Genomic analyses reveal a conserved glutathione homeostasis pathway in the invertebrate chordate Ciona intestinalis

Abstract

The major thiol redox buffer glutathione (l-γ-glutamyl-l-cysteinylglycine, GSH) is central to cell fate determination, and thus, associated metabolic and regulatory pathways are exquisitely sensitive to a wide range of environmental cues. An imbalance of cellular redox homeostasis has emerged as a pathologic hallmark of a diverse range of human gene-environment disorders. Despite the central importance of GSH in cellular homeostasis, underlying genetic regulatory pathways remain poorly defined. This report describes the annotation and expression analysis of genes contributing to GSH homeostasis in the invertebrate chordate Ciona intestinalis. A core pathway comprising 19 genes contributing to the biosynthesis of GSH and its use as both a redox buffer and a conjugate in phase II detoxification as well as known transcriptional regulators were analyzed. These genes exhibit a high level of sequence conservation with corresponding human, rat, and mouse homologs and were expressed constitutively in tissues of adult animals. The GSH biosynthetic genes Gclc and Gclm were also responsive to the prototypical antioxidant tert-butylhydroquinone. The present evidence of a conserved GSH homeostasis pathway in C. intestinalis together with its phylogenetic position as a basal chordate and lifestyle as a filter feeder constantly exposed to natural marine toxins introduces this species as an important animal model for defining molecular mechanisms that potentially underlie genetic susceptibility to environmentally associated stress.

the thiol-containing tripeptide glutathione (l-γ-glutamyl-l-cysteinyl-glycine; GSH) provides the most abundant redox buffer in eukaryotic cells and contributes to a key pathway of phase II detoxification (49). Glutathione biosynthesis is catalyzed by two ATP-dependent enzymes, γ-glutamyl-cysteine ligase (GCL; EC 6.3.2.2) and glutathione synthase (GSS; EC 6.3.2.3) (30, 49). The rate-limiting GCL is a heterodimeric enzyme composed of a catalytic subunit (GCLC; 73 kDa) and a modifier subunit (GCLM; 31 kDa). It catalyzes the reaction of the carboxyl group of glutamate with the amino group of cysteine to form the dipeptide γ-glutamylcysteine. The enzyme GSH synthetase (GSS; EC 6.3.2.3) then catalyses the formation of GSH from γ-glutamylcysteine and l-glycine (30, 37, 54). The cysteine thiol in GSH acts as a nucleophile in reactions with both exogenous and endogenous electrophilic species. Thus, GSH can be used to reduce reactive oxygen species via spontaneous and catalytic reactions (23, 49). As a major component of phase II detoxification, the nucleophilic addition of the GSH thiol group to electrophilic centers of various endobiotic and xenobiotic substances by an extensive family of glutathione transferases renders them more water soluble and thereby facilitates their excretion (23). Given the importance of GSH for the maintenance of the intracellular redox environment and the role of this ubiquitous thiol-containing tripeptide in detoxification, it is not surprising that perturbations of GSH homeostasis have been linked with numerous gene-environment disorders including diabetes, cancer, and neurodegenerative and inflammatory disorders (23). Despite the central importance of glutathione in the pathophysiology of human diseases, many aspects of its biosynthesis, metabolism, and responsiveness to environmental cues remain poorly understood.

The present study used a comparative genomic approach with the evolutionarily divergent model Ciona intestinalis for the study of GSH metabolism and its regulation. C. intestinalis is an invertebrate member of the chordate clade Urochordata (tunicates), which diverged from the last common ancestor of all chordates at least 500 million years ago (14, 15). Their critical evolutionary position as the closest living relative of vertebrates and the simplicity of their embryogenesis have attracted developmental and evolutionary biologists since the turn of the last century. Because of this, the genomes of C. intestinalis (15, 59) as well as its close relative C. savignyi (62) have been sequenced.

C. intestinalis lives attached to submerged substrates in the subtidal zone of marine waters. Here they are constantly filtering sea water at 20–30 ml/min (28, 58), which contains high concentrations of complex polyphenols, halogenated aromatics, methylated sulfides, and some heavy metals (33). Structurally, the marine toxins that C. intestinalis continuously encounters fall into chemical classes that are well characterized for their induction of many detoxification genes across the vertebrate chordates. However, little is known about the mechanisms by which tunicates detoxify or otherwise tolerate these natural toxins in the marine environment.

The evolutionary distance separating C. intestinalis and humans provides another distinct advantage of this animal model. Specifically, this phylogenetic relationship has provided more time for mutation and random evolutionary events to remove neutral sequences, thereby improving prospects for distinguishing functional sequences. Although methodologically challenging, the value of including divergent evolutionary constraints in the process of discovering cis-regulatory elements and transcriptional networks (73), as well as human disease genes (17, 45), is becoming increasingly documented. The accurate identification of genes and their boundaries is an important first step of executing a comparative genomics strategy for the study of metabolic pathways and associated regulatory modules. Through the combined implementation of manual annotation, phylogenetic methods, and expression analyses, the present data reveal that core genes contributing to GSH biosynthesis and metabolism are conserved in C. intestinalis, supporting the use of this invertebrate model for in vivo toxicogenomic studies.

MATERIALS AND METHODS

Manual annotation of a GSH homeostasis pathway.

