1. Introduction

      The last three decades have seen a mounting realization of the impact of trace gas emissions on atmospheric chemistry, and an attendant increase in the scientific interest in these emissions. Organo-halogen and -sulfur gases are of particular importance in this regard, the former because of their effects on the depletion of stratospheric ozone (Prather and Watson, 1990), and the latter for their contribution to the formation of acid precipitation (Nriagu et al., 1987). Chlorofluorocarbons (CFC), a group of synthetic halocarbons, are thought to be the principal cause of ozone destruction in the stratosphere (Prather et al., 1984; Prather and Watson, 1990). This severity of the impact of CFC is mainly because they have long atmospheric lives of several years to many centuries (Prather and Watson, 1990). Volatile sulfur compounds originating from industrial activities are the main agents responsible for acid precipitation (Wine et al., 1981).

      Although human activities are apparently responsible for perturbing the balance of the above atmospheric components, their basal concentrations in the atmosphere are primarily determined by natural activities. For example, yearly biogenic emissions of up to 7 Mt of monohalomethanes (CH3X) [chloromethane (CH3Cl), bromomethane (CH3Br) and iodomethane (CH3I)] represent a considerable fraction of the total input of halogens to the atmosphere (Harper, 1993). Similarly, reduced sulfur compounds of biogenic origin, such as dimethyl sulfide (CH3SCH3), methanethiol (CH3SH) and hydrogen sulfide (H2S), are major carriers of sulfur in the environment (Cullis and Hirschler, 1980). For example, CH3SCH3 alone, for which reliable estimates have been made, accounts for about 50% of the total biogenic sulfur in the atmosphere. Consequently, attempts to understand the complex processes affecting the chemical transformations in the atmosphere must include an understanding of the physiological and biochemical processes controlling exchange of the above gases between the biosphere and the atmosphere.

      Marine organisms are the principal source of biogenic emissions of halomethanes and reduced sulfur gases (Aneja and Cooper, 1989; Lovelock, 1975; Lovelock et al., 1972; Rasmussen, 1974; Rasmussen et al., 1980). Terrestrial sources include bacteria, fungi and a few higher plants (Harper, 1993; Rennenberg, 1991). Despite the diversity of sources of CH3X and reduced sulfur gases, and their widespread occurrence in the environment, very little is known about their metabolic origins and physiological relevance to the organisms that produce them. Enzymatic synthesis of CH3I, supposedly through I- methylation in marine algae, followed by a spontaneous replacement of I- with Br- or Cl- in seawater was thought to be the main route for the synthesis of the CH3X (Zafiriou, 1975). Another mechanism involving a haloperoxidase-mediated reaction was later proposed for the synthesis of CH3X (Neideleman and Geigert, 1983). However, no evidence for this has ever been given, although haloperoxidases have been found in algae, plants, and animals (Neideleman and Geigert, 1983). These haloperoxidases catalyze X- incorporation into two-carbon metabolites, but not into one-carbon compounds like methane (CH4) to produce CH3X (Harper, 1993). Recently, an S-adenosyl-L-methionine (AdoMet)-dependent methyl chloride transferase (MCT), that could methylate halide ions (X-) to CH3X, was detected in a marine alga, a wood-rotting fungus and a halophytic higher plant (Wuosmaa and Hager, 1990). The presence of this enzyme activity in two organisms that grow in highly saline environments suggested that it might be involved in salt tolerance. This speculation triggered further work in our laboratory to determine the presence of such an enzyme in higher plants (Saini et al., 1995). Using I- as the marker halide substrate, a halide methyltransferase (HMT) activity was detected in 87 of the 118 plants tested, including many salt tolerant species. However, the activity did not correlate with the relative salt tolerance of the species, nor was it induced by salinization of selected species under controlled conditions (Saini et al., 1995). Instead, the highest activities were detected in the Brassicaceae and Resedaceae families of the order Capparales. Leaf disks or extracts of plants that contained this enzyme activity were also able to methylate bisulfide ions (HS-) to methanethiol (CH3SH) (Saini et al., 1995). Preliminary biochemical characterization of these two enzyme activities from Brassica oleracea suggested that both X- and HS- methylations may be catalyzed by the same enzyme, which was tentatively named halide/bisulfide methyltransferase (H/BMT) (Attieh, 1993). The enzyme was partially purified from cabbage. At this point, nothing was known about the natural metabolic and physiological roles of this enzyme, except that it could methylate X- and HS- ions. Therefore, the present work was undertaken with an aim to elucidate the role of H/BMT in plant metabolism. This research comprised the following milestones:

  1. Purification of a protein capable of methylating X- and HS- ions, and demonstration that these methylations are catalyzed on the same active site of a single protein, H/BMT.
  2. Purification and characterization of several isoforms of the enzyme.
  3. Determination of the subcellular localization, kinetic properties, and physical attributes of H/BMT.
  4. Determination of the substrate specificity of H/BMT, which showed that thiocyanate, organic thiols, and HS- are the probable natural substrates of the enzyme.

2. Review Of Literature

      The enzyme studied in the present work was first suspected to catalyze the synthesis of environmentally important gases, CH3X. It was later found that it also produced CH3SH, an equally important trace gas emitted by living organisms. We undertook the purification of this H/BMT from Brassica oleracea, assuming that its natural substrates are X- and HS- ions. However, after elaborate biochemical studies, we concluded that the enzyme was in fact a thiol methyltransferase involved in glucosinolate metabolism. The enzyme probably detoxifies reactive thiocyanate and thiol groups that are produced upon glucosinolate degradation.

      The halide/bisulfide methyltransferase is a novel enzyme. The products of various reactions it can catalyze have diverse effects, ranging from impacts on atmospheric chemistry to toxicity to cellular components. This review presents a synthesis of the knowledge about different processes in which the products of H/BMT could be involved. The review starts with a comprehensive discussion of CH3X and CH3SH, the original reason for embarking on this research. Special attention is given to the enzymatic basis for the synthesis and physiological significance of these gases. The literature on the sources, effects, and biochemical origins of these two distinct classes of gases is discussed in separate sections. Sulfide is the final product of the sulfate reduction pathway in plants. It is also produced upon degradation of the sulfur-containing secondary metabolites, glucosinolates (Brown and Morra, 1997). Glucosinolates are found in plants of the family Brassicaceae (Kjaer, 1966), whose members can methylate HS- to CH3SH at high rates (Saini et al., 1995). Therefore, given the relationship of sulfate reduction and glucosinolate metabolism to the subject of this thesis, a section of the review is devoted to the current understanding of the biochemistry of these two sectors of sulfur metabolism.


2.1. Monohalomethanes

      Monohalomethanes are Cl-, Br- or I-containing trace gases that are generated by natural and anthropogenic sources (Lovelock, 1975). These gases play important roles in the environment because they are the primary carriers of halides in the atmosphere (Singh et al., 1983). They, along with CFCs, participate in the degradation of stratospheric ozone (Lovelock, 1975; Prather et al., 1984; Prather and Watson, 1990).


2.1.1. Sources Of Ch3x

      It is generally accepted that the atmospheric CH3X concentrations over the northern hemisphere represent a mixture of anthropogenic and natural emissions, whereas those over the southern hemisphere arise mainly from natural sources. Hence, measurements of interhemispheric concentra-tions have been used to assess the extent of natural versus anthropogenic contribution to the total atmospheric budget of halomethanes (Penkett et al., 1985). The analysis of the differences between the concentrations over the two hemispheres indicates that natural processes contribute a significant, often dominant, part of total halomethane emissions. For example, natural emissions contribute up to 5 Mt y-1 of CH3Cl, the most abundant halomethane in the atmosphere (Rasmussen et al., 1980; Singh et al., 1983). By comparison, the average yearly anthropogenic production of CH3Cl of 30 Kt is negligible (Harper, 1985). Natural sources of CH3Br and CH3I also account for well over half the total atmospheric burden of these gases (Harper, 1985; Penkett et al., 1985; Rasmussen et al., 1982; Rasmussen et al., 1980). The natural emissions of CH3X arise from oceans, volcanic activity, and forest fires (Palmer, 1976; Rasmussen et al., 1980). Measurements of CH3X concentration in seawater show that these gases originate primarily from the high biological activity in the oceans. The concentration of CH3I in the water surrounding beds of the large kelp Laminaria digitata was approximately one thousand times higher than in the open sea (Lovelock, 1975). Furthermore, the highest concentration of CH3Cl has been recorded in the atmosphere above the equatorial regions of the oceans, where biological activity is the highest (Rasmussen et al., 1980). Identification of a number of marine macroalgae and microorganisms as CH3X producers (Manley and Dastoor, 1987; Pedersén et al., 1996; Scarratt and Moore, 1996; Wuosmaa and Hager, 1990) has confirmed the above conclusion from the correlations between concentrations and biomass. In view of these observations and the large surface area of oceans, the latter appear to be the dominant source of CH3X (Lovelock, 1975; Singh et al., 1983). Terrestrial organisms also contribute to CH3X biosynthesis. Many species of Fomes, a widely distributed white-rot fungus, produce CH3Cl, presumably as an allelopathic agent against plants, bacteria and other fungi (Cowan et al., 1973). Phellinus pomaceus, also a white-rot fungus, emits CH3Cl, CH3Br, and CH3I in the presence of the corresponding X- ion (Harper, 1985; White, 1982b). Among higher plants, freshly harvested potato tubers produce appreciable amounts of CH3Cl during suberization (Varns, 1982). Whole cells and crude extracts of the ice plant, Mesembryanthemum crystallinum, emit CH3Cl in the presence of Cl- ions (Wuosmaa and Hager, 1990). Recently, the ability to produce CH3I from I- ions was shown in a survey of a wide variety of herbaceous plant species (Saini et al., 1995).


2.1.2. Biosynthesis Of Ch3x

      Although considerable advances have been made in the knowledge of the effects of halomethanes on the chemical integrity of the atmosphere, research to understand the nature of the reactions governing their biosynthesis has lagged behind. A known biological route for the formation of halometabolites is a haloperoxidase-mediated reaction in the presence of hydrogen peroxide (Neideleman and Geigert, 1983).

      
Carbon substrate + X-+ H++ H2O2 ^ halometabolite + 2H2O (1)

      The carbon substrate may be an alkyne, cyclopropane, phenol, or aniline. The X- ion can be I-, Br-, Cl-, or even the pseudohalide thiocyanate ion (SCN-). It is obvious from the nature of the carbon substrates that for this incorporation to take place, an intermediate with two or more carbons must be produced. This intermediate is then decomposed to a monohalomethane (Harper, 1993). Such a mechanism has been invoked by Theiler and co-workers (1978), but only for the production of polyhalogenated methanes. These workers demonstrated that a bromoperoxidase enzyme isolated from the tropical red alga, Bonnemaisonia hamifera, catalyzed the incorporation of Br- ions into a number of organic compounds such as ß-keto acids. The pH of the reaction medium was important in determining the nature of the final products. In the presence of H2O2, Br-, and 3-oxooctanoic acid (CH3(CH2)4COCH2COOH) at pH 5.8, the major reaction products were 1-bromo-, 1,1-dibromo-, and 1,1,1-tribromoheptanone.

      
CH3(CH2)4COCH2COOH + Br- + H2O2 ^ CH3(CH2)4COCH2Br + CO2 (2)
CH3(CH2)4COCH2Br + Br- + H2O2 ^ CH3(CH2)4COCHBr2 (3)
CH3(CH2)4COCHBr2+ Br- + H2O2 ^ CH3(CH2)4COCBr3 (4)

      At pH 7.3, CH2Br2, CHBr3, and 1-bromopentane were formed in high amounts (Theiler et al., 1978). This observation was explained by the multiple bromination of oxooctanoic acid into the above brominated ketones, which at pH 7.3 undergo spontaneous or enzymatic degradation to give bromoalkanes.

      
CH3(CH2)4COCHBr2 ^ CH3(CH2)4COOH + CH2Br2 (5)
CH3(CH2)4COCBr3 ^ CH3(CH2)4COOH + CHBr3 (6)

      Based on these findings, Theiler et al. (1978) suggested that bromo- and iodo-methanes found in seawater and in algae are the products of enzyme-catalyzed halogenation of ketones. However, the production of monohalomethanes through such a process has never been demonstrated. This is despite the fact that the production of a large number of organohalogen compounds from the haloperoxidase route in marine organisms has been reported (Gschwend et al., 1985; Neideleman and Geigert, 1983; Pedersén et al., 1996).

      A logical alternative route for CH3X biosynthesis is a direct addition of a methyl group to an X- ion. The first suggestion of such a mechanism was put forward for the production of CH3I from marine algae (Lovelock et al., 1973). Once released into the seawater, CH3I was thought to undergo spontaneous transformation to more stable CH3Cl and CH3Br through the following reactions (Zafiriou, 1975):

      
CH3I + Cl- ^ CH3Cl + I- (7)
CH3I + Br- ^ CH3Br + I- (8)
CH3Br + Cl- ^ CH3Cl + Br- (9)

      Based on relative reactivities of Cl- (9.4), Br- (0.12) and I- (10-5) in seawater, CH3I was predicted to react predominantly by reaction 7. This prediction was supported by experimental evidence when CH3I, added to a saturated NaCl solution, reacted to completion with NaCl, liberating I- ions and CH3Cl (Zafiriou, 1975). However, Singh and co-workers (1983) found much fluctuation in the concentration of CH3Cl and CH3I in seawater. In different parts of the ocean, high levels of CH3I coexisted with relatively low levels of CH3Cl and vice versa. This indicated that there were independent oceanic sources for the two gases. The only demonstration for the independent sources of CH3Cl and CH3I came from the work of White (1982b) with a terrestrial organism, the fungus P. pomaceus, using radiolabeled metabolites. He also presented evidence for the involvement of methionine as the precursor of the methyl group in the formation of both CH3Cl and CH3I. Without any experimental evidence, White speculated that CH3Cl may be spontaneously formed by reversion of a sulfonium compound, such as AdoMet, in the presence of a halide, to an alkyl halide and a sulfide. A known analogous chemical transformation occurs upon warming a solution containing S-methylmethionine (SMM, a simple chemical homologue of AdoMet) and Cl-, resulting in the formation of CH3Cl and CH3SCH3 (Stirling, 1977). However, this speculation was soon ruled out with the demonstration of an enzymatic basis for the biosynthesis of all CH3X via a high affinity methylating system in the fungus P. pomaceus (Harper, 1985). A direct experimental proof for the existence of such a system was recently given through the report of an AdoMet-dependent MCT in the fungus P. pomaceus, a marine red alga Endocladia muricata, and a halophytic higher plant Mesembryanthemum crystallinum (Wuosmaa and Hager, 1990). The enzyme catalyzed the following reaction:

      
X- + AdoMet ^ CH3X + S-adenosylhomocysteine (AdoHcy) (10)

      Only the algal enzyme was partially purified and further characterized. Work in our laboratory detected a similar halide methyltransferase enzyme activity in a number of higher plants (Saini et al., 1995). Subsequently, an enzyme that could methylate X- ions was partially purified and characterized from cabbage (Attieh, 1993). The MCT from the fungus P. pomaceus was also recently characterized (Saxena et al., 1998). Thus, three halide methylating enzymes have been studied, one each from an alga, a fungus and a higher plant.


2.1.2.1. Halide Methylation In Algae

      The algal MCT enzyme was partially purified 800-fold from extracts of E. muricata (Wuosmaa and Hager, 1990). The purification involved an 80 to 100% (NH4)2SO4 precipitation step, followed by gel filtration high performance liquid chromatography (HPLC). The enzyme had a native molecular mass between 20 and 25 kDa, a sharp pH optimum at 7.5-7.6 and no prosthetic group or cofactor requirement. Its Michaelis-Menten constants (Km) for Br-, Cl- and AdoMet were 40 mM, 5 mM and 16 µM, respectively. The reactivity of MCT with different X- ions decreased in the order I-, Br-, and Cl-. Crude extract of the alga methylated I- at a rate which was 135-and 270-fold higher than for Br- and Cl-, respectively.


2.1.2.2. Halide Methylation In Fungi

      White (1982b) demonstrated, using L-(CH3-2H3)methionine, that this amino acid was the source of the methyl group in CH3Cl produced by cultures of P. pomaceus. Harper and Hamilton (1988b), also using whole cultures of P. pomaceus, later provided evidence for the involvement of methionine in the synthesis of CH3Cl through a transmethylation reaction. However, they were unable to detect a chloride methyltransferase activity in the fungal cell extracts (Harper and Hamilton, 1988a). Wuosmaa and Hager (1990), on the other hand, did detect this activity in P. pomaceus, albeit at the very low level of 25 fmol min-1 mg-1 protein. These workers did not further characterize the MCT activity from P. pomaceus.

      Harper (1985) had indicated that the high efficiency Cl--methylating system in P. pomaceus may be useful in biotechnology to improve crop salt tolerance, thus triggering interest in isolating the fungal MCT (Saxena et al., 1998). Cultures of P. pomaceus, grown on a liquid glucose mycological peptone (GMP) medium containing 18 mM KCl, produced 325 nmol day-1 g-1 FW CH3Cl (Saxena et al., 1998). This rate was 46,000 fold greater than that recorded for the same organism grown on 5% malt extract and 100 mM KCl (Wuosmaa and Hager, 1990), and approximately 800 fold higher than those observed from cabbage leaves (Saini et al., 1995) and the kelp Macrocystis pyrifera (Manley and Dastoor, 1987). This placed P. pomaceus far ahead of any other organism in terms of X- methylation efficiency, thus corroborating Harper's suggestion (Harper, 1985).

      Attempts to extract the enzyme from P. pomaceus showed that it was membrane-bound and highly labile to mechanical damage (Saxena et al., 1998). Addition of any one of a number of common detergents to solubilize membrane proteins, led to a total loss of enzyme activity. A membrane fraction, prepared by centrifuging the crude extract at low-speed and desalting the supernatant on Sephadex G-25 PD-10 columns, was used to further characterize this enzyme. The enzyme had a sharp pH optimum between 7 and 7.2, and a high calculated Km value of 300 mM for Cl-. The extremely high efficiency of in vivo CH3Cl production from P. pomaceus even at low Cl- concentrations, made it unlikely that this value reflected the real Km for this substrate. This was interpreted as an indication that the CH3Cl produced is rapidly channeled by a coupled reaction since this product was known to act as methyl donor in the methylation of carboxylic acids and phenols in this fungus (McNally et al., 1990). This interpretation is also indirectly supported by the fact that the Km values for the methylation of Br- and I- were comparatively low at 11 mM and 250 µM, respectively (Saxena et al., 1998); the products of these reactions, CH3Br and CH3I, are much poorer methyl donors for the coupled reactions (Harper et al., 1989), and would not be drained as quickly as CH3Cl. This MCT exhibited a low Km of 4.5 µM for the methyl donor AdoMet (Saxena et al., 1998).


2.1.2.3. Halide Methylation In Higher Plants

      The presence of MCT activity in a halophyte and a marine alga, both of which grow in highly saline environments, suggested that this enzyme might be involved in salt tolerance via volatilization of Cl- to CH3Cl. The presence of a halide methylating activity in higher plants and its involvement in Cl- detoxification were further examined in a survey of 118 plants, including 21 that are known to be halophytes or salt tolerant (Saini et al., 1995). Using an in vivo leaf-disc assay with 100 mM KI, the halide methylation ability was detected in 87 species representing 44 families and 33 orders. Contrary to the expectation, most of the salt tolerant plants had relatively low CH3I emission rates. Moreover, salinization of three of these plants, including M. crystallinum in which the MCT was first detected (Wuosmaa and Hager, 1990), did not cause any increase in X- methylation. These findings strongly argued against the hypothesis that X- methylation was a Cl- tolerance mechanism in higher plants. Instead, there was a clear taxonomic trend in the distribution of the activity; the highest CH3I production rates were recorded in plants from the order Capparales, represented by members of the families Brassicaceae and Resedaceae. This extended the scope of the possible role of the X- methylating activity in higher plants, and highlighted the need for a thorough investigation of the biochemical nature of this enzyme. Since leaf extracts of red cabbage had one of the highest rates of X- methylation, this plant was chosen for further experimentation and purification of the X- methyltransferase. The enzyme was partially purified approximately 800-fold from Brassica oleracea cv. April Red by (NH4)2SO4 precipitation between 60 and 85% saturation, followed by gel filtration chromatography on Sephadex G-100 and affinity chromatography on adenosine-agarose (Attieh, 1993). The enzyme had a native molecular mass of 29.5 kDa as determined by gel filtration, and a pI of 4.8. Kinetic characterization of the enzyme preparation purified through the gel filtration step, suggested a sequential substrate binding mechanism and gave Km values of 1.3 mM, 29 mM, 85 mM and 30 µM for I-, Br-, Cl- and AdoMet, respectively. The enzyme exhibited the same order of preference for the substrates as the algal MCT. Its specific activity of 139 nmol min-1 mg-1 protein for methylation of I- was 16-fold higher than that for Br- and over 2500 fold higher than that for Cl-.


