MATERIALS AND METHODS
C- Terminal of CSD Protein May Determines Its Activation by CCS or the Unknown Factor An important difference between CCS-dependent and -independent pathways concerns the
status of the CSD disulfide bond (Leitch et al., 2009b). CCS can activate only disulfide-reduced CSD proteins, because the interaction between CCS and CSD involves formation of an intermolecular disulfide bond (Fig. 20A; Culotta et al., 2006). Thus, intramolecular disulfide-oxidized CSD would preclude the interaction (Fig. 20B). In contrast, both disulfide-reduced and -oxidized CSD can be activated by the CCS-independent pathway (Leitch et al., 2009b). This observation led us to suggest that the interaction of CSD and the unidentified
factor differs from that of CSD and CCS, which does not require the formation of the intermolecular disulfide bond (Fig. 20D), so the disulfide-oxidized CSD can still interact with the unidentified factor (Fig. 20E).
As mentioned previously, the presence of proline at position 144 of ySOD1 prevents activation by the CCS-independent pathway (Fig. 17; Carroll et al., 2004; Jensen and Culotta, 2005). The most recent model indicates that this proline results in a conformational restriction that prevents disulfide bond formation within CSD in the absence of CCS (Leitch et al., 2009b). This restriction on disulfide bond formation can be overcome by the action of CCS but not the CCS-independent mechanism (Fig. 20, C and F, respectively). Here, we suggest another reason for an essential unidentified factor in CCS-independent pathway interacting with the disulfide-oxidized CSD. In this model, the molecular conformation adopted by CSD containing proline 144 may prevent interaction of the unidentified factor with the CSD protein. This situation would explain why proline 144 blocked only CCS-independent activity but not that conferred by CCS. Further investigation is required to characterize the unidentified factor and elucidate the complete mechanism.
PERSPECTIVES
In this thesis, we address an important finding of the CCS-independent activation of three CSDs in the Arabidopsis. Thus, our future research will be to reveal the detail CCS-independent activation mechanism, and its effect to plants. The following are the major aspects: (1) Identify the unknown factor which involved in the CCS-independent pathway; (2) Test the effects of the oxygen on the CCS-independent activation; (3) Investigate the disulfide status of the three CSDs in two activation pathways; (4) Identify the phenotype of the SOD-knockout lines during different oxidative stress. Detail descriptions are as follows:
Search of the unknown factor which involved in the CCS-independent pathway
In an attempt to understand the CCS-independent activation mechanism, we will identify the unknown factor(s) which cooperates with GSH to activate CSD. In this effort, the Atccs-cellular extract will be fractionated by chromatography of ion exchange, gel filtration, and reverse-phase HPLC. Apo-CSD1 will be treated with the fractions to see activity recovery in vitro, the most effective fractions will be sent for LC-MS/MS analysis to identify the potentials factors.
The factors identified by LC-MS/MS will then be tested in various methods to confirm their relations to CCS-independent activation. Experiments can be performed with the recombinant protein, and the protoplast overexpression system to analyze the variation of the CSDs activities.
And, we can use yeast-two hybrid system and BiFC to observe its interaction with CSDs. We also can observe the phenotype of the Arabidopsis knockout/overexpression mutants of each gene, or analysis the domain function of the factors as further investigations.
Effect of the molecular oxygen in the CCS-independent activation
It has proved that CCS activation of SOD1 requires molecular oxygen (Brown et al., 2004).
However, there is no similar oxygen dependence with CCS-independent activation in human SOD1 (Culotta et al., 2009). Inhibition of CCS-independent activation of Arabidopsis chloroplastic CSD2 might also be a hint for a key role of oxygen in determining the dependence of CSD activation (see our discussion). Hence, we are curious about the relation between oxygen conditions to the CCS-independent activation in Arabidopsis. Thus, we will overexpress the CSDs in WT and Atccs protoplasts, and measure the changes of SOD activity under hypoxic and anoxic treatments.
