Possible role of reactive chlorine in microbial antagonism and organic matter chlorination in terrestrial environments


  • Per Bengtson,

    1. Department of Microbial Ecology, Lund University, Ecology Building, SE-223 62 Lund, Sweden.
    Search for more papers by this author
  • David Bastviken,

    1. Department of Geology and Geochemistry, Stockholm University, 106 91 Stockholm, Sweden.
    2. Department of Water and Environmental Studies, Linköping university, 581 83 Linköping, Sweden.
    Search for more papers by this author
  • Wietse De Boer,

    1. NIOO-KNAW, Centre for Terrestrial Ecology, P.O. Box 40, 6666 ZG Heter, en The Netherlands.
    Search for more papers by this author
  • Gunilla Öberg

    Corresponding author
    1. Institute for Resources, Environment and Sustainability, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z4.
      *E-mail goberg@ires.ubc.ca; Tel. (+1) 604 822 3010; Fax (+1) 604 822 9250.
    Search for more papers by this author

*E-mail goberg@ires.ubc.ca; Tel. (+1) 604 822 3010; Fax (+1) 604 822 9250.


Several studies have demonstrated that extensive formation of organically bound chlorine occurs both in soil and in decaying plant material. Previous studies suggest that enzymatic formation of reactive chlorine outside cells is a major source. However, the ecological role of microbial-induced extracellular chlorination processes remains unclear. In the present paper, we assess whether or not the literature supports the hypothesis that extracellular chlorination is involved in direct antagonism against competitors for the same resources. Our review shows that it is by no means rare that biotic processes create conditions that render biocidal concentrations of reactive chlorine compounds, which suggest that extracellular production of reactive chlorine may have an important role in antagonistic microbial interactions. To test the validity, we searched the UniprotPK database for microorganisms that are known to produce haloperoxidases. It appeared that many of the identified haloperoxidases from terrestrial environments are originating from organisms that are associated with living plants or decomposing plant material. The results of the in silico screening were supported by various field and laboratory studies on natural chlorination. Hence, the ability to produce reactive chlorine seems to be especially common in environments that are known for antibiotic-mediated competition for resources (interference competition). Yet, the ability to produce haloperoxidases is also recorded, for example, for plant endosymbionts and parasites, and there is little or no empirical evidence that suggests that these organisms are antagonistic.


For several decades, the prevailing paradigm in the environmental field was that all chlorinated organic compounds are of anthropogenic origin. However, the Wood Blewitt (Fig. 1) is only one of a wide variety of organisms that are known to produce organic chlorine and more than 4500 natural chlorinated compounds have been identified to date, the first few over a century ago, and the list is growing (Gribble, 2003; 2004; Vaillancourt et al., 2006). Detailed mechanisms and descriptions of enzymes involved in catalysing chlorination of organic compounds were published by Shaw and Hager as early as 1959. Still, and perhaps as a result of poor communication among disciplines, chlorine was considered to be inert as it was assumed to only be present in the environment as chloride until a decade or so ago, when the debate about the natural chlorine cycle was initiated (Müller and Schmitz, 1985; Asplund and Grimvall, 1991). Since then, improvements in methodology have resulted in several studies demonstrating that chlorine is actively cycled in the environment, and that extensive formation of organically bound chlorine occurs both in soil and in decaying plant material (Myneni, 2002; Öberg and Sanden, 2005; Bastviken et al., 2007; Ortiz-Bermudez et al., 2007). In fact, the amount of organic chlorine in soil is frequently similar to or even exceeds that of inorganic chloride, at least in boreal forest soils (Asplund and Grimvall, 1991; Öberg and Gron, 1998; Johansson et al., 2003; Öberg et al., 2005).

Figure 1.

Wood Blewitt (Lepista nuda) grows in fairy rings and is known to produce organically bound chlorine (Hjelm et al., 1996).