Candidate sequences were identified initially by searching reference sequences [human, rat, and mouse nucleotide and protein sequences retrieved from the highly curated National Center for Biotechnology Information (NCBI) database] against the C. intestinalis genome (http://www.ensembl.org/Ciona_intestinalis/index.html) via the basic local alignment search tool (BLAST) similarity search of Ensembl BLASTView (12). Sequence alignments with the highest score, lowest expectation value (E-value), and their percentages of identity were archived. Reference and genomic sequences as well as predicted transcripts and amino acid sequences were retrieved from the Ensembl project website (12) for additional analysis of gene homology in C. intestinalis. A systematic approach was followed to infer gene homology. Orthology assignment between genes was derived from hit-clustering methods (70) using the results (hits) from sequence similarity searches among highly annotated proteins from human, rat and mouse. First, the widely used reciprocal (bidirectional) best hit (RBH) (22, 68) method was performed. BLASTP (1) analysis was used to back-search the full retrieved amino acid sequence against RefSeq (NCBI Reference Sequence Project) protein sequences. Second, the Pfam algorithm (21) was used for the identification of conserved domains within the predicted C. intestinalis sequences. These results were confirmed by reverse position-specific (RPS) BLAST analysis performed with the amino acid sequences against the NCBI Conserved Domain Search (47) to detect structural and functional domains in protein sequences. Third, the OrthoMCL algorithm was used to construct orthologous groups across multiple eukaryotic taxa (8). Finally, the highly annotated and curated SWISS-PROT and TrEMBL protein databases (72) were used to assess statistical significance of pairwise sequence similarity for each candidate gene using probability of random shuffle (PRSS) and SSEARCH analyses (56, 57). With this systematic process, the accuracy of the annotations was validated and consistency with previously published data was tracked. All candidate genes contributing to GSH homeostasis (Table 1) were annotated from the C. intestinalis genome following this systematic approach.

Table 1.

Summary of genes contributing to GSH homeostasis pathway in Ciona intestinalis

Gene Symbol	Chromosomal location	Name and Function	Transcriptional Regulators*
Gclc	Chr 2q_549	glutamate-cysteine ligase catalytic subunit; ligates	AP-1 (7, 14, 22, 23), NFKB (22), NRF1 (22), NRF2 (19, 22), RXR (21)
Gclm	Chr 13q_104	glutamate-cysteine ligase modifier subunit; heterodimerizes with GCLC	AP-1 (23), NRF2 (11, 17, 19)
Gss	Chr 3q_254	glutathione synthetase; catalyzes 2nd step in GSH synthesis	AP-1 (23), NRF2 (9, 17)
Ggt1	Chr 3q_234	γ-glutamyltransferase; catalyzes 1st step in GSH degradation	NRF2 (17)
Gpx1	Chr 12p_118	glutathione peroxidase; catalyzes GSH-dependent reduction of H2O2	NRF2 (8, 19), RXRa (21)
Gsr	Chr 7q_437	glutathione reductase; reduces GSSG to GSH	NRF2 (8, 17, 19)
Gsta1†	Scaffold 114_38	glutathione transferase activity; catalyzes the reaction R-X + GSH = H-X + R-S-SG. R may be an aliphatic, aromatic or heterocyclic group; X may be a sulfate, nitrile, or halide group.	AP-1 (13), CREB-1 (10), GATA-1 (16), LXR (18), MAFG (3), MAFK (3, 4), NRF2 (1, 4, 8, 17), PXR (6), RXRa (2, 12, 21)
Gstm1	Chr 1q_68
Gstp1	Scaffold 1275_1
Glrx	Chr 14p_149	glutaredoxin; catalyzes GSH-dependent disulfide oxidoreduction reactions in a coupled system with NADPH, GSH, and GSR	(unknown)
Slc7a11 (xCT)	Chr 1q_31	with Slc3a2 (4F2hc) forms the heterodimeric amino acid transport system Xc(-), which mediates cystine-glutamate exchange and regulates intracellular glutathione concentrations	ATF4 (15), NRF2 (17)
Abcc2 (Mrp2)	Chr 1q_334	ATP-binding cassette, subfamily C (CFTR/MRP), member 2; export pump involved in excretion of endogenous and exogenous organic anions, including bilirubin, glutathione, drugs, and toxic chemicals	CAR (5), FXR (5), NRF2 (17, 20), PXR (5)

^*Among those experimentally validated for human, rat, and mouse. References listed in supplemental text file (ST1).

^†Three well-annotated Gst genes are listed from 24 that have been identified in the C. intestinalis genome (not shown).

Experimental animals.

Adult C. intestinalis specimens were obtained from Marine Biology Laboratory (Woods Hole, MA) and maintained in circulating sea water tanks (average sea water temperature 15°C) at Mount Desert Island Biological Laboratory (MDIBL; http://www.mdibl.org/). Animals were adjusted to the environmental conditions at MDIBL for at least 6 days before experiments were performed.

In silico cloning of GSH biosynthesis and metabolism genes in C. intestinalis.

Each of the chosen C. intestinalis genes contributing to GSH homeostasis was validated via in silico cloning analysis [prediction of a gene product sequence using only genomic and expressed sequence tags (ESTs) data] (5, 60). Transcript sequences of the genes annotated with the manual approach described above were retrieved from the Ensembl project website (12) and were used as queries for BLASTN analysis (75) against the nonmouse and nonhuman EST database in GenBank [NCBI; (4)].

Expression of GSH biosynthesis and metabolism genes in adult C. intestinalis.

Total RNA was extracted from the branchial sac, gastrointestinal tract, and heart dissected from replicate adult animals using the RNeasy kit (Qiagen, Valencia, CA) following the manufacturer's instructions. Primers were designed for the following GSH-related genes and known transcriptional regulators from nucleotide sequences annotated manually as described above: Gclc, Gclm, Gss, Ggt1, Gpx1, Gsr, Gstm1, Glrx, Slc7a11 (xCT), Abcc2 (Mrp2), Nfe2l2 (Nrf2), Keap1, Maff, Bach1, Nr1h3 (Lxra), Nr1h4 (Fxr), and Rxra (Supplemental Table S1).¹ A one-step RT-PCR kit (Qiagen) was used to examine mRNA expression. Reactions were performed in a total volume of 25 μl with ∼40 ng/l of total RNA. First, cDNA synthesis and predenaturation were performed in single cycles at 50°C for 30 min and 95°C for 15 min, respectively. Next, PCR amplification was performed for 35 cycles: 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s with a final extension cycle at 72°C for 6 min. The housekeeping gene β-actin (Actb) was used as a constitutively expressed control and RT-PCR amplifications targeting Actb in which reverse transcriptase was omitted did not generate amplicons, demonstrating that the RNA samples were not contaminated with genomic DNA (data not shown). In addition, two negative controls, no template and no reverse transcriptase reactions, were included in each RT-PCR amplification to avoid false positive results. PCR products were visualized by agarose gel electrophoresis and ethidium bromide staining.

