expression to a reporter gene, whereas only the GTAC core of the CuRE is required for the nutritional copper response (Quinn et al., 2000a).

View larger version (25K): [in this window] [in a new window]   Figure 6.   Oxygen deficiency expression of Cpx1-Ars2 reporter gene constructs. Strains containing the indicated Cpx1 5'-upstream sequences fused to the Ars2 reporter gene were grown to 2 × 106 cells mL1 and were bubbled with the indicated concentrations of air for 24 h prior to preparation and analysis of total RNA. CO2 was kept constant at 2%. Endogenous Cpx1 was probed as a positive control for the efficacy of the oxygen deprivation, and RbcS was probed as a control for loading (not shown). Copper-responsive expression of the same constructs was analyzed by Quinn et al. (2000a). Multiple, independent transformants were generated for each construct. The data shown are from a single representative transformant.

The CRR1 Locus, Which Encodes a trans-Acting Component Required for Adaptation to Copper Deficiency, Is Also Required for Oxygen Deficiency Responsive Expression

Genetic analysis of the copper deficiency response led to the identification of the CRR1 locus, which is required for adaptation to copper deficiency (M Eriksson and S Merchant, unpublished data). A crr1 mutant grows at a reduced rate in copper-deficient medium and cannot induce any of the target genes under Cu-deficient conditions (Crd1, Cyc6, or Cpx1 in Fig. 7A). We tested whether the CRR1 locus might also represent a shared signal transduction component in the copper- versus oxygen-sensing pathways. The crr1 strain was blocked in the oxygen-deficiency response for all target genes tested (Fig. 7B). To rule out the possibility that CRR1 is a "general" factor required for multiple nutritional stresses, we analyzed the expression of target genes of other nutrient-responsive pathways. The crr1 mutant was capable of responding to other nutritional deficiencies: Under sulfate deficiency, Ars2, encoding a sulfatase, is induced (de Hostos et al., 1989; Fig. 7C), and under iron deficiency, Ftr1, encoding a ferric transporter, is induced (Quinn et al., 2000b; Fig. 7D). The sulfur and iron deficiency responses occur even in copper-deficient crr1 cells that are growth compromised, and the extent of activation is comparable with that of sulfate- or iron-deficient wild-type cells (not shown). Thus, the specific absence of a response to oxygen deficiency in the crr1 strain indicates that there is a mechanistic connection between the copper and oxygen deficiency responses. We conclude that the oxygen deficiency response of the Cyc6, Cpx1, and Crd1 genes requires a cis- and trans-component of the nutritional copper signaling pathway, i.e. a CuRE and Crr1. This oxygen deficiency response is distinct from mechanisms operating to control hydrogenase production in anaerobic C. reinhardtii cells because crr1 can activate Hyd1 expression under anaerobic conditions (Fig. 7E).

View larger version (39K): [in this window] [in a new window]   Figure 7.   The crr1 mutant is blocked in the oxygen deficiency response. Total RNA isolated from the crr1 mutant and a wild-type strain (CRR1) grown under the indicated copper- (A) or oxygen-deficient (B) conditions was analyzed for Crd1, Cpx1, and Cyc6 expression. Total RNA from copper-deficient crr1 was analyzed for Ars2 expression in response to sulfur deficiency (C) or Ftr1 (encoding a ferric transporter) in response to iron deficiency (D). Comparable expression of these genes in wild-type cells is not shown. Total RNA was analyzed for Hyd1 expression (E) in wild-type versus crr1 cells. A time course of response to oxygen deprivation (achieved by transfer to 98% N2/2% CO2) is shown with Cyc6 expression as an internal reference.

An Antagonist of the Copper Deficiency Response Blocks the Oxygen Deficiency Response

Genes that are activated in copper-deficient conditions are deactivated upon provision of mercuric salts at subtoxic concentrations (Hill et al., 1991; Quinn et al., 2000a). If oxygen deficiency signaling shares factors involved in copper deficiency signaling, Hg^II salts might deactivate the oxygen deficiency response just as they do the copper deficiency response. When HgCl₂ is added to C. reinhardtii cells simultaneously with the shift to oxygen-deprived conditions, activation of the Cyc6 and Cpx1 genes is blocked (Fig. 8A, compare b and a). If the target genes are already induced by oxygen deficiency, the response is turned off when HgCl₂ is added (Fig. 8A, compare 2 and 4 h 1% air in c and a).