There are indications of both biotic and abiotic formation of chlorinated organic matter in terrestrial environments (Keppler et al., 2000; Winterton, 2000), but evidence suggests that the chlorination rate is largely controlled by biotic processes (Bastviken et al., 2007). Two types of biotic chlorination can be distinguished: (i) chlorination resulting in specific chlorinated compounds produced within cells and (ii) chlorination occurring outside cells as a result of production of reactive chlorine (e.g. hypohalous acid, HOCl) (Hewson and Hager, 1979; Neidleman and Geigert, 1986; Gribble, 1994; Öberg, 2002; van Pee and Unversucht, 2003).

A synthesis of literature from a number of different fields revealed that there is substantial support for chlorination taking place outside the cell. In summary, a large number of organisms are known to produce enzymes that are capable of producing reactive chlorine (Neidleman and Geigert, 1986; Hunter et al., 1987), and it has been found that soil contains such biocatalysts (Asplund et al., 1993). Pulp and paper research has shown that chlorination takes place when reactive chlorine and organic matter are mixed. Products similar to those caused by chemical chlorination (i.e. during pulp bleaching and water disenfection) have been detected in the presence of chlorinating enzymes (Ortiz-Bermudez et al., 2003). Taken together, the results of previous studies suggest that enzymatic formation of reactive chlorine outside cells is a major source of the large amounts of chlorinated organic matter present in soil.

In the following, we discuss extracellular chlorination in light of microbial antagonism. The purpose is to assess whether or not the literature supports the hypothesis that extracellular chlorination is involved in interference competition, that is, direct antagonism against competitors for the same resources.

Microbial antagonism

Antagonistic interactions are common among microorganisms and are often expressed by producing and releasing antimicrobial compounds that decrease the activity of other microbes, also referred to as antibiosis (Raaijmakers et al., 2008). Antagonistic microorganisms express traits that enable them to interfere with other organisms' growth, survival and infection (Berg et al., 2005). An astonishing number of different microbial antibiotics have been described, ranging from simple (e.g. cyanide) to very complex structures (Lorenzen and Anke, 1998; Raaijmakers et al., 2002; Odds et al., 2003; Montesinos, 2007). The actual importance of antibiotics during in situ competition has been an issue of considerable debate for decades (Gottlieb, 1976; Fravel, 1988; Raaijmakers et al., 2002). Analysis techniques were previously often not sensitive enough to detect the antibiotics in the environment. However, improvements in detection techniques as well as the introduction of methods based on gene expression have clearly demonstrated the possible importance of antibiotics during competitive microbial interactions (Raaijmakers et al., 2002; Haas and Keel, 2003). It is today a commonly accepted theory that chemical warfare is used as a major strategy when organisms are competing for common resources (Wiener, 1996; Czárán et al., 2002; Haas and Keel, 2003). The benefit of antibiotic production during competition has, however, also resulted in counteraction, namely the evolution of genes that provide resistance against antibiotics (D'Costa et al., 2006). This is probably the reason why the variety in antibiotics is so huge, because selection will be in favour of organisms that produce a new antibiotic compound or new antibiotic strategy for which resistance has not yet been developed. Extracellular chlorination may be such an alternative strategy in microbial antagonism. We established three criteria (adapted from the International Allelopathy Society and Macias et al., 2007) that must be met to make the claim that production of reactive chlorine in terrestrial environments can be explained by microbial antagonism/antibiosis:

  • i. The antagonistic microorganism(s) have the ability to produce reactive chlorine.
  • ii. The bioactive chlorine compounds must have biocidal or biostatic effects.
  • iii. The concentrations of the bio-reactive chlorine compounds should be high enough to produce effects on the growth and/or survival of neighbouring microorganisms.

In the following sections we will discuss how well the current knowledge about chlorination in soil and decaying plant material fits these criteria.

Do antagonistic microorganism(s) have ability to produce reactive chlorine?