Response of GSH biosynthetic genes to oxidative stress in C. intestinalis.

Both semiquantitative end-point and RT-qPCR were used to examine the extent to which core genes of the GSH biosynthesis pathway in C. intestinalis are responsive to the prototypical antioxidant tert-butylhydroquinone (tBHQ). Three independent experiments were performed. First, a total of five adult animals per group were exposed to tBHQ (100 μM) or vehicle (DMSO) for 8 h in sea water tanks at MDIBL. Heart rate and response to contact were measured after exposure to the experimental treatments and immediately before dissection of the tissues as physiological vital signs. Total RNA was extracted from the heart of animals exposed to control and tBHQ treatments using the RNeasy kit (Qiagen). Semiquantitative end-point RT-PCR was performed in a total volume of 25 μl with ∼37 ng/l of total RNA, and Actb was used as a constitutively expressed control. The semiquantitative RT-PCR reactions were performed as described above with the exception that the number of PCR cycles was gradually decreased. Twenty-seven cycles were required to reveal differences in RNA concentration between treatments.

Second, because the semiquantitative RT-PCR analysis revealed that tBHQ increased Gclc and Gclm expression in the heart of C. intestinalis, and it is well established that treatment with tBHQ increases steady-state RNA expression of these glutathione biosynthesis genes in mouse (63), rat (7, 43, 44), and human cells (26, 42), two additional independent experiments were performed to corroborate the previous findings. For these two experiments, a total of seven adult animals per group (n = 4 and n = 3) were exposed to tBHQ (100 μM) or vehicle (DMSO) for 12 h in sea water tanks. Total RNA was extracted from the branchial sac, gastrointestinal tract, and heart as described above, and RT-qPCR (OneStep RT-PCR Kit, Qiagen) analysis was performed with the protocol described above with the following modifications: PCR volumes were adjusted to a total of 10 μl. PCR denaturation cycles were performed at 95°C for 15 min, and a slow temperature-ramping dissociation stage (95°C for 15 s, 60°C for 15 s, and 95°C for 15 s) was added after the final PCR extension cycle to monitor the specificity of amplification. Absolute quantification was performed using a pool of RNA from different tissues for the standard curve. Target gene RT-qPCR signals were normalized with the Actb housekeeping gene signal to standardize the quantity of starting RNA. Analyses were run in triplicate. For estimation of fold-change, RT-qPCR signals from the control group were set as 1.

Statistical analysis of gene expression.

Differences in fold change of gene expression between control and treatment groups were analyzed for statistical significance using the Mann-Whitney U-test (Statview, version 5.0.1; SAS Institute, Cary, NC). P < 0.05 was accepted as statistically significant. Results of RT-qPCR are shown in box plots, which include their five-number summaries (the smallest observation, lower quartile, median, upper quartile, and largest observation).

Phylogenetic analysis of the rate-limiting enzyme of the GSH biosynthetic pathway.

To examine the homology of the C. intestinalis GCL subunits to other highly annotated GCLC and GCLM proteins, sequences were retrieved from Ensembl for human and a minimum number of National Institutes of Health model organisms (rat, mouse, zebrafish, Caenorhabditis elegans; www.nih.gov/science/models/index.html). It was confirmed that the GCLC and GCLM proteins from these flanking species correspond to consensus sequences archived in SWISS-PROT, TrEMBL, NCBI, and model organism databases. These sequences were aligned using M-Coffee software, a meta-method for assembling multiple sequence alignments (69). Jalview software analysis (9) was used to calculate percentage of identity, degree of conservation, and consensus among all selected sequences. To confirm homology among GCLC and GCLM sequences as revealed by RBH and alignment methods, phylogenetic trees were constructed with the maximum parsimony and maximum likelihood methods (18) using the MEGA software version 4.0 (65). Trees were rooted using C. elegans homologs as out-group protein sequences. The statistical significance of branch order was estimated by the generation of 1,000 replications of bootstrap re-sampling of the originally-aligned amino acid sequences.

RESULTS

Manual annotation of a GSH homeostasis pathway.

Twelve genes that contribute to GSH homeostasis in C. intestinalis were annotated including core genes involved in GSH biosynthesis and its use as a redox buffer and conjugate in phase II detoxification reactions (Table 1). This analysis also included seven transcriptional regulators, which have been experimentally confirmed in human, rat or mouse to regulate expression of the GSH homeostasis genes (Table 2). Although the Ensembl project website provides an automatic preannotation of eukaryotic genomes, manual annotation was performed for each gene of interest to increase the accuracy of these predictions (12).

Table 2.

Summary of experimentally verified transcription factors regulating GSH homeostasis in human, rat, or mouse shown to be expressed in C. intestinalis

Gene Symbol	Chromosomal Location	Name and Function
Nfe2l2 (Nrf2)	Chr 1p_604	nuclear factor (erythroid-derived 2)-like 2 (NRF2); member of cap ‘n’ collar (CNC) family of basic leucine zipper (bZIP) transcription factors that binds to antioxidant response element (ARE) and thereby upregulates expression of genes responsive to oxidative stress
Keap1	Chr 8q_623	Kelch-like ECH-associated protein 1; retains NRF2 in cytosol in a redox-sensitive manner and thereby represses ARE-mediated gene expression under homeostatic conditions
Maff	Chr 12q_52	v-Maf musculoaponeurotic fibrosarcoma oncogene homolog F; most closely related to the small Maf protein family, which are required as obligatory heterodimeric partners with NRF2 for upregulation of ARE-dependent gene expression in response to oxidative stress
Bach1	Chr 6q_172	BTB and CNC homology 1, bZIP transcription factor 1; CNC transcription factor that functions typically in association with small Maf proteins to repress expression of oxidative stress genes
Nr1h3 (Lxra)	Chr 2q_696	nuclear receptor subfamily 1, group H, member 3, a nuclear oxysterol sensor involved in lipid metabolism
Nr1h4 (Fxr)	Chr 2q_540	nuclear receptor subfamily 1, group H, member 4, the bile acid sensor in mammals, which regulates genes involved in bile acid synthesis, metabolism and transport
Rxra	Chr 9p_200	retinoid X receptor-α, the obligate partner for heterodimeric NR subfamily 1 members, including FXR, LXR, PXR, CAR and VDR