View larger version (55K): [in this window] [in a new window]   Figure 8.   Mercuric ions specifically block the oxygen deficiency response of Cpx1 and Cyc6. A, Wild-type cells were grown to 2 × 106 cells mL1 and were divided into three subcultures. Each subculture was bubbled with 1% air (2% CO2, balance N2) for the indicated times and HgCl2 was added to 10 µM final concentration at the indicated times. A, No added HgCl2; B, HgCl2 added at t = 0 h; C, HgCl2 added 2 h after initiation of hypoxic treatment. B, Response of Tub2 (encoding -tubulin). A culture of CC125 was grown to 2 × 106 cells mL1 and was divided into six subcultures. Each subculture was bubbled with 1% air for the indicated times prior to sampling for RNA preparation. HgCl2 was added to the final indicated concentrations after 2 h of treatment with 1% air. C, Response of Hyd1 (encoding Fe-hydrogenase). Mercuric chloride was added to a final concentration of 1 µM 2 h before or immediately prior to transfer to 0% air (2% CO2/98% N2; A and B). As a positive control, no addition was made before or after transfer to 0% air (2% CO2/98% N2; C).

The deactivation of the oxygen deficiency response by mercuric ions cannot be attributed to toxicity because the accumulation of other transcripts such as Pcy1 (encoding plastocyanin) and RbcS2 is not affected under these conditions (Hill et al., 1991; Quinn et al., 2000a). Nevertheless, we tested the effect of mercuric chloride on short-lived transcripts whose accumulation might be more sensitive to cytotoxic agents. HgCl₂, at concentrations from 1 to 10 µM, had no effect on the abundance of Tub2 transcripts [which has a t_1/2 similar to that of Cpx1 (Baker et al., 1984)]. At these concentrations, HgCl₂ effectively inhibited hypoxic induction of Cpx1 (Fig. 8B). Also, expression of the Hyd1 gene, which is induced in oxygen-depleted cells by a different pathway (see above), is unaffected by HgCl₂ addition (Fig. 8C), verifying the selectivity of the mercuric response.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Function of the Oxygen Deficiency Response in C. reinhardtii

C. reinhardtii, like other organisms, responds to changes in oxygen supply with alteration of metabolism (Harris, 1989). Some species in this genus are found in naturally oxygen-deficient habitats like peat bogs and sewage lagoons (Harris, 1989) where a hypoxic response is probably critical for survival. Three oxygen deficiency-induced genes are described in this work: Cyc6, Cpx1, and Crd1. Increased Cpx1 expression in oxygen-deficient conditions has already been observed in other organisms. In Saccharomyces cerevisiae, coproporphyrinogen oxidase, an oxygen-dependent enzyme, becomes rate limiting for heme synthesis in oxygen-depleted growth conditions, and the organism responds by synthesis of more enzyme (Zagorec et al., 1988). Many bacteria have two versions of this enzyme, an aerobic form that uses oxygen as a substrate and an anaerobic form that is oxygen independent (Keithly and Nadler, 1983; Xu et al., 1992). Therefore, the dramatic regulation of Cpx1 in C. reinhardtii is well precedented. Increased expression of Crd1 might be explained on the same basis. Crd1 is proposed to catalyze an oxygen-dependent oxidation in an Fe deficiency-sensitive pathway for cofactor biosynthesis in the plastid (Moseley et al., 2000; Pinta et al., 2002). Perhaps Crd1 becomes rate limiting for this pathway in oxygen-deficient cells. Crd1 function is known to be required in oxygen-deficient cells because the crd1 mutant exhibits iron deficiency-type chlorosis when it is grown at low oxygen conditions (Fig. 4). The modulation of cofactor biosynthetic pathways through changes in Cpx1 and Crd1 expression might also serve to signal oxygen status to the nucleus. This is precedented by the heme-dependent O₂-sensing system in S. cerevisiae (Zhang and Hach, 1999) and by the Mg-protoporphyrin IX-dependent plastid-to-nucleus signaling system in C. reinhardtii and plants (Kropat et al., 1997; Mochizuki et al., 2001).