2.1.3. Environmental And Physiological Significance Of Ch3x

      Much work in recent years has been focused on assessing the effects of various forms of human interference on the chemical equilibrium of the atmosphere. Halogenated organic compounds are biologically recalcitrant in terrestrial ecosystems, presumably because terrestrial organisms have not evolved with them and, consequently, have not developed metabolic means to degrade them (Gschwend et al., 1985). Increasing atmospheric concentrations of CFCs and other halocarbons are thought to be the main cause of the appearance of the Antarctic ozone hole in the late 1970s, and the more modest ozone depletion observed over the Northern Hemisphere (Prather and Watson, 1990). However, natural halocarbons, particularly CH3X, are also important halogen-carriers in the atmosphere (Chameides and Davis, 1980; Lovelock, 1975; Manley and Dastoor, 1987; Rasmussen et al., 1980). This is especially true for CH3Cl, which, despite its short atmospheric life (1.2 years), is capable of reaching the stratosphere at concentrations high enough to make it the principal Cl carrier in the upper atmosphere (Lovelock, 1975). There, it competes with the major CFCs -- CF2Cl2 and CCl3F -- as a source of Cl atoms (Rasmussen et al., 1980). CH3Cl undergoes photolysis in the presence of OH radicals to form chlorine monoxide (ClO) and atomic Cl, each of which plays a major role in the conversion of stratospheric O3 to O2 (Lovelock, 1975; Prather and Watson, 1990; Stolarski and Cicerone, 1974) through the following reactions:

      
Cl + O3 ^ ClO + O2 (11)
ClO + O ^ Cl + O2 (12)

      The net reaction being, O3 + O ^ 2O2

      Natural sources of CH3Br account for over 35% of the total Br burden in the atmosphere (Prather et al., 1984). This gas also has a short atmospheric lifetime of about 1.5 to 3 years, but could still contribute Br and BrO to the atmosphere, which destroy O3 as follows:

      
Br + O3 ^ BrO + O2 (13)
BrO + O ^ Br + O2 (14)

      The net reaction being, O3 + O ^ 2O2

      In the presence of both halogens, O3 destruction becomes even more efficient:

      
Cl + O3 ^ ClO + O2 (15)
Br + O3 ^ BrO + O2 (16)
ClO + BrO ^ Cl + Br + O2 (17)

      The net reaction being, 2O3 ^ 3O2

      Despite abundant natural emission of CH3I, its ability to contribute significantly to ozone destruction is considered minor because of its very short atmospheric life of 5 days (Rasmussen et al., 1982). Instead, CH3I plays a major role in the mobilization and transport of heavy metals in the environment (Brinkman et al., 1985).

      It has been speculated that, because halomethanes can cause mutations in bacteria, they may have played a role similar to that of radiation in the biological evolution (Singh et al., 1983). Lovelock (1975) suggested that the significance of atomic halogens and their monoxides in the atmosphere could represent an active process involved in the maintenance of an atmospheric equilibrium. In this context, the large-scale biosynthesis of CH3Cl and other CH3X could be a natural process for ozone depletion. This author introduced the interesting theory that the biosynthesis of these compounds could respond to some function of stratospheric ozone density and act as a regulator of the thickness of the ozone layer (Lovelock, 1975).

      The metabolic function of natural volatile organo-halogens in the organisms that produce them is generally poorly understood and subject of speculations. The only exception is the established role of CH3Cl in the metabolism of a number of wood-rotting fungi (Harper, 1993). CH3Cl is a methyl donor in the biosynthesis of methyl esters of benzoic and furoic acids in P. pomaceus (Harper et al., 1989), and in the synthesis of veratryl alcohol in Phanerochaete crysosporium and other lignin-degrading fungi (Harper et al., 1990). Veratryl alcohol is a secondary metabolite that induces the ligninolytic system and prevents inactivation of lignin peroxidase in wood-degrading fungi. Among higher plants, freshly harvested potato tubers produce CH3Cl during their suberization (Varns, 1982). The origin of this CH3Cl is not known. Varns (1982) had suggested that CH3Cl could be concentrated by the tubers from the surrounding air, then released upon damage to the tissue. However, Harper (1993), in disagreement with this explanation, proposed that CH3Cl is produced as a metabolic intermediate or a byproduct of the biosynthesis of some other, unidentified metabolite. The succulent ice plant, M. crystallinum, which grows abundantly in highly saline soils, also produces CH3Cl upon incubation with 100 mM KCl (Wuosmaa and Hager, 1990). These authors suggested that CH3Cl emission may be a salt tolerance mechanism in plants, and emphasized the need to confirm this through a broader survey of CH3Cl emission from other halophytes. The significance to algae of the large quantities of halomethanes they produce also remains obscure. Gschwend and co-workers (1985) reported that halomethanes could be formed by nonspecific reactions of some haloperoxidases or other oxidases. They also found that some algal species avoided by the herbivore snail Littorina littorea contain diiodomethane which, when added to macerated algae, deterred feeding by the snail. Thus, halomethanes could act as protectants from herbivore feeding.


2.2. Reduced Sulfur Gases

      The chemistry of sulfur compounds in the environment has taken on a new significance since the discovery of their involvement in the formation of atmospheric aerosols and acid precipitation. Sulfur emission from natural sources has been a well-known process for over a century. However, the identities, quantities, and metabolic origins of compounds in these emissions were determined only in the last three decades. Atmospheric sulfur originates from both anthropogenic and natural sources. Burning of fossil fuels, refining of petroleum and smelting of ores produce large amounts of sulfur dioxide (SO2), which is a major anthropogenic sulfur carrier to the atmosphere (De Kok, 1990). On the other hand, biological reduction of sulfur compounds is generally believed to be the primary natural source of atmospheric sulfur (Aneja and Cooper, 1989). Reduced sulfur gases, such as CH3SCH3, H2S, CH3SH and CH3SSCH3, are major representatives of sulfur from natural origin. For the purpose of the present text, the origins, biosynthesis, and significance of CH3SH emissions will be discussed in detail, whereas CH3SCH3 and H2S will only be treated in relation to their effects on the synthesis and metabolism of CH3SH. Further information on CH3SCH3 and H2S can be obtained in several reviews (Cooper and Matrai, 1989; Dunnette, 1989; Keller et al., 1989; Rennenberg, 1989; Saltzman and Cooper, 1989; Taylor and Kiene, 1989).


2.2.1. Origins Of Biogenic Ch3sh

      All models for the natural cycle of sulfur in the environment require some volatile or gaseous compound as a vehicle for sulfur transfer from the sea through the air to the land. Originally, H2S was thought to be this vehicle (Eriksson, 1963). Theoretically, an atmospheric concentration of H2S averaging 0.2 parts per billion (ppb) was considered satisfactory for the mass transfer needs of the models (Lovelock et al., 1972). However, no measurements ever detected such high concentrations of H2S (Lovelock et al., 1972). More importantly, the ocean surface waters are far too oxidizing to permit the development of such concentrations (Rasmussen, 1974). Since these concentrations are indeed necessary to sustain an atmospheric equilibrium of sulfur, the attention turned toward other possible sulfur carriers to the atmosphere. Volatile methyl derivatives of sulfur have attracted much attention in this regard, and it is now established that CH3SCH3 and CH3SH surpass H2S as the major volatile, biogenic intermediates in the biogeochemical cycling of sulfur (Kelly et al., 1994; Kiene, 1996; Rasmussen, 1974). Methylated sulfur gases are produced by a variety of living organisms, and are also liberated upon breakdown of sulfur-containing organic compounds such as dimethylsulfoniopropionate (DMSP) and methionine (Kiene and Capone, 1988). It has been widely accepted that CH3SCH3 is the principal methylated sulfur gas transferred to the environment (Calhoun and Bates, 1989; Kelly et al., 1994; Lovelock et al., 1972). Recent studies show that CH3SH is also a major product of DMSP degradation in surface seawater, sometimes in amounts comparable to those of CH3SCH3 (Kiene, 1996; van der Maarel and Hansen, 1997; Visscher and Taylor, 1993).

      Slurries of anoxic sediment from Spartina salt marsh contained endogenous CH3SCH3 which was volatilized in the same form for 24 h during incubation under anaerobic conditions (Kiene, 1988). CH3SCH3 production decreased to undetectable levels after this initial period, and was replaced by CH3SH evolution. Presence of molybdate severely inhibited CH3SH production, indicating that anaerobic sulfate-reducing bacteria were responsible for its synthesis. S-methylcysteine, cysteine, methionine, and DMSP, added to the sediments, were also readily metabolized by sediment microflora (Kiene and Capone, 1988). CH3SH was the dominant methyl sulfur product of the degradation of both methionine and S-methylcysteine. DMSP yielded mostly CH3SCH3, and cysteine degradation did not produce any of the above species. Aerobic pathways for the degradation of CH3SCH3 also include CH3SH as an intermediate, but its rapid chemical oxidation to CH3SSCH3 makes it hard to detect (Kiene, 1988; Kiene and Capone, 1988; Zinder et al., 1977).

      CH3SH is the major volatile from decomposition of algal mats (Zinder et al., 1977). It is also produced by several species of the widely distributed bacterium, Pseudomonas (Rasmussen, 1974; Tanaka et al., 1976), actinomycetes and filamentous fungi (Segal and Starkey, 1969), and marine and freshwater algae (Drotar et al., 1987a; Drotar et al., 1987b; Rasmussen, 1974). A considerable proportion of heterotrophic bacteria isolated from soil, water, and vegetation produce CH3SH (Drotar et al., 1987a; McCarty et al., 1993). Several cheese bacteria, such as Lactococcus lactis also produce high amounts of CH3SH during the maturation of cheddar and Camambert cheese (Bruinenberg et al., 1997; Ferchichi et al., 1985; Kubickova and Grosch, 1997). CH3SH is also produced in mammalian digestive tracts (Weisiger et al., 1980).

      Among higher plants, CH3SH is produced by many species of the family Alliaceae (Saghir et al., 1966), and during decomposition of crucifers in soil (Lewis and Papavizas, 1970). CH3SH is also formed by cell suspension cultures of Catharanthus roseus (Schwenn et al., 1983), and emitted from leaf discs of Cucurbita pepo (Schmidt et al., 1985). The ability to produce CH3SH was recently detected in a number of other higher plants, among which members of the family Brassicaceae exhibited highest rates of CH3SH emissions (Saini et al., 1995).


2.2.2. Enzymatic Synthesis Of Ch3sh

      Two major routes for the formation of CH3SH in biological systems are well known. These are degradation of methionine (Kadota and Ishida, 1972; Segal and Starkey, 1969) and of CH3SCH3 or its precursor, DMSP (Kiene, 1988; Kiene, 1996). A third possible way for the formation of this compound is through the methylation of sulfide (S2-) (Drotar et al., 1987b; Saini et al., 1995).


2.2.2.1. Degradation Of Methionine

      The production of CH3SH from methionine in microorganisms has been extensively studied (Bruinenberg et al., 1997; Ferchichi et al., 1985; Kadota and Ishida, 1972; Ohigashi et al., 1951; Tanaka et al., 1976). The sequence of reactions leading from methionine to CH3SH in bacteria is as follows: methionine first is oxidatively a-deaminated to produce a-ketomethionine (a-keto-g-methylmercaptobutyric acid) and NH3. a-Ketomethionine is then g-demethiolated to yield CH3SH and a-ketobutyric acid (Segal and Starkey, 1969). Anaerobic incubation of bacteria greatly enhances CH3SH production (Ohigashi et al., 1951). An enzyme, methionine-a-deamino-g-mercaptomethane-lyase (methioninase), catalyzing the production of CH3SH through this route was first partially purified from acetone-dried cells of Escherichia coli (Ohigashi et al., 1951), then purified to homogeneity from Pseudomonas ovalis (Tanaka et al., 1976). The occurrence of methioninase has also been demonstrated in several other bacterial strains (Tanaka et al., 1976). The same enzyme catalyzes the a,g-elimination reactions of methionine and of its analogs, S-methylcysteine and cysteine, producing CH3SH in the former case and H2S in the latter. Methioninase requires pyridoxal phosphate for activity, has a molecular mass of 180 kDa, a pH optimum around 8.0, and a Km of 1.3 mM for L-methionine. In many fungi, methionine degradation gives a-ketomethionine and a-hydroxy-g-methylmercaptobutyric acid in the absence of a carbon source. It is only in the presence of glucose that all the sulfur from methionine is converted to CH3SH. This indicates that a source of energy is necessary for the demethiolation reaction, and that independent enzymes are probably responsible for deamination and demethiolation reactions in these fungi (Kadota and Ishida, 1972).

      Cell cultures of Catharanthus roseus produced CH3SH from DL-methionine through a reaction chain that was initiated by oxidative deamination of the amino acid (Schwenn et al., 1983). However, the demethiolation reaction was not investigated, and it is not known whether or not this reaction involves a separate g-methylmercapto-lyase activity. Schmidt and co-workers (1985) reported the presence of an inducible degradation system in pumpkin leaves, that specifically degrades L-methionine to CH3SH. Pretreatment of leaf extracts with L-methionine for 40 h enhanced the degrading activity by about 18-fold. In view of the similarities between this and the bacterial system, these workers suggested that a methioninase similar to the bacterial enzyme could be induced in pumpkin (Schmidt et al., 1985). However, no direct experimental proof exists to support this hypothesis.


2.2.2.2. Degradation Of Ch3sch3 And Dmsp

      Aerobic catabolism of CH3SCH3 has been described in hyphomicrobia (De Bont et al., 1981; Suylen et al., 1986) and thiobacilli (Visscher and Taylor, 1993 and references therein). In these organisms, at least two mechanisms exist for CH3SCH3 degradation. The first mechanism involves oxidation of CH3SCH3 by direct incorporation of molecular oxygen into its methyl groups through a reaction catalyzed by an NADH-dependent CH3SCH3 monooxygenase,

      
CH3SCH3 + O2 + NADH + H+ ^ CH3SH + CH2O + H2O + NAD+ (18)

      The second mechanism, shown below, operates via a methyltransferase and involves a cobalamin carrier:

      
CH3SCH3 + Carrier ^ CH3SH + CH3-Carrier (19)

      The carrier apparently transfers the methyl group to tetrahydrofolate, with subsequent oxidation of this bound CH3 unit to formic acid (Visscher and Taylor, 1993). Because molecular oxygen is not a substrate, this route can drive CH3SCH3 degradation under anaerobic conditions where alternative electron acceptors, such as NO3- or NO2-, are available. Chloroform (CHCl3) reacts with cobalamin carriers and thus inhibits cobalamin-dependent methyltransferases (Wood et al., 1968). CHCl3 strongly inhibits CH3SCH3 utilization in natural samples of seawater (Visscher and Taylor, 1993). This inhibition indicates that, in the sea environment, the dominant pathway for CH3SCH3 degradation operates through the methyltransferase and not the monooxygenase reaction.

      The pathway for CH3SH production from DMSP in marine surface water involves initial demethylation of this sulfonium compound to yield 3-methiolpropionate (MMPA), a fraction of which is demethiolated to give acrylate and CH3SH as the final sulfur-containing product (Kiene, 1996; van der Maarel and Hansen, 1997):

      
(CH3)2S+CH2CH2COO-^CH3SCH2CH2COO-^CH3SH + CH2=CH-COO- (20)

      Addition of 50 nM MMPA significantly stimulates CH3SH production, as does the addition of methionine. This indicates that demethiolation of these natural sulfur compounds may occur readily in the oceans (Kiene, 1996).


2.2.2.3. Direct Methylation Of Sulfide

      Reduction of sulfate and degradation of sulfur-containing amino acids, cysteine and homocysteine, by anaerobic bacteria, are two major sources of sulfide ions in several environments. Part of this sulfide is methylated to CH3SH by microbes in freshwater environments and in the ocean, and in animal digestive organs (Drotar et al., 1987a; Drotar et al., 1987b; Weisiger et al., 1980). Thiol methyltransferases responsible for this transformation have also been described from mammalian tissues (Weisiger and Jakoby, 1979). They catalyze the following reaction:

      
HS- + AdoMet + H+ ^ CH3SH + AdoHcy (21)

      No such reaction was known to occur in plants, until a similar AdoMet-dependent methylation of HS- ions to CH3SH was recently detected in a number of herbaceous species (Saini et al., 1995).


2.2.2.3.1. Mammalian s-Methyltransferase

      This microsomal enzyme was purified to homogeneity from rat liver (Weisiger and Jakoby, 1979). The enzyme was solubilized by repeated freeze-thaw cycles, and purified through precipitation with (NH4)2SO4, and the following chromatographic steps: anion exchange on DEAE-cellulose, two hydroxyapatite steps, isoelectric focusing and gel filtration on Sephadex G-75. The enzyme is a monomer of 28 kDa, has a pH optimum of 7.5 and Km values of 0.6 µM and 64 µM for AdoMet and HS-, respectively (Weisiger and Jakoby, 1979; Weisiger et al., 1980).


2.2.2.3.2. Microbial s-methyltransferases

      In a survey of a number of microorganisms, Drotar and co-workers (1987a) found that a majority of heterotrophic bacteria isolated from soil, water, sediment, vegetation, and cultures of marine algae methylated HS- to CH3SH. The reaction was catalyzed by thiol methyltransferase enzymes that used AdoMet as the methyl donor. Molecular masses of partially purified enzymes from two Pseudomonas strains were more than 10 kDa apart (Drotar et al., 1987a), indicating that there was probably considerable polymorphism among such methyltransferases. An enzyme with similar substrate specificity was partially purified from cell extracts of the freshwater protozoan Tetrahymena thermophila using gel filtration and anion exchange chromatography (Drotar et al., 1987b). It had a Km of 1.4 mM for HS-. No further information as to its structure and properties was published.


2.2.2.3.3. Plant s-methyltransferases

      Leaf extracts of cabbage and 19 other higher plants were able to catalyze the AdoMet-dependent methylation of HS- to CH3SH (Saini et al., 1995). The HS- methyltransferase enzyme, responsible for this reaction, was partially purified from cabbage leaves (Attieh, 1993). It had a Km of 4.7 mM for HS- and 226 µM for AdoMet, and a molecular mass of 29.5 kDa as determined by gel filtration chromatography on Sephadex G-100. The HS- methylating activity co-purified with the X- methylating activity from this plant in all the steps, listed earlier for the partial purification of the latter (see section 2.1.2.3. above). Furthermore, the methyl acceptor substrates X- and HS- exhibited mutual competitive inhibition in kinetic studies using a partially purified extract of cabbage. These results strongly suggested that both X- and HS- methylation activities existed on the same protein. Consequently, this putative dual-function enzyme was tentatively named halide/bisulfide methyltransferase (H/BMT) (Attieh, 1993).


2.2.3. Environmental And Physiological Significance Of Ch3sh Production

      Oxidation of reduced sulfur compounds of biogenic origin, such as CH3SH, CH3SCH3 and CH3SSCH3, is an important source of SO2, SO42- and methanesulfonic acid (MSA) in the troposphere (Plane, 1989). Global emissions of SO2, SO42-, and MSA from biogenic sources are comparable to anthropogenic emissions of SO2. Therefore, the former constitute a significant factor in the formation of acid precipitation and climate change (Charlson and Rodhe, 1982). Clouds of liquid water droplets can form only in the presence of cloud condensation nuclei (CCN) (Charlson et al., 1987). Most CCN are composed of water-soluble materials, and it is widely accepted that CCN in air over oceans consist of SO42- submicrometer particles (SO42- aerosols) (Charlson et al., 1987). The concentration of these CCN affects the concentration of cloud droplets, the radiative properties of clouds, and hence the climate system (Charlson et al., 1987). Thus SO42- has important impact on the global environment (Brasseur and Chatfield, 1991).

      The most important oxidants in the troposphere are the OH radical, the NO3 radical and various halogen species (Plane, 1989). CH3SH and OH undergo the following condensation reaction:

      
CH3SH + OH ^ CH3S(OH)H (22)

      The product is either thermally decomposed to yield the alkyl thiyl radical (CH3S) and H2O, or subjected to further attack by NO3 radicals to give CH3SNO. The latter is then rapidly photolyzed to give CH3S and NO (Plane, 1989). Thiyl radicals are very important intermediates in the oxidation of all reduced sulfur compounds. The CH3S further reacts with NO2 to form SO2 as shown below (Hatakeyama, 1989):

      
CH3S + NO2 ^ CH3SO + NO (23)
CH3SO + O2 ^ SO2 + CH2O (24)

      The resulting SO2 undergoes a series of oxidations in the gas phase, producing H2SO4 (Anderson et al., 1989):

      
SO2 + OH ^ HOSO2 (25)
HOSO2 + O2 ^ HO2 + SO3 (26)
SO3 + H2O ^ H2SO4 (27)

      CH3SH can also act as precursor for CH3SCH3 in anoxic sediments (Kiene and Capone, 1988; Kiene and Visscher, 1987), and in aerobic bacteria through methyl transfer reactions (Drotar et al., 1987a; Drotar et al., 1987b). CH3SCH3 is the major biogenic sulfur compound emitted by planktonic algae in marine environments (Kiene and Visscher, 1987; Lovelock et al., 1972). In the troposphere, it is rapidly oxidized by OH radicals to form sulfate aerosols (Charlson et al., 1987). Since these aerosols scatter sunlight back to space, they are an important factor in the global radiation budget and, thus, in climate determination.

      CH3SH can be converted photochemically to carbonyl sulfide (COS) in surface seawater (Kiene, 1996). COS is the most abundant sulfur gas in the stratosphere, where its oxidation would also result in the production of sulfate aerosols (Charlson et al., 1987).

      Another, and equally important, role of CH3SH is its involvement in methanogenesis (van der Maarel and Hansen, 1997). CH3SH is demethylated in anaerobic intertidal sediments through the following reaction described for a methanogenic bacterium (Finster et al., 1992):

      
4 CH3SH + 3 H2O ^ 3 CH4 + HCO3- + 4 HS- + 5 H+ (28)

      The CH4 and CO2 thus produced contribute to the greenhouse effect.

      The physiological significance of CH3SH formation has been studied with some detail only in mammals (Weisiger et al., 1980). In these organisms, CH3SH is a detoxification product of H2S. H2S is produced in the colon mainly upon the degradation of proteins by anaerobic bacteria. This gas is extremely toxic to higher animals, causing death in the same range of concentrations (300-500 ppm) as does hydrogen cyanide. It must, therefore, be quickly detoxified. A thiol methyltransferase, localized in the gut mucosa and many other tissues, detoxifies H2S (becoming HS- in solution) by methylating it to CH3SH. A similar role probably exists for its biosynthesis in microorganisms, however, CH3SH has to be further methylated to CH3SCH3 in order to significantly reduce toxicity (Drotar et al., 1987b).

      Among plants, CH3SH evolved by Catharanthus cell cultures acts as the substrate for a replacement reaction which produces either S-methylcysteine in the presence of cysteine, or methionine in the presence of homocysteine (Schwenn et al., 1983). Therefore, CH3SH is viewed as an intermediate in the biosynthesis of methionine in these cultures (Schwenn et al., 1983). Pumpkin leaves degrade methionine to CH3SH but instead of emitting the latter to the atmosphere, translocate it downwards in the phloem (Schmidt et al., 1985). Here, CH3SH could be a signal regulating the uptake of sulfate by the root according to the sulfur needed for growth. Volatiles from Brassica, containing CH3SH and CH3SSCH3, inhibited the growth of fungi that cause root rot in a number of plant species (Lewis and Papavizas, 1971). Upon attack by the herbivorous larvae of Pieris brassicae, brussels sprout leaves produce volatiles, including CH3SSCH3, that attract natural enemies of the herbivore and thus reduce its impact on the plant (Mattiacci et al., 1994).