C. intestinalis candidate amino acid sequences were used to perform RBH (68) analysis against the entire RefSeq database (Supplemental Table S2). The RBH analysis revealed that C. intestinalis genes have close sequence homology to other eukaryotic genes. For Gclc, Abbc2, Nfe2l2, Keap1, and Rxra, the best hits have E-values of zero. For the remainder of the annotated genes, the E-values range from 6E-156 to 5E-22 (Supplemental Table S2).

Table 3 presents the data for searches against human, rat, and mouse RefSeq databases. All candidate sequences retrieved from the C. intestinalis genome database have E-values <10⁻⁶, a cut-off threshold used commonly for initial assignment of sequence homology (27, 29). Supplemental Table S3 summarizes the construction of orthologous groups across multiple eukaryotic taxa. The use of the OrthoMCL algorithm generated a list of putative orthologous relationships between multiple genomes by reciprocal best similarity pairs. This analysis revealed that all candidate genes of the GSH homeostasis pathway in C. intestinalis clustered with known orthologous groups from multiple eukaryotic taxa. In all cases, the probabilities of these predictions were lower than the cut-off threshold (10⁻⁵) established for the prediction of orthologous groups (8). Moreover, the presence of orthologous genes in at least seven different taxa, the high percentage of sequence identity, and the high percentage of sequence coverage provide additional support of the accuracy of the annotation process.

Table 3.

Reciprocal best hits of C. intestinalis amino acid sequences against human, rat, and mouse RefSeq database

Gene Symbol	Best Hit Human*	E-value	Best Hit Rat	E-value	Best Hit Mouse	E-value
Gclc	NP_001489, GCLC	0	NP_036947, GCLC	0	NP_034425, GCLC	0
Gclm	NP_002052, GCLM	7E-44	NP_059001, GCLM	4E-44	NP_032155, GCLM	8E-44
Gss	NP_000169, GSS	5E-77	NP_037094, GSS	1E-73	NP_032206, GSS	4E-74
Ggt1	NP_005256, GGT1	5E-152	NP_446292, GGT1	2E-149	NP_032142, GGT1	3E-147
Gpx1	NP_002075, GPX3	2E-25	NP_110453, GPX1	3E-25	NP_032186, GPX1	1E-25
Gsr	NP_000628, GSR	4E-136	NP_446358, GSR	9E-118	NP_034474, GSR	8E-135
Gsta1	NP_000838, GSTA3	6E-44	NP_001009920, GST-YC2	1E-48	NP_034486, GSTA3	4E-48
Gstm1	NP_000840, GSTM3	6E-60	NP_742035, GSTM5	3E-57	NP_034490, GSTM5	4E-58
Gstp1	NP_055300, PTGD2	3E-23	NP_113832, PTGDS2	5E-20	NP_062328, PTGDS2	4E-19
Glrx	NP_932066, GLRX2	2E-16	NP_001013052, GLRX2	7E-15	NP_001033681, GLRX2	3E-16
Slc7a11	NP_001070253, SLC7A6	2E-134	NP_112631, SLC7A7	9E-130	NP_848913, SLC7A6	1E-136
Abcc2	NP_004987, ABCC1	0	NP_542148, ABCC3	0	NP_032602, ABCC1	0
Nfe2l2	NP_003195, NFE2L1	7E-34	NP_001012224, NFE2	5E-29	NP_032712, NFE2L1	2E-33
Keap1	NP_036421, KEAP1	0	NP_476493, KEAP1	0	NP_057888, KEAP1	0
Maff	NP_036455, MAFF	3E-24	NP_663706, MAFK	4E-23	NP_034887, MAFK	3E-23
Bach1	NP_056982, ZBTB7A	6E-23	NP_446454, ZBTB7A	1E-22	NP_034861, ZBTB7A	6E-23
Nr1 h3	NP_005684, NR1H3	3E-112	NP_113815, NR1H3	8E-112	NP_038867, NR1H3	4E-112
Nr1 h4	NP_005114, NR1H4	7E-56	NP_068513, NR1H4	2E-54	NP_033134, NR1H4	6E-56
Rxra	NP_002948, RXRA	2E-141	NP_036937, RXRA	1E-141	NP_035435, RXRA	1E-141

^*Accession numbers and sequence annotations in the human, rat, and mouse RefSeq database are provided.

Because the identification of domains that occur within proteins can provide insights into their function (3, 64), conserved domains occurring in the candidate genes from C. intestinalis were identified. Pfam domains within C. intestinalis amino acid sequences of the GSH biosynthesis and metabolism pathway genes are shown in Table 4. For all the proteins analyzed, the E-values were <3 × 10⁻⁵, the cut-off value used for accurate prediction of protein family members (55). Finally, the statistical significance of pairwise sequence similarity for each candidate gene using PRSS and SSEARCH analyses (56, 57) was calculated (Supplemental Table S4). These analyses revealed that for all candidate genes of the GSH homeostasis pathway in C. intestinalis, the probability of sequence homology at the primary structure level was highly significant. All predicted E-values are below the cut-off level of 10⁻⁵ for these methods of statistical analysis (56, 71).

Table 4.

