INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

In a rapidly maturing technology called phytoremediation, plants serve as valuable tools for extracting, sequestering, and detoxifying harmful environmental pollutants. Native and genetically engineered plants can be used to degrade toxic organic pollutants such as trinitrotoluene and trichloroethylene or to manage immutable elemental pollutants such as the heavy metals mercury and cadmium (Doty et al., 2000; Meagher, 2000; Hannink et al., 2001; Persans et al., 2001). Methylmercury is one of the most hazardous elemental pollutants in our environment and a major focus for phytoremediation research (Meagher et al., 2000). Methylmercury is extremely toxic and is efficiently biomagnified by several orders of magnitude in long, aquatic food chains. It is the primary source of human mercury poisoning from consuming fish. Using a metabolic engineering approach, we have demonstrated the extraordinary effectiveness of both MerB (organomercurial lyase) and MerA (mercuric reductase) enzymes at detoxifying hazardous mercury compounds {i.e. ionic mercury [Hg(II)], methylmercury, and phenylmercury [PMA]} in transgenic plants (Rugh et al., 1998; Bizily et al., 1999; Bizily et al., 2000) as shown in Reactions 1 and 2 (see below). MerB transforms methylmercury to less toxic, non-biomagnified ionic mercury [Hg(II), and MerA electrochemically reduces Hg(II) to the least toxic metallic mercury, Hg(0)]. When the MerB and MerA enzymes are co-expressed in transgenic plants, the coupled reaction transforms methylmercury to (Hg(0)) (Bizily et al., 2000). In this current study, we examine the impact of subcellular protein targeting on mercury phytoremediation.
(1)
(2)
Previous work showed that the MerB catalyzed protonolysis of organic mercury limited the efficiency of the coupled reaction in engineered plants even when MerB enzyme levels were relatively high (Bizily et al., 2000). We propose two mechanisms that could cause this problem. First, the cell wall (CW) and cell membrane present semipermeable barriers to methylmercury ions, which are probably bound by free thiol ligands at the cell periphery. As a result, the toxin may accumulate in the CW and diffuse slowly into the cytoplasm. Second, it is possible that methylmercury inside the cell is partitioned into hydrophobic microenvironments, such as the membrane-rich secretory pathway. The evidence for this proposal is the nonlinear relationship between MerB concentration and the rate of the two-step conversion of organomercurials to Hg(0) (Reactions 1 and 2). Ten-fold increases in cytoplasmic MerB levels result in only 2-fold increases in the protonolytic cleavage rate, even when MerA levels are greatly in excess over MerB levels (Bizily et al., 2000). In contrast, the rate at which ionic mercury Hg(II) is reduced to Hg(0) (Reaction 2) is linearly correlated with merA expression levels (Rugh et al., 1996). Although the substrates for both enzymes are highly reactive, organic mercurial compounds like methylmercury are much more hydrophobic than free metal ions and this may affect their cellular behavior. The methyl group on methylmercury makes it far more membrane soluble than ionic mercury; thus, in higher animals it concentrates in the membrane rich organs and causes rapid lysosomal damage in cells (Eto et al., 1997; Dare et al., 2001). This hydrophobicity suggests that organomercurial substrates may not diffuse efficiently to the cytoplasmically expressed MerB enzyme.

We proposed that by modifying the bacterial merB gene and targeting the MerB protein (Bizily et al., 2000) for accumulation in the endoplasmic reticulum (ER) and for secretion to the CW, we will improve the efficiency of organic mercury processing in plants. By analogy, native eukaryotic cytochrome P-450 hyroxylases that process hydrophobic organic substrates are targeted to the hydrophobic environments of the secretory pathway, primarily the ER, and do not appear at significant levels in the cytoplasm ( Monier et al., 1988; Szczesna-Skorupa and Kemper, 2000). Similarly, we are targeting MerB to the hydrophobic ER and secretory pathway, where access to organic mercury substrates may be improved. We demonstrate that Arabidopsis plant lines with ER-targeted or CW-secreted versions of MerB more efficiently resist and detoxify organic mercury than plants expressing MerB in the cytoplasm.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

MerB, CW-MerB, and ER-MerB Expressed in Escherichia coli Confer Resistance to PMA