Comparing the ratio of the CSD disulfide status between two activation pathways
Disulfide of SOD proteins is firstly shown to be oxidized by Cu-CCS during activation processes (Furukawa et al., 2004). However, the disulfide of C. elegans CuZnSOD is retained in the oxidized state regardless of copper or CCS conditions, and the disulfide of human SOD1 is also
50% oxidized in the absence of CCS and copper (Culotta et al., 2009). It is hypothesized that CCS
can only activate disulfide-reduced SOD, and CCS-independent activation can activate SOD with either reduced or oxidized disulfide bond, which make the SOD disulfide bond status be a good indicator for researches in the two activation pathway. Since the three Arabidopsis CSDs show different preference to the CCS-dependent and -independent activation, their ratio of the disulfide status might be very different. For example, the disulfide bond status of CSD2 (which is suggested to fully depend on CCS) should show more reduced form than CSD1, and CSD3 (which is suggested to be CCS-independent) should show more oxidized form than CSD1. In this effort, we can use AMS (4-acetamido-4’-maleimidylstilbene-2,2’-disulfonic acid) treatment to compare the disulfide status between CSDs in WT and Atccs, including three Arabidopsis CSD proteins (CSD1, CSD2, CSD3), CSD proteins delivered to different compartments (TP2-CSD1, TP-CSD2, CSD3-AKL), and variants of CSD1 (CSD1-AA, CSD1-SL, and CSD1-PP). We can also observe the variation of the CSDs disulfide status after different concentration of GSH, or copper treatments.
These results should be supporting evidences of our CSDs activation preference model, and also provide information for researches in the CCS-independent activation mechanism.
Phenotype observations of the SOD knockout lines during different oxidative stress
Superoxide dismutases are conserved in evolution, which suggests important functions of these proteins. However, it is surprising that there are no obvious phenotypes in Atcsd1 and Atccs
(Fig. 15, 16; Cohu et al., 2009; Chu et al., 2005). It can be explained by the functional complement of the different SODs, or other purposes (such as copper pool) for the highly expression SODs level under normal conditions (our discussion; Cohu et al., 2009). To answer this, we need to collect more information, especially in phenotypes of various SOD-knockout mutants. First, we can compare the phenotypes between different CSD-knockout lines, double-knockout lines under different oxidative stresses. We also can observe the phenotype of the fsd1/Atccs double-knockout mutant, which loses both the FSD1 and CSD2 activity in chloroplast, to help us understand the role of SOD in the chloroplast. Comparing the Atcsd1/Atccs double-knockout mutant (with no CSD1 protein and loses both CSD1 and CSD2 activities, which has CSD2 protein only) with Atcsd1 (with no CSD1 protein and activity, which has both the CSD2 protein and activity) to show the unknown function of CSD2 protein. Comparing the phenotype of Atccs/TP-CCS (has CSD2 proteins but with no CSD2 activity) with Atcsd2 (with no CSD2 protein and activity) can help us to discriminate the function of the SOD protein itself (but not its activity). Similarly, comparing the Atccs/CCS-SKV (which may or may not has CSD3 activity but has the CSD3 protein) with Atccs (with normal CSD3 protein and activity) and Atcsd3 (with no CSD3 protein and activity) may help us to confirm the CSD3 activation preference and its unknown function.
Figure 1. Residual CSD activities in Arabidopsis AtCCS-deleted flowers.
Crude protein extracts of 45-d-old Arabidopsis wild-type (WT) and Atccs flowers were analyzed for SOD activity (top) and CSD1 and CSD2 protein levels (bottom). No CSD3 protein was detected and the data is not presented here. A 26S proteasome regulatory subunit RPN8 was used as a loading control. Thirty (lanes 1 and 4), 60 (lanes 2 and 5) and 120 g (lanes 3 and 6) of total protein were loaded, as indicated. The percentage of CCS-independent activity represents an average of at least four replicates and was calculated relative to the wild type activity.
Figure 2. Transient expression of CSD genes in the Arabidopsis WT and Atccs protoplasts.
(A) Characterization the transiently expressed CSD1 (15 kD), CSD2 (22 kD) and CSD3 (16.9 kD) proteins in the WT and Atccs protoplasts by immunoblotting. The input amount protoplasts of Atccs (7.5105 cells) was 3-fold higher than that of the WT (2.5105 cells). (B) The activity (top panel) and protein amount (bottom panel) of CSD1, CSD2 and CSD3 were analyzed in the WT protoplasts.
Lane 4 was the protoplast extracts without transfection as a control.
Figure 3. The activities of three CSDs expressed in yeast sod1 and ccs.
(A) Lysates of yeast expressing CSD1, CSD2 and CSD3 were analyzed by the in-gel SOD activity assay (top) and immunoblotting (bottom) with 150 g and 15 g of total protein, respectively.