Conserved domains in C. intestinalis amino acid sequences identified by Pfam database

Gene Symbol	Accession Number	Pfam Domain	Pfam Accession	Pfam Description	E-value
Gclc	ENSCING00000003707	GCS	PF03074	glutamate-cysteine ligase	3.50E-238
Gclm	ENSCING00000014616	Aldo_ket_red	PF00248	aldo/keto reductase family	7.5E-10
Gss	ENSCING00000009385	GSH_synth_ATP	PF03917	eukaryotic glutathione synthase, ATP binding domain	7.10E-93
Ggt1	ENSCING00000001905	G_glu_transpept	PF01019	γ-glutamyl transpeptidase	1.60E-196
Gpx1	ENSCING00000001683	GSHPx	PF00255	glutathione peroxidase	2.00E-18
Gsr	ENSCING00000003772	Pyr_redox_2	PF07992	pyridine nucleotide-disulphide oxidoreductase	1.70E-65
Gsta1	ENSCING00000000098	GST_N	PF02798	glutathione S-transferase, N-terminal domain	1.90E-16
Gstm1	ENSCING00000015785				5.70E-17
Gstp1	ENSCING00000004369				2.10E-19
Glrx	ENSCING00000014681	Glutaredoxin	PF00462	glutaredoxin	1.80E-22
Slc7a11	ENSCING00000001642	AA_permease	PF00324	amino acid permease	1.3E-10
Abcc2	ENSCING00000001934	ABC_membrane	PF00664	ABC transporter transmembrane region	3.90E-49
Nfe2l2	ENSCING00000003373	bZIP_1	PF00170	bZIP transcription factor	6.20E-08
Keap1	ENSCING00000008356	BTB	PF00651	BTB/POZ domain	5.80E-40
Maff	ENSCING00000014271	bZIP_Maf	PF03131	bZIP Maf transcription factor	7.70E-42
Bach1	ENSCING00000014767	BTB	PF00651	BTB/POZ domain	7.40E-28
Nr1 h3	ENSCING00000007203	Hormone_recep	PF00104	ligand-binding domain of nuclear hormone receptor	4.50E-48
Nr1 h4	ENSCING00000003634	Hormone_recep	PF00104	ligand-binding domain of nuclear hormone receptor	5.80E-39
Rxra	ENSCING00000001836	Hormone_recep	PF00104	ligand-binding domain of nuclear hormone receptor	9.20E-64

In silico cloning of GSH biosynthesis and metabolism genes in C. intestinalis.

Production of EST libraries is a common approach for characterizing protein-coding sequences within fully sequenced genomes (4). Thus, in silico cloning (5, 60) was performed for the chosen C. intestinalis genes involved in GSH biosynthesis and metabolism using the nonmouse and nonhuman EST database in GenBank. This approach identified at least one homologous EST clone for all of the 19 genes contributing to GSH homeostasis and known transcriptional regulators (Table 5). All of the identified EST clones were obtained from C. intestinalis cDNA clone libraries produced from embryos or various tissues of juvenile and adult animals (Table 5).

Table 5.

In silico cloning of genes contributing to GSH homeostasis in C. intestinalis

Gene Symbol	Best Hit	Description	E-value
Gclc	BW441631.1	cDNA library, juvenile whole animal	0
Gclm	BW312931.1	cDNA library, heart	0
Gss	BW452049.1	cDNA library, juvenile whole animal	0
Ggt1	BW350478.1	cDNA library, embryo whole animal	0
Gpx1	BW371823.1	cDNA library, mature adult whole animal	0
Gsr	BW481411.1	cDNA library, mature adult whole animal	0
Gsta1	BP005088.1	cDNA library, young adult	0
Gstm1	BW295476.1	cDNA library, neural complex	0
Gstp1	BW385835.1	cDNA library, adult digestive gland	0
Glrx	BW300911.1	cDNA library, neural complex	3E-164
Slc7a11	BW028716.1	cDNA library, blood cells	0
Abcc2	BW003593.1	cDNA library, blood cells	0
Nfe2l2	BW481489.1	cDNA library, mature adult whole animal	0
Keap1	AL667683.1	directional larval cDNA library	0
Maff	BW310785.1	cDNA library, heart	0
Bach1	AV976319.1	cDNA library, egg	0
Nr1 h3	BW150081.1	cDNA library, gonad	0
Nr1 h4	BW067324.1	cDNA library, cleaving embryo	0
Rxra	BW093519.1	cDNA library, tailbud embryo	0

Expression of GSH biosynthesis and metabolism genes in adult C. intestinalis.

Primers were designed from each of the annotated sequences and used for end-point RT-PCR analysis to confirm the expression of core genes contributing to GSH biosynthesis and metabolism. Actb was used as a housekeeping gene to ensure the presence of amplifiable RNA in all tissue samples before the analysis of candidate pathway genes. Single amplicons of predicted size were obtained for each target gene.

Each of the genes encoding functions in the GSH homeostasis pathway including the transcriptional regulators Nfe2l2, Keap1, Maff, Bach1, Nr1h3, Nr1h4, and Rxra (Tables 1 and 2) was expressed constitutively in branchial sac, gastrointestinal tract and heart of adult animals (Fig. 1). Relative to Actb RNA expression, considerable interindividual variation was observed in all three tissues for at least five of the GSH-related genes (Fig. 1). Of these, constitutive expression of Abcc2, Slc7a11, and Gss exhibited the greatest variation among replicate animals. The consistent level of Actb RNA expression for the three tissues indicates that variable expression of the GSH-related genes likely reflects physiological differences. However, the literature contains little related information regarding variability in the expression of these genes in other animal models, and this observation warrants additional research.

Fig. 1.