Two merB gene constructs targeting modified MerB proteins are diagrammed in Figure 1A. The CW-merB gene encodes CW-MerB with the N-terminal 21-amino acid signal sequence from tobacco (Nicotiana tabacum) extensin after the initiator Met codon. Thus, the modified protein targets the secretory pathway, which carries it through the plasma membrane to the CW. The ER-merB gene encodes a protein containing the same N-terminal signal sequence as CW-MerB. This gene also encodes, just before the stop codon, an ER retention signal (KDEL) characteristic of a large number of ER-accumulated proteins. The goal of the ER-MerB construct was to trap MerB activity within the secretory pathway. Before examining the effectiveness of these modified MerB proteins in plants, we compared their enzymatic and immunological activities in E. coli with a wild-type (WT) MerB protein that accumulates in the cytoplasm.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Modification of the bacterial merB gene for CW- and ER-specific expression in plants. A, WT 642-bp bacterial coding sequence of merB (Bizily et al., 1999) was altered by two cycles of nested PCR to make gene constructs targeting the protein product to the CW (CW-merB) and ER (ER-merB). Arrows show the position of the mutagenic primers used. B, Bacterial disc sensitivity assays with PMA were used to compare the functionality of MerB, CW-MerB, and ER-MerB expressed in E. coli as compared with the a bacterial strain containing only the empty vector plasmid [() merB] or the WT bacterial sequence (merB; see "Materials and Methods"). C, Western-blot analysis of MerB, CW- MerB, and ER-MerB expression showed that three different MerB antibodies all react similarly to the modified MerB proteins. Blots of the bacterial extracts were probed with rabbit polyclonal antibody (pAb-MerB) and mouse monoclonal antibodies mAb10E2 and mAb2H8. CW-MerB and ER-MerB ran at slightly higher Mrs than MerB as would be predicted from their increased sequence lengths. CW-MerB and ER-MerB were quantified (see "Materials and Methods") and normalized relative to MerB levels in strains expressing WT merB. D. MerB, CW-MerB, and ER-MerB protein levels were determined on western blots performed on parallel samples of the crude protein extracts examined in the p-chloromercuribenzoic acid (PCMB) spectrophotometric MerB enzyme assays (Table I). The filter was reacted with pAb-MerB and 2-min exposure is shown. MerB, CW-MerB, and ER-MerB bands were quantified using a film exposed for 20 s (see Table I). The E. coli protein polynucleotide phosphorylase (PNPase), probed with a polyclonal antibody, is shown to confirm the equal loading and transfer of samples to the membrane. Band intensities were normalized relative to WT MerB protein levels from three repetitions of this experiment (see Table I).

The three constructs to be compared were cloned under the control of a bacterial promoter and transformed into an E. coli strain already expressing merA (see "Materials and Methods"). Thus, MerB activity would be coupled with MerA to give maximum bacterial resistance to organic mercury (i.e. coupled Reactions 1 and 2). Enzyme functionality was first verified by a disc sensitivity assay in which 6-mm paper filter discs loaded with a source organic mercury (PMA) were placed on top of freshly plated bacterial lawns. The E. coli strain containing the WT form of merB was the most organic mercury resistant and cells grew very near to the PMA disc with a small zone of clearance (Fig. 1B). Cells expressing merA and an empty vector plasmid, pBS-SKII⁺, were highly sensitive to PMA and had a large zone of clearance. The strains expressing the CW-merB and ER-merB genes had clearance zones similar in size to cells with the WT merB gene. Hence, both modified forms of enzyme are highly active.

To quantitatively compare expression levels among MerB, CW-MerB, and ER-MerB proteins, we had to exclude the possibility that the available antibodies differed in their affinity to the modified forms of the protein. Equal amounts of crude bacterial protein extracts from the same four strains were separated by SDS-PAGE, blotted, and probed with a polyclonal (pAb-MerB) and two monoclonal (mAb-2H8 and mAb-10E2) antibodies that react with MerB (Fig. 1C). All three western blots detected a single band at approximately 30 to 32 kD and showed similar in expression levels among strains. The polyclonal antibody pAb-MerB appeared to react slightly more strongly with the modified proteins CW-MerB and ER-MerB than the monoclonal antibodies. Equal loading and transfer of samples among the western blots was confirmed by probing the high-M_r region of the same blots with antibodies to PNPase (Fig. 1D, top). Among the strains, MerB was expressed at the highest level, followed by CW-MerB, and then ER-MerB at the lowest level. Most noteworthy, it appears that minor differences in PMA resistance (Fig. 1B) result from significant differences in the steady-state levels of protein in E. coli as described previously for MerB levels in plants. The interpretation of these three assays as independent measurements of MerB levels implicitly assumes that all three antibodies recognize different epitopes. This is a reasonable assumption because one of the antibodies is a rabbit (Sylvilagus robustus) polyclonal and the two monoclonal antibodies were produced from hybridoma cell lines derived from different mice (Mus musculus).