2.2.4. Sulfate Reduction In Higher Plants

      Sulfur is an essential macronutrient for all living organisms, including plants. Its importance is directly linked to the crucial roles that sulfur-containing amino acids cysteine and methionine, and their numerous derivatives, such as glutathione, AdoMet and many secondary metabolites play in growth and metabolism (Leustek, 1996; Schmidt and Jäger, 1992). Sulfur is generally taken up in its oxidized form, SO42-, found at fairly high concentrations (ca. 0.5 mM) in arable soils (Cram, 1990). Roots are a major site of SO42- uptake, but a considerable supplement is also provided by the uptake of atmospheric SO2 through stomata (Cram, 1990). Sulfate has to be reduced to S2- prior to its incorporation into plant constituents. The sequence of reactions from SO42- to S2- is called assimilatory sulfate reduction as opposed to dissimilatory sulfate reduction which occurs in a number of anaerobic microorganisms (Brunold, 1990). These microorganisms use SO42- as the ultimate electron sink during oxidation of organic substrates. Reduced forms of sulfur, arising from the dissimilatory pathway, are excreted into the surroundings of these microorganisms. The assimilatory SO42- reductive machinery exists almost exclusively in the photosynthetic organs. Further, most of the enzymes in the pathway are localized in the plastids (Joyard et al., 1988). There, photosynthetically derived ATP and reductant directly supply the high energy requirement of the SO42- reduction process (Leustek, 1996). Assimilatory sulfate reduction pathway in plants can be divided into four successive parts: uptake, activation, reduction to S2-, and incorporation of sulfide into cysteine. The individual reactions of each part are still not fully understood (Brunold, 1990; Schmidt and Jäger, 1992), and there are two competing hypotheses (Figure 1).

      Sulfate is actively transported into plant cells, its transport being stimulated by sulfate starvation and repressed by feeding sulfate, cysteine or glutathione (Kylin and Hylmo, 1957; Legget and Epstein, 1956). This regulation

      

Figure 1. Possible pathways for the reduction of sulfate to cysteine.

      The only established steps in the pathway are sulfate uptake and activation to APS. The question marks signify that either one of two hypothetical pathways can be operating in higher plants. The shaded reactions illustrate the pathway that exists in bacteria, and could function in higher plants. The sequence to the left indicates the widely cited pathway for sulfate reduction in higher plants. In this pathway, sulfate is transferred to a carrier and is reduced in a bound form. The proteins and enzymes that are involved in the two possible pathways are as follows: sulfate permease (1), ATP sulforylase (2), APS sulfotransferase (3), thiosulfonate reductase (4), APS kinase (5), PAPS sulfotransferase (6), sulfite reductase (7), serine acetyltransferase (8), and O-acetylserine(thiol)lyase (9). APS, 5'-adenosinephosphosulfate; PAPS, 3'-phosphoadenosine-5'-phosphosulfate; GSH, reduced glutathione; Trx, reduced and oxidized thioredoxin; Fr, reduced and oxidized ferredoxin. Sulfite reductase in the bacterial pathway uses NADPH instead of ferredoxin as a reductant. Adapted from Leustek (1996) and Schmidt & Jäger (1992).

      of sulfate uptake is mediated by transcriptional control of the transporter proteins (Hart and Filner, 1969). The kinetics of SO42- uptake into isolated membrane vesicles from Brassica napus indicates the presence of only one transporter, the abundance of which is increased by sulfur starvation (Hawkesford et al., 1993). Sulfite, selenate, and arsenate, which all compete against SO42- for binding to the transporter, inhibit SO42- uptake in plants (Vange et al., 1974). Sulfate transporters, known as sulfate permeases, were recently characterized from the legume Stylosanthes hamata (Smith et al., 1995). They consist of a single polypeptide chain with 12 membrane-spanning regions, and are either predominantly expressed in leaves or exclusively in roots, indicating that plant SO42- permeases are functionally heterogeneous.

      After uptake, SO42- undergoes the first reaction in the reduction pathway. This reaction is catalyzed by ATP sulfurylase which produces adenosine 5'-phosphosulfate (APS) and inorganic pyrophosphate from SO42- and ATP (Brunold, 1990; Wilson and Reuveny, 1976). APS contains a high-energy sulfatophosphate bond which potentiates the SO42- moiety for subsequent metabolic transformations, thus the term activated sulfate. Similar to SO42- uptake, its activation by ATP sulfurylase is transcriptionally regulated by sulfur status of the plant, being induced by sulfur starvation and repressed by sulfur feeding (Reuveny and Filner, 1977). In spinach leaves, the enzyme is predominantly localized in the chloroplast and requires divalent cations for activity (Lunn et al., 1990).

      Based on the paradigms of sulfate reduction in bacteria and yeast, it was widely accepted in the past that the step after the formation of APS was its phosphorylation to 3'-phosphoadenosine-5'-phosphosulfate (PAPS) by APS-kinase (Brunold, 1990). However, this point always remained a subject of controversy (Leustek, 1996). Despite the presence of APS-kinase and PAPS in higher plants (Jain and Leustek, 1994), much evidence pointed to APS as the real active substrate for SO32- production (Setya et al., 1996; Wilson and Reuveny, 1976). Firstly, experimental evidence (Schiff, 1983) indicated that for PAPS to be a SO42- donor for further reduction, it must be converted to APS by the action of a 3'-nucleotidase, which is present in plants. Moreover, an APS-sulfotransferase necessary for transferring SO42- from APS to a carrier, was found in spinach leaves (Schmidt, 1975), and was recently purified from the alga Euglena gracilis (Li and Schiff, 1991). This enzyme could use reduced glutathione (GSH) as a carrier of SO42-, thus forming a thiosulfonate intermediate (Figure 1). However, the same enzyme from A. thaliana was also found to have APS kinase activity which produced PAPS (Schiffmann and Schwenn, 1994). This finding brought back the central question whether sulfate reduction in higher plants was any different from that in bacteria (Figure 1). Indeed, the rest of the enzymes involved in the bacterial pathway, PAPS sulfotransferase and sulfite reductase have homologues in higher plants (Gisselmann et al., 1992; Krueger and Siegel, 1982). Most recently, Setya and co-workers (1996) answered this question when they cloned an APS sulfotransferase gene family from A. thaliana, and confirmed the existence of an APS-dependent sulfate reduction pathway for the first time in higher plants. They demonstrated that sulfotransfer is not the catalytic function of the enzyme. Instead, the enzyme catalyzes sulfonucleotide reduction in a manner similar to that of bacterial PAPS reductase (Figure 1). Because the enzyme uses thioredoxin as the electron donor, and produces free SO32-, they named it APS reductase.

      The last enzyme in the pathway to S2- formation, sulfite reductase, is well characterized from plants (Krueger and Siegel, 1982). This hemoprotein is a homo-oligomer which uses reduced ferredoxin as the electron donor. Sulfite reductase was recently cloned from A. thaliana (Leustek, 1996). Its open reading frame contains the sequence of a potential signal peptide that targets the protein to the chloroplast. Whether the plant enzyme carries out SO32- reduction to S2- via a carrier-bound or free SO32- is still a subject of much controversy. Free S2-, the end-product of the pathway, is generally found at low concentrations in the cell because it is a highly reactive and toxic compound (Wilson and Reuveny, 1976). It is mainly assimilated into cysteine, homocysteine, methionine and other sulfur-containing cell components. Sulfur is also accumulated in secondary products by many plant families, such as Alliaceae known for its organic disulfides (Harborne, 1973), Liliaceae which accumulates S-alkylcysteine sulfoxides, and Brassicaceae which is rich in glucosinolates (Cram, 1990).


2.2.5. Glucosinolates

      Formerly known as mustard oil glucosides, glucosinolates are hydrophilic, nonvolatile, organic anions which occur in a limited number of plant families. They all have the same general structure, consisting of a thioglucose group linked to a side chain (R-group) through a sulfonated oxime group (Figure 2). Over 100 side chains, and thus glucosinolates, have been identified (Sørensen, 1991). They are found in all plant parts, but their quantities may vary considerably among organs (Kjaer, 1976). Glucosinolates are localized in the vacuole within the cell (Halkier and Du, 1997). Trivial names have been given to different glucosinolates, but a more precise nomenclature system exists that uses the prefixes to designate the composition of the side chain present in the molecule (Ettlinger and Kjaer, 1968). Table I gives an example of the nomenclature system with the corresponding trivial name for a number of glucosinolates.

      

Figure 2. General structure of glucosinolates

      
Table I. Trivial names of some glucosinolates with the corresponding side chain (R) composition.
Trivial name R-glucosinolate
Sinigrin 2-Propenyl
Gluconapin 3-Butenyl
Glucobrassicanapin 4-pentenyl
Progoitrin 2-Hydroxy-3-butenyl
Gluconapoleiferin 2-Hydroxy-4-pentenyl
Glucoiberin 3-Methylsulfinylpropyl
Glucocheirolin 3-Methylsulfonylpropyl
Glucoerucin 4-Methylthiobutyl
Glucoberteroin 5-Methylthiopentyl
Glucobrassicin 3-Indolylmethyl
Neoglucobrassicin 1-Methoxy-3-indolylmethyl
4-Hydroxyglucobrassicin 4-Hydroxy-3-indolylmethyl
4-Methoxyglucobrassicin 4-Methoxy-3-indolylmethyl
Glucobrassicin-1-sulfonate 1-sulfonate-3-indolylmethyl

      Glucosinolates have only been detected in a restricted number of dicotyledonous angiosperms (Fenwick et al., 1983). They are prevalent throughout the order Capparales, including the families Brassicaceae, Capparaceae, Resedaceae, Tovariaceae, and Moringaceae. They also occur occasionally outside this order in some members of the families Caricaceae, Gyrostemonaceae, Salvadoraceae, Tropaeolaceae, Limnathaceae, and Euphorbiaceae. Over 80% of all glucosinolates so far identified are found in the family Brassicaceae, and their presence has been used as an important chemotaxonomic criterion for classification within this family (Kjaer, 1976). Their presence in all members of Brassicaceae is of great economic importance, because members of this family are a source of condiments, relishes, vegetables, oil, and forage crops (Poulton and Møller, 1993).

      Structural similarities between naturally occurring amino acids and the aglucone moiety of glucosinolates suggested that the latter are amino acid derivatives, which was subsequently confirmed using isotopically labeled amino acids (Kutacek and Kefeli, 1968). Glucosinolates are classified as aliphatic, aromatic and indole glucosinolates, depending on whether they are derived from methionine, phenylalanine or tyrosine, or tryptophan, respectively. They, however, share similar biosynthetic pathways which involve three different stages: the first one is the synthesis of chain-elongated amino acids, the second involves the synthesis of the common aglucone moiety, and the third is the side chain modifications (Halkier and Du, 1997). Despite some uncertainties, formation of the common aglucone moiety, outlined in Figure 3, is the best characterized part of glucosinolate synthesis (Duncan, 1991; Fenwick et al., 1983; Halkier and Du, 1997). With the exception of indole acetaldoxime formation from tryptophan, the aglucone formation involves an initial N-hydroxylation step which yields an N-hydroxy amino acid, followed by oxidative decarboxylation of the amino acid to form an aldoxime. This chain of reactions is catalyzed by a single monooxygenase enzyme. Biochemical evidence suggests that different monooxygenases are involved in different plant species or for different glucosinolates within a species: these include cytochrome P450s and flavin-dependent enzymes (Bennett et al., 1996; Bennett et al., 1997; Du and Halkier, 1996; Du et al., 1995). The transformation of tryptophan to indole acetaldoxime is catalyzed by a plasma membrane-bound peroxidase (Ludwig-Muller and Hilgenberg, 1988). Further incorporation of sulfur in the aldoxime results in the formation of thiohydroximic acid. The nature of the sulfur donor is still unknown, although radiolabeling experiments (Matsuo et al., 1972) showed that sulfur from cysteine was, by far, the most readily incorporated, producing an S-alkylthiohydroximate (Figure 3). The enzyme responsible could be a glutathione-S-transferase, since such enzymes can substitute glutathione with cysteine. A C-S lyase activity

      

Figure 3. The biosynthesis of glucosinolates.

      Reactions in the shaded area are still not proven experimentally. Adapted from Halkier & Du (1997), Poulton & Möller (1993), and Fenwick et al. (1983).is, however, necessary to cleave the S-alkylthiohydroximate to yield a thiohydroximate (Halkier and Du, 1997). This lyase still remains to be isolated from glucosinolate-containing plants, despite the well-established presence of its reaction product, thiohydroximate. Thiohydroximate is S-glucosylated to desulfoglucosinolate via a uridine-diphospho-glucose (UDPG) thiohydroximate glucosyltransferase, an enzyme that has been isolated from oilseed rape (Brassica napus) (Reed, 1993). The enzyme has high specificity for thiohydroximate structures but little, if any, for the side chain composition. The final step in this pathway involves the addition of SO42-, which requires the presence of PAPS as the sulfur donor (Halkier and Du, 1997). A sulfotransferase with absolute substrate requirement for the desulfoglucosinolate structure, but none for the side chain, has been isolated from cress (Lepidium sativa) (Glendening and Poulton, 1988).

      All families containing glucosinolates also contain a thioglucoside glucohydrolase enzyme, called myrosinase (Bones and Rossiter, 1996). Myrosinase hydrolyzes the thioglucoside bond and yields a glucose molecule and an aglucone which spontaneously degrades into one of a number of toxic compounds (Figure 4). Depending on several variables such as the pH, temperature, presence of protein co-factors and nature of the glucosinolate side chain, the major products formed are isothiocyanates, thiocyanates, nitriles or

      

Figure 4. Products of glucosinolate degradation by myrosinase.

      After cleavage of the S-glucose bond, the transformation of the unstable intermediate is dependent on the nature of the side chain and the surrounding conditions. Isothiocyanates are formed at pH>7, nitriles are formed at pH<4. Oxazolidine-2-thiones are formed in the presence of ß-hydroxylated side chains. A terminal double bond in the side chain, the presence of an epithiospecifier protein, and Fe2+ ions result in the formation of an epithionitrile. R, variable side chain. Adapted from Halkier & Du (1997) and Poulton & Möller (1993).epithioalkanes (Bones and Rossiter, 1996; Duncan, 1991; Fenwick et al., 1983). In the intact plant, myrosinase is located in specialized myrosin cells, and thus glucosinolates and myrosinase are physically separated (Duncan, 1991). Myrosin cells are dispersed throughout the seedlings, mature plants, and seeds, but during germination, they are present in large proportion in the outermost part of the cotyledons, and are most abundant in the cortex area of the radicles. These locations afford myrosin cells a role as a  toxic mine  when the tissue is destroyed (Thangstad et al., 1993). Upon injury and cell rupture, the enzyme and its glucosinolate substrates come in contact, and hydrolysis of these substrates occurs.

      Glucosinolate hydrolysis products make a significant contribution to the typical flavor of vegetables from the Brassicaceae family. They also possess chemical properties rendering them powerful allelochemical agents against microbes, fungi, and plants (Angus et al., 1994; Lewis and Papavizas, 1970; Lewis and Papavizas, 1971; Teasdale and Taylorson, 1986; Wolf et al., 1984). Isothiocyanates interact nonspecifically and irreversibly with sulfhydryl groups, disulfide bonds and amino groups in proteins, thus forming stable products with proteins and amino acids (Fenwick et al., 1983). Interest in the chemistry of isothiocyanates has increased over the last two decades with the realization of their beneficial effects on human health (Verhoeven et al., 1997). Isothiocyanates and indoles have antineoplastic properties because they are able to induce detoxification mechanisms in animals, thereby influencing several processes related to chemical carcinogenesis. These mechanisms include phase 1 and phase 2 biotransformation enzyme activities. Phase 1 involves oxidation, reduction and hydrolysis reactions, which make the carcinogens more hydrophilic, and more susceptible to detoxification. The most important enzymes of this phase are cytochrome P450s. Phase 2 metabolism involves detoxification, including conjugation reactions which facilitate excretion of suspected carcinogens. Enzymes such as glutathione-S-transferases and UDP-glucuronyl transferases are representatives of this phase (Verhoeven et al., 1997). In contrast to the beneficial properties of glucosinolate hydrolysis products, they can also have deleterious effects on animals, reducing fertility and inducing goiter (Brown and Morra, 1997; McDanell et al., 1988). Goitrogenicity is caused by thiocyanate ion and 5-vinyloxazolidine. Thiocyanate is a pseudohalide, and therefore, shares several chemical properties with halides (Hughes, 1975). This is particularly true for iodide, with which the thiocyanate ion competes for uptake by the thyroid gland. Thiocyanate acts as a competitive inhibitor of iodide uptake as well as its transformation to diodotyrosine and thyroxine (Wood, 1975).

      Glucosinolate hydrolysis products, particularly thiocyanate ion and isothiocyanates, are not only toxic to animals, fungi, microbes and other plants but also to the plant that produces them (Duncan, 1991). They must, therefore, be detoxified. This is achieved by further metabolism to a variety of simple compounds that are either recycled into other metabolites or excreted in the form of volatiles such as methylisothiocyanate, methylthiocyanate, and large amounts of CH3SH, CH3SSCH3, and CH3SCH3 (Brown and Morra, 1997).


3. Materials And Methods


3.1. Plant Material

      Plants of Brassica oleracea cv. April Red were grown in a greenhouse under a photoperiod of at least 12 hours. They were watered on alternate days and given 2 g L-1 of a commercial fertilizer (N:P:K, 20:20:20) once every week. The two basal most-expanded leaves of three to six-month old plants were used for enzyme purification. Young, expanding leaves of 4 month-old plants were used for subcellular localization experiments. Spinach (Spinacia oleracea) plants, used as a control for the chloroplast isolation procedure, were purchased from a local market. Brassica oleracea cv. Botrytis, Raphanus sativus, Iberis sempervirens, Cleome hassleriana, Reseda odorata, Reseda luteola, Carica papaya, Meconopsis cambrica, Papaver orientale, Chelidonium japonicum, and Allium cepa, used for the survey of H/BMT activity, were collected from the Montreal Botanical Garden.

      3.2. Chemicals

      Chemicals were purchased from the following sources:

      
Chemical Source
Authentic CH3Cl Liquid Carbonic (Scarborough, Canada)
Calf intestinal alkaline phosphatase (CIP) New England Biolabs (Beverly, MA, USA)
Continued on the next page
AdoMet (SO42- salt), AdoHcy Boehringer Mannheim (Laval, Canada)
Sodium [1-14C]acetate (14.7 µCi/µmol), sodium [14C]bicarbonate (0.2 mCi/mmol) Amersham (Oakville, Canada)
[Methyl-3H]AdoMet (80 Ci/mmol, SO42- salt) DuPont NEN (Boston, MA, USA)
Sephadex G-25, G-100, the gel filtration calibration kit Pharmacia Biotech (Montreal, Canada)
AdoMet (Cl- salt), 5'-AMP-agarose, phenol, cytochrome c, ATP, NADPH, NADH, DL-glyceraldehyde-3-phosphate, glycerophosphate dehydrogenase, ribulose 1,5-diphosphate (Na+ salt), CoA, DTT, 2,6-dichlorophenolindophenol (DCPIP), succinate, phenasine methosulfate (PMS), L-methionine, phenyl isothiocyanate, L-cysteine, SMM, Percoll Sigma Chemical Co. (St. Louis, MO, USA)
Ammonium sulfide, thiosalicylic acid, 4,4'-thiobisbenzenethiol, 4-nitrophenol, aniline, 4-hydroxymercuribenzoic acid, iodoacetamide, S-methyl-L-cysteine, DL-homocysteine, DL-methionine, L-serine, thiophenol, sodium thiocyanate, benzyl isothiocyanate, phenethyl isothiocyanate, propyl isothiocyanate, Quercetin, and authentic CH3SH, CH3Br, CH3I Aldrich (Milwaukee, WI, USA)
The protein dye reagent, N,N,N',N'-tetramethyl-ethylenediamine (TEMED), ammonium persulfate, bis-acrylamide Bio-Rad (Mississauga, Canada)

      All chemicals used in this study were of analytical reagent or higher grade.


3.3. Buffers

      
Buffer name Composition
A 100 mM Tris-acetate, pH 7.5 containing 10% glycerol (v/v) and 14 mM ß-mercaptoethanol
B 25 mM Tris-acetate, pH 7.4 containing 10% glycerol and 14 mM ß-mercaptoethanol
C 25 mM Tris-acetate, pH 7.4 containing 14 mM ß-mercaptoethanol
D 0.1 M HEPES (pH 6.8), 0.35 M sorbitol, 5 mM DTT and 0.5% BSA
E 0.1 M HEPES (pH 6.8), 0.35 M sorbitol and 5 mM DTT

      Buffers were filtered before use through a 0.22 µm membrane.


3.4. Measurement Of Halide/bisulfide Methyltransferase Activity


3.4.1. Gas Chromatographic Assay

      Since I- was the most efficiently methylated halide by H/BMT, it was used as the halide substrate throughout the purification procedures and the subcellular localization experiments. The enzyme activity was assayed in a 1-ml or 0.5-ml mixture containing 0.5 mM AdoMet, 50 mM KI (X- methylation) or 20 mM (NH4)2S (HS- methylation), all prepared in buffer A. Enzyme preparations containing up to 1 mg protein were used to start the reaction. The mixture was incubated in a 5-ml glass vial sealed with a screw cap fitted with a Teflon-lined septum (Supelco, Oakville, Canada), and maintained on an orbital shaker (150 rpm) at the room temperature. The reaction rate was linear for at least 45 min at all enzyme and substrate concentrations. A standard incubation time of 30 min was adopted. In vivo H/BMT activity was measured as described earlier (Saini et al., 1995). Briefly, 10 leaf disks were placed on a Whatman filter paper (4.25 cm) in a 50-mL Erlenmeyer flask and incubated with 2 mL of 100 mM KI solution. The flask was sealed with a rubber stopper and CH3I formed in the headspace was quantified after 2 h by gas chromatography (GC).