Ciona intestinalis gene expression analysis. Representative electrophoresis gels of RT-PCR products from 5 replicate C. intestinalis specimens. Gels demonstrate the expression of Gclc, Gclm, Gss, Ggt1, Gpx1, Gsr, Gstm1, Glrx, Slc7a11 (xCT), Abcc2 (Mrp2), Nfe2l2 (Nrf2), Keap1, Maff, Bach1, Nr1h3 (Lxra), Nr1h4 (Fxr), and Rxra genes. Schematic representation of adult animal anatomy (drawing modified from Ref. 13) illustrating the branchial sac, gastrointestinal tract, and heart. Initially, the target genes β-actin (Actb), Gclc, Gclm, Gstm1, Keap1, and Nrf2 were analyzed in 3 tissues from 12 animals, and the Actb signal was used as an indicator of RNA quality to identify 5 replicate animals for investigation of additional gene targets. This approach required that some of the gels be spliced to show only 1 target gene within a given tissue for the 5 replicate animals and to position the 5 lanes of interest adjacent to the ladder used for that gel. Targets affected include: Branchial sac, Gss, Ggt1, Gpx1, Gsr, Gstm1, Abcc2, Keap1, Bach1, Nr1h3, Rxra: Heart, Gss, Ggt1, Gsr, Abcc2, Keap1, Nr1h3, Rxra; Gastrointestinal tract, Gss, Gpx1, Gstm1, Abcc2, Bach1, Nr1h3, Rxra. The same 5 animals were used for all of the gels depicted in this figure.

Response of GSH biosynthetic genes to oxidative stress in C. intestinalis.

Replicate adult animals were exposed to the prototypical phenolic antioxidant tBHQ to assess responsiveness of C. intestinalis to oxidative stress. Semiquantitative RT-PCR analysis of heart revealed that tBHQ increased steady-state abundance of RNA transcripts for the two genes encoding the rate-limiting enzyme of GSH biosynthesis, Gclc and Gclm (Fig. 2) as reported for other species. In contrast to Gclc and Gclm, the expression Gss did not exhibit responsiveness to tBHQ with end-point RT-PCR analysis (Fig. 2).

Fig. 2.

Gene expression analysis in C. intestinalis heart after exposure to tert-butylhydroquinone (tBHQ). Semiquantitative end-point RT-PCR was performed with RNA extracted from the heart of 5 replicate adult animals (see materials and methods). Shown are relative differences in the level of expression of Gclc, Gclm, and Actb (constitutively expressed gene) in response to 100 μM tBHQ exposure for 8 h in circulating sea water tanks.

The responsiveness of Gclc and Gclm to tBHQ was corroborated by RT-qPCR analysis. Box plots illustrating the distribution of RT-qPCR data are presented in Fig. 3. Treatment with tBHQ significantly increased the expression of Gclc and Gclm genes in branchial sac (4.4 and 17.7 mean fold change), gastrointestinal tract (9.3 and 51.1) and heart (6.9 and 65.8), respectively.

Fig. 3.

Expression of Gclc and Gclm in C. intestinalis after exposure to the prototypical antioxidant tBHQ. Relative fold changes (x-axes) in the level of expression of C. intestinalis Gclc (A–C) and Gclm (D–F) genes in response to tBHQ. We exposed 7 replicate animals to 100 μM tBHQ for 12 h in circulating sea water tanks. Branchial sac (BS; A and D), gastrointestinal tract (GIT; B and E), and heart (C and F) were collected after exposure and gene expression was analyzed via RT-qPCR. Gclc and Gclm values were normalized with the C. intestinalis Actb gene. Shown are box plots with relative fold change (setting controls as 1), and 5-number summaries (smallest observation, lower quartile, median, upper quartile, and largest observation). Gene expression values were compared by Mann-Whitney U-test. P < 0.05 was accepted as statistically significant.

Phylogenetic analysis of GCL subunits.

Protein alignments and pairwise sequence identities for GCLC and GCLM were estimated for human, rat, mouse, zebrafish, C. intestinalis, and C. elegans. This analysis revealed a relative high degree of conservation among these proteins (Fig. 4, A and B). Pairwise sequence identities within each group ranged from 50.7 to 98.3% for GCLC (Table 6) and from 29.7 to 98.9% for GCLM (Table 6). Alignment of GCLC residues revealed complete conservation of a glycine-rich motif MGFGMGXXCLQ (alignment positions 264-274), cysteine (Cys)-272 (alignment position; Cys249 in human) and Cys591 (alignment positions; Cys553 in human) across human, rat, mouse, zebrafish, C. intestinalis, and C. elegans (Fig. 4A). For GCLM, the protein alignment revealed complete conservation of Cys212 and Cys213 (alignment positions; Cys213 and Cys214 in Drosophila melanogaster, Fig. 4B). The glycine-rich motif of GCLC and conserved Cys residues in both GCLC and GCLM contribute to structural modifications required for GCL activity (7, 25, 26, 34, 35, 43). Specifically, the single glycine-rich motif of GCLC lies near the catalytic active site and is thought to contribute to the transfer of the ATP-phosphate group to glutamate to form an intermediate, to which cysteine binds to form the dipeptide γ-glutamylcysteine (31). The conserved Cys residues form molecular disulfide linkages between GCLC and GCLM to form the GCL holoenzyme in response to many cellular insults, particularly oxidative stress (24, 25, 61, 66). In addition, feedback inhibition of GCL activity is mediated by GSH reduction of the disulfide linkages between GCLC and GCLM (24, 25, 61).

Fig. 4.

Phylogenetic analysis of the rate-limiting enzyme of the GSH biosynthetic pathway in C. intestinalis and other eukaryotes. A: GCLC protein alignment revealed complete conservation of a glycine-rich motif MGFGMGXXCLQ (alignment positions 264-274), Cys272 and Cys591 (alignment positions; denoted by *; Cys249 and Cys553 in human, respectively) across human, rat, mouse, zebrafish, C. intestinalis, and Caenorhabditis elegans. B: GCLM protein alignment revealed complete conservation of Cys212 and Cys213 (alignment positions; denoted by *; Cys213 and Cys214 in Drosophila melanogaster). For A and B, the level of sequence conservation is shown on a color scale with red residues being the most conserved and yellow least conserved; nonshaded residues are not conserved across species. Phylogenetic rooted trees based on the amino acid sequences of GCLC (C) and GCLM enzymes (D). Phylogenetic trees were constructed with the maximum parsimony method using the MEGA software version 4.0. The statistical significance of branch order (numbers on branches) was estimated by the generation of 1,000 replications of bootstrap resampling of the originally aligned amino acid sequences. Scale bar indicates the number of changes over the whole sequence.

Table 6.