Specific Activities of MerB, CW-MerB, and ER-MerB Expressed in E. coli Are Similar

Even though the PMA resistance levels of the modified enzymes in bacteria appear similar to WT and approximately proportional to enzyme expression levels, it was still possible that that the specific activities (enzyme activity per unit of protein) of the modified forms of protein were no longer equivalent to WT MerB. To measure the relative enzymatic activities of WT MerB, CW-MerB, and ER-MerB, bacterial protein extracts from four Top 10F' E. coli strains containing an empty vector or plasmids encoding merB, CW-merB, and ER-merB were added to enzyme assays containing 100 µM PCMB. PCMB is a spectrophotometrically active, alternative substrate for the MerB enzyme. PCMB assays were performed for each bacterial strain and the values were averaged and are shown in Table I. Extracts expressing WT MerB had the strongest activity, followed again by CW-MerB, and ER-MerB had the weakest activity. Relative protein concentrations were determined for these same protein extracts (Table I) by quantifying and averaging duplicate bands on western blots (Fig. 1D). Specific activities for MerB (enzyme activity per unit protein) were calculated by dividing change in substrate absorbance by the relative amount of MerB, CW-MerB, or ER-MerB protein added to the reaction chamber, presented in Table I. These data show that there are no significant differences in specific activities among MerB, CW-MerB, and ER-MerB enzymes for an organic mercury substrate.

PMA Resistance Phenotypes of Plants Expressing Modified Forms of MerB

The CW-merB and ER-merB genes were cloned under control of a constitutive plant promoter in a plant transformation vector in a manner similar to earlier expression studies on WT merB (Bizily et al., 2000). The same 35S promoter/nopaline synthase (NOS) terminator combination was used for each gene construct. Five independently isolated transgenic plant lines containing each new transgene in a constant strong merA transgenic plant background (see "Materials and Methods") were analyzed for resistance to toxic organic mercury as shown in Figure 2. Two previously characterized lines expressing transgenes for WT merB in this same merA background (merA/merB lines AB-1 and AB-5; Bizily et al., 2000) and the merA parent line lacking merB were used as positive and negative controls, respectively. AB-1 is the most resistant merA/merB line that we have thus far isolated from a screen of 60 transformants and AB-5 is a moderately resistant line (Bizily et al., 2000). The plant lines indicated (Fig. 2A) were grown on standard germination plates in the absence of PMA and with 1 and 5 µM PMA (Fig. 2, plates B-D, respectively). Lines with the CW-merB and ER-merB constructs varied greatly in their resistance to PMA as we had found earlier for a variety of independent merA/merB lines with cytoplasmically expressed MerB (Bizily et al., 2000). Lines with the CW-targeted construct CW-1, -2, and -5 and lines with the ER-targeted construct ER-1, -2, -3, and -5 appeared as resistant as the positive controls, AB-1 and AB-5, when grown on 1 µM PMA. At 5 µM PMA, CW-2, ER-1, and ER-2 showed comparable growth to the most resistant positive control line, AB-1. CW-5 and ER-3 grew similarly to AB-5, the moderately resistant positive control that shows approximately one-third of the organic mercury processing capacity of AB-1 (Bizily et al., 2000). The most resistant MerB targeted plant lines, CW-2, ER-1, and ER-2, grew as rapidly and appeared as healthy on 5 µM PMA as on mercury-free control plates. AB-5, CW-5, and ER-3 appeared healthy, although they grew at least 50% more slowly on 5 µM PMA. Other lines germinated on 5 µM PMA but were severely stunted, suffered bleaching, and did not survive past 4 weeks. Root growth of the more resistant ER and CW targeted lines was equivalent or superior to the best cytoplasmically expressed MerB lines (Fig. 2, E and F).