Phosphoglycerate kinase 1 (PGK1) is a loading control for the yeast extract. Strains expressed on sod1 or ccs background were represented as yCCSySOD1 and yCCSySOD1, respectively.
(B) and (C) The viability of yeast deletion strains sod1 and ccs expressing CSD1, CSD2 or CSD3 under lysine lacking conditions. Experimental procedures describing the plate (B) and liquid (C) assays can be found in Materials and Methods. Values of cell density were expressed relative to the WT level. All data were from at least four independent tests (mean SD).
Figure 4. The viability of yeast ccs expressing CSD2 with or without transit peptide under
lysine lacking condition.
Full-length and chloroplast transit peptide deleted CSD2 (TP-CSD2) were expressed in yeast ccs, and then the plate (A) and liquid (B) assays were performed as described in Fig. 3. Yeast WT and ccs were used as references.
Figure 5. Localization and SOD activity of CSD1 and CSD2 overexpressed in Atccs
protoplasts.
(A) and (C) YFP fusion proteins of full length- and chloroplast transit peptide deleted- (TP)-CSD2 (A), and YFP fusion proteins of full length-CSD1 and that fused to the chloroplastic transit peptide (TPCSD2-CSD1) of CSD2 (C), were expressed in Arabidopsis WT protoplasts for localization
(B) and (D) Full length- and TP-CSD2 (B), and full length- and TPCSD2-CSD1 (D) were overexpressed with or without AtCCS co-expression in Atccs or WT protoplasts, after which CSD activity (top) and protein amount (bottom) were analyzed. The CSD genes used here contained no YFP fusion. Because CSD protein expression was reduced in Atccs (Fig. 1, bottom), we loaded threefold more Atccs protoplasts (7.5105 cells) than WT protoplasts (2.5105 cells), in order to present a similar protein level in the result.
Figure 6. Localization and SOD activity of full length and AKL-deleted CSD3 overexpressed
in Arabidopsis protoplasts.
(A) GFP fusion proteins of full length and C-terminal AKL deleted (AKL)-CSD3 were expressed in Arabidopsis WT protoplasts for localization analysis. DsRed-SKL is a peroxisomal marker. The autofluorescence of chlorophyll is depicted in blue. Bars, 10 m. (B) Full length- and AKL-CSD3 were overexpressed with or without AtCCS co-expression in Atccs or WT protoplasts, after which
contained no GFP fusion.
Figure 7. The activity of CSD1 expressed in yeast ccs with different glutathione
concentrations.
One mM CDNB, 1 mM BSO or 20 mM GSH were used in the treatments as indicated. Equal volumes of 95 ethanol (EtOH) or H2O were added as mock treatments of CDNB and GSH. (A) Cellular concentrations of total and reduced glutathione in the yeast lysates were analyzed after treatments. Data represent results of three independent experiments (means SD). Values are relative to the H2O control (treatment 5). , P 0.05 (Student’s t test). (B) CSD1 activity (top) and
respectively. The values below the activity gel are values relative to the mock control. Bands were quantified using ImageQuant software. PGK1 was used as a loading control.
Figure 8. The effect of glutathione level on CCS-independent CSD activity in Atccs flowers.
Flowers taken from Atccs plants were soaked in CDNB or GSH solutions at the indicated concentrations for 3 or 1 h, respectively. Equal volumes of 95 ethanol (EtOH) or H2O were added
(A) and (B) Cellular concentrations of total glutathione (gray bar) and GSH (white bar) in Atccs flowers treated with CDNB (A) or GSH (B). The values are relative to those of the mock treatments.
Data represent results of three independent experiments (means SD). , P 0.05 (Student’s t test).
(C) SOD activity (top) and CSD1 and CSD2 protein amounts (bottom) were analyzed using 150 g and 30 g total protein, respectively. RPN8 was used as a loading control. A white arrow in the lower part of the GSH-treatment activity gel indicates the location of the major CSD activity band.
Figure 9. CSD activities in Atccs flower protein extract treated with GSH in vitro.
The Atccs flower protein extract was treated with 10 mM GSH for different incubation time (A) or with GSH at different concentrations for 1.5 h (B). Lane 1 presents a mock treatment in panels (A) and (B). RPN8, an internal reference protein, exhibited decreased accumulation andor stability high GSH concentrations.