The ability of terrestrial microorganisms to produce reactive chlorinated compounds has been known for decades and appears to be common, but the conversion rate varies considerably among species. In 1940, Clutterbuck and colleagues investigated the ability to convert chloride among soil fungi. They investigated 139 species and found that most were able to convert chloride to organic chlorine. In 1996, Verhagen and colleagues investigated basidiomycetes and found that approximately 50% of 191 tested strains produced organic halogens. Our research group, finally, investigated 10 white rot fungi and found that eight of the investigated fungi converted significant amounts of chloride to organic chlorine. The amounts varied from 10 to 55 μg g−1 wood, suggesting that the rate of chlorination varied to the same extent (Öberg et al., 1997a).

The first identified chlorinating enzyme, a haloperoxidase-denoted chloroperoxidase (E.C., was first described in 1959 by Shaw, Hager and coworkers (Shaw et al., 1959). It was isolated from the caldariomycin (2,2-dichlorocyclopentane-1,3-diol)-producing fungus Caldariomyces fumago and was identified as a haeme-dependent enzyme carrying out halogenation by the following equation:


It is now generally accepted that he chlorination occurs via diffusible intermediates, such that HOX produced at the active site of the enzyme (Eq. 2) diffuses into the media and then reacts with substrates in free solution according to Eq. 3:


where X is chlorine, bromine or iodine and A is an organic nucleophilic acceptor (Murphy, 2003; van Pee and Unversucht, 2003; Manoj, 2006). Since then, haeme- and non-haeme-dependent haloperoxidases have been isolated from terrestrial organisms among ascomycetes, basidiomycetes, various bacteria, lichen, bryophytae and even insects (Neidleman and Geigert, 1986; van Pee and Unversucht, 2003; Speicher et al., 2003).

It is thus clear that several microorganisms that live in soil and litter and on plant surfaces are able to produce extracellular reactive chlorine.

Do reactive chlorine compounds have biocidal or biostatic effects?

Reactive chlorine has been used in various forms as a disinfectant for over 100 years, and its antimicrobial effect has been thoroughly studied and described (Koski et al., 1966; Linden and Oliver, 2004). For example, studies on chlorination of potable water, hospital sanitization and household cleaning where chlorinating agents are used as disinfectants yield ample evidence that reactive chlorine has an antimicrobial effect.

The most likely mechanism by which haloperoxidases have detrimental effects is by causing oxidative stress. HOCl has a pKa of 7.53 and exists as a mixture of the undissociated acid and the hypochlorite ion at this pH. At lower pH, such as in many soil environments, HOCl predominates and can react with excess chloride to form molecular chlorine (Cl2). These are both highly reactive oxidizing agents that can attack the microorganisms at a variety of chemical sites (Hewson and Hager, 1979). Essentially, any oxidizable group in microorganisms, for example, sulfhydryl groups, iron-sulfur centres, sulfur-ether groups, haeme groups and unsaturated fatty acids, can be oxidized (Klebanoff, 2005; Pattison and Davies, 2006). Hypohalic acids also react readily with amino acids, proteins, nucleobases etc., resulting in unstable halamine formation (Pattison and Davies, 2006; Yap et al., 2007). The reactions may result in loss of microbial membrane transport, interruption of the membrane electron transport chain, dissipation of adenylate energy reserves and suppression of DNA synthesis, eventually leading, for example, to disruption of membrane structure and cell lysis (Klebanoff, 2005; Pattison and Davies, 2006; Yap et al., 2007). The detrimental effect of HOCl has thus been clearly demonstrated.

Are the concentrations of the bio-reactive compound in/on soil, litter and plants high enough to produce effects on the growth and survival of neighbouring microorganisms?