View larger version (81K): [in this window] [in a new window]   Figure 2.   Resistance of Arabidopsis merB, CW-merB, and ER-merB lines to organic mercury. Arabidopsis seeds were germinated on 0.8% (w/v) agarose plates containing standard one-half-strength Murashige and Skoog plant growth medium (Bizily et al., 2000). The various plant lines discussed in the text are distributed on the plates (B-D) as shown in A. The plates were dosed with 0 µM PMA (B), 1 µM PMA (C), or 5 µM PMA (D). Plants were grown for 16 d at 22°C with 16 h of light per day. The slight growth of the merA negative control line on 1 µM PMA results from the heavy seed density used in this experiment. Plant lines "a" and "b" are unrelated to these experiments. E and F, Root growth of WT, AB-1, CW-2, and ER-3 lines on one-half-strength Murashige and Skoog medium without (E) and with (F) 2 µM PMA. Seeds were germinated on plates and plants were grown vertically for 10 d.

Rates of PMA Conversion to Hg(0)

The organic mercury conversion rate was determined for each independently derived plant line by measuring gaseous Hg(0), the final product of the coupled MerA- and MerB-catalyzed reactions. Groups of 10 seedlings from each line were immersed in medium with PMA. Hg(0) emissions were sampled at 0, 5, and 10 min and used to calculate a PMA-to-Hg(0) conversion rate for each sample, as shown in Figure 3 and summarized in Table II. AB-1 and ER-2 transformed PMA substrate to Hg(0) product at a rate of 800 ng Hg(0) min¹ g¹ fresh weight tissue and consistently showed the highest levels of coupled enzyme activity. CW-2 produced Hg(0) at approximately 50% of this rate, although this was more rapid than any of the remaining lines examined, including AB-5. AB-5, CW-5, ER-1, and ER-3 had conversion rates close to 200 ng Hg(0) min¹ g¹ fresh weight tissue. Conversion rates approximately reflected levels of PMA resistance with one exception: ER-1 consistently showed a resistance phenotype comparable with lines that had 2 to 4 times higher volatilization rates.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Organic mercury to Hg(0) conversion by selected Arabidopsis merB, CW-merB, and ER-merB lines. PMA to Hg(0) conversion (Reactions 1 and 2) rates were measured for each transgenic line. The enzyme activity [nanograms of Hg(0) per minute per gram plant tissue [fresh weight]) for each line is reported as the average of three separate assays with 10 seedlings each and the SE among these three assays.

Thirty 2-week-old Arabidopsis plants from each line were assayed by western blot for MerB protein levels in crude protein extracts, as shown in Figure 4 and summarized in Table II. Relative to the modified merB lines, 10 times less crude protein was loaded for WT merB lines (AB-1 and AB-5) because pilot experiments showed that WT MerB was expressed at much higher levels. PEP-carboxylase levels were also measured on these same western blots as an indicator of total protein loading and transfer to the membrane (Bizily et al., 2000). The 10-fold reduction in protein loading for the AB-1 and AB-5 lines is confirmed in the PEP carboxylase assays (top of each set of assays, Fig. 4, A and B). MerB and the modified forms of the protein ran as a single band at 32 kD in plants. AB-1 expressed 300% to 500% higher levels of MerB as compared with AB-5. CW-2 contained approximately 14% as much MerB as AB-5. ER-2, ER-3, and ER-1 contained 13%, 3%, and 1%, respectively, of the amount of MerB in AB-5. CW-1, -3, -4, and -5, and ER-4 and -5 did not produce detectable levels of MerB in three repetitions of this experiment. MerB was undetectable in CW-5 despite having volatilization and resistance phenotypes similar to ER-3. ER-1 had a very low level of MerB compared with CW-2 and ER-2, reflecting ER-1's lower volatilization rate; however, the disproportionately high level of PMA resistance for ER-1 is still enigmatic. Similar results were obtained in several repetitions of these assays.