Pairwise amino acid sequence identities for subunits of the rate-limiting enzyme for GSH biosynthesis in C. intestinalis and other eukaryotes

	C. elegans	C. intestinalis	Zebrafish	Mouse	Rat
Glutamate-cysteine ligase catalytic subunit (GCLC)
Human	53.3	58.8	82.1	94.5	94.2
Rat	52.6	58.6	80.8	98.3
Mouse	52.9	58.8	80.8
Zebrafish	55.4	58.8
C. intestinalis	50.7

	C. elegans	C. intestinalis	Zebrafish	Mouse	Rat
Glutamate-cysteine ligase modifier subunit (GCLM)
Human	31.0	36.8	63.4	96.4	96.0
Rat	31.3	36.8	61.5	98.9
Mouse	31.0	36.8	62.3
Zebrafish	29.7	29.7
C. intestinalis	30.8

Phylogenetic analysis with the maximum parsimony method using protein sequences from flanking model organisms confirmed the homology of C. intestinalis GCLC and GCLM as revealed by the RBH and alignment methods (Fig. 4, C and D). A similar tree topology was observed for both GCLC and GCLM with sequences from the five model organisms. Phylogenetic trees constructed with the maximum likelihood method generated comparable topologies (data not shown). While the combined data from the GCLC and GCLM alignments and phylogenetic trees confirms homology for the C. intestinalis GCL protein subunits, a more extensive analysis with an expanded sequence dataset would yield greater insight into the evolution of the GSH biosynthesis pathway among eukaryotic taxa.

DISCUSSION

The present study reveals that core genes in the C. intestinalis genome contributing to the biosynthesis of glutathione (Gclc, Gclm, Gss, Ggt1, Slc7a11), its use as a redox buffer (Glrx, Gpx1, Gsr) and as a conjugate in phase II detoxification (Gsta1, Gstm1, Gstp1, Abcc2), as well as known transcriptional regulators (Nfe2l2, Keap1, Maff, Bach1, Nr1h3, Nr1h4 and Rxra), exhibit a high level of homology to corresponding genes in human, rat, mouse, and other eukaryotic taxa (Fig. 5). The data are also the first to demonstrate that these genes are expressed constitutively in branchial sac, gastrointestinal tract, and heart of adult animals and that C. intestinalis GSH biosynthetic genes are responsive to tBHQ, an electrophilic phenol that is used commonly as an experimental reagent to examine redox-sensitive genes and metabolic pathways (34).

Fig. 5.

Summary of glutathione (l-γ-glutamyl-l-cysteinylglycine, GSH) homeostasis pathway identified in C. intestinalis including gene products (highlighted in gray) contributing to the biosynthesis of glutathione [GCLC, GCLM, GSS, GGT1, xCT (SLC7A11)], its use as a redox buffer [GLRX, GPX1, GSR], and as a conjugate in phase II detoxification [GSTA1, GSTM1, GSTP1, MRP2 (ABCC2)], as well as known transcriptional regulators [NRF2 (NFE2l2), KEAP1, MAFF, BACH1, LXRA (NR1H3), FXR (NR1H4), and RXRA]. ARE, antioxidant response element; HRE, hormone response element.

C. intestinalis exhibits a number of anatomical and environmental features that contribute to its value as a model organism for comparative toxicogenomics, specifically as it relates to maintenance of redox homeostasis. Adult animals are filter feeders, and replicate specimens can thus be easily exposed to xenobiotic solutions. Moreover, adult animals are relatively simple anatomically (28). The branchial sac, gastrointestinal tract, heart, and gonads are the most prominent organs in adult specimens. The incurrent or buccal siphon is the anterior opening into the gastrointestinal tract, while the rectum and gonads expel through the atrial or excurrent siphon. The intestine is dominated by an enlarged pharynx or branchial sac preceding the stomach whose wall is perforated by numerous tiny gill slits. The branchial sac is both a respiratory organ and filter-feeding device. Water and food particles enter the branchial sac, and water passes through the gill slits and then out the atrial siphon, while food remains in the gastrointestinal tract and passes posteriorly to be digested.

Surprisingly little is described regarding the anatomy and cell biology of xenobiotic detoxification and excretion in C. intestinalis. This tunicate does not have a structure recognizable as a kidney but possesses a pyloric gland comprising a system of tubules that form a single pyloric duct that opens into the digestive tract at the junction of the stomach and intestine (28). The pyloric gland has been proposed to be an excretory organ in C. intestinalis; however, this function has not been conclusively demonstrated in the literature (28, 50). The minute size of this tissue precluded its isolation in the present study. There is also evidence that nephrocytes and orange pigmented cells, two of the eight blood cell types in the circulatory system, may function as a storage and transport system for metabolic waste products, which are stored in vesicles (so-called “renal vesicles”) located near the pericardium (28). The present evidence for expression of a comprehensive pathway of genes contributing to GSH homeostasis in the heart indicates that this organ may also contribute to xenobiotic metabolism in C. intestinalis.

Whereas C. intestinalis has been used extensively to define ancestral mechanisms of development that are shared among the chordates (13), the present data are among the first to examine a key environmentally responsive pathway in this animal model. It has been concluded that with respect to gene content, the genome of C. intestinalis has experienced an extensive loss of ancestral genes and that these derived features may reflect adaptation to their specific ecological niche (38). However, regarding the GSH pathway, the unique ecological niche occupied by C. intestinalis may have provided optimal selection pressure for preservation of this central pathway of cellular homeostasis. Specifically, in the subtidal marine environment, C. intestinalis is exposed continuously to high concentrations of numerous xenobiotics that fall into chemical classes known to induce various detoxification genes across the chordates (16). Of particular relevance are the many secondary phenolic compounds that are used by marine plants and macrophytic algae (seaweed) as chemical defense against herbivorous invertebrates and bacterial degradation (2, 20). Polyphenols in various dietary plants comprise a versatile group of phytochemicals that are increasingly studied because of their antioxidant properties and induction of genes contributing to xenobiotic metabolism. Numerous studies demonstrate that plant polyphenols may enhance cellular resistance to electrophilic xenobiotics through direct modulation of genes involved in glutathione homeostasis (51, 52). For instance, Myhrstad et al. (52) demonstrated that quercetin, a particularly bioactive polyphenol found in a wide variety of dietary plants, increased GSH concentrations as well as steady-state mRNA expression of GCL subunit genes Gclc and Gclm (49).