In order for the enzymatic systems to produce high enough concentrations of reactive chlorine in soil, litter and plants, the concentration of the precursors, i.e. chloride, hydrogen peroxide and the activity of chlorinating enzymes, must be ‘sufficient’. Dose–response effects of microbially produced reactive chlorine on microorganisms have to our knowledge been investigated only for the haloperoxidase system of Curvularia sp., moulds belonging to the ascomycetes (Hansen et al., 2003; 2004; 2005; Klebanoff, 2005; Pattison and Davies, 2006; Yap et al., 2007; Renirie et al., 2008). The studies show that the Curvularia system ‘has a rapid antimicrobial effect against a broad spectrum of bacteria, yeasts and filamentous fungi’ (Hansen et al., 2003). Concentrations of 0.9 mM hydrogen peroxide, 4 mM halide and 0.8 mg chlorinating enzyme l−1 result in an almost instantaneous lethal effect and twofold log reduction of Escherichia coli density by the HOX-producing Curvularia haloperoxidase system (Hansen et al., 2004). Other studies on the effect of the same haloperoxidase system on a number of Gram-positive and Gram-negative bacteria, as well as several yeasts and filamentous fungi, showed that concentrations as low as 0.05 mM hydrogen peroxide, 0.05 mM halide and 0.03 mg enzyme l−1 result in a significant reduction in colony-forming units (Hansen et al., 2003; 2005).

The in situ distribution and concentration of chlorinating enzymes in soil, litter and plants have been little studied. To our knowledge, there are only two studies on chloroperoxidases in soil (Asplund et al., 1993; Laturnus et al., 1995). The enzyme activity in these two studies varied from non-detectable to 14 enzyme units kg−1 soil (dw), which means that some of these soils have the theoretical potential to produce up to ∼1 g HOCl kg−1 soil (dw) day−1, given sufficient substrate concentrations. It thus appears as if in situ chloroperoxidase concentrations are not limiting the production rate of HOCl in soils, and that HOCl production should follow first-order reaction kinetics and be dependent on the concentration of the other precursors, that is, chloride and hydrogen peroxide.

Rain contains up to 0.2 mM chloride (Clin) (Winterton, 2000), and concentrations in soil water vary from 0.03 to 0.2 mM (Öberg and Sanden, 2005). The typical Km of the Curvularia chloroperoxidase system varies from 0.25 mM at pH 4.5 to 116 mM at pH 8 for Cl (Vanschijndel et al., 1993), reinforcing the observation above that HOCl production should follow first-order reaction kinetics. Chloride concentrations of up to 0.2 mM are also well above that required by the Curvularia peroxidase system to have a detrimental effect in laboratory environments (Hansen et al., 2003; 2005). Numerous soil enzymatic systems produce hydrogen peroxide, and even though the concentration at any given time is rather low, the formation rate is comparably high. The concentration of hydrogen peroxide in rainwater varies between a < 1 to > 60 μm, with the higher concentrations generally found in spring and summer (Kok, 1980; Yoshizumi et al., 1984; Römer et al., 1985; Olszyna et al., 1988; Padilla et al., 2007). Given that concentrations of 0.05 mM hydrogen peroxide are sufficient for the same haloperoxidase system to be lethal to at least some microorganisms, as long as enzyme and chlorine concentrations are high enough (Hansen et al., 2003; 2005), in situ concentrations of hydrogen peroxide (up to > 0.06 mM) should be sufficient to support chlorination having detrimental effects on target microorganisms.

As H2O2 and chloride concentrations in soil and rainwater are close to or even exceed the concentrations proved to be lethal in laboratory environments, extrapolating from the laboratory results to field situations appears to be valid. Barnett and colleagues (1998), for example, found a Km (H2O2) of 60 μm for chloroperoxidase isolated from Embellisia didymosphora. Our assessment that HOCl production should follow first-order reaction kinetics and be substrate- rather than enzyme-limited is supported by the data on the H2O2 concentration, chloride concentration and chloroperoxidase kinetics presented above.