View larger version (70K): [in this window] [in a new window]   Figure 4.   Western-blot analysis of MerB, CW-MerB, and ER-MerB expression levels in transgenic Arabidopsis lines. For each line, 30 2-week-old seedlings were removed from mercury-free growth medium and ground in liquid nitrogen. Crude cell protein extracts were resolved on 12.5% (w/v) PAGE (Bizily et al., 2000). Western blots were probed with the pAb-MerB and an anti-phosphoenolpyruvate (PEP)-carboxylase polyclonal antisera. Crude protein loading was reduced 10-fold for lines AB-1 and AB-5. CW-MerB and ER-MerB bands were quantified and normalized relative to the MerB level in the line AB-5.

In summary, the CW- and ER-targeted, merB-expressing lines were highly resistant to PMA and volatilized large amounts of Hg(0), and yet they expressed very low levels of MerB protein relative to lines expressing WT MerB with similar levels of resistance. Using the coupled conversion of PMA to Hg(0) as a measure of MerB activity (Bizily et al., 2000) and western blots to assay relative MerB protein levels, the relative specific activities of MerB in the various plant line were estimated, as shown in Table II. Our analysis of the one CW-targeted and three ER-targeted lines, where MerB levels could be quantified, indicated that these plants had much higher specific activities for MerB than plants with cytoplasmic MerB. Specific activities ranged from 12-fold (CW-2) to 67-fold (ER-1) higher than AB-5. AB-5 and AB-1 had nearly identical specific activities as expected. The ER-targeted lines ER-1 and ER-2 and one of the CW-targeted lines, CW-2, had resistance levels equivalent to the best WT merA/merB line, AB-1. One of these lines, ER-2, volatilized an equal amount of Hg(0) as AB-1. ER-1 and CW-2 volatilized Hg(0) at approximately 25% and 50% the rate of AB-1. Remarkably, ER-1, ER-2, and CW-2 all had less than 5% of the MerB protein expressed in AB-1. Furthermore, lines such as ER-5, CW-1, and CW-5 showed significant PMA resistance (Fig. 2C) despite having undetectable levels of MerB (Fig. 4).

Subcellular Localization of MerB and ER-MerB

Leaf cells from Arabidopsis seedlings containing merA alone, and the highly resistant AB-1, CW-2, and ER-2 lines, were fixed and reacted with pAb-MerB primary antibodies and Texas red-conjugated goat-anti-rabbit secondary antibodies (Molecular Probes, Inc., Eugene, OR). Visualization of the labeled cells from the merB line, AB-1, with the confocal microscope show strong expression of MerB in the cytoplasm (Fig. 5A). The large oval opaque regions without MerB label represent chloroplasts, which are abundant in leaf cells. Under equivalent laser light settings, cells from the negative control (merA) line showed no apparent fluorescence signal or exhibited a very weak autofluorescence (Fig. 5B). Cells from line ER-2 expressed relatively low levels of ER-MerB relative to AB-1 showed strong staining (Fig. 5, C and D), but in a pattern that was entirely different from that observed for AB-1. The signal is restricted to large vesicular structures that are consistent in shape and size with components of the ER and Golgi network. Confirmation of their subcellular location would require observation by electron microscopy. Most of the ER-2 cells examined showed this pattern of MerB localization (not shown). We were unable to observe any consistent strong signal in several attempts to localize CW-MerB in the line CW-2 (data not shown), consistent with the low protein levels detected by western. There was no specific localization of CW-MerB in the CW, and very little MerB detected anywhere in these cells despite repeated attempts to detect this protein. This is likely due to the fact that the CW was partially digested with cellulase and pectinase during preparation of single cells for immunolabeling and any protein contained on the cell surface could be lost in this step. It is also possible that CW-MerB was secreted through the CW and lost before fixation. In any case, for the CW-MerB protein to produce the strong PMA resistance and transformation phenotypes observed it must have been functioning as it moved through the secretory pathway.

View larger version (103K): [in this window] [in a new window]   Figure 5.   Immunofluorescence localization of MerB and ER-MerB in fixed Arabidopsis leaf cells. A, MerB WT line. B, MerA line with no MerB (negative control). C and D, ER-MerB line. The confocal images shown are reconstructed from stacks of 0.5-micron photographs. Under equivalent laser parameters, MerB (A) and ER-MerB (C and D) showed different patterns of localization. The dimension of each cell is approximately 25 × 40 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

It may be beneficial to target some enzymes to specific subcellular microenvironments when engineering metabolic pathways (Burbulis and Winkel-Shirley, 1999; Facchini, 2001). Clearly, localizing organomercury lyase activity to the ER, and possibly through it in the case of CW-MerB, allowed more efficient detoxification of organic mercury than cytoplasmic expression of the MerB enzyme. Perhaps localization to nucleus, mitochondria, chloroplasts, and peroxisomes also will have beneficial effects on organic mercury processing and resistance.