At present, an explanation for why Cyc6 expression is activated in oxygen-deficient cells eludes us. The gene product, cytochrome c₆, which accumulates in oxygen-depleted cells following the increase in mRNA (Fig. 2), is an electron transfer catalyst whose function is apparent and necessary only when plastocyanin is absent (Wood, 1978; Merchant and Bogorad, 1987a, 1987b). Because anaerobic cells accumulate holoplastocyanin, cytochrome c₆ is not necessary for its replacement (Fig. 3). One possibility is that in anaerobic cells, cytochrome c₆ is used for a metabolic pathway other than photosynthesis. Another possibility is that Cyc6 expression in anoxic cells is simply fortuitous, being a consequence of the gene containing two CuREs that have the same core as the candidate HyRE (see below).

Mechanism of the Oxygen Deficiency Response

The commonality of the copper deficiency and oxygen deficiency responses might be trivially explained if oxygen-deficient cells are internally copper deficient despite plentiful supply in the medium. Nevertheless, this is not the case because oxygen-deprived cells are capable of de novo synthesis of the copper protein holoplastocyanin, indicating that copper is available intracellularly (Fig. 3). Also, even 100-fold excess copper provided in the medium cannot repress the oxygen deficiency response (data not shown). Furthermore, comparison of the oxygen and copper deficiency responses of each target gene indicates that the two responses are not identical (Fig. 5). If the oxygen deficiency response worked by creating internal copper deficiency, we would expect that the oxygen deficiency response would recapitulate the copper deficiency response. In this context, we note the distinct pattern of the oxygen deficiency response of Cyc6 versus Cpx1 and Crd1 (Fig. 1B). The Cpx1 and Crd1 genes are much more sensitive to hypoxia than is the Cyc6 gene, whereas the opposite is true of the copper deficiency response (Moseley et al., 2000; Quinn et al., 2000a) The oxygen deficiency responses of Cyc6, Cpx1, and Crd1 are a separate physiological response, but with a mechanistic and perhaps evolutionary connection to the copper deficiency response.

Because the same three genes are regulated by both copper and oxygen deficiency, we tested whether components of the nutritional copper-sensing pathway are involved in oxygen deficiency signaling. One component is the CuRE, which was defined previously by mutational analysis (Quinn et al., 2000a). The core of the CuRE is the sequence GTAC, which is absolutely essential for CuRE activity. This sequence is also essential for the oxygen deficiency response (Fig. 6), indicating that the two signaling pathways have some common components. Yet, the oxygen deficiency signaling pathway is not exactly the same, and this is demonstrated by the observation that although the copper deficiency response of Cpx1 is dependent on a single CuRE (defined by the upstream GTAC core), a second element is required for the hypoxic response of Cpx1. Mutation of the core GTAC of the second element abolishes hypoxic expression of the Cpx1-Ars2 reporter gene (Fig. 6), but does not affect its response to copper deficiency (Quinn et al., 2000a). Perhaps the much greater degree of activation of Cpx1 in oxygen deficiency relative to its activation in copper deficiency can be explained by the presence of more than one cis-regulatory element for the former response. We refer to the second element as a HyRE. The two elements are regulated by virtue of the common GTAC core, but are distinct because they are not functionally identical. The distinction resides probably in flanking nucleotides that are part of the element beyond the GTAC core.

Another regulatory component of the copper deficiency response is Crr1, defined by a mutation at the CRR1 locus (M. Eriksson and S. Merchant, unpublished data). The crr1 strains are not able to activate any of the copper deficiency responses and hence are modeled to be defective in a "master" regulator of the nutritional copper response. In this work, we show that a functional CRR1 locus is required for the oxygen deficiency response, but not for other nutrition starvation responses (Fig. 7), indicating that Crr1 is a specific rather than a general regulatory factor.

A heme- or iron-based sensor is the favored model for direct oxygen or indirect redox sensing in various hypoxia-responsive pathways in eukaryotes (for review, see Semenza, 1999; Wenger, 2000; Zhu and Bunn, 2001). There is considerable precedence for the use of iron in biological reactions where oxygen is a substrate and iron is also a common cofactor in enzymes that catalyze redox reactions, which influences our thinking about mechanisms for oxygen and redox sensing. Yet bacterial sensors are designed with a variety of redox-sensitive cofactors besides iron (for review, see Bauer et al., 1999), so other non-iron based redox or oxygen sensors are likely to occur in eukaryotic cells as well. The parallels between copper and iron biochemistry (Kaim and Schwederski, 1984) raised the possibility of a mechanistic connection between the copper deficiency and the oxygen deficiency responses.