      Assays for alternative substrates or inhibitors of H/BMT contained 20 mM KI (the known substrate) and at least equimolar concentration of the compound to be tested. When thiol-group inhibitors were tested, DTT was removed by desalting the enzyme preparation on a Sephadex G-25, PD-10 column, and ß-mercaptoethanol was omitted from the assay mixture.

      The products of the assay were analyzed by GC using a flame ionization detector. One-mL headspace samples were injected in a 210 x 0.3 cm stainless steel column packed with 80/100 mesh Porapak Q (Supelco) in a Hewlett Packard 5890 series II gas chromatograph. Column temperatures were 160°C for CH3I, 145°C for CH3SH, and 130°C for CH3Cl. The carrier gas was ultra pure He at a flow rate of 40 ml min-1. The column was purged by heating to 200°C between injections. Products were quantified by peak area and identified by comparison of their retention times with those of authentic CH3X or CH3SH, used to calibrate the instrument. Recoveries of 75 nL CH3I and 28 nL CH3SH injected to vials containing the assay mixture were 69% and 87%, respectively, after 30 min. The data presented here were not corrected for recoveries. The identity of CH3I and CH3SH in the headspace of enzyme assays was confirmed by mass spectrometry using a KRATOS MS 50 mass spectrometer operated in the Electron Impact mode (Attieh, 1993).

      3.4.2. Radiometric assay

      This method, adapted from thiol methyltransferase assay of Borchardt and Cheng (1978), was based on the ability of purified H/BMT to transfer 3H-CH3 group of [methyl-3H]AdoMet to thiol compounds. The standard assay mixture was done in buffer A, and contained 0.14 µCi [methyl-3H]AdoMet (SO42- salt) diluted in 200 µM total AdoMet (SO42- salt), and 200 µM methyl acceptor in a total volume of 100 µL. The reaction was started by adding the purified enzyme (up to 6 µg protein). After 30 min at room temperature, the reaction was stopped by adding 50 µL of 10 N NaOH. The mixture was transferred to 16 x 125 mm culture tubes, and the methylated products were extracted with 5 mL toluene (5 mL hexane for thiocyanate). In assays containing organic acid substrates, the mixture was acidified with 50 µL of 10 N HCl prior to extraction. After centrifugation for 5 min at 3000 x g, a 3-mL aliquot of the organic layer was mixed with an equal volume of scintillation cocktail and counted for radioactivity using a Model 1214 Rackbeta liquid scintillation counter (Pharmacia). Assay mixtures with boiled enzyme or without the methyl acceptor were used as blanks, and the data were corrected accordingly.

      The kinetic parameters of the enzyme and its isoforms toward thiol compounds were determined in a reaction mixture which contained the above reagents and various concentrations of the thiol substrates. The tubes were incubated for 5 min, which was within the linear phase of the reaction with all substrate concentrations used. The reactions were stopped and processed as described above. All enzyme assays were done in duplicate and experiments were repeated at least twice. The results were always consistent, and representative data from individual experiments are presented.

      All manipulations described in the following sections were carried out at 0-4°C unless stated otherwise.


3.5. Purification Of H/bmt

      Halide and HS- methylating activities were monitored simultaneously throughout the purification procedure. The partial purification of H/BMT was described earlier (Attieh, 1993) and is presented briefly in the next paragraph.

      Leaves were cut into small squares, frozen with liquid nitrogen and homogenized in buffer A (1:3 w/v) and polyvinylpolypyrrolidone (PVPP, 10% w/w) for 5 min. The homogenate was filtered through 8 layers of cheesecloth and centrifuged for 20 min at 15,000 x g. The supernatant (crude extract) was fractionated with solid (NH4)2SO4, and the proteins that precipitated between 60 and 85% saturation, containing H/BMT activity, were recovered by centrifugation at 12,000 x g for 20 min. The pellet was resuspended in 15 ml of buffer B, and applied to a Sephadex G-100 gel filtration column which had been previously equilibrated in the same buffer. The active fractions from this step were pooled and concentrated by diafiltration in a stirred Amicon cell fitted with a YM-30 membrane (Amicon, Danvers, MA, USA). The concentrated sample was applied to an adenosine-agarose column that had been previously equilibrated with buffer B. The column was washed at a constant flow rate of 10 ml hr-1 with 15 ml of buffer B, followed by 8 ml of the same buffer containing 300 mM (NH4)2SO4. The salt was then removed by washing the column with buffer B. H/BMT activity was recovered by elution with a 40-ml linear gradient of 0-4 mM AdoMet in buffer B. Further purification was done as part of the present thesis project and is described below in detail.


3.5.1. Anion Exchange Hplc

      The active eluate from the affinity step was submitted to HPLC on a 1x10 cm Protein Pak Q (PP-Q) anion exchange column (Millipore, Milford, MA, USA). The column had been equilibrated with buffer C, and was operated with a Waters 640 Advanced Protein Purification System (Millipore). The sample was loaded at a flow rate of 0.5 ml min-1, and the column was washed with the same buffer until no further UV-absorbing material was eluted. Enzyme activity was then eluted with NaCl in buffer C in a 2-step linear gradient of 0-200 mM in 50 min, and then 200-500 mM in 20 min. The eluate was collected in 0.5 ml fractions. H/BMT activity eluted in four distinct peaks.


3.5.2. Analytical Gel Filtration Hplc

      An aliquot (200 µL) from the fraction containing the first peak of H/BMT activity from the previous step was loaded on a 1 x 30 cm Superdex 75 (Pharmacia) gel filtration HPLC column. The column had been equilibrated with buffer C containing 100 mM NaCl. Proteins were eluted in the same buffer at a flow rate of 0.5 ml min-1, and 0.5-ml fractions were collected. After this step, the H/BMT preparation was apparently free from contaminating proteins.


3.6. Purification Of H/bmt Isoforms

      Anion exchange HPLC in the above purification procedure resolved multiple peaks of H/BMT activity, suggesting the presence of isoforms of the enzyme. This was further investigated in the following set of experiments.


3.6.1. Standard Procedure For The Purification Of H/bmt Isoforms

      The extraction, (NH4)2SO4 precipitation and the Sephadex G-100 gel filtration steps were as summarized in section 3.5, and described earlier in detail (Attieh, 1993). After diafiltration on a YM-30 membrane, the protein preparation was often stored at -80°C, where it was stable for several months.


3.6.1.1. Affinity Chromatography

      The previous procedure (Attieh, 1993) to prepare adenosine-agarose affinity matrix was slightly modified to use 5'-AMP-agarose as the starting material instead of ADP-agarose, which was the only commercially available matrix during the previous study (Attieh, 1993). The greater ease of removal of phosphate from AMP-agarose resulted in a higher protein binding capacity of the column. The resin (5 ml) was thoroughly washed under vacuum with deionized water to remove the stabilizing lactose. The resin was preincubated for 1 h in 5 ml of 100 mM ethanolamine buffer, pH 9.0, containing 4% BSA (w:v) at 37°C. The dephosphorylation reaction was then started by adding 2000 p-nitrophenyl phosphate units of calf intestinal alkaline phosphatase (CIP). The reaction vial was continuously rotated overnight in a hybridization oven at 37°C. The resin was packed into a column and washed successively with deionized water and buffer B containing 2 M NaCl, and was then equilibrated with buffer B.

      The concentrated sample was loaded on the adenosine-agarose affinity column and washed as described in section 3.5. Finally, H/BMT was specifically eluted with an AdoMet gradient of 0 to 7 mM prepared in 30 mL of buffer B. One-mL fractions of the eluate were collected.


3.6.1.2. First Anion Exchange Hplc Step

      Active fractions from the affinity step were pooled and loaded on a 1 x 10 cm PP-Q anion exchange HPLC column equilibrated with buffer C. The column was washed with the same buffer until no UV absorbing material was detected in the flow through. Bound proteins were then eluted with buffer C containing 500 mM NaCl, at a flow rate of 0.5 mL min-1. One-mL fractions were collected, and the fractions exhibiting H/BMT activity were pooled.


3.6.1.3. Preparative Gel Filtration Hplc

      The resulting concentrated preparation was loaded on a Superdex-75 HiLoad gel filtration HPLC column that had been previously equilibrated with buffer C. The column was developed in the same buffer at 0.5 mL min-1, and 1-mL fractions were collected.


3.6.1.4. Second Anion Exchange Hplc Step

      The active fractions from the above step were pooled and loaded on the PP-Q column as in the first anion exchange step. The column was washed with buffer C until no significant UV absorbance was recorded. Proteins were finally eluted with an NaCl gradient in two steps: 0 to 160 mM in 50 min, and then 160 to 200 mM in 15 min. At this point, H/BMT activity was resolved in 5 distinct peaks, each of which was free from other contaminating proteins.


3.6.2. Alternative Procedure For H/bmt Purification

      This procedure was followed to verify whether the multiple peaks observed in the standard procedure represented true isoforms, or they were artifacts caused by processing of the protein during the purification.


3.6.2.1. Extraction

      Leaf tissue (150 g) frozen in liquid N2 was homogenized for 5 min in a blender with buffer A (1:2, w:v) containing 10% PVPP. The homogenate was filtered through 8 layers of cheesecloth, and centrifuged at 20,000 x g for 30 min in a Beckman JA-14 rotor. The resulting supernatant was ultracentrifuged at 110,000 x g for 1 h in a Beckman 70.1Ti rotor. The supernatant was concentrated by diafiltration using an Amicon cell fitted with a YM-30 membrane.


3.6.2.2. Affinity Chromatography

      The concentrated sample was loaded onto an adenosine-agarose affinity column previously equilibrated with buffer B. The column was washed with the same buffer, and H/BMT activity was eluted with 2 column volumes of a 7 mM AdoMet solution in buffer B.


3.6.2.3. Anion Exchange Hplc

      Active fractions from the affinity step were pooled and applied on a PP-Q anion exchange HPLC column that was previously equilibrated with buffer C. The column was then washed and eluted as described in the second anion exchange step of the standard purification (section 3.6.1.4.).


3.7. Molecular Weight Determination

      The molecular weight of the native protein was first determined by gel filtration on a Sephadex G-100 column, then, for the isoforms, by gel filtration HPLC on a Superdex 75 HiLoad column. In both cases, the columns were equilibrated with buffer C and calibrated with the following standards: Ribonuclease A (Mr 13,700), chymotrypsinogen (Mr 25,000), ovalbumin (Mr 43,000) and albumin (Mr 67,000). The column void volume (V0) was determined with blue dextran 2000. One-mL samples were applied to the columns which were operated at 20 ml h-1 (Sephadex G-100) and 30 mL h-1 (Superdex 75). In order to achieve optimal calibration, the standards were chromatographed in two batches, the first included ribonuclease A and ovalbumin, and the second chymotrypsinogen and albumin. The molecular weights were determined by plotting logMr values against the partition coefficient (Kav) values calculated as follows:

      

      Where Ve is the protein elution volume and Vt is the column bed volume.


3.8. Protein Estimation

      Total soluble protein content was measured by the method of Bradford (1976) using the Bio-Rad protein reagent and following the manufacturer's microassay procedure. Bovine serum albumin was used as a standard.


3.9. Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (sds-Page)

      At each step of the purification procedure, protein aliquots were denatured and separated by SDS-PAGE according to Laemmli (Laemmli, 1970), using stacking and separating gels with 5% and 12% acrylamide, respectively. Proteins were visualized by staining with Coomassie Blue R-250 or with silver nitrate (Gottlieb and Chavko, 1987). Molecular masses of the denatured H/BMT proteins were calculated by fitting their relative migration distance to a plot of log molecular mass of standard markers against their relative migration distance on the gel.


3.10. Kinetic Studies

      The kinetic properties of H/BMT were first determined using a preparation purified through a Sephadex G-100 gel filtration step, and then using the 3 most abundant H/BMT isoforms purified to homogeneity.

      Substrate interactions and product inhibition studies were done using KI as the CH3-acceptor substrate because of the ease of detection of the product by GC. Substrate interaction studies were done by fixing the concentration of one substrate while varying that of the other. The kinetic constants were fitted to equation 2 representing a sequential bireactant mechanism,

      

      where V is the maximum velocity, A and B are the substrates, KA and KB are the Michaelis-Menten constants for the two substrates and KiAKB is a combined Michaelis-Menten constants term. Product inhibition kinetics were done with AdoHcy as the inhibiting product. Data were presented as double reciprocal plots of initial velocity (v) versus substrate (S) concentrations. Product inhibition data were fitted to the equations for competitive (Equation 3) and noncompetitive (Equation 4) inhibition,

      

      

      where S is the varied substrate and I is the inhibiting product, V' and K's are maximum velocity and the apparent Michaelis-Menten constant, respectively, at each concentration of the fixed substrate in the presence of the product. Kis is the slope constant and Kii the intercept inhibition constant. Michaelis-Menten constants (Km) for substrates were determined by varying the concentration of the substrate in question, and using a fixed saturating concentration of the other. Lineweaver-Burk plots were used to determine Km values.


3.11. Subcellular Localization Of H/bmt


3.11.1. Isolation Of Intact Chloroplasts

      Intact chloroplasts were purified according to the method of Mills and Joy (1980). Cabbage plants were kept in darkness for two days in order to minimize starch accumulation in the chloroplasts. Fresh, young leaf tissue (30 g) was cut into small squares and homogenized in buffer D (4 mL g-1) at high speed in a prechilled blender for 5 x 5 s with 2 s intervals. The homogenate was filtered through 8 layers of cheesecloth and then centrifuged at 4,000 x g for 3 min. The resulting pellet was resuspended in minimal volume of buffer D, and 1-mL aliquots of it were immediately applied on top of a 3-step discontinuous Percoll gradient, consisting of 5 mL of 20%, 5 mL of 40% and 1 mL of 90% Percoll. Each step also contained 0.35 M sorbitol. The gradient was centrifuged at 6,000 x g for 20 min in a Beckman JA-17 rotor. The intact chloroplast fractions collected from the 40-90% interface from each tube were pooled, washed with an equal volume of buffer D, pelleted by centrifugation, then gently resuspended in fresh buffer D. To verify the reliability of this purification technique, the ability of cabbage chloroplasts to carry out the marker enzyme assays was compared to that of purified spinach chloroplasts. The integrity of the purified chloroplasts from cabbage and spinach was determined by phase-contrast light microscopy (Walker et al., 1987).


3.11.2. Cell Fractionation

      Fresh, young cabbage leaf tissue (100 g) was homogenized as described above with buffer E containing 10% polyvinylpyrrolidone (PVP) (w/w). After filtration, the homogenate was centrifuged at 500 x g for 10 min in a Beckman JA-17 rotor. The 500g supernatant was carefully removed and centrifuged at 10,000 x g for 30 min in the same rotor. The resulting supernatant was subjected to ultracentrifugation at 100,000 x g for 1 h using a Beckman 70.1-Ti rotor. The pellet at each step was resuspended in a minimum volume of buffer E. The pellets and the 100,000 x g supernatant were assayed for enzyme activities.


3.11.3. Measurement Of Marker Enzyme Activities And Chlorophyll


3.11.3.1. Triose Phosphate Isomerase

      The marker for the cytosol, triose phosphate isomerase (TPI), was assayed according to Beisenherz (1955) as described below:

      Stock solutions

      Triethanolamine-HCl (TRA) buffer, 1 M, pH 7.6
KCN, 20 mM
NADH, 1.2 mM
DL-glyceraldehyde-3-phosphate (GA3P), 30 mM
Glycerophosphate dehydrogenase (GPDH), 4 U mL-1

      Procedure

      The following were added to a 1-mL disposable cuvette:

      0.1 mL each of TRA, KCN, NADH, GPDH
5-25 µL enzyme extract (or H2O for blank)
H2O up to 0.9 mL total volume
0.1 mL GA3P to start the reaction

      NADH oxidation was monitored by a decrease in absorbance at 340 nm (visible). Extinction coefficient (E) for NADH at 340 nm is 6.22 cm2 µmol-1. The assay involves a first reaction, catalyzed by TPI, which produces dihydroxyacetone phosphate (DHAP) from GA3P, and a second reaction, catalyzed by GPDH, producing glycerol-3-phosphate (G3P) and NAD+ from DHAP and NADH.

      
GA3P × DHAP (29)
DHAP + NADH × G3P + NAD+ (30)


3.11.3.2. Succinate Dehydrogenase

      Succinate dehydrogenase (SDH) activity, the marker for mitochondria, was measured by the method of Davis & Merret (1974). It involved the following procedure:

      Stock solutions

      Potassium phosphate buffer, 165 mM, pH 7.6
2,6-Dichlorophenolindophenol (DCPIP), 5 mM
Phenazine methosulfate (PMS), 1.65% (w:v)
Sodium succinate, 33 mM
KCN, 16.5 mM

      Procedure
The following were mixed in a disposable 1-mL cuvette:
200 µL each of phosphate buffer, KCN, sodium succinate
20 µL DCPIP
H2O to 1 mL (Blank), or H2O and extract (10-50 µL) to 980 µL total reaction volume

      The reaction was started with 20 µL PMS, and the activity was monitored by decreasing absorbance of DCPIP at 600 nm. The extinction coefficient for DCPIP at 600 nm is 21x103 cm2 mmol-1. Reactions involved are as follows:

      
Succinate + PMS ^ Fumarate + reduced PMS (31)
Reduced PMS + DCPIP ^ Reduced DCPIP + PMS (32)


3.11.3.3. Catalase

      The microsome marker, catalase (Cat), was assayed following the procedure of Lück (1965). The assay involved the following:

      Stock solutions

      Potassium phosphate buffer, 67 mM, pH 7.0
H2O2 in phosphate buffer, 0.05% (v:v)

      Procedure

      The following were mixed in a Quartz cuvette:
1 mL phosphate buffer and 5-10 µL enzyme extract (Blank)
or
1 mL H2O2 solution and 5-10 enzyme extract (Reaction).

      The reaction was started by adding the extract and H2O2 degradation was monitored by decreasing absorbance at 240 nm (UV). The extinction coefficient for H2O2 at 240 nm is 6.7x104 cm2 mol-1.


3.11.3.4. Nadph Cytochrome C Reductase

      The assay for NADPH cytochrome C reductase (CCR), the marker for endoplasmic reticulum (ER), was done according to Lord et al. (1973), as described below:

      Stock solutions

      Potassium phosphate buffer, 100 mM, pH 7.2
Cytochrome C (oxidized form), 0.1 mM
NADPH, 1 mM
KCN, 50 mM

      Procedure

      Equal quantities from the above stocks were mixed
A 0.8 mL aliquot of the mixture was dispensed in a 1-mL disposable cuvette
10-50 µL extract (reaction), or 10-50 µL H2O (Blank)
H2O up to a total assay volume of 1 mL

      The reaction was started by addition of the enzyme, and the reduction of cytochrome C was monitored by the increase in absorbance at 550 nm. The extinction coefficient of cytochrome C at 550 nm is 21x103 cm2 mmol-1.


3.11.3.5. Photosynthetic Co2 Fixation

      Photosynthetic carbon fixation was measured using the method of Lilley and Walker (1974) under non saturating light conditions (45 µmol s-1 m-2). The assay involved the following:

      Stock solutions

      Tris-HCl buffer, 250 mM, containing 25 mM MgCl2, pH 7.5
NaH14CO3 (0.2 mCi/mmol), 200 mM
Ribulose 1,5-diphosphate (RUDP), 50 mM
All reagents were prepared in water that had been boiled and cooled in order to drive off all CO2 dissolved therein.

      Procedure

      Screw-cap 5-mL glass vials (Supelco) were purged with N2 for 5 min in order to remove O2, then sealed with caps fitted with Teflon-lined septa. The following stocks were injected through the septa and preincubated for 5 min at 30°C:
300 µL Tris-HCl/MgCl2
125 µL NaH14CO3
10 µL RUDP
15 µL CO2-free water

      The reaction was started by injecting 50 µL of the chloroplast preparation into the vial. The reaction was allowed to proceed for 30 min at 30°C, then stopped by adding 500 µL glacial acetic acid. The latter, and all subsequent extraction steps were done under a fumehood. The contents of the vial were transferred to test tubes and boiled for 10 min. After cooling, scintillation cocktail was added to the mixture and the sample was counted for incorporated radioactivity. Blank assays were done without the chloroplast preparation.


3.11.3.6. in Vitro Lipid Synthesis

      Lipid synthesis from [1-14C]acetate and subsequent extraction were done as described by Sparace et al. (1988). Synthesis of polar lipids from radioactive fatty acids in intact, isolated chloroplasts was done at 25°C under light (45 µmol s-1 m-2), in a total volume of 1 mL containing the following: 37.5 mM HEPES buffer, pH 7.9, 300 mM sorbitol, 10 mM KHCO3, 4 mM glycerol-3-phosphate 2 mM each of MgCl2, sodium pyrophosphate, ATP, and CoA, 1 mM DTT, and 0.15 mM sodium [1-14C]acetate.

      The reaction was started by adding 100 µL of chloroplast preparation, and was allowed to proceed for 30 min, after which it was stopped by the addition of 3 mL chloroform:methanol:acetic acid (1:2:0.1, v:v). Lipids were extracted by adding 1 mL each of chloroform and 1 M KCl to the reaction mixture, vortexing for 10 s, and centrifuging at 1000 x g. The chloroform phase was recovered and washed three times with chloroform:methanol:water (3:48:49, v:v) and then dried under a N2 current. The dried sample was resuspended in 1 mL chloroform, transferred into a scintillation vial, and dried again under N2. The sample was finally redissolved in 4.5 mL of scintillation liquid and counted for radioactivity. Blank assays lacked chloroplast preparation.


3.11.3.7. Chlorophyll Determination

      Chlorophyll content was determined using Arnon's procedure (1949). Aliquots (10-25 µL) of the purified chloroplasts were added to 3 mL of cold 80% aqueous acetone, vortexed, and left on ice for 10 min. Samples were then centrifuged and their absorbance was measured at 652 nm. The value of the specific absorbance coefficient for chlorophyll a and b at 652 nm is 36.


3.11.3.8. Calculation Of Marker Enzyme Activities

      Activities of the marker enzymes TPI, SDH, CCR, and Cat were calculated using the following equation:

      

      Where DA is the difference in absorbance between the blank and the reaction assays; V is the reaction volume (mL); T is the reaction time (s); E is the extinction coefficient at the given wavelength; v is the extract volume used in the assay (mL).