There are a few possible interpretations of our results demonstrating more efficient organic mercury processing when MerB proteins are targeted. First, we had previously demonstrated that the metabolic flow (Kacser and Porteous, 1987) through the coupled mercury transformation pathway (Reactions 1 and 2) is most limited by flux through organomercury lyase (Bizily et al., 2000), but that increasing flux required disproportionately large increases in cytoplasmic MerB concentrations. Herein, we have demonstrated that movement of MerB into or through the membrane-rich ER and secretory pathways improves enzyme-substrate reaction kinetics, perhaps because these cellular compartments have higher substrate concentrations. Increased substrate concentrations would certainly improve MerB activity because it appears to have a low substrate or high product affinity. The in vitro K_m values of MerB are in the range of 0.5 mM for most substrates (Begley et al., 1986), whereas environmental and laboratory organic mercury concentrations are generally in the low micromolar range. Rapid movement of even small amounts of CW-MerB through the secretory pathway may be sufficient to transform toxic organic mercury and flush it out of this system. Second, these subcellular compartments may provide a chemical environment that favors a higher substrate-to-product turnover rate for MerB. For example, the high thiol content of the ER system or particular thiol compounds like glutathione might favor more efficient removal of Hg(II) from the MerB-Hg(II) enzyme product complex during the reaction (Begley et al., 1986; Pitts and Summers, 2002). Third, it appears that with ER- and CW-targeted enzymes we have changed the problem of flux through organomercury lyase from one of substrate limitation to one with greater direct dependence on enzyme concentration. This is a highly positive outcome because it is quite feasible to isolate more active plants or to further engineer MerB for increased expression in the secretory system. Indirect evidence for this is the fact that 60 transgenic plants were screened to find the line AB-1 (Bizily et al., 2000), whereas only five lines of each of the targeted lines were screened herein to find plants with equal organic mercury processing activities. Finally, the fact that both the CW and ER forms of the enzyme produce similarly efficient resistance to and processing of organic mercury suggests organic mercury is most toxic early in the ER and secretory pathways.

	CONCLUSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

This research on increasing the efficiency of MerB catalysis is motivated by our interest in creating an optimal tool for detoxifying methylmercury and remediating mercury contaminated marshlands and other aquatic ecosystems, where methylmercury is most abundant and biomagnified. We have demonstrated that targeting MerB protein to the secretory pathway increased the efficiency of processing toxic organic mercury by more than an order of magnitude. In addition, targeting greatly reduced the total number of transformed plants that were screened to achieve the same high levels of toxic substrate processing relative to plants expressing untargeted cytoplasmic MerB. A recent publication suggests that native trees along riverbanks transmit from mercury-contaminated sediments dangerous levels of methylmercury through their fallen leaves into the adjacent aquatic environments (Balogh et al., 2002). Wetland trees expressing enzymes that efficiently degraded methylmercury to Hg(II) or Hg(0) could be a solution to this problem. Future efforts will apply the enhanced effects of MerB protein targeting to environmentally relevant species such as cottonwood (Populus deltoides), willow (Salix nigra), and aquatic plants that could eliminate sources of methylmercury before they enter the food chain.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Modification of merB and Cloning for Expression