Less is known about the distribution and occurrence of chlorinating enzymes and the precursors in litter and on plants, in comparison with the concentration and occurrence of chloride and H2O2. As mentioned earlier, a series of studies were conducted during the late 1950s and early 1960s using the chloroperoxidase system of the fungus C. fumago (Shaw et al., 1959; Shaw and Hager, 1959a; b;c;d;e;f). Caldariomyces fumago is a greenhouse mould that lives on plants and is known to excrete considerable amounts of a chlorinating enzyme. The C. fumago enzyme system has been thoroughly investigated in laboratory studies, but we are not aware of any studies dealing with the natural distribution of this system in and on plants or litter. However, X-ray spectroscopic investigation of the distribution of organic chloride in decaying leaves has revealed that such compounds occur in sparsely distributed but extremely intense hotspots (Leri et al., 2007). The hotspots are scattered among diffuse areas of low concentration of organic chloride, and the results suggest that fungi play a role in the production of organic chlorine. Inoculation of leaves with the chloroperoxidase-producing Fusarium oxysporum gave further support for the assertion that biological chlorination reactions occur on leaves, probably catalysed by chloroperoxidase (Leri et al., 2007). Taken together, these results suggest that chloride and hydrogen peroxide concentrations, as well as enzyme concentrations, are sufficient to support formation of HOCl on decaying leaves, although it should be pointed out that data on enzyme distribution and H2O2 on leaves are not sufficient to be conclusive.

In summary, the ability to produce chlorinating enzymes that are active outside the cells is common among terrestrial microorganisms, it is known that at least some of these enzyme systems may convert chloride to detrimental amounts of reactive chlorine, and the concentrations of chloride and hydrogen peroxide in soil are sufficient to support such formation. Taken together, the literature shows that it is by no means rare that biotic processes create conditions that render biocidal concentrations of reactive chlorine compounds.

The observations above suggest that two of the three criteria for microbial antagonism outlined above are clearly met, and that although not conclusive, the evidence points to fulfilment of the third criterion. Our conclusion is that extracellular production of reactive chlorine may have an important role in antagonistic microbial interactions. To test the validity of the assertion that production of reactive chlorine is used to compete with other microorganisms, we first explore empirical and theoretical evidence regarding the environments in which we find organisms that are able to produce haloperoxidases. Thereafter, we evaluate whether or not the literature suggests that microbial antagonism is likely to occur in environments where the ability to produce reactive chlorine seems to be common. Finally, we synthesize empirical evidence that suggests that production of reactive chlorine can be explained by means of microbial antagonism.

In what terrestrial environments are microorganisms found that have the ability to produce reactive chlorine?

An overview of the state-of-the-art knowledge on microorganisms that are known to produce haloperoxidases was generated by making a search for chloroperoxidase/haloperoxidase in the UniprotPK database 2008-05-20. The search revealed that many of the identified haloperoxidases from terrestrial environments are produced by organisms that are plant pathogens, symbionts, litter-degraders or rhizosphere inhabitants. For example, the bacterial no-haeme no-vanadium haloperoxidases seem to occur in microorganisms commonly found in the rhizosphere/rhizoplane, such as Streptomyces and Pseudomonas spp., and in endosymbiotic or parasitic Bradyrhizobium, Rhizobium, Agrobacterium, as well as in Burkholderia spp. Haeme-dependent haloperoxidases occur in many organisms associated with dead plant material (e.g. wood and litter decomposers), such as actinomycetes, various ascomycetes and basidiomycetes. In fact, there are indications that the ability to produce organohalogens is almost ubiquitous among litter-decomposing basidiomycetes (Verhagen et al., 1996). Vanadium-dependent haloperoxidases are, again, found mainly in bacteria and fungi associated with living plants or decomposing plant material. In summary, the results of the search in the UniprotPK database suggest that organisms with ability to chlorinate are in many cases associated with living plants or decomposing plant material.

Several studies support the results of our search in UniprotPK as they show that chlorination does take place in at least some of the above-mentioned environments (deJong and Field, 1997; Reina et al., 2004; Laturnus et al., 2005; Bastviken et al., 2007; Ortiz-Bermudez et al., 2007); that is, the environments where the ability to produce haloperoxidase is common among bacteria and/or fungi. Regarding environments where decomposition of plant material occurs, chlorination in the organic layer of boreal forest soils has been confirmed by, for example, 36Cl tracer studies (Bastviken et al., 2007). Less information is available regarding potential chlorination on living plant surfaces; however, a study conducted in Klosterhede, Denmark showed that the organic chlorine to carbon ratio in throughfall was considerably higher than in rain (Öberg and Gron, 1998). This infers that the organic matter washed from the leaves was highly chlorinated, which in turn suggests that chlorination was taking place on the leaves. The results of the study in Klosterhede are supported by more recent studies, which showed co-occurrence of high organochlorine concentrations and moulds on leaf surfaces (Reina et al., 2004). Likewise, organochlorine production in decomposing wood/litter is high and likely a result of the activity of microbial degraders (Myneni, 2002; Reina et al., 2004; Leri et al., 2007; Ortiz-Bermudez et al., 2007). To our knowledge, there are no studies of chlorination rates in the rhizosphere, but it has been hypothesized that biotic chlorination reactions taking place in the rhizosphere are a major source of chlorinated volatile compounds released from forest soils (Laturnus et al., 2005).