Figure 10. The CuZnSOD activities of WT flower protein extract treated with CDNB in vitro.
(A) and (B) CDNB was added into the protein extract at different concentrations for 1.5 h incubation (A) or with 10 mM for different incubation time (B). After incubation, CDNB solutions were removed, and the flowers were washed with water 3 times, then the WT flower protein extract was analyzed for CSD activities. Lane 1 in both (A) and (B) was a mock treatment. RPN8 was a loading control.
Figure 11. CSD1 and CSD2 protein levels in WT flower protein extract treated with GSH in vitro.
The WT flower protein extract was treated without or with 10 mM GSH and incubated at 30C for 18 h, then protein levels were analyzed by use of -CSD1 and -CSD2 antibodies. An aliquot of the same plant extract was stored at 20C and then used as a control (lane 1).
Figure 12. The effect of glutathione concentration on the activity of CSD1 transiently
expressed in Arabidopsis protoplasts.
(A) Atccs protoplasts overexpressing CSD1 were treated with 5 mM GSH or 5 mM CDNB for 2 h as indicated. After incubation, extracts of these protoplasts were analyzed for their cellular
concentrations of total glutathione and GSH, SOD activity, or amount of CSD1 protein. Actin (ACT) was used as an internal control. (B) CSD1 was transiently expressed in Atccs protoplasts together with ROXY1 or GRXcp. Extracts of these protoplasts were analyzed for their cellular concentrations of total glutathione and GSH, SOD activity, or amount of CSD1 protein. ACT was used as an internal control. Data represent results of three independent experiments (means SD). , P 0.05 (Student’s t test).
Figure 13. Purification and characterization of recombinant CSD1 proteins.
(A) The CSD1 gene was cloned into the pGEX-6P-1 vector for fusion with a GST tag. Twenty L of the samples at each step was subjected to SDS-PAGE and Coomassie Blue staining. I, IPTG induced cellular extract. FL, flow through. W1W3, wash. E1E3, eluate. (B) Fifty ng each of purified GST-, and GST digested Holo- and Apo-CSD1 was analyzed by immunoblotting. (C) The
activity of 870 ng Apo-CSD1 incubated with 0.1 M Cu was analyzed after the indicated treatments.
Buffers used for each pH were: 0.2 M Na2HPO4 and 0.1 M citrate for pH 27, and 0.2 M glycine-NaOH for pH 910.
Figure 14. Activation of Apo-CSD1 by Cu, GSH and Atccs leaf cellular extract.
The activity of Apo-CSD1 was analyzed by the in-gel SOD activity assay following the indicated treatments (top). Each reaction contained 870 ng Apo-CSD1 protein and 20 M ZnSO2, except for lane 5, which contained only 15 g Atccs cellular extract and served as a control for extract-containing samples. For GSH, Cu or Atccs cellular extract treatments, 1 mM, 0.1 M and 15
g of these additions were used, respectively. Both the native-gel and the running buffer contained 0.1 mM EDTA. A replicate with the same treatments on 50 ng Apo-CSD1 was used for analyzing the amount of CSD1 protein by immunoblotting (bottom).
Figure 15. Characterization and seed germination rate of the WT, Atccs and Atcsd1.
(A) T-DNA insertion site of the CSD1-knockout plant, Atcsd1, can be found at the third exon of the
CSD1 gene (572th nucleotide of the genomic sequence). (B) Gene expression levels of CSD1,
CSD2 and CSD3 in 7-d-old WT and Atcsd1 seedlings. Actin2 (ACT2) was used as a loading control.
(C) CSD1 and CSD2 protein levels in WT and Atcsd1 seedlings. RPN8 was used as a loading control. (D) SOD activities of the WT, Atccs and Atcsd1. Thirty g of the protein extract from 7-d-old seedlings was analyzed. The CuZnSOD activity level is generally lower in seedlings, thereby the residual activity in lane 1 is unable to be detected. (E) Superoxide anion level of the WT,
Atccs and Atcsd1 plants was analyzed by measuring the NBT-formazan production level in 7-d-old
seedlings. Data represent results from three independent experiments (means SD). (F) The seed germination rate of the WT, Atccs and Atcsd1 plants grown on 12 MS plates with 100 M BH, 0.008 or 0.04 M MV treatments was analyzed. Data represent results from five independent experiments (means SD). a, b and c indicate different statistic groups between WT, Atccs and Atcsd1 in each treatment at P 0.05 (Student’s t test).