A 21-amino acid plant signal sequence (Met Gly Lys Met Ala Ser Leu Phe Ala Thr Phe Leu Val Val Leu Val Ser Leu Ser Leu Ala) derived from the Nicotiana plumbaginifolia ext (extensin) gene (De Loose et al., 1991) efficiently directs extensin to the ER and through to the CW. These 21 codons were added de novo to the 5' end of merB using two rounds of nested PCR with primers positioned as diagrammed in Figure 1A. Two sense (S) primers, ext-1S (5'AGTGTCACTTAGCTTAGCTATGAAGCTCGC CCCATATATTTTAGAACTTCTCACTTCGGTCAATCGT) and ext-2S (5'ACGGTCGGATCCTAAGGAGGAGAAACCATGGGAAAAATGGCTTCTCTATTTGC CACATTTTTAGTGGTTTTAGTGTCACTTAGCTTAGCT), encoded the N-terminal signal sequence and an in-frame stop codon, a Shine-Delgarno sequence for bacterial expression, and sequence of six nucleotides typical of plant translation signals in front of the initiation codon. Ext-1S also contains 18 nucleotides at its 3' end homologous to the merB N-terminal sequence. The ext-2S primer contains restriction sites and its 3' end overlaps the 5' end of ext-1S by 18 nucleotides for the second round of PCR. Two antisense (A) primers, 1A (5'AGTATCCTCGAGGAATTCAAGCTTAT CAGATATCCGGTGTCCTAGATGACATGGTCTGCAACAGATGTCGATTAAACT) and 2A (5'AGTATCCTCGAGGAATTCAAGCTTATCAGATATCTAGCTCATCTTTCTCAGACG GTGTCCTAGATGACATGGTCTGCAACAGATGTCGATTAAACT), add the 3' end sequences to CW-merB and ER-merB, respectively. Both were designed to overlap the 3' end of merB by 18 nucleotides and contained suitable restriction sites for cloning. Primer 2A also contains added codons for the ER retention sequence, (SE)KDEL, immediately in front of the stop codon.

After two rounds of nested PCR, the appropriately sized fragments for each modified merB gene were gel purified, digested with BamHI and XhoI, and ligated into digested pBluescript SKII⁺ plasmids. The BamHI/XhoI fragment from the WT merB gene was similarly subcloned (Bizily et al., 1999). These plasmids, named pBS-CW-merB and pBS-ER-merB, were transformed into Top10F' competent Escherichia coli (Invitrogen, Inc., Carlsbad, CA) by electroporation and their sequences were confirmed. Plasmids containing WT and modified merB gene sequences for plant expression, pBS-35S/CW-merB/NOS and pBS-35S/ER-merB/NOS, were prepared by subcloning the same BamHI/XhoI fragments into the multilinker replacement region of pBS-VST1. To make pBS-VST1, the 35S cauliflower mosaic virus promoter and NOS 3' terminator from the plant binary vector pVST1 (Malik and Wahab, 1993) were subcloned into the SacI-KpnI replacement region of Bluescript SK^II+. The entire promoter, modified merB coding sequence, and terminator fragments were then subcloned into the corresponding regions of pCAMBIA (Hajdukiewicz et al., 1994) vector by digesting at an SacI site upstream of the promoter and a KpnI site downstream of the NOS terminator. pCAMBIA-35S/CW-merB/NOS and pCAMBIA-35S/ER-merB/NOS were used for Agrobacterium tumefaciens-based plant transformations. The pCAMBIA vector confers kanamycin resistance and hygromycin resistance for bacterial and plant selections, respectively.

Assays of Bacterially Expressed MerB

Bacterial strains were tested for resistance to PMA using a disc assay described previously (Bizily et al., 1999). Each paper disc contained 2 µL of 100 mM PMA and the diameter of the zone of clearing was measured from the plate after 16 h of incubation at 37°C. Plasmid pDU202 encodes the narrow spectrum mer operon with several genes for drug and mercury resistance (MerA). E. coli strain pDU202/SK1592 (Foster, 1983; Hamlett et al., 1992; no. 890), which is resistant to inorganic mercury but lacks MerB, was transformed with pBS-CW-merB and pBS-ER-merB. The zone of clearance (average diameter) and SEs presented were calculated from at least six repetitions. Repetitions within experiments on a single strain varied by no more than 0.5 mm. Bacterial protein was isolated from these strains, as well as from positive (pBS-merB) and negative (pBSKSII) control strains, in the form of crude extract by sonication (Bizily et al., 1999) and mixed immediately with 2× sample buffer (4% [w/v] SDS, 125 mM Tris-HCl [pH 6.7], 30% [v/v] glycerol, 2% [v/v] -mercaptoethanol, and 0.002% [w/v] bromphenol blue; O'Farrell et al., 1977). Clarified protein extracts were resolved by 12.5% (w/v) PAGE and blotted. Protein loading was equalized based on Coomassie blue staining of pilot gels and confirmed by running samples in parallel along with those blotted for this experiment. PNPase levels were used as an independent assessment of the efficiency of equivalent loading (not shown). Mouse monoclonal antibodies mAb2H8 and mAb10E2 were used at 1 µg mL¹ concentrations, whereas rabbit polyclonal serum pAb-MerB was used at a 1:500 (v/v) dilution (Bizily et al., 2000). Primary reactions were carried out for 1 h after which the blots were washed 3 × 10 min. Secondary reactions with horseradish peroxidase-linked anti-mouse or anti-rabbit antibodies (Amersham Pharmacia Biotech, Piscataway, NJ) were performed at 1:2,000 (v/v) dilutions for 1 h. Blots were washed 3 × 10 min before development using a horseradish peroxidase luminol-based kit (Amersham Pharmacia Biotech). The blots were checked for equal loading by staining them directly with Coomassie blue after exposing them to film. Quantification of MerB bands on western blots was performed by scanning the exposed film in a densitometer loaded with the ImageQuant Software (Molecular Dynamics, Sunnyvale, CA).