      Activities of the chloroplast markers, total carbon fixation and lipid synthesis, were determined using the following equation:

      

      Where S is the amount of radiolabeled substrate (nmol); L is the amount of radiolabel (µCi); T is reaction time (s); 1 µCi = 2.22x106 dpm.

      The amount of chlorophyll in the different fractions from the Percoll gradient was determined by dividing the value of the absorbance by the absorbance coefficient, 36. The units, expressed in mg chlorophyll mL-1 acetone solution, are then normalized to the volume of each gradient fraction.


3.12. Enzyme Extraction For The Survey Of H/bmt Activity

      Leaves (2 g), from a number of higher plants, were cut into small squares, frozen in liquid nitrogen, and ground with acid-washed sea sand (Fischer Scientific, Ottawa, Canada) (50% w:w) and insoluble PVPP (10% w:w) into a fine powder with a pestle and mortar. Proteins were then extracted with buffer A. The homogenate was centrifuged at 15,000 x g, and the supernatant was recovered, desalted on a PD-10 column (Pharmacia) and used for enzyme assays.


3.13. Immunological Methods


3.13.1. Production Of Polyclonal Antibodies For H/bmt

      Polyclonal antibodies against H/BMT were raised at a commercial facility (Cocalico Biologicals, Reamstown, PA, USA). Approximately 3 kg of cabbage leaf tissue were used to obtain the necessary amount of protein. Proteins exhibiting H/BMT activity in the first PP-Q step of the isoform purification procedure were pooled (1.5 mg protein) and subjected to SDS-PAGE using a Protean II xi Slab Cell (Bio Rad). The gel was stained with 300 mM CuCl2, and the protein bands corresponding to H/BMT were cut and used without further elution. Polyclonal antibodies were produced by making a subcutaneous injection of ca. 100 µg denatured protein emulsified with Freund's complete adjuvant in a rabbit (2 animals were used to choose the best antisera). Boosts (50 µg) were given after 2, 3 and 7 weeks, using incomplete Freund's adjuvant. Serum collected after the last injection did not contain enough specific antiserum, therefore, boosting injections were continued for another 5 weeks. The serum collected one week after the last injection was aliquoted and stored at -80°C.


3.13.2. Western Blotting

      Proteins from SDS-PAGE were transferred to a nitrocellulose membrane for 90 min (1 mA cm-2) in an LKB semi-dry transfer unit (Pharmacia), and visualized by staining with Ponceau Red (0.2% in water). Prior to hybridization, blots were blocked for at least 1 h with Tris-buffered saline (TBS, 25 mM Tris-HCl pH 7.4, 150 mM NaCl) containing 0.1% Tween 20 and 5% nonfat milk powder. For immunodetection, the blots were incubated at room temperature for 2 h with antiserum diluted (1:100) in TBS containing 0.1% Tween 20 and 3% nonfat milk powder, and then with goat anti-rabbit immunoglobulin G conjugated either to alkaline phosphatase (AP, Sigma) or to horseradish peroxidase (HRP, Amersham). The AP reaction was detected colorimetrically with nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-indolyl-phosphate (BCIP) as the substrates. The HRP reaction was detected by chemiluminescence. Light produced upon oxidative degradation of Luminol by HRP was captured on a Kodak X-OMAT film (Kodak).


3.13.3. Purification Of Antibodies From Immunoblots

      H/BMT-specific antibodies were purified as described by Harlow and Lane (1988). A purified preparation of H/BMT was subjected to SDS-PAGE and transferred to nitrocellulose. The bands corresponding to H/BMT were cut and blocked as described in section 3.12.2. The blots were incubated for 90 min with antiserum (diluted 10-fold in TBS plus 0.1% Tween 20 and 5% nonfat milk powder). The blots were washed with TBS containing 0.1% Tween 20 and then incubated for 10 min in 100 mM glycine buffer, pH 2.5. The solution, containing eluted antibodies, was finally neutralized with 150 µl of 1 M Tris-HCl, pH 8.5, and stored at 4°C.


3.14. Analysis Of H/bmt Amino Acid Sequence

      Sequencing of H/BMT polypeptide was done at a commercial service (Harvard MicroChem Facility, MA, USA). The purified H/BMT was electrophoretically transferred from an SDS-PAGE gel onto a polyvinylidene difluoride (PVDF) membrane (Bio Rad) as described by Moos et al. (1988). The blotted protein was subjected to in situ tryptic digestion according to Aebersold et al. (1987). The digestion was done overnight (24 h) at 37°C in 0.1 M Tris-HCl buffer containing 1% RTX and 10% acetonitrile, pH 8.2, and 0.3 µg of Promega modified trypsin (Promega) (Dr. Renee Robinson, Harvard MicroChem, personal communication). Tryptic fragments were then subjected to reverse phase HPLC with a Zorbax C18 (1 x 150 mm) Microtech column, and isolated by elution with a 0-56% linear gradient of acetonitrile in 0.05% trifluoroacetic acid at a flow rate of 100 µl min-1. Peptide elution was monitored at 205, 277 and 292 nm. Five peaks were further analyzed by matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI-TOF MS) using a Finnigan Lasermat 2000 mass spectrometer (Hemel, UK). Three peptides, P77, P90 and P97 were then chosen for microsequencing and were subjected to automated Edman degradation using an Applied Biosystems 477A gas-phase sequencer (Harvard MicroChem).


3.15. Molecular Methods


3.15.1. Isolation And Purification Of Total And Poly(a)+ Rna

      Total RNA was extracted from young cabbage leaves as follows: Leaf tissue was frozen with liquid nitrogen and ground in a mortar into a fine powder. The powder was then mixed with 1.5 volumes of extraction buffer (1 M Tris-HCl pH 7.5, containing 10 mM EDTA and 1% SDS) and 1 volume of a phenol:chloroform:isoamyl alcohol mixture (50:48:2). The mixture was homogenized for 15 min by vigorous shaking (300 rpm) on an orbital shaker at 4°C. After centrifugation, the aqueous phase was recovered and nucleic acids were precipitated by adding 1/10 volume of a 3 M sodium acetate (NaOAc) solution, pH 5.2, and 2 volumes of 100% ethanol. The suspension was kept for 20 min at -20°C. Nucleic acids were resuspended in a minimum volume of deionized water. Total RNA was selectively precipitated by adding an equal volume of a 4 M lithium chloride solution, and incubating the mixture overnight on ice. After final centrifugation, RNA was stored as an ethanol suspension at -20°C for further use. Polyadenylated RNA was purified by oligo(dT)-cellulose chromatography using the mRNA Separator Kit (Clontech, CA, USA), according to the manufacturer's instructions.


3.15.2. Rna Electrophoresis

      Up to 20 µg total and 5 µg poly(A)+ RNA were denatured by heating at 70°C for 15 min, and separated by electrophoresis on 1.2% agarose gels. The gels contained 2.2 M formaldehyde, and were run in 1X MOPS (50 mM MOPS, pH 7.3 plus 1 mM EDTA) containing 2 M formaldehyde. Nucleic acids were visualized by staining with ethidium bromide.


3.15.3. Rna Transfer

      Gels were washed with several changes of deionized water for at least 2 h, then equilibrated for 1 h with 10X SSC buffer (1.5 M NaCl, 0.15 M sodium citrate, pH 7.0). RNA was transferred by capillarity onto GeneScreen Plus positively charged nylon membranes (NEN, MA, USA) in 10X SSC for 16 h. Transferred RNA was stained with methylene Blue (0.02% in 0.3 M NaOAc pH 5.5). The nucleic acids were fixed to the membrane by baking it at 80°C for 90 min in a vacuum oven.


3.15.4. Northern Hybridization

      Oligonucleotide probes were synthesized from the Arabidopsis EST and labeled by random priming with 50 µCi of [32P]dCTP according to Feinberg and Vogelstein (1984) using the High Prime labeling kit (Boehringer Mannheim). The radiolabeled probe was separated from unincorporated nucleotides by gel filtration on Sephadex G-50 Quick Spin columns (Boehringer Mannheim). Specific activities of probes ranged between 108 and 109 cpm µg-1 DNA.

      RNA blots were prehybridized overnight with 250 mM phosphate buffer pH 7.0 containing 7% SDS (w:v), 1% BSA (w:v) and 1 mM EDTA, at 42°C, in continuously rotating tubes in a hybridization oven. Hybridization was performed overnight under the same conditions. The membranes were then washed with SSC containing 0.1% SDS (w:v) as follows: the first wash was done with 5X SSC for 1 h at 65°C, the second with 1X SSC for the same time at 50°C, and the third with 2 changes of 0.1X SSC, each for 15 min at 50°C. The hybridization was detected by autoradiography by exposing the membranes to Kodak X-OMAT films for 24 to 36 h and developing the films with a Curix 60 automated developing machine (Agfa).


4. Results

      An AdoMet-dependent halide/bisulfide methyltransferase capable of catalyzing the methylation of X- or HS- ions to CH3X or CH3SH, respectively, was purified to homogeneity from leaves of Brassica oleracea. Earlier kinetic analysis of a partially purified preparation of this enzyme had revealed mutual competitive inhibition between the alternative substrates I- and HS-, strongly suggesting that the H/BMT catalyzed the methylation of both these substrates (Attieh, 1993). Purification to homogeneity of a single protein that possesses these two activities confirmed the dual catalytic nature of H/BMT. The purification was achieved through a novel procedure employing an unusual affinity matrix. The enzyme existed in at least five isoforms, which were all purified to apparent homogeneity. The three most abundant of these isoforms were biochemically characterized. H/BMT was localized in the soluble, cytosolic fraction of the cell, and was shown to be involved in sulfur metabolism as a thiol methyltransferase.


4.1. Purification And Characterization Of H/bmt


4.1.1 Enzyme Purification

      Prior to this work, the enzyme had been extracted from B. oleracea by subjecting leaf tissue to a freeze-thaw cycle follwed by homogenization. H/BMT was partially purified using (NH4)2SO4 precipitation, gel filtration chromatography on Sephadex G-100 and affinity chromatography on adenosine-agarose (Attieh, 1993). At this level of purification, the enzyme preparation was enriched but was still contaminated with a number of other proteins as shown in Figure 5, Lane E.

      The present work completed the purification of the enzyme to homogeneity through two additional steps, involving HPLC on anion exchange and high-resolution gel filtration columns. The anion exchange chromatography step on Protein Pak Q column not only removed a large number of the contaminating proteins, but also resolved the H/BMT activity into at least four peaks (Figure 6). The first two of these peaks eluted between 100 and 180 mM NaCl, and the second two between 200 and 300 mM NaCl (Figure 6). This suggested that the enzyme has multiple isoforms. The first, most prominent, peak of activity, eluted from this column at approximately 120 mM NaCl, and contained only two proteins as observed after separation through SDS-PAGE (Figure 5, Lane F). Further purification of this peak by HPLC on an analytical Superdex 75 gel filtration column separated these two proteins, only one of which contained H/BMT activity. This peak appeared free of contaminants when visualized by SDS-PAGE (Figure 5, Lane G). The specific activity of the enzyme after this step was enriched by more than 1000-fold over the crude preparation, with a recovery of 0.35%. Table II summarizes the purification procedure, and Figure 5 shows a typical pattern of proteins from various purification steps separated

      
Table II. Purification of the first anion exchange peak of halide/bisulfide methyltransferase.
Step Total activity Total protein Specific activity CH3I. CH3SH Ratio Purification   Recovery
  CH3I CH3SH   CH3I CH3SH   CH3I CH3SH   CH3I CH3SH
  picokatalsa mg picokatals/mg   -fold   %
Crude extract 10889 7944 4276.0 2.5 1.9 1.37 1 1   100 100
85%(NH4)2SO4 8667 7333 860.5 10.1 8.5 1.18 4.0 4.43   79.6 92.3
Sephadex G-100 4611 3278 300.7 15.3 11.0 1.41 6.1 5.87   42.3 41.3
Ad-agarose 2305 1972 0.722 3192 2731 1.17 1277 1437   21.2 24.8
PP-Q peak 1 169.4 130.5 0.041 4132 3183 1.29 1653 1675   1.55 1.65
Superdex 75 34.6 29.17 0.0135 2563 2148 1.19 1025 1137   0.3 0.36

      aOne katal of halide/bisulfide methyltransferase is defined as the amount of enzyme which catalyzes the conversion of one mole of substrates per second under the assay conditions.

      

Figure 5. SDS-PAGE of fractions with halide/bisulfide methyltransferase activity from the successive steps of purification.

      Lane A, crude extract (~10 µg); Lane B, 60-85% (NH4)2SO4 (~5 µg); Lane C, Sephadex G-100 (~5 µg); Lane D, Amicon YM-30 (~5 µg); Lane E, adenosine-agarose (~5 µg); Lane F, Protein Pak Q (~5 µg); Lane G, Superdex 75 (~1µg). The molecular mass markers are indicated to the left in kDa. The gel was stained with high sensitivity Coomassie brilliant Blue G-250.

      

Figure 6. Elution of the halide/bisulfide methyltransferase from a Protein Pak Q column upon anion exchange high performance liquid chromatography.

      Bound proteins were eluted in a 0-200 mM and 200-500 mM two-step linear gradient of NaCl in buffer C. Halide (G) and bisulfide (J) methylating activities were assayed in each fraction. Proteins were monitored by UV absorbance at 280 nm (--). nkat, nanokatals.by SDS-PAGE and visualized by Coomassie Blue staining. The apparent molecular mass of the purified H/BMT, calculated from its migration upon SDS-PAGE, was 28 kDa.


4.1.2. Confirmation Of The Dual Activity Of H/bmt

      The I-- and HS--methylating activities in cabbage leaf extracts co-purified to homogeneity in a constant ratio (1.26 ± 0.14); at no step in the procedure, including the strong anion exchange chromatography, were they even slightly separable from each other (Table II, Figure 6). Moreover, it was previously shown through the analysis of double reciprocal plots of the data from competition kinetics between I- and HS-, that the two substrates competed for the same active site on the enzyme (Attieh, 1993). These results, taken together, demonstrate that the halide and bisulfide methylations are catalyzed by a single protein on the same active site.


4.1.3. Purification Of H/bmt Isoforms

      The anion exchange step had resolved H/BMT activity in four peaks after NaCl gradient elution (0-500 mM) (Figure 6). However, by using a shallower two-step NaCl gradient of 0 to 160 mM (50 min) and 160 to 200 mM (15 min), H/BMT activity was resolved in five peaks all of which eluted between 120 and 200 mM NaCl (Figure 7). This pattern was consistently reproducible. Therefore, the purification procedure was modified to purify these five proteins to homogeneity. The

      

Figure 7. Elution of halide/bisulfide methyltransferase isoforms from a Protein Pak Q column upon anion exchange HPLC.

      The solid circles represent H/BMT activity after the standard purification procedure, and the open circles represent the activity after an alternative procedure not involving salt precipitation. Bound proteins were eluted in a 2-step NaCl gradient from 0 to 160 mM and from 160 to 200 mM. nkat, nanokatals.enzyme preparation was processed through the steps up to Sephadex G-100 gel filtration chromatography as described in section 4.1.1. Steps from affinity onwards were modified as follows: the affinity step on adenosine-agarose was optimized to bind a larger amount of proteins. This was achieved by dephosphorylating AMP-agarose to yield more adenosine ligands than what was obtained from the dephosphorylation of ADP-agarose (Attieh, 1993). After the affinity step, the enzyme preparation was subjected to an anion exchange HPLC step on PP-Q, followed by a preparative gel filtration HPLC step on Superdex 75 Hiload column, and finally a second anion exchange HPLC step on PP-Q.

      

Figure 8. Elution of halide/bisulfide methyltransferase from an adenosine-agarose affinity column.

      Specifically bound proteins were eluted in a 0-7 mM AdoMet gradient in buffer B. Fractions of the eluate were assayed for halide methyltransferase activity (J). Proteins were continuously monitored by absorbance at 280 nm (--). Most of the increase in the absorbance with the gradient elution was due to the presence of AdoMet. nkat, nanokatals.

      The alkaline phosphatase was able to remove 62% of the bound PO43- groups from the AMP-agarose support as compared to a maximum of 40% from the ADP-agarose. However, a noticeable difference between the two resulting adenosine-agarose supports was that, using the AMP-agarose dephosphorylation product, a higher concentration of AdoMet had to be used in the gradient (0-7 mM) in order to efficiently elute the H/BMT activity. The AdoMet gradient used in the previous column was from 0 to 4 mM (Attieh, 1993). The new adenosine-agarose column bound all H/BMT activity and substantially reduced the number of contaminating proteins. H/BMT activity eluted as a single peak between 3 and 5 mM AdoMet (Figure 8). Application of the preparation to the first anion exchange HPLC step removed the high concentration of positively-charged AdoMet from the sample, and the subsequent high NaCl elution (500 mM) concentrated it. As a result, the whole enzyme preparation from this step, concentrated in 2 mL, was loaded onto the preparative Superdex 75 Hiload gel filtration column. This column separated the low molecular weight H/BMT isoforms from other contaminating proteins and resolved the activity in two peaks at 31 and 26 kDa (Figure 9). Five H/BMT proteins were finally resolved in homogeneous state through the second anion exchange chromatography using a 2-step NaCl gradient for elution. Isoforms I, II, III and IV eluted between 120 and 140 mM NaCl, and isoform V at 190 mM NaCl (Figure 7). Figure 10 shows the profiles of proteins separated by SDS-PAGE after various purification steps, and visualized with Coomassie Blue. Each of the five H/BMT proteins purified through the final step, migrated as a single band. This protocol, summarized in Table III, resulted in 1600- to 18615-fold enrichment in specific activities of different isoforms. The five proteins were all able to carry out the typical H/BMT reactions, thus confirming the previous indication that H/BMT had multiple isoforms. The apparent molecular masses of these proteins, determined from migration upon denaturing SDS-PAGE, were 28, 26, 30, 31 and 31 kDa. In contrast, the proteins had eluted in only two peaks from the earlier gel filtration step in this procedure (Figure 9).

      
Table III. Purification of H/BMT isoforms from 500 g of Brassica oleracea leaf tissue.
Step Activity Protein Specific activity Purification Recovery
  picokatals mg picokatals/mg -fold %
Crude extract 4355.7 3350.2 1.3 1 100
60-85% (NH4)2SO4 2928.4 380.3 7.7 5.9 67.2
Sephadex G-100 2677.7 239.1 11.2 8.6 61.5
Adenosine-agarose 1713.2 0.870 1969.2 1514.8 39.3
Protein Pak Q I 1003.3 0.435 2306.4 1774.1 23.0
Superdex-75 1001.3 0.241 4154.8 3196 22.9
Protein Pak Q II          
Isoform I 64.4 0.017 3788.2 2914 1.5
Isoform II 211.5 0.057 3710.5 2854.2 4.8
Isoform III 152.2 0.073 2084.9 1603.8 3.5
Isoform IV 24.2 < ;0.001 > ;24200 18615.4 0.55
Isoform V 176.7 0.0112 15776.8 12136 4.0

      

Figure 9. Elution of halide/bisulfide methyltransferase from a Superdex-75 HiLoad gel filtration HPLC column.

      Halide methylating activity (J) and the protein content (--) were measured in each fraction. The estimation of native molecular mass of the isoforms is shown in the inset (J). The column was calibrated as described in section 3.7. The native molecular mass of H/BMT was determined by comparison to a standard curve made with Kav values of the standards against their log Mr. pkat, picokatals.

      The likely reason for this difference is that the closeness of the molecular sizes of the proteins would tend to bunch them together during gel filtration chromatography.

      

Figure 10. SDS-PAGE analysis of proteins at different steps of isoform purification.

      Lane A, crude extract (20 µg); lane B, Sephadex G-100 (20 µg); lane C, adenosine-agarose (5 µg); lanes D, E, F, G and H, isoforms I (~0.1 µg), II, (2 µg), III (1 µg), IV (~0.1 µg) and V (~0.5 µg), respectively. The molecular mass markers are indicated to the left in kDa. The gel was stained with Coomassie Brilliant Blue R.


4.1.4. Alternative Purification Of H/bmt Isoforms

      Protein precipitation with (NH4)2SO4 could have altered the charge distribution on protein molecules, and thus resulted in partial charge modifications. This could have led to a differential elution of the same protein from an ion exchange column. A possibility existed that the apparent isoforms of H/BMT could have resulted from such modifications because the purification procedure for this enzyme involved precipitation with an extremely high (NH4)2SO4 concentration of 85%. Therefore, the crude extract was also processed through an alternative procedure, in which the (NH4)2SO4 precipitation step was omitted. The crude extract was directly subjected to affinity chromatography on adenosine-agarose and then resolved by PP-Q anion exchange HPLC. The elution profiles of H/BMT activity from the anion exchange steps in this and the standard procedure were very similar (Figure 7). This indicated that charge modifications were not the cause of the multiple peak pattern observed with H/BMT upon anion exchange chromatography, and confirmed that the five purified proteins were indeed true isoforms


4.1.5. Stability Of The Purified H/bmt

      After Sephadex G-100 gel filtration, H/BMT was stable for over 2 months at -80°C in buffer B. However, after the affinity chromatography step, the enzyme became extremely labile, losing all activity after overnight storage at 4°C or -80°C. Addition of 20% glycerol to the preparation conserved 12% of the activity after storage at -20°C for 48 hours. In contrast, when stored after the PP-Q anion exchange step at 4°C in buffer C containing 175 mM NaCl, the enzyme retained more than 70% and 55% of its activity after 24 and 48 hours, respectively.


4.2. Characterization Of H/bmt

      The characterization of H/BMT was first done using an enzyme preparation purified through the Sephadex G-100 gel filtration step, which contains all the isoforms. In later studies, the purified isoforms were characterized separately. In the latter case, only the three most abundant isoforms II, III and V, were used; the quantities of the other two were too low to allow further characterization.


4.2.1. Determination Of The Molecular Mass Of H/bmt

      The native H/BMT had a molecular mass of 29.5 kDa as determined by conventional gel filtration on Sephadex G-100 (Attieh, 1993). The higher resolution of Superdex 75 column separated 2 molecular mass species of 26 and 31 kDa. As all these values are close to the molecular mass values of 26, 28, 30, and 31 kDa of the denatured proteins determined by SDS-PAGE (Figure 10), all the active isoforms exist as monomers.