Hyd1 Expression Is Independent of Copper and of CRR1

The genetic handle on a putative hypoxia-signaling pathway in C. reinhardtii presented us with the opportunity to test whether there might be a second, Crr1-independent mechanism. Hydrogenase production in anaerobic C. reinhardtii cells occurs, in part, by increased accumulation of Hyd1 mRNA (J.M. Quinn, unpublished data). The anaerobic response of the Hyd1 gene is the same in a crr1 mutant compared with wild-type cells (Fig. 7E) Also, Hg²⁺, which blocks the hypoxic response of Cyc6, Cpx1, and Crd1, has no effect on the hypoxic accumulation of Hyd1 transcripts (Fig. 8C, compare b and c). This indicates that a second signaling pathway activates Hyd1 through a different (copper-independent) signal transduction mechanism. The existence of at least one other oxygen deficiency signaling pathway may allow the organism to perceive and respond to a broad range of oxygen supply (for discussion, see Poyton, 1999).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Chlamydomonas reinhardtii Strains and Culture Conditions

C. reinhardtii wild-type strain CC125, mutant strains crr1-1, crd1, and transformants of strain CC425 were typically cultured under 100 to 125 µmol m² s¹ illumination in copper-supplemented or copper-deficient Tris-acetate-phosphate media (Quinn and Merchant, 1998). For growth under specified oxygen concentrations, cultures were bubbled with mixtures of indicated concentrations of air plus 2% CO₂ with the balance as N₂ (Quinn et al., 2000a). For experiments involving assessment of gene expression (e.g. to assay reporter gene constructs), cultures were bubbled with 98% N₂/2% CO₂ to maximize differences, but when the experimental design required cell growth and division (e.g. to monitor de novo protein synthesis), the cultures were bubbled with 2% air/2% CO₂ in N₂. Where indicated, cultures were supplemented with HgCl₂ or AgCl from stock solutions. All experiments were performed at least twice, and were often repeated multiple times.

Amplification of Ftr1 and Hyd1 cDNAs

Expressed sequence tags corresponding to C. reinhardtii hydrogenase (Hyd1) were identified by BLAST search using the N terminus of the protein as the input sequence. An expressed sequence tag corresponding to a putative ferric transporter (Ftr1) was identified by a BLAST search with the Saccharomyces cerevisiae Ftr1p sequence as input. Primers for amplification of these sequences were designed with BamHI sites to facilitate cloning of the reverse transcriptase-PCR products (Ftr1, 348 bp, and Hyd1, 372 bp) into the BamHI site of pBluescriptIIKS⁺ (Stratagene, La Jolla, CA).

Nucleic Acid Analysis

Total RNA was prepared and analyzed by hybridization as described by Quinn and Merchant (1995) for Ars2 or as described by Hill et al. (1991) for all other transcripts. Five micrograms of total RNA was loaded per lane. Probes for Cpx1, Cyc6, Ars2, and RbcS2 (encoding the small subunit of Rubisco) were prepared as described (Quinn et al., 1999). The RbcS2 hybridization signal is used for normalization between samples (not shown in every figure). For detection of Crd1 transcripts, the 55 × 10²-bp insert from pCrd1-5 (Moseley et al., 2000) was used. The 358-bp insert of pFtr358 and the 372-bp insert of pHyd372 (described above) were used for detection of Ftr1 and Hyd1 transcripts, respectively. The 1,227-bp insert from pGPX9 (Leisinger et al., 1999) was used for detection of Gpxh RNA. Specific activities of probes ranged from 3 to 6 × 10⁸ cpm µg¹ DNA. Blots were exposed at 80°C to film (XRP-1; Eastman-Kodak, Rochester, NY) with two intensifying screens, and were typically developed after overnight exposure.

Immunoblot Analysis

Total soluble protein was prepared (Li et al., 1996) and separated on 12% (w/v) SDS-polyacrylamide gels (Li and Merchant, 1992) or 15% (w/v) anionic gels for immunoblot analysis (Hill et al., 1991; Merchant et al., 1991). 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 IgG (Southern Biotechnology Associates, Birmingham, AL) as the secondary antibody. Bound antibody was detected using the alkaline phosphatase color reaction (Sambrook et al., 1989).