MerB kinetics assays were performed using an assay solution containing 50 mM Tris-HCl (pH 7.5), 100 µM PCMB, and 1 mM L-Cys based on a modification of MerB assays described previously (Boyer, 1954; Begley et al., 1986). PCMB has an absorbance peak at 250 nm that shifts when MerB cleaves it into mercury and chlorobenzoate. Enzymatic rates were measured as A₂₅₀ absorbance with time. Protein extracts were prepared from 250 mL of Top 10F' E. coli cultures grown in Luria broth medium for 2 h to an optical density of approximately 0.3 at A₅₆₀, induced with 1 mM isopropylthio--galactoside, and grown for an additional 4 h. Cultures were spun down at 5,000 rpm for 15 min, washed with protein extraction buffer (50 mM Tris-HCl [pH 7.5], 1 mM -mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 0.5 M EDTA, and 5% [v/v] glycerol), and spun down again. Final cell pellets were resuspended in 2 mL of protein extraction buffer, sonicated for 15 s, and spun down at 15,000 rpm for 10 min. Twenty microliters of crude protein supernatant was added to and mixed with 1 mL of assay solution in a quartz cuvette to initiate an assay. A₂₅₀ readings were taken at 1-min intervals over 20 min and compared against a blank containing 1 mL of assay solution. Four independent assays were averaged for each value presented in Table I.

Analysis of Transgenic Plants

MerA Arabidopsis RLD plants (line A9-1A) already expressing mercuric ion reductase (Rugh et al., 1996) were grown up and transformed using vacuum infiltration to deliver A. tumefaciens carrying the various gene constructs (Bariola et al., 1999). T₁ progeny were selected from the seed population for hygromycin resistance and allowed to self fertilize. T₂ plants were used for all experiments described in this paper.

Organic mercury (PMA) resistance assays on seedlings and plants were described previously (Bizily et al., 2000). Protein extracts and western blots were prepared as described previously for MerB (Bizily et al., 2000) or with a direct SDS protein extraction procedure (McLean et al., 1990). Equal amounts of protein for all CW-merB and ER-merB lines and merA negative control plant lines were resolved on SDS-PAGE, transferred to membranes, and reacted with antibodies. Measurements of PMA to Hg(0) conversion rates were done as previously described with the exception that samples consisted of 10 2-week-old seedlings rather than individual plants (Bizily et al., 2000).

Immunocytochemistry

Seedlings from the AB-1, CW-2, and ER-2 lines were grown for 2 to 3 weeks on agar plates and prepared by rapid freeze fixation and freeze substitution for immunofluorescence microscopy as described previously (Kandasamy et al., 1999). The dissociated cells were labeled with pAb-MerB primary antibody at 1:100 (v/v) dilution for 16 h. Cells were washed 3 × 15 min in phosphate-buffered saline (PBS) and labeled with Texas red-conjugated goat anti-rabbit Ig secondary antibody (Amersham Life Science, Cleveland) at 1:100 (v/v) dilution for 3 h. After washing again in PBS (3 × 15 min) the samples were mounted in 90% (v/v) glycerol in PBS containing 0.1% (w/v) para-phenylenediamene and visualized with an MCR-600 confocal microscope (Bio-Rad, Hercules, CA).