4.2.2. Effect Of Sh Group Inhibitors On H/bmt Activity

      The effect of thiol-group inhibitors on H/BMT activity is shown in Table IV. 4-Hydroxymercuribenzoic acid (10 mM) severely inhibited H/BMT activity, and this inhibition was completely reversed by the presence of 20 mM cysteine. Ten millimolar cystine or iodoacetamide reduced H/BMT activity by 55%. The inhibition by cystine was fully reversed by the simultaneous presence of 5 mM DTT. This indicated that H/BMT contains at least two SH groups that must remain reduced in order to preserve activity.


4.2.3. Effect Of Ph On H/bmt Activity

      
Table IV. Effects of -SH group inhibitors on the I- methylation activity of halide/bisulfide methyltransferase.
-SH Inhibitor (mM) Activator (mM) CH3I Production(% of control)
None DTT (5) 100a
Iodoacetamide (10) None 45
Iodoacetamide (10) DTT (5) 98
4-hydroxymercuribenzoate (10) None 5
4-hydroxymercuribenzoate (10) Cysteine(20) 104
Cystine (10) None 45
Cystine (10) DTT (5) 102

      aThe rate of control activity was 1.2 pkat mg protein-1

      The crude enzyme preparation was desalted on a PD-10 column to remove DTT. The desalted preparation was preincubated for 30 min with the inhibitor before adding the substrates.

      

Figure 11. Effect of pH on halide/bisulfide methyltransferase activity.

      The pH dependence of halide (A) and bisulfide (B) methylating activities was investigated using 100 mM of the following buffers: phosphate-citrate (J), phosphate (E) and Tris-acetate (I). nkat, nanokatals.

      Figure 12. Effect of pH on the activity of halide/bisulfide methyltransferase isoforms.

      

      The pH dependence of isoforms II (A), III (B) and V (C) was determined using 100 mM of the following buffers: citrate-phosphate (J), phosphate (E), and Tris-acetate (I). nkat, nanokatals.


4.2.4. Substrate Specificity Of H/bmt


4.2.4.1. Methyl-Group Acceptor Specificity Of H/bmt

      Other than the ability of H/BMT to methylate X- and HS- ions, nothing was known about its substrate specificity. In order to gain further understanding of the physiological role of this novel enzyme, a wide variety of compounds had to be tested as possible substrates. This was done using a simple and quick assay based on the reasoning that if a compound was a substrate for H/BMT, it should compete against I- for the enzyme's active site and thus inhibit I- methylation. Therefore, the extent to which a compound inhibited I- methylation in the standard H/BMT assay was considered a measure of its being a putative substrate (Table V).

      Whether I- was supplied as Na or K salt had no effect on I- methylation activity of the enzyme (KI, 1 pkat mg-1 protein; NaI, 0.9 pkat mg-1 protein). Consistent with the rationale for this assay, the known substrates of H/BMT HS-, Br- and Cl-, inhibited I- methylation (Table V) in proportion to the relative preference by H/BMT for them (Attieh, 1993; Saini et al., 1995). In contrast, a number of monovalent or divalent anions had no effect on I- methylation, indicating that these were neither substrates nor inhibitors of H/BMT.

      
Table V. A survey of potential substrates for the halide/bisulfide methyltransferase.
Substrate Concentration (mM) CH3I Production (% of control)
inorganic anions    
KI 20 100a
NaBr 20 98
  10 101
NaCl 200 81
  100 94
  50 98
NaF 40 95
  20 120
(NH4)2S 20 53
  4 81
Na2Se 20 61
  5 94
KCN 40 73
  20 82
KNO3 20 95
KH2PO4 20 106
NaHCO3 20 112
KPPi 20 116
K2SO3 20 111
Na2SO4 20 111
Na2SeO3 20 34
  5 30
K2SeO4 20 111
  Continued on the next page
Na2O5S2 20 115
Na2MoO4 20 116
AsHNa2O4 20 123
NaN3 40 110
  20 103
H2O2 40 41
  20 99
Organic compounds    
Iodoaniline 20 73
  10 106
Iodobenzoic acid 20 78
  10 117
CH3COOK 20 116
Quercetin 20 109
Salicylic acid 20 107
Aniline 20 119
Phenol 20 121
4-Nitrophenol 20 104
Thiols    
Thiocyanate 20 0
  1 0
  0.1 54
Thiosalicylic acid 20 44
Thiobenzoic acid 20 48
  10 77
4,4'Thiobisbenzenethiol 20 0
  1 38
Thiophenol 20 0
  1 0
  0.1 91
Thiosuccinic acid 20 92
Thioglycolic acid 20 102
6-Mercaptopurine 20 107
  Continued on the next page
Methanethiol 15 110
Isothiocyanates (ITC)  
Phenyl ITC 5 149
Phenethyl ITC 5 157
Benzyl ITC 5 178
Propyl ITC 5 151
Biological thiols    
L-homocysteine 20 103
L-cysteine 20 96
Glutathione 20 94
L-methionine 20 103
D-methionine 20 105

      aThe rate of control activity was 1 pkat mg protein-1

      This was tested by the ability of various compounds to inhibit the methylation of 20 mM I- (5 mM with isothiocyanates) in the standard assay for H/BMT activity using a crude enzyme preparation

      Among such ions, only HSe-, CN-, and SeO32- partially inhibited I- methylation. The oxidizing agent, H2O2, also partially inhibited I- methylation by H/BMT. Several of a variety of thiols and related phenols, anilines and organic acids significantly inhibited I- methylation (Table V). The most potent inhibitor was thiocyanate, which completely abolished I- methylation at a concentration 20-fold lower than that of I- in the assay. Isothiocyanates, on the other hand, had no inhibitory effect on I- methylation. The aromatic thiols 4,4'-thiobisbenzenethiol and thiophenol were also highly inhibitory (Table V). In general, short-chain thiols were considerably weaker inhibitors of I--methylation. Partial inhibition of I- methylation was also observed with iodoaniline and iodobenzoic acid, but O- or N-substituted derivatives of the above inhibitory chemicals, such as aniline, salicylic acid, phenol, p-nitrophenol and quercetin, did not inhibit I- methylation. Methanethiol, 6-mercaptopurine and the biological thiols, homocysteine, cysteine, glutathione, and methionine had no effect on H/BMT activity.

      Having thus narrowed down the number of potential substrates, a direct radiometric assay with a completely purified preparation of the enzyme was used to determine whether the compounds that inhibited I- methylation were in fact substrates for H/BMT. Thiocyanate ion as well as all the inhibitory thiols were efficiently methylated by the Superdex 75 HiLoad-purified enzyme preparation that contained all the isoforms (Table VI). Among these compounds, thiocyanate was the most preferred substrate, with a Km of 11 µM (Figure 13A), followed by 4,4'-thiobisbenzenethiol (Km 51 µM; Figure 13B), thiophenol (Km 250 µM; Figure 13C), and thiosalicylic acid (Km 746 µM; Figure 13D). Other inhibitory compounds, such as HSe-, CN-, and SeO32-, did not form any methylated products in this assay, indicating that they were not substrates for the enzyme.

      After the demonstration that a purified preparation of pooled H/BMT isoforms methylated thiocyanate and aromatic thiols (Table VI), isoforms II, III, and V were individually tested for their ability to methylate these substrates. All the isoforms methylated thiocyanate and the organic thiols 4,4'-thiobisbenzenethiol, thiophenol, and thiosalicylic acid to their methylthio derivatives (Table VII). O- or N-substituted equivalents of these thiols were not substrates for methylation by any of the purified isoforms (Table VI). Moreover, the biological thiols -- sulfur-containing amino acids and glutathione -- were not methylated. Kinetic studies with I-, HS-, thiocyanate, 4,4'-thiobisbenzenethiol, thiophenol and thiosalicylic acid showed that each isoform exhibited a significantly different set of Km values for these substrates (Table VII). Overall, the lowest Km values were for thiocyanate, i.e. 32 nM, 80 nM, and 3.5 µM for isoforms V, III, and II, respectively. The lowest Km values for 4,4'-thiobisbenzenethiol (3 µM), thiophenol (21 µM), and I- (370 µM) were found with isoform V. The lowest Km value for thiosalicylic acid (1.5 mM) was with isoform II, and for HS- (1.1 mM) with isoform III. The three isoforms had fairly comparable Km values for AdoMet (Table VII).

      

Figure 13. Effect of substrate concentration on the activity of the purified, pooled isoforms of halide/bisulfide methyltransferase.

      The insets represent Lineweaver-Burk plots of the data. The ability of H/BMT to methylate the substrates was measured radiometrically using 3H-CH3-AdoMet. The Km values were 11 µM for thiocyanate (A), 51 µM for 4,4'-thiobisbenzenethiol (B), 250 µM for thiophenol (C), and 746 µM for thiosalicylic acid (D). nkat, nanokatals.

      
Table VI. Methylating activity of the halide/bisulfide methyltransferase against various organic thiols and other analogous substrates.
Substrate Activity (nkat mg-1 protein)
Thiocyanate 7620
4,4'-thiobisbenzenethiol 1850
Thiosalicylic acid 1000
Thiophenol 460
Thiobenzoic acid 6
4-Nitrophenol N.D.
Quercetin N.D.
Aniline N.D.
Phenol N.D.
Salicylic acid N.D.

      Enzyme activity was assayed by measuring the amount of radiolabel in the 3H-methylated product formed from [3H-methyl]AdoMet and the substrate tested (200 µM), using a pooled preparation of purified isoforms II, III and V; N.D., not detected.

      
Table VII. Effect of different methyl donors on the I- methylation activity of H/BMT.
Methyl donor Concentration (mM) CH3I Production (pkat mg-1protein)
None 0 N.D.
AdoMet 0.5 1.1
SMM 1 N.D.
  5 0.4
S-methyl cysteine 1 N.D.
  5 N.D.
Choline 1 N.D.
Dimethylsulfoniopropionate 1 N.D.
  5 N.D.
L-Met 1 N.D.
  5 N.D.
D-Met 1 N.D.

      AdoMet was replaced with the alternative methyl donors in the standard enzyme assay using a crude enzyme preparation; N.D., Not detected.


4.2.4.2. Methyl-Group Donor Specificity Of H/bmt

      Among a number of potential methyl donors tested, AdoMet was the only one that supported any appreciable production of CH3I by H/BMT (Table VIII). S-methylmethionine acted as a weak methyl donor, and only at a very high concentration of 5 mM; at this concentration of SMM, the I- methylation rate was only 37% of the rate with 0.5 mM AdoMet. None of the other compounds tested was a methyl donor for H/BMT-catalyzed CH3I production (Table VIII).


4.3. Kinetic Analysis Of H/bmt

      It has been already shown that H/BMT methylated I- and HS- at the same active site ((Attieh, 1993), Table II, Figure 5). A similar situation appears to exist for the methylation of thiocyanate and the thiol substrates, because these compounds efficiently inhibit I- methylation by H/BMT (Table V), and they are methylated by all the purified isoforms of the enzyme (Table VII). Thus, methylation of all these substrates by any of the H/BMT isoforms should be governed by the same mechanism. Therefore, substrate interaction and product inhibition experiments were done using I- as the methyl acceptor substrate, because the GC-detectable product of this reaction is the easiest to analyze.

      
Table VIII. Effect of different methyl donors on the I- methylation activity of the halide/bisulfide methyltransferase.
Methyl donor Concentration(mM) CH3I Production(pkat mg-1protein)
None 0 N.D.
AdoMet 0.5 1.1
S-methylmethionine 1 N.D.
  5 0.4
S-methyl cysteine 1 N.D.
  5 N.D.
Choline 1 N.D.
Dimethylsulfoniopropionate 1 N.D.
  5 N.D.
L-Met 1 N.D.
  5 N.D.
D-Met 1 N.D.

      AdoMet was replaced with the alternative methyl donors in the standard enzyme assay using a crude enzyme preparation; N.D., Not detected.


4.3.1. Substrate Interactions

      Substrate interaction studies are useful in determining the general order of substrate binding for enzymes with more than one substrate. Double reciprocal plots of the initial velocity against various concentrations of one substrate at fixed concentrations of the other can have two patterns: The first, where lines intersect on the left of the vertical axis, indicates a sequential substrate binding mechanism, whereas the second, with parallel lines, indicates a Ping Pong mechanism (Cleland, 1970). Substrate interactions for the isoforms II, III and V at fixed concentrations of AdoMet and variable concentrations of I-, gave double reciprocal plots in which the lines intersected to the left of the vertical axis (Figure 14). These data are, thus, consistent with a sequential binding mechanism, and rule out any possibility of a Ping Pong mechanism. The data obtained were subsequently fitted to equation 4 (Section 3.11.), which represents the rate of a sequential two-substrate reaction.

      A sequential binding mechanism can be ordered or random. Therefore, the order of substrate binding and product release was determined from product inhibition experiments.

      fig014

      The effect of varying iodide concentrations on activities of isoforms II (A), III (B) and V (C), at different fixed S-adenosyl-L-methionine (AdoMet) concentrations is shown. nkat, nanokatals.


4.3.2. Product Inhibition

      For all three isoforms, double reciprocal plots of data for competition between AdoHcy and AdoMet, obtained with saturating concentrations of non-varied substrate, gave lines that intersected on the vertical axis (Figure 15). This indicates a competitive inhibition pattern between the reaction partners. Competition between AdoHcy and I- gave lines that intersected to the left of the vertical axis in double reciprocal plots (Figure 16). This is consistent with a noncompetitive inhibition pattern between AdoHcy and I-. Put together, the data generated from these studies indicated an Ordered Bi Bi mechanism for substrate binding and product release (Morrison and Ebner, 1971), whereby AdoMet is the first substrate to bind H/BMT, followed by the methyl acceptor substrate. The methylated product is the first to leave the enzyme, followed by AdoHcy (Figure 17). Product inhibition data were fitted to the equations 3 or 4 (Section 3.10.) to determine Ki values for AdoHcy (Table IX).

      
Table IX. Product inhibition kinetics for H/BMT isoforms.
Isoform Variable substrate   Constant substrate   Kis Kii
    mM     mM   µM µM
II AdoMet 0.025-0.2   I- 50   4.5  
  I- 2-50   AdoMet 0.5   3.5 7.1
                 
III AdoMet 0.025-0.2   I- 50   1.7  
  I- 2-50   AdoMet 0.5   3.1 12.3
                 
V AdoMet 0.025-0.2   I- 50   1.3  
  I- 2-50   AdoMet 0.5   0.7 23

      Kis and Kii, the slope and intercept inhibition constants, were determined by fitting data to Equation 3 or 4 in section 3.10. The inhibitor product used was AdoHcy (0-40 µM).


4.4. Intracellular Localization Of H/bmt

      Since H/BMT catalyzes the methylation of HS- and I- on the active site (Table II), only the I- methylation activity of H/BMT was monitored and used as the index of H/BMT activity in these experiments.

      Cabbage chloroplasts were purified using a method that is regularly used for spinach. Purified, intact cabbage chloroplasts showed no H/BMT activity, neither under the standard conditions for H/BMT assay nor when tested under different light conditions (Table X). Cabbage chloroplasts were considered physiologically active since they were physically intact (Figure 18) and were capable of carrying out light-dependent 14CO2 fixation and fatty acid biosynthesis (Table X). The latter, like CO2 fixation, occurs in the plastids (Stumpf, 1984). The procedure for chloroplast purification was further validated by the ability of spinach chloroplasts, purified through the same procedure, to carry out the marker enzyme assays (Table X).

      
Table X. Activity of H/BMT as compared to the rates of 14CO2 fixation and lipid synthesis in various chloroplast fractions.
  Crude
homogenate
Cabbage
chloroplasts
Intact spinach
chloroplasts
Activity
(pkat mg-1 chl)
  Intact Broken  
H/BMT 80.67 0 0 0
14CO2 fixed N.D. 116.67 5 158.33
Lipid synthesized N.D. 20.8 1.5 108.33

      Enzyme activities were not expressed on protein basis because of the presence of BSA in the chloroplast extraction medium. H/BMT, halide/bisulfide methyltransferase; N.D., not determined.

      

Figure 15. Product inhibition kinetics of halide/bisulfide methyltransferase reaction with respect to S-adenosyl-L-methionine (AdoMet).

      Inhibition of the methylation reaction for isoforms II (A), III (B) and V (C), by S-adenosylhomocysteine (AdoHcy) was done by varying AdoMet concentrations at different fixed concentrations of AdoHcy. Iodide concentration was fixed at 50 mM. nkat, nanokatals.

      

Figure 16. Product inhibition kinetics of halide/bisulfide methyltransferase reaction with respect to iodide, presented as double reciprocal plots.

      Inhibition of the methylation reaction for isoforms II (A), III (B) and V (C), by S-adenosylhomocysteine (AdoHcy) was done by varying iodide concentrations at different fixed concentrations of AdoHcy. S-adenosyl-L-methionine (AdoMet) concentration was fixed at 0.5 mM. nkat, nanokatals.

      

Figure 17. Reaction mechanism of the halide/bisulfide methyltransferase.

      Substrate binding and product release of H/BMT obey an Ordered Bi Bi mechanism whereby the donor substrate and the methylated product are the first to bind and leave the enzyme, respectively.

      Table XI shows the activities of H/BMT and marker enzymes in various fractions after differential centrifugation of the crude homogenate. As expected, the greatest activities of triose phosphate isomerase, NADH cytochrome c reductase, succinate dehydrogenase, and catalase (the marker enzymes) were in the cytosol (100,000g Sup.), endoplasmic reticulum (100,000g Pellet and sup.), mitochondria (10,000g Pellet) and microbodies (500g pellet and 100,000g sup.), respectively. By far, the highest H/BMT activity was in the 100,000g supernatant, along with the cytosolic triose phosphate isomerase. The low amounts of H/BMT activity in the pellets were probably due to cytosolic contamination. Further immunolocalization studies using polyclonal antibodies raised against the purified H/BMT in rabbit, could not be attempted because of the extremely low specificity of the antiserum. Purification of the H/BMT-specific antibodies from immunoblots containing the purified enzyme as the ligand, and using the purified antisera did not reduce the background levels in subsequent Western blotting experiments. Furthermore, these antibodies failed to immunoprecipitate H/BMT activity even at antiserum concentrations as high as 1:10 in buffer (v:v).

      
Table XI. Activity of H/BMT as compared to those of marker enzymes after differential centrifugation.
 
Crude
Subcellular fraction
Enzyme activities homogenate 500g pellet 10,000g pellet 100,000g pellet 100,000g supernatant
H/BMT
(pkat)

438 (100)

6 (1.3)

6 (1.4)

3 (0.6)

227 (52)
TPI
(nkat)

7898 (100)

362 (4.6)

94 (1.2)

83 (1)

12724 (161)
CCR
(µkat)

48 (100)

2 (4.6)

9 (18.5)

11 (23.1)

57 (120)
SDH
(nkat)

127 (100)

6 (4.5)

9 (7.3)

1 (0.8)

0 (0)
Cat
(µkat)

328 (100)

171 (52)

65 (20)

50 (15.2)

317 (96.6)
Protein
(mg)

286 (100)

48 (16.7)

41 (14.3)

8 (2.7)

155 (54.3)

      H/BMT, halide/bisulfide methyltransferase; TPI, triose phosphate isomerase; SDH, succinate dehydrogenase; CCR, NADPH cyt C reductase; Cat, catalase. Numbers in brackets represent % recoveries compared to crude homogenate.

      

Figure 18. Phase microscopy of chloroplast preparations.

      Chloroplasts from cabbage (A) and spinach (B) were purified on a Percoll gradient and viewed under differential interference contrast microscopy.


4.5. Association Of H/bmt Activity With Glucosinolates

      H/BMT accepted thiocyanate ion and a number of aromatic thiols as substrates (Tables V & VI). The thiocyanate ion, and analogues of these thiols, are known metabolites in glucosinolate-accumulating plants (Brown and Morra, 1997). Since Brassica oleracea, the source of this enzyme, is a glucosinolate accumulator, a strong possibility existed that H/BMT could be involved in glucosinolate metabolism. This possibility was further investigated by determining the presence of H/BMT activity in members of different plant families known to accumulate glucosinolates. H/BMT activity was present in all the surveyed plants belonging to glucosinolate-containing families Brassicaceae, Capparaceae, Resedaceae, and Caricaceae (Table XII). Representatives of the family Papaveraceae which lacks glucosinolates but is known for its active secondary metabolism, contained no measurable H/BMT activity. Furthermore, this activity was absent from a representative of the family Alliaceae, the members of which lack glucosinolates but accumulate disulfides and emit CH3SH (Harborne, 1973).

      
Table XII. A survey of higher plants showing the association of the halide/bisulfide methyltransferase activity with the presence of glucosinolates.
Species Family
(Glucosinolates)
In vivo H/BMT
(nmol h-1 g-1)
Extractable H/BMT
(pmol min-1 mg-1 protein)
Brassica oleracea, Capitata Brassicaceae (+) 255 115.7
B. oleracea, Botrytis Brassicaceae (+) 238.3 32.1
Raphanus sativus Brassicaceae (+) 72.5 11.1
Iberis sempervirens Brassicaceae (+) 8.8 2.7
Cleome hassleriana Capparaceae (+) 72.5 7.2
Reseda luteola Resedaceae (+) 87.5 12.0
R. odorata Resedaceae (+) 67.0 14.5
Carica papaya Caricaceae (+) 39.3 5.0
Meconopsis cambrica Papaveraceae (-) N.D. N.D.
Papaver orientale Papaveraceae (-) N.D. N.D.
Chelidonium japonicum Papaveraceae (-) N.D. N.D.
Allium cepa Alliaceae (-) N.D. N.D.