Figure 16. The antioxidant abilities in WT, Atccs and Atcsd1 with paraquat and BSO
treatments.
(A) and (B) Root length of WT (white bar), Atccs (gray bar) and Atcsd1 (black bar) grown on 12 MS plates with 0, 0.008, 0.04 and 0.2 M paraquat treatment (A) or additionally with 0.4 mM BSO (B) for 4 d. Different statistic groups with P 0.05 are indicated as a, b and c and was assessed by two-tailed, unequal-variance Student’s t test. (C) and (D) Total (C) and reduced (D) glutathione
concentration of plants grown on 12 MS plates without or with 0.4 mM BSO treatment were measured and presented as values relative to that of WT without treatment. All data was determined in 3 independent experiments from 16 plants per line of each treatment (means SD). Under the normal condition, the reduced glutathione concentration in Atccs was significantly higher than that of the WT (panel D).
Figure 17. Expression of Arabidopsis CSD1 variants in yeast sod1 and ccs.
(A) Amino acid sequence alignment of CuZnSOD genes in A. thaliana (CSD1, CSD2 and CSD3),
H. sapiens (hSOD1), S. cerevisiae (ySOD1) and C. elegans (wSod-1). Residues corresponding to
141G and 143V of CSD1 are indicated in the open boxes with arrows. Black boxes highlight identical residues. (B) Yeast strains expressing the CSD1 variants in sod1 and ccs were analyzed as described in Fig. 3A for SOD activity (top) and CSD1 protein levels (bottom). (C) and (D) The viability of yeast strains sod1 and ccs expressing the CSD1 variants under lysine lacking conditions. Plate (C) and liquid (D) assays were performed as described in Fig. 3B and 3C. The data were taken from at least four independent tests (mean SD).
Figure 18. Summary of different activation dependence of the three Arabidopsis CSDs.
Assuming that the total activity of each CSD in the WT is 100, the fraction activated by AtCCS (CCS) is represented in green and that activated through the CCS-independent (CI) pathway is represented in orange. Solid lines stand for the activation which was demonstrated experimentally, and the dashed lines indicate undetectable activation. Based on our results, we propose that cytosolic CSD1 is activated mainly by CCS and partially by a CCS-independent pathway. The level of CCS-independent CSD1 activation was 36 of that in WT (Fig. 5D, lanes 2 and 3). CSD2 in chloroplasts should be activated completely by CCS (Fig. 5B, lanes 2 and 3), whereas most activity of peroxisomal CSD3 might be conferred by the CCS-independent pathway (see Discussion).
Figure 19. Two possible models for CCS-independent activation of CSD.
(A) The unidentified factor (unknown), which acts like a scaffold protein, first interacts with both GSH and the CSD protein, and the GSH-bound Cu cofactor is then transferred to CSD. (B) The Cu cofactor is first transferred from GSH to the unidentified factor and then transferred to a CSD
Figure 20. CSD interacts with the CCS versus the unidentified factor.
(A) to (C) The CCS interacts with disulfide-reduced CSD and CSDPP but not disulfide-oxidized CSD. (D) to (F) The unidentified factor (unknown) interacts with disulfide-reduced and -oxidized CCS but not CSDPP
Table 1. Accession numbers of genes described in this article.
Gene Genus species Accession number
AtCCS Arabidopsis thaliana NP_563910.2 CSD1 Arabidopsis thaliana NP_001077494.1 CSD2 Arabidopsis thaliana NP_565666.1 CSD3 Arabidopsis thaliana NP_197311.1 ROXY1 Arabidopsis thaliana NP_186849.1 GRXcp Arabidopsis thaliana NP_191050.1
hSOD1 Homo sapiens NP_000445.1
ySOD1 Saccharomyces cerevisiae NP_012638.1 wSod-1 Caenorhabditis elegans NP_001021956.1
Table 2. Primer pairs used in this research.
The mutation sites of CSD1 are underlined.
Construct Primer Set Sequence (5 to 3)
G141AV143A-Fw AACGCAGGCGCCCGTGCTGCTTGCGGC TPCSD2-CSD1 CSD1-Fw-NcoI TCTCCATGGCGAAAGGAGTTGCAGTTTT
CSD1-Rv TTAGCCCTGGAGACCAATGATGC