In the present study, adult specimens of C. intestinalis were exposed to tBHQ, a well-characterized phenolic antioxidant, also known to induce expression of GCL subunit genes and its enzyme activity (7, 26, 42–44, 63). Similar to observations for these genes in mouse (63), rat (7, 43, 44), and human cells (26, 42), treatment with tBHQ increased steady-state mRNA expression of both Gclc and Gclm in branchial sac, gastrointestinal tract, and heart. Responsiveness of the Gss gene in C. intestinalis to tBHQ was more variable, which has also been reported for a human cell line (39). Thus, genes contributing to the biosynthesis of GSH in C. intestinalis respond in a manner similar to homologous genes in human and rodent models.

Phylogenetic comparison of GCLC and GCLM proteins from C. intestinalis with those from C. elegans, zebrafish, rat, mouse, and human revealed that the primary structure for both subunits is also conserved, with pairwise sequence identities ranging from 50 to 94% for GCLC and 29 to 96% for GCLM. A similar degree of conservation for GCLC was found in other evolutionary studies of eukaryotic taxa that did not include C. intestinalis (11, 25). Also conserved in the GCL catalytic and modifier subunits of C. intestinalis are cysteine residues Cys272 and Cys591 in GCLC, and Cys212 and Cys213 in GCLM. The disulfides formed between GCLC and GCLM by these cysteine residues are susceptible to reducing environments and modify GCL activity in response to changes in intracellular GSH concentrations (6, 24, 25, 35, 36, 46). Although GCLC alone can catalyze the formation of γ-glutamylcysteine, its binding with GCLM enhances enzyme activity by lowering the K_m for glutamate and ATP, and increasing the K_i for GSH inhibition (25, 35). Apparently, only some eukaryotic species have evolved the GCL modifier subunit to regulate GSH synthesis. It has been hypothesized that the presence of conserved cysteines in GCLC is a good indicator of GCLM regulation (24). For example, eukaryotic species containing the conserved Cys591 in GCLC (e.g., chicken, zebrafish, Xenopus, Anopheles gambiae, C. elegans, Saccharomyces pombe, and Neurospora crassa) also have a GCLM homolog (24). In contrast, Trypanosoma brucei and S. cerevisiae lack the conserved Cys591 residue in their GCL monomer, and their genomes do not encode a GCLM homolog (24). While the two GCL subunits in C. intestinalis also share critical features with homologous proteins of human and rodent models, further study will be required to determine the extent to which the GCLM subunit of C. intestinalis regulates the GCL holoenzyme.

Regarding transcriptional regulation, the present data demonstrate that Nfe2l2, Keap1, Maff, Bach1, as well as the nuclear receptors Nr1h3, Nr1h4, and Rxra, each of which has also been linked to GSH homeostasis, are expressed in various tissues and different developmental stages of C. intestinalis. Of these transcription factors, NRF2 is the best characterized in eukaryotic species for its regulation of genes contributing to GSH homeostasis. Under constitutive conditions, NRF2 activity is repressed through the binding of the cytoskeleton-associated protein KEAP1, and its activation requires the interference of NRF2/KEAP1 interactions (40, 41, 53, 67). Although several models exist regarding the molecular mechanisms by which NRF2-KEAP1 interactions are regulated, transcriptional activation by NRF2 is mediated by its binding to the antioxidant response element found in upstream promoter regions of genes responsive to a wide range of chemical insults including GSH homeostasis genes (40, 41, 53, 67). There is also evidence that other transcriptional regulators such as small MAF proteins and BACH1 modulate NRF2 activity (40). Notably, with the exception of the Maf gene, the C. intestinalis genome contains a single ortholog of the core GSH-related transcription factors, which typically comprise multiple paralogs in vertebrate genomes (74). The Maf gene, which was observed to be constitutively expressed in branchial sac, gastrointestinal tract, and heart of adult C. intestinalis, is most similar to Maff (small MAF family member). The single paralog of this sequence in the C. intestinalis genome appears to be more closely related to members of the large MAF family. Thus, similar to gene networks regulating development (13, 14), it appears that the core complex of transcription factors regulating GSH homeostasis in C. intestinalis also exhibits minimal genetic redundancy. This feature reinforces the value of C. intestinalis as a model organism for identification of environmentally responsive gene regulatory networks.

In summary, the present evidence of conservation of a core GSH homeostasis pathway in C. intestinalis together with its phylogenetic position and lifestyle as a filter feeder constantly exposed to natural marine toxins introduces this species as an important animal model for comparative toxicogenomics. The accurate identification of genes and their boundaries and the inclusion of sequences with greater phylogenetic distances such as C. intestinalis facilitate the discovery of functional elements within genomes because neutral sequences and patterns of substitution, insertion, and deletion can be more precisely detected (19). Recent studies have revealed the value of this strategy for discovering conserved cis-regulatory elements in higher eukaryotic species (10, 32, 48). Moreover, the majority of genes involved in genetic diseases apparently are of ancient origin (17). Given the critical role of glutathione in cellular homeostasis and xenobiotic detoxification, continued study of this and other redox-sensitive pathways in the ancestral chordate C. intestinalis may facilitate the identification of molecular mechanisms that contribute to human gene-environment disorders.

NOTE ADDED IN PROOF

The channels revealing points where gel photographs were cut and repositioned did not appear in Fig. 1 in the Articles in Press version of the article. The revised figure and legend appear in this final published version of this article.

GRANTS

Financial support was provided, in part, by a New Investigator Award (H. R. Gaskins) with funds from the National Institute of Environmental Health Sciences Center for Membrane Toxicity Studies (ES-03828-19) at Mount Desert Island Biological Laboratory.