      Extractable H/BMT activity was measured using crude extracts desalted on Sephadex G-25 columns; N.D., not detected


4.6. Molecular Studies


4.6.1. Amino Acid Sequence Analysis

      Isoform V, purified through the PP-Q anion exchange step, was subjected to Edman degradation in order to obtain N-terminal sequence information (Sheldon Biotechnology Centre, McGill University, Montreal, Canada). This reaction did not release any free amino acids, suggesting that the N-terminus of the protein was blocked. The denatured H/BMT was consequently subjected to tryptic digestion (Harvard MicroChem Facility, MA, USA), and the resulting peptides were isolated by reverse phase HPLC (Figure 19). Five polypeptides were further analyzed using MALDI mass spectrometry. Three of them, P77, P90 and P97 were chosen for microsequencing based on the symmetry of their respective spectra and their convenient size between 1 and 4 kDa (Figure 20). Internal amino acid sequences of these polypeptides were obtained (Table XIII). Peptides P77, P90 and P97 were composed of 16, 26 and 20 amino acids, respectively. Consistent with trypsin catalytic specificity, the terminal amino acid of each of the three peptides was basic. The sequence of these peptides were used to search the GenBank data base. No similarities with any methyltransferase or any other protein of known function were found. However, the polypeptides P77, P90, and P97 shared 81%, 88%, and 90% identity, respectively, with the deduced amino acid sequence of an Arabidopsis 493 base-long expressed sequence tag (EST) of unknown function (Newman et al., 1994). Alignment of the sequences of the polypeptides with that of the Arabidopsis EST indicated that the three peptides constituted a continuous stretch of 45 amino acids (Figure 21), with an overlap of 17 residues between P90 and P97 (Table XIII, Figure 21).

      

Figure 19. HPLC profile of peptide fragments from tryptic digestion of the halide/bisulfide methyltransferase.

      The PP-Q-purified isoform V was subjected to digestion with trypsin. The resulting fragments were separated by reversed phase HPLC on a Zorbax C18 column as described in section 3.14. The peaks corresponding to peptides P77, P90, and P97 are marked by asterisks.

      

Figure 20. Matrix-assisted laser desorption mass spectrum of one tryptic fragment (P77) of the halide/bisulfide methyltransferase.

      Three HPLC-purified peptides from H/BMT tryptic digestion were subjected to MALDI-MS and their peak integrity and molecular masses determined. Peptide P77, shown here as an example, was pure as judged by the symmetry of its peak, and had a suitable size (1.7226 kDa) between 1 and 4 kDa.

      
Table XIII. Amino acid composition of three internal tryptic digests of isoform V of the halide/bisulfide methyltransferase.
Peptide (kDa) Amino acid sequence
P77 (1.7226) Val-Ala-Val-Ser-Thr-Tyr-Glu-Asp-Val-Leu-Val-Pro-Val-Gly-Phe-Lys
   
P90 (2.9142) Glu-Leu-Leu-Lys-Pro-Asp-Gly-Glu-Leu-Ile-Thr-Leu-Met-Tyr-Pro-[Ile]-Thr-Asp-His-Asp-His-Asp-[Gly]-[Gly]-Pro-[Pro]-Tyr-(Lys)
   
P97 (3.2941) Ser-Met-Tyr-Glu-Leu-Leu-Lys-Pro-Asp-Gly-Glu-Leu-Ile-Thr-Leu-Met-Tyr-Pro-[Ile]-Thr

      [aa], amino acid identified with reasonable confidence, i.e. probable; (aa), amino acid identified with low confidence, i.e. possible.


4.6.2. Expression Of H/bmt In Cabbage Leaves

      The Arabidopsis EST used in this study was a gift from Dr. K. R. Davis (The Arabidopsis Biological Resource Center, Ohio State University, OH. USA). The expression of H/BMT was examined using Northern blots of total and poly(A)+ RNA with the Arabidopsis EST as a probe. No transcripts were detected in the Northern blots of total RNA from either cabbage or Arabidopsis. However, one transcript at 1 kb was observed with poly(A)+ RNA (0.5-2 µg) from cabbage (Figure 22). No signal was detected from blots of poly(A)+ RNA from Arabidopsis.

      

Figure 21. Similarity in amino acid sequence between the halide/bisulfide methyltransferase and other proteins.

      Alignment of partial amino acid sequence from cabbage H/BMT and the deduced amino acid sequence of an EST from Arabidopsis thaliana. Dots represent residues in the H/BMT sequence that are identical to those in the EST. Residues in the shaded area are overlapping between peptides P90 and P97.

      

Figure 22. Northern blot analysis of the halide/bisulfide methyltransferase transcript.

      Different amounts of cabbage poly(A+) RNA was blotted onto a nylon membrane and probed with a 493-base-long Arabidopsis EST labeled by random priming with [32P]dCTP. A, 2 µg; B, 1 µg; C, 0.5 µg.


5. Discussion

      Wuosmaa and Hager (1990) reported that a halophytic plant, M. crystallinum, and a marine alga, E. muricata, contain an AdoMet-dependent methyl chloride transferase that can methylate X- ions to the CH3X gases. Because the above two organisms grow in highly saline environments, the methylation of Cl- to CH3Cl was viewed as a Cl--detoxifying mechanism in salt tolerant plants (Wuosmaa and Hager, 1990). Thus, it presented a potentially novel possibility for improving salt tolerance by engineering the gene for this enzyme into Cl--sensitive crop species (Saini et al., 1995). Such a use of a fungal halide methylating enzyme was also suggested earlier by Harper (1985). These considerations triggered a search in our laboratory for halide methylating enzymes in higher plants, whereby it was shown that 87 of the 118 species surveyed contained a halide methyltransferase activity (Saini et al., 1995). The rate of methylation varied over four orders of magnitude among these plants. However, the highest activities were not associated with salt tolerance (Saini et al., 1995). Instead, plants from the Brassicaceae family exhibited the highest X- methylation rates. A bisulfide (HS-) methylating activity, which produced CH3SH, was also detected in all the plants exhibiting X- methylation activity (Saini et al., 1995). Biochemical characterization of the X- and HS- methylating activities was started by partial purification of the proteins from cabbage leaves (Attieh, 1993). The activities co-purified at a constant ratio through the three-step protocol used for their partial purification (Attieh, 1993). Further, the kinetics of competition between I- and HS- strongly suggested that the two methylating activities resided on the same enzyme protein (Attieh, 1993). Hence, the enzyme was tentatively named halide/bisulfide methyltransferase (H/BMT). The H/BMT had a high Km value of 85 mM for Cl- ion, its only physiologically relevant halide substrate (Attieh, 1993). In contrast, its Km of 4.7 mM for HS- was relatively low. This, together with the detection of the highest activities in the sulfur-rich family Brassicaceae (Saini et al., 1995), indicated that the natural role of this enzyme was more likely to be in sulfur metabolism rather than Cl- detoxification.

      The present work was undertaken as a continuation of the previous investigations (Attieh, 1993; Saini et al., 1995). It was aimed at understanding the role of H/BMT in plant metabolism, and at assessing its environmental significance in contributing to the emissions of halomethanes and methanethiol. The present study showed that the natural role of H/BMT is likely to be the methylation and consequent detoxification of thiols, specifically the thiocyanate ions and other thiol groups produced upon degradation of the sulfur-containing secondary metabolites, glucosinolates. This conclusion on the role of H/BMT was reached following a multistep study involving: (1) the purification of the enzyme and its 5 isoforms to homogeneity, (2) the determination of its subcellular location, (3) the study of its substrate specificity, and (4) the demonstration of its ubiquity in glucosinolate-containing plants.

      The X- methylation by H/BMT is apparently due to the structural similarities between these ions and the natural substrates of the enzyme, and would take place only if X- concentrations in the cell were extremely high.


5.1. Purification And Characterization Of H/bmt

      The present work was started with the assumption that X- ions might be natural substrates for H/BMT. Thus, I-, the most preferred substrate known at that time, was routinely used as the methyl acceptor for the assays of this enzyme activity, because its methylation product could be easily detected by GC. This was also the reason for continuing to use I- in several subsequent experiments, although other, more efficiently methylated substrates were later discovered.

      Partial purification of H/BMT was earlier achieved through (NH4)2SO4 precipitation, followed by conventional gel filtration chromatography on Sephadex G-100 and affinity chromatography on adenosine-agarose (Attieh, 1993). In the present work, the homogeneous protein was obtained through two additional chromatography steps, anion exchange HPLC on PP-Q and analytical gel filtration HPLC on Superdex 75. The 5-step procedure resulted in an overall purification of approximately 1000-fold with a recovery of 0.35% (Table II). The most dramatic enrichment of activity occurred during affinity chromatography on adenosine-agarose, which preferentially bound H/BMT while allowing most of the contaminating proteins to wash through during sample loading. Adenosine, linked via its adenine-C8 to an agarose matrix (Ag-adenosine, Pharmacia), has been used to purify several plant methyltransferases (De Carolis and Ibrahim, 1989; Dumas et al., 1988). Since this product was no longer available from its commercial source, a similar support was first prepared by dephosphorylating ADP linked to agarose via the adenine-C8 group (Attieh, 1993). However, the efficiency of the dephosphorylation of this resin (40%) was compromised by the need to remove two PO43- groups bound to each adenosine molecule. The efficiency of dephosphorylation was improved (62%) by using AMP-agarose which required the removal of only one PO43- per molecule of adenosine. The resulting adenosine-agarose resin was highly stable and was repeatedly used without any loss of its binding efficiency. Columns prepared through this procedure have since been successfully used to also purify other plant AdoMet-dependent methyltransferases (James et al., 1995; Wang et al., 1997). This technique should, therefore, be widely useful for purifying such enzymes.

      Although the enzyme preparation was considerably purified after affinity chromatography, many contaminants of different molecular masses still remained (Figure 5, Lane E). The homogeneous enzyme was obtained only after HPLC anion exchange on PP-Q and gel filtration on Superdex 75. The NaCl gradient used to elute the enzyme from the PP-Q column, clearly separated H/BMT from other contaminating proteins, and resolved the enzyme activity into four peaks, presumably representing four isoforms (Figure 6). All of these putative isoforms were able to carry out the methylation of X- and HS- (Figure 6). Analysis of the protein profile using SDS-PAGE indicated that the first peak of H/BMT activity had only one contaminating protein of about 42 kDa (Figure 5, Lane F). Since the other peaks of H/BMT activity were more contaminated than the first one, the latter was initially chosen for further purification of the enzyme. The high capability of the analytical Superdex 75 HPLC column to resolve proteins in the molecular mass range of 3 to 70 kDa, was the key to separating these two proteins. The resulting preparation was homogeneous for a 28 kDa protein with H/BMT activity (Figure 5, Lane G). The X- and HS- methylating activities co-purified to homogeneity in a constant ratio (Table II). This, together with the mutual competitive inhibition pattern exhibited by the substrates I- and HS- (Attieh, 1993), unequivocally establish that both activities exist on the same active site of a single enzyme protein.

      The above procedure was modified to purify the remainder of the H/BMT isoforms (Figure 6). The key to the purification of these forms was the use of two anion exchange HPLC steps separated by a preparative gel filtration HPLC step (Table III). The first anion exchange chromatography on PP-Q column concentrated the sample, thus allowing the entire preparation to be applied in the subsequent preparative Superdex 75 Hiload gel filtration step. Upon gel filtration, H/BMT activity was separated from the contaminating high molecular weight proteins, and was resolved in two peaks of 26 and 31 kDa that were pooled prior to application on the last column (Figure 9). Homogeneity was finally achieved through the second PP-Q chromatography using a shallow, 2-step NaCl gradient which resolved five protein species, all of which exhibited the H/BMT activity (Figure 7). To confirm that these five peaks were in fact isoforms and not the result of experimental artifacts, an alternative purification procedure was developed in which the harsh (NH4)2SO4 precipitation step was omitted. The profile of H/BMT activity after anion exchange HPLC in this procedure closely matched that in the standard protocol used to purify the isoforms (Figure 7). This indicated that the separation of multiple peaks of H/BMT activity through anion exchange was not due to charge modifications which could have occurred during salt precipitation. Further, the molecular masses of the proteins were different, ranging from 26 to 31 kDa (Figure 10). They also had significantly different kinetic properties toward each of their substrates (Table VII). Thus, each of these proteins is distinct in many physical and catalytic properties. These results, put together, confirm that the five proteins were indeed true isoforms. However, it remains to be determined if these isoforms are the products of more than one gene. The variety of molecular sizes of these isoforms suggests that they probably are. Internal amino acid sequence information from isoform V is now available. Although this sequence has no matches with any known protein, it has over 85% identity with the deduced amino acid sequence of an EST from Arabidopsis thaliana. This EST specifically recognized a message of 1 kb from cabbage, which closely approximates the size of the putative H/BMT transcript (Figure 22). The fact that the message was detected in Northern blots of poly(A+) RNA but not of total RNA suggests that the H/BMT mRNA is extremely low in abundance. This assertion is further supported by the fact that no message was detected in 10 µg of total RNA or 2 µg of poly(A)+ RNA from Arabidopsis, in which the H/BMT activity is over 20-fold lower than in cabbage (Saini et al., 1995).

      Many properties of H/BMT resembled those of other small molecule methyltransferases. The specific activities of the purified isoforms (2 to 24 nkat mg-1 protein) compared well with the typical 1-10 nkat mg-1 protein range for small molecule methyltransferases (Edwards and Dixon, 1991; Preisig et al., 1989; Takemura et al., 1992). The molecular mass of H/BMT isoforms (26-31 kDa) also fell within the 20-45 kDa range for most methyltransferases, which are typically monomers or dimers (De Carolis and Ibrahim, 1989; Fujioka, 1992; Weisiger and Jakoby, 1979; Wuosmaa and Hager, 1990). Since native and denatured H/BMT isoforms had similar molecular masses, they are all likely to be monomers.

      Three of the H/BMT isoforms shared the same sequential substrate binding mechanism. This conclusion was deduced from the pattern of converging lines obtained in double reciprocal plots of substrate interaction data (Figure 14). Product inhibition kinetics were used to determine the reaction mechanism of H/BMT. Since the inhibition by AdoHcy was competitive with respect to AdoMet (Figure 15) and noncompetitive with respect to the X- (Figure 16), AdoMet appears to be the first substrate to bind to the enzyme. The methylated product would be the first to be released and AdoHcy the last. This is consistent with an Ordered Bi Bi kinetic mechanism in transmethylation reactions, whereby the methyl acceptor substrate is the last to bind the enzyme, and the methylated product is the first to be released (Figure 17) (Morrison and Ebner, 1971). This mechanism is typical of many plant methyltransferases (De Carolis and Ibrahim, 1989; James et al., 1995; Khouri et al., 1988).

      The thiol-group inhibitors, cystine, iodoacetamide, and 4-hydroxymercuribenzoate (Bendall, 1963), severely inhibited H/BMT activity in the absence of the reducing agents DTT or cysteine (Table IV). This effect was fully reversed by the addition of either of the reducing agents. Inhibition by iodoacetamide and 4-hydroxymercuribenzoic acid indicated that reduced thiol groups in the enzyme molecule were necessary for the activity. Since cystine, an oxidizing agent that acts upon pairs of thiol groups (Bendall, 1963), also strongly inhibited H/BMT activity, the H/BMT protein must have at least two thiol groups that have to be in a reduced state for the enzyme to function. This seems to be a common feature of many small molecule methyltransferases which apparently have 4-6 cysteine residues, at least two of which are adjacent in the three-dimensional structure of the enzyme and are located in a catalytically important area (Fujioka, 1992). A reduced thiol group present in or near the active site of the enzyme molecule could act as base group which interacts with the nucleophile of the acceptor substrate in order to facilitate the reaction of the latter with the methyl group of AdoMet (Fujioka, 1992)

      A Sephadex G-100-purified preparation of H/BMT, containing all the isoforms, exhibited a wide pH optimum of 5-7.5 for I- methylation (Figure 11A), and a rather sharp optimum between pH 7 and 8 for HS- (Figure 11B). The most likely reason for this difference is the pH-dependence of the HS- ion concentration. Since H2S has a pKa1 of 7.02 (Crampton, 1974), HS- concentration would be expected to fall sharply as the pH drops below 7. Therefore, in further experiments, the pH optimum for I-, the concentration of which is not pH-dependent, was used as a measure of the real optimum for H/BMT activity. The finding that three of the isoforms of the enzyme, that were fully characterized, had different pH optima which together covered the range from pH 5 to 8 (Figure 12) explains the wide optimum observed earlier for the partially purified enzyme.


5.2. H/bmt Is A Thiol Methyltransferase

      The realization that H/BMT may be involved in sulfur- instead of X--metabolism, considerably broadened the spectrum of its potential substrates. Therefore, we developed a simple assay to search for potential substrates. A compound which can inhibit an enzyme reaction by competing against the substrate for binding to the active site could either be a substrate or a competitive inhibitor for the enzyme. Compounds with such properties usually have similarities in structure and charge distribution with the substrate. Their properties, such as hydrogen bonding, ionic or hydrophobic interactions, allow them to mimic the fit of the substrate within the active site. If the complex thus formed is further hydrolysable, the compound is considered a substrate, and if not, it is considered a competitive inhibitor (Dixon and Webb, 1979). Using this logic as a preliminary indication that a given compound could be a substrate for H/BMT, we examined the ability of a number of compounds to inhibit H/BMT-catalyzed I- methylation. Inhibition of I- methylation by the known alternative substrates, HS-, Br- and Cl-, in proportion to their relative preference by the enzyme (Attieh, 1993; Saini et al., 1995), confirmed that this approach was conceptually sound. Thus the inhibition of I--methylation by a compound could be used as a rapid screen for potential substrates. Since this method employed a crude enzyme extract, it was possible for halomethanes to be produced via an alternative peroxidation reaction, as has been reported in some marine algae (Pedersén et al., 1996). Our data, however, showed that this pathway did not contribute to the CH3I produced in the present assay because NaN3, a strong peroxidase inhibitor, did not affect CH3I production (Table V). Moreover, H2O2, a peroxidase inducer, did not promote CH3I production, and instead inhibited it at high concentration (Table V). This inhibition could be explained by the fact that H2O2, an oxidizing agent, would have oxidized the reduced thiol groups in the enzyme molecule, which are necessary for activity (Table IV).

      A variety of inorganic anions were screened as potential substrates or inhibitors for H/BMT. Among these, HSe- inhibited I- methylation. Since the methylation of I- and HS- take place on the same active site of H/BMT, HSe- could be competing against I- for binding to the enzyme. However, despite the ability of HSe- to inhibit CH3I production, it did not form any detectable amounts of methane selenol (CH3SeH) in the reaction headspace. This observation suggests that HSe- is a competitive inhibitor, but not a substrate, for H/BMT. Indeed, selenium and sulfur have very similar chemical properties, and biochemical systems do not always distinguish between the two elements (Shamberger, 1983). The protozoan Tetrahymena thermophila, which produces CH3SH and CH3SeH, has distinct HS- and HSe- methyltransferases (Drotar et al., 1987b). However, cabbage does not appear to have the ability to volatilize selenium through direct methylation. Selenite, but not sulfite or selenate, severely inhibited H/BMT activity; yet no volatile selenium species were detected in the reaction headspace. The reason for this inhibition is not clear, but could also include a blockage of the active site by direct interaction of the ion with the enzyme. The inhibition by CN- could also be explained by the structural similarity of this pseudo-halide to I-, allowing it to inhibit CH3I production by blocking the I- binding site on the enzyme, and acting as a competitive inhibitor.

      Thiol-substituted phenols and organic acids strongly inhibited CH3I production (Table V). By far, the most potent inhibitor of CH3I production was thiocyanate ion, followed by the aromatic thiol compounds, 4,4'-thiobisbenzenethiol, thiophenol, and thiosalicylic acid. The inhibition by thiocyanate could not have been due to the toxicity of the CN moiety of this molecule because CN- alone only slightly inhibited CH3I production. Similarly, the inhibition of CH3I production by aromatic thiols could not have been caused by hydrophobic interactions with the enzyme molecule, since the equally hydrophobic O- and N-substituted equivalents of these compounds had no effect on H/BMT activity (Table V). These results indicated that the thiol groups in these compounds were the key factors in their inhibitory action on CH3I production, and that H/BMT probably methylated these thiol groups (Table V).

      To determine if the thiol groups of the above inhibitors of CH3I production were indeed methylated by H/BMT, a more accurate and direct radiometric assay using 3H-CH3 AdoMet along with the completely purified enzyme was developed. This enzyme preparation, containing all H/BMT isoforms (purified through Superdex 75 Hiload), efficiently methylated the organic thiols that inhibited I- methylation. Since radio-labeled products were formed only in the presence of active enzyme and the inhibitory thiols, it was confirmed that these thiols were indeed potent substrates for H/BMT (Table VI). Kinetic studies done on this enzyme preparation gave the low Km values of 11 µM for thiocyanate, 51 µM for 4,4'-thiobisbenzenethiol, 250 µM for thiophenol, and 746 µM for thiosalicylic acid (Figure 13). These are the lowest Km values of H/BMT for any substrates so far tested (Attieh, 1993). The possibility that methylthiocyanate, and methylthio derivatives of thiols could be formed spontaneously by a nucleophilic replacement by thiols with a CH3X that could have been formed by H/BMT and an X- ion is also ruled out since AdoMet was used as a SO42- salt, and the buffers were halide-free.

      Characterization of the three most abundant isoforms of H/BMT (II, III and V) revealed that each one of them methylated thiocyanate and the above thiols, in addition to X- and HS-. They had significantly different kinetic properties toward each of these substrates (Table VII). Overall, different Km values of the isoforms for various substrates indicate that each isoform probably has a specific natural substrate or a class of substrates, and thus, plays a specific metabolic role. All the isoforms had the lowest Km values for thiocyanate (Km = 3.5 µM, 80 nM, and 32 nM for isoforms II, III, and V, respectively). However, since thiocyanate is not always present in the cell, H/BMT isoforms could also methylate other substrates. For example, The Km values of isoform V for the two aromatic thiols 4,4'-thiobisbenzenethiol and thiophenol were quite low, suggesting an involvement in the methylation of hydrophobic sulfur compounds. In addition, the low Km values of the isoforms III (1.1 mM) and V (2.1 mM) for HS- could mean that they contribute to CH3SH emission from cabbage under conditions that favor the accumulation of HS- ions in the cell. Isoform II, on the other hand, could be active in methylating the thiosalicylate-type of metabolites, for which its Km (1.5 mM) was the lowest among the isoforms. This isoform could not be involved in X- or HS- volatilization, because its Kms for these substrates are unusually high (Km > 90 mM for Cl-, 34 mM for I-, and 44 mM for HS-). H/BMT is not an O- or N-methyltransferase since none of the O- or N-substituted analogs of the above compounds inhibited CH3I production (Table V) or formed radiolabeled products (Table VI). These results, put together, strongly suggest that H/BMT is a thiol methyltransferase.

      Despite its apparent broad specificity for methyl acceptors, H/BMT was highly specific for the methyl donor AdoMet (Table VIII). In Catharanthus roseus cell cultures, SMM is thought to serve as a depository for methyl groups derived from methionine or other unidentified sources (Schwenn et al., 1983). The presence of SMM has been detected in higher plants, and its accumulation has been shown in cabbage and a few other species (James et al., 1995; White, 1982a). Recently, Hanson and co-workers (1994) showed that SMM is an intermediate in the biosynthesis of DMSP in Wollastonia biflora. Thus, it was possible that SMM could serve as a source of the methyl group in the reaction catalyzed by H/BMT. Although a high concentration (5 mM) of SMM supported a low level of methylation by H/BMT, a lower, more physiological concentration (1 mM; (Mudd and Datko, 1990) did not (Table VIII). Therefore, SMM is unlikely to be a natural methyl donor in this reaction. Some marine algae can synthesize CH3I by transferring one of the sulfonium methyl groups of a dimethyl sulfonium compound, such as DMSP, to I- ions (Coward and Sweet, 1971). However, no CH3I was formed from I- in H/BMT assay containing DMSP as the methyl donor, indicating that the above mechanism was not operative in cabbage. Neither L- nor D-methionine acted as direct methyl donors for H/BMT, consistent with White's observation (1982b) that the amino acid must be activated before it can be used in this kind of transmethylation reactions.


5.3. H/bmt Is A Cytosolic Enzyme Involved In Glucosinolate Metabolism

      The existence of H/BMT in multiple isoforms, and the widely different pH optima of these isoforms could mean that they are differentially compartmented within the cell or have different specificities for other, untested substrates. Since HS-, the end product of the sulfate reduction pathway, is a substrate for all H/BMT isoforms, one could expect some of these isoforms to be located in the chloroplast, the primary site for sulfate reduction in plants (Joyard et al., 1988). Physiologically active chloroplasts from cabbage had no H/BMT activity, whether tested under the standard H/BMT assay conditions or under those that favor sulfate reduction (Table X). Breaking the chloroplasts to mimic the conditions of the standard assay for H/BMT, also did not cause I- methylation. Furthermore, many known modulators of sulfate reduction such as phosphate, selenate, molybdate, arsenate, sulfite, pyrophosphate and bicarbonate (Vange et al., 1974), failed to affect H/BMT activity in the crude homogenates (Table V). Taken together, these observations show that none of the H/BMT isoforms is localized in the chloroplast, and that the function of this enzyme is probably not linked directly to the sulfate reduction pathway.

      A bulk of H/BMT activity was localized in the cytosol along with TPI, the marker for this compartment (Table XI). The small amount of H/BMT activity in the various pellets is most likely to be due to cytosolic contamination, as is also the case for TPI. The cytosolic location of H/BMT is identical to that of the thiol methyltransferase from Euglena gracilis (Drotar and Fall, 1985) and the MCT form E. muricata (Wuosmaa and Hager, 1990). The latter two enzymes appear to be involved in general thiol and halide detoxification mechanisms, respectively. Their cytosolic location gives them rapid access to their substrates before any further compartmentation could occur. This also seems to be the case for H/BMT. Substrates of H/BMT, such as HS-, thiocyanate, and aromatic thiols, are highly toxic to plant metabolism (Brown and Morra, 1997; Harborne, 1973; Leustek, 1996) and must be rapidly eliminated. The fungal MCT, which is not involved in detoxification reactions (Harper, 1993), is membrane-bound (Saxena et al., 1998). Although the overall H/BMT activity was found in the cytosol (Table XI), the method used could not distinguish between the vacuole and the cytosol. Further immunohistochemical work is needed to determine whether any of H/BMT isoforms, especially isoform II whose pH optimum was acidic, is in the vacuole. First attempts in this direction were frustrated by the extremely low specificity of polyclonal antibodies raised against the purified H/BMT. Experiments of this nature will probably have to await cloning of the gene(s) encoding H/BMT, and its expression in bacterial systems in enough quantities for the production of isoform-specific antibodies.

      All members of the Brassicaceae family accumulate sulfur-containing thioglucosides, glucosinolates (Figure 2). These secondary metabolites also occur in the families Capparidaceae and Resedaceae, related to Brassicaceae, and in some members of the Caricaceae family which is unrelated to the Brassicaceae (Harborne, 1973). Significant amounts of H/BMT activity were found in all plants tested from these families (Table XII). In contrast, members of Papaveraceae, originally associated to Brassicaceae because of many similarities in ovary and fruit structure (Ettlinger and Kjaer, 1968; Rodman et al., 1996), lacked H/BMT activity. Plants from this family lack glucosinolates, but are known for their active secondary metabolism. Allium cepa and many other members of the family Alliaceae, characterized by the presence of organic disulfides but an absence of glucosinolates (Harborne, 1973; Saghir et al., 1966), also had no H/BMT activity (Saini et al., 1995). These observations seem to correlate the presence of H/BMT and glucosinolates in plants, and therefore, favor a role of this enzyme in glucosinolate metabolism.


5.4. Comparison Of H/bmt To Other s-Methyltransferases

      To my knowledge, this is the first time that a thiol methyltransferase enzyme has been purified to homogeneity from a plant. Very few plant enzymes that catalyze S-methylation reactions are known. An S-methyltransferase that is part of a pesticide-metabolizing complex, was found in onion and a number of spruce species (Lamoureux and Rusness, 1980; Lamoureux et al., 1993). However, no further information on its physiological role or substrate specificity were reported. This methyltransferase is catalytically different from H/BMT because onion does not methylate X- ions (Saini et al., 1995). An AdoMet-dependent methionine methyltransferase that produces SMM was purified to homogeneity from the leaves of Wollastonia biflora (James et al., 1995). It is a homotetramer with a native molecular mass of 450 kDa, distinctly different from the monomeric, low molecular mass H/BMT. Moreover, many plants that contain the former enzyme (James et al., 1995) do not exhibit the capability to produce CH3X (Saini et al., 1995). Enzymes that catalyze methyl transfer to thiol groups are better characterized in mammals (Drummer et al., 1983; Weisiger et al., 1980) and microorganisms (Drotar et al., 1987a; Drotar and Fall, 1985). A thiol methyltransferase from rat liver has been purified to homogeneity (Weisiger and Jakoby, 1979), and other thiol methyltransferases have been partially purified from Pseudomonas fluorescens and from the green alga, Euglena gracilis (Drotar et al., 1987a; Drotar and Fall, 1985). In animals, sulfur methylation is important in the biotransformation of aliphatic and aromatic sulfhydryl drugs (Carrithers and Hoffman, 1994). Methylation is also involved in the detoxification of xenobiotics and of endogenous sulfhydryl compounds, mainly H2S (Weisiger et al., 1980). Similarly, microbial thiol methyltransferases are involved in the methylation of xenobiotic thiols and endogenous reduced sulfur compounds, such as H2S and CH3SH (Drotar et al., 1987a; Drotar and Fall, 1985). A common feature of all these enzymes is that they detoxify chemically reactive endogenous or foreign thiols by forming less reactive methylthio derivatives, but none of them methylates the biological thiols cysteine, methionine, and glutathione (Borchardt and Cheng, 1978; Drotar et al., 1987a; Drotar and Fall, 1985; Drotar and Fall, 1986). H/BMT has very similar properties. This analogy between H/BMT and enzymes with recognized detoxification functions reinforces the idea that H/BMT is involved in thiol detoxification in sulfur-accumulating plants. The high affinity of this enzyme toward toxic compounds like thiocyanate and aromatic thiols, and its broad substrate range that excludes the biological thiols, are all supporting evidence for such a role. Also consistent with such a role is the broad pH range covered by H/BMT isoforms, which would facilitate efficient catalysis under different pH conditions that could arise upon cell rupture. Interestingly, like many other thiol methyltransferases involved in detoxification (Drotar et al., 1987b; Drotar and Fall, 1986; Weisiger et al., 1980), H/BMT uses a single methyl donor compound to methylate a wide spectrum of methyl acceptors.


5.5. Possibility Of Halide Methylation By H/bmt

      Although H/BMT methylates X- ions to CH3X, it has a very high Km value of 85 mM for Cl- and, consequently, could not be naturally involved in the methylation of this ion. Since Cl- is the only physiologically significant X- substrate, the X- methylation activity of H/BMT is likely to be an artifact caused by similarities in chemical and physical properties between X- and the natural substrate thiocyanate, a known pseudo-halide (Hughes, 1975). Similar situations exist in animals. In fact, the strong goitrogenic activity of thiocyanate is due to its ability to act as a competitive inhibitor of I- absorption by the thyroid (Duncan, 1991; Verhoeven et al., 1997; Wood, 1975). Therefore, if thiocyanate is absent, or I- is present at high enough concentrations, the latter could compete against thiocyanate and act as substrate for methylation by H/BMT. The fact that I- is the chemically closest halide to thiocyanate (Hughes, 1975) now explains why this ion was the most preferred X- substrate for H/BMT (Attieh, 1993).

      Although the X- methylation capability of H/BMT may not be a natural feature, it appears to be unique for this enzyme among such S-methylating plant methyltransferases. Onion, known to have a herbicide-metabolizing system which involves an S-methyltransferase (Lamoureux and Rusness, 1980; Lamoureux et al., 1993), had no H/BMT-like activity (Saini et al., 1995). Similarly, lettuce and maize, both of which contain an SMM-synthesizing S-methyltransferase (James et al., 1995), could not produce CH3I in an in vivo leaf-disc assay (Saini et al., 1995). The halide methyltransferase activity of H/BMT could also have an agronomic significance: cabbage is known for its high tolerance to I--containing herbicides (Wain et al., 1966), a property that is probably due to the presence of high levels of H/BMT activity in this species.


5.6. Physiological Significance Of H/bmt

      The data so far presented indicates that H/BMT is a thiol methyltransferase which exists as multiple isoforms. The activity is soluble and likely located in the cytosol. It has a rather broad substrate specificity for the methyl acceptors, and is probably involved in glucosinolate metabolism. Thiocyanate ion, aromatic thiols and HS- were the most preferred substrates by H/BMT (Table VII).


5.6.1. Methylation Of S-Containing Compounds And The Relationship With Glucosinolates

      Indole glucosinolates, which are biosynthetically derived from tryptophan, are the predominant type of glucosinolates in Brassica vegetables (Kutacek and Kefeli, 1968; Verhoeven et al., 1997). At pH>7, myrosinase degrades these compounds to unstable isothiocyanates, which are further transformed to indoles and thiocyanate ion (Figure 23). The latter was shown here to be the best substrate for H/BMT. At lower pH values, indole glucosinolates form acetonitriles and elemental sulfur (Verhoeven et al., 1997). The excess sulfur produced can then be further reduced to HS- (Legris-Delaporte et al., 1987), another substrate for H/BMT (Attieh, 1993), and volatilized as CH3SH as is the case in many members of the Brassicaceae (Buttery et al., 1976; Forney et al., 1991; Mattiacci et al., 1994; Saini et al., 1995). Interestingly, the highest H/BMT activities recorded in an earlier survey were in cabbage (Saini et al., 1995).

      Total glucosinolate concentrations in cabbage can reach values as high as 3 mM (~1.1 g/kg) per tissue fresh weight (McDanell et al., 1988; Verhoeven et al., 1997). In situ, these values can be much higher since these compounds are sequestered in the vacuole. Thiocyanate concentration rises to comparable levels upon glucosinolate hydrolysis. Thiocyanate is a toxic metabolite and has been used in many herbicidal formulations (Brown and Morra, 1997). In fact, it is the glucosinolate degradation product most studied in relation to herbicidal potential of glucosinolates. Thiocyanate kills tobacco and beans at concentrations between 100 and 400 µM (Ju et al., 1983). However, it only causes growth inhibition of cabbage at the high concentration of 1.5 mM (Ju et al., 1983). The data presented here clearly show that thiocyanate is by far the most preferred substrate for H/BMT. All three of the H/BMT isoforms studied have very low Km values for this substrate (II, 32 nM; III, 80 nM; V, 3.5 µM). The high preference of the enzyme for this substrate seems logical if thiocyanate is to be eliminated before it accumulates to toxic levels (Figure 23). Methylthiocyanate, the product of thiocyanate methylation, has been identified as a high-boiling-point volatile from Brassica oleracea (Hansen et al., 1992). This explains the ability of cabbage to resist the herbicidal activity of thiocyanate even at concentrations as high as 1 mM. Thiocyanate (-S--C+N) can also be isomerized to isothiocyanate (S=C=N-), but the greater charge on the S means that thiocyanate ion would be the dominant form under normal conditions (Hughes, 1975; Witczak, 1986). The free electron on sulfur is likely to be the key feature which makes thiocyanate a good substrate for H/BMT. This is particularly supported by the fact that isothiocyanates, in which both electrons on sulfur are bound, do not act as substrates for the enzyme (Table V).

      

Figure 23. Degradation of indole glucosinolates by myrosinase.

      At pH>7, indole glucosinolate degradation forms unstable isothiocyanates (I) which spontaneously degrade to the corresponding alcohol (II) and thiocyanate ions. At more acidic pH, indole glucosinolates form indole-3-acetonitrile (III) and elemental sulfur. Sulfur can then undergo reduction to HS-. Thiocyanate and HS- ions are detoxified via methylation by H/BMT. AdoMet, S-adenosyl-L-methionine; AdoHcy, S-adenosylhomocysteine.

      Unlike thiocyanate and all simple alkyl thiocyanates, methylthiocyanate is highly unstable and undergoes immediate isomerization to the more stable and considerably more volatile product, methyl isothiocyanate (Hughes, 1975). Furthermore, methylthiocyanate is transformed into the iso form upon heating, a condition always present in GC used for their analysis (Hughes, 1975). This could explain why workers describing the role of volatile production from Brassica have often found methyl isothiocyanate, but not methylthiocyanate (Angus et al., 1994; Brown et al., 1994; Buttery et al., 1976; Lewis and Papavizas, 1971; Mattiacci et al., 1994). Our radiometric assay overcomes this difficulty by measuring the actual incorporation of [3H-CH3] from AdoMet into a methylated product which could only be that of the substrate added to the pure enzyme preparation.

      The fact that H/BMT accepted aromatic thiols as substrates is also of interest, although a number among those tested in the present study are not known to occur naturally in plants. Thiobenzoic acid and thiosalicylic acid are structurally analogous to intermediates in the glucosinolate degradation pathway. Glucosinolate metabolism often leads to the formation of organic thiocyanates (Figure 4) (Wood, 1975). Further degradation of organic thiocyanates rarely leads to the full loss of SCN grouping. Instead, cleavage of thiocyanates at the RS--C bond occurs, and gives products containing the thiol RS- groups (Hughes, 1975). These are highly reactive groups that must be kept at low intracellular concentrations because they are known to inhibit enzymes of aerobic respiration, and oxidative enzymes such as cytochrome oxidase, catalase and peroxidase (Wilson and Reuveny, 1976). One way to detoxify them is to form their methyl derivatives, as H/BMT does.


5.6.2. Functions Of S-Containing Volatiles From Plants

      Several other functions of volatile sulfur production from plants have also been suggested. These include antimicrobial and allelopathic effects. Methanethiol, its oxidation product, CH3SSCH3, and methyl isothiocyanate suppress the pea root-rot disease when used as soil fumigants (Lewis and Papavizas, 1971). They do so by adversely affecting the morphological properties, mycelial growth, and oospore development of the fungus Aphanomyces euteiches that causes this disease. Amendment of soil with leaf tissue of cabbage and other Brassicas has similar effects on the pea, bean and sesame root-rot fungi (Kirkegaard et al., 1996), and on the wheat take-all fungus (Angus et al., 1994). Gas chromatographic analysis of headspace constituents of these amended soils revealed the presence of CH3SH and CH3SSCH3 as the most abundant components (Lewis and Papavizas, 1971). The term  biofumigation  has been given to this effect of volatiles from plant tissues on microorganisms (Angus et al., 1994). Methyl isothiocyanate is the stable product of thiocyanate methylation, which is catalyzed by H/BMT. Volatile isothiocyanates are suspected to be the major allelochemicals responsible for the inhibitory effects of Brassica on growth of other plants. Methyl isothiocyanate inhibits germination and growth of a number of weeds such as pigweed, dandelion, crabgrass and common ragweed (Beekhuis, 1975; Teasdale and Taylorson, 1986). This product is also active against wire worms, nematodes and a number of insects (Beekhuis, 1975). Hence, these observations also relate to the physiological significance of H/BMT. In addition to its detoxifying role, H/BMT could also be involved in defense reactions by providing protective agents against competition by other plants, or invasion by insects.

      Plant volatiles offer the fastest means of responding to invasion by pests (Brown and Morra, 1997; Doughty et al., 1996), by signaling the invasion either to neighboring plants (Bruin et al., 1995; Shulaev et al., 1997), or to enemies of the invaders (Geervliet et al., 1994; Mattiacci et al., 1994; Mattiacci et al., 1995). This is best illustrated by the recent finding that methyl isothiocyanate and CH3SSCH3 emitted by herbivore-infested cabbage plants, function as airborne signals that attract natural enemies of the herbivores (Mattiacci et al., 1994). H/BMT is a constitutive enzyme, and its localization in the cytosol is understandable because glucosinolate hydrolysis products are released in this compartment after myrosinase action. Regardless of whether H/BMT is involved in thiol detoxification or in defense mechanisms, it must have immediate access to the glucosinolate degradation products. This task is best achieved if the enzyme is present in the cytosol.

      Methanethiol, methylthiocyanate and its isomer methyl isothiocyanate are all known volatiles from cabbage and many other Brassica species. However, very little information was available with respect to their synthesis. The present report gives the first documentation for the presence of a single thiol methyltransferase enzyme which could produce these metabolites by methylating reactive thiol compounds produced upon glucosinolate degradation. The understanding of the extent to which H/BMT is involved in detoxification and defense mechanisms is just beginning. Further attempts in this direction will rely on the analysis of H/BMT-deficient antisense plants, and is awaiting the cloning of its gene(s).


6. Conclusion And Future Prospects

      The present report is the first to describe the purification and biochemical characterization of a plant thiol methyltransferase. The work also implicates this enzyme in the detoxification of sulfur-containing compounds released upon glucosinolate catabolism.

      Beginning with the observation that cabbage leaf extracts could methylate halides to monohalomethanes and bisulfide to methanethiol, this work identified a novel plant enzyme, H/BMT, and developed an original procedure for its purification. The procedure involved a combination of conventional and high performance liquid chromatographic techniques, which resulted in the isolation of five functional isoforms of the protein. Affinity chromatography on adenosine-agarose was an important step in enriching the H/BMT in the extracts, and a combination of anion exchange with preparative gel filtration HPLC was the key to separating the isoforms in active state. Substrate specificity studies indicated that thiocyanate and a number of aromatic thiols were the most efficiently methylated substrates by H/BMT. All three of the isoforms that were characterized followed the same Ordered Bi Bi mechanism for substrate binding and product release. However, these isoforms had significantly different kinetic parameters toward various substrates. Km values for halide methylation were too high to permit any significant production of monohalomethanes under physiological conditions, indicating that the halide methylation ability of H/BMT is likely to be a chemical artifact.

      The finding that H/BMT activity was restricted to glucosinolate-accumulating members of a number of plant families, indicated that its natural role could be linked to some aspect of glucosinolate metabolism. Glucosinolate degradation often leads to the formation of toxic compounds such as thiocyanate, aromatic thiols, and HS- ions. These were the most efficiently methylated substrates of H/BMT, thus favoring a role of this enzyme in thiol detoxification. The localization of the enzyme in the cytosol, its inability to methylate S-containing amino acids, and the broad pH optimum range covered by the isoforms were all consistent with a role in the detoxification of reactive compounds. Many of the methylation products of the reactions catalyzed by H/BMT were also volatile compounds. Volatile emissions from plants play important roles in processes such as defense against pests, and allelopathy in plant communities.

      Purification of the protein provided the necessary tool to initiate molecular studies which will improve the understanding of the extent to which H/BMT is involved in detoxification or defense mechanisms. Partial amino acid sequence information was useful in identifying an EST from Arabidopsis, which has high sequence identity with the enzyme. This EST specifically recognized a 1-kb transcript in Northern blot analysis experiments, and should, therefore, prove useful in the screening of a cabbage cDNA library to clone the H/BMT-encoding gene(s). An H/BMT clone, once available, will provide the basic tool for new experimental strategies in order to clearly assign a role for H/BMT in either of the above two processes. These strategies include:

  1. Determination of the number of genes that encode H/BMT.
  2. Expression of the gene(s) in bacteria, which should allow raising antibodies that are more sensitive and specific than those raised against the enzyme purified from cabbage.
  3. Accurate determination of the subcellular compartmentation of H/BMT between the cytosol and the vacuole using immunocytochemical techniques.
  4. Determination of the effects of over-expression or suppression of H/BMT activity in transformed plants on their ability to deal with the toxicity of glucosinolate degradation products, or to resist herbivore attacks.

      The availability of such information on H/BMT, especially regarding its involvement in defense or detoxification processes, will be of great significance for the development of crop plants with improved food quality and insect-resistance. Many members of the Brassicaceae family are important crop plants, and the glucosinolate degradation products contribute to the distinctive flavor and aroma characteristic of this family. However, these products may also have undesirable effects due to their pungency and goitrogenic activity. Consequently, the suitability of such important plants for human consumption or as animal foodstuffs is often compromised by their glucosinolate content. Goitrogenicity is primarily caused by the production of thiocyanate ions upon glucosinolate degradation, which sequester thyroxine causing its depletion in the thyroid gland. Thiocyanate was the most preferred among all the substrates tested for H/BMT. Its methylation product, methylthiocyanate is a volatile that easily escapes the plant. Over-expressing H/BMT in plants that have high tissue concentrations of thiocyanate should prove useful in improving the food quality of these plants. The same strategy could also be used to test insect resistance of such plants. Methylthiocyanate is the active component in a number of synthetic pesticides. The minimization of synthetic pesticide use, and the reduction of environmental contamination are but some of the obvious economic and environmental advantages of the production of insect-resistant crop plants.


[Précédent] [Suivant]