Is microbial antagonism likely to occur in environments where the ability to produce reactive chlorine seems to be common?

The microbial diversity in most terrestrial ecosystems is very high, with most of the microbial species being organotrophic (Kent and Triplett, 2002). Growth of organotrophic microorganisms in soils is limited mostly by the availability of carbon. Also, for a carbon-rich environment such as litter, the availability of easy degradable carbon limits microbial growth (Ekblad and Nordgren, 2002). Plants exude organic compounds via their roots and leaves, and in such microenvironments nutrients other than carbon, mostly nitrogen, may be limiting microbial growth (Cheng et al., 1996). Hence terrestrial microorganisms are commonly confronted with a situation of carbon limitation and occasionally also with nitrogen limitation. Consequently, they have to compete with other microorganisms for organic or inorganic nutrients. The release of antibiotics is a commonly used strategy to increase the competitive ability of microorganisms (Czárán et al., 2002; Kent and Triplett, 2002). Antibiotic production is taxonomically widely distributed among terrestrial fungi and bacteria. A huge variety of antibiotics have been described that can be involved in both intra-kingdom (bacteria versus bacteria) and inter-kingdom (bacteria versus fungi and vice versa) competition (Kent and Triplett, 2002; de Boer et al., 2008). In conclusion, the occurrence of antibiotic-mediated competition for resources (interference competition) is common in most terrestrial environments where the ability to produce reactive chlorine seems to be especially common.

Can production of reactive chlorine be explained by means of microbial antagonism?

Literature on the role of reactive chlorine as a candidate for microbial antagonism in terrestrial environments is, to our knowledge, virtually nonexistent; consequently, experimental evidence to support such a hypothesis is scarce. Several bacterial strains belonging to the genera Bacillus, Burkholderia and Pseudomonas that have the potential to suppress competing organisms have the ability to produce reactive chlorine through chloroperoxidase. The mode of action of these antagonistic microorganisms is generally thought to be a combination of several processes or mechanisms, including the production of one or several secondary antimicrobial metabolites, hydro cyanide, lytic enzymes, and not yet identified effectors secreted by the bacterial type III secretion system (Raaijmakers et al., 2008). It is possible that the production of HOCl represents an overlooked way to explain the antagonistic activities of these organisms. This hypothesis is supported by observations that transgenic tobacco plants expressing a bacterial (Pseudomonas pyrrocinia) non-haeme chloroperoxidase gene had increased disease resistance against the pathogenic fungi Aspergillus flavus and Colletotrichum destructivium (Borchardt et al., 2001). Similarly, reactive halogenated compounds seem to be involved in defence against microorganisms in marine algae (Ohsawa et al., 2001; Manley, 2002; Potin et al., 2002). Rajasekaran and colleagues (2000) have suggested that this is not only due to the direct toxic effects of HOCl/HOBr, but also to their ability to react with microbial signal-molecules involved in biofilm formation/dispersal, which results in decreased fouling and dispersal of already established biofilms. A similar mechanism could be effective in terrestrial environments. HOCl is, as previously discussed, a highly reactive compound that reacts with numerous compounds essential for cell functioning. In fact, one of the major defence mechanisms against invading microorganisms in mammals is the production of HOCl by myeloperoxidase. Myeloperoxidases are, together with a variety of proteins, present as granules in mature neutrophiles, which are essential for optimal host response to microbes and an integral part of the phagocytosis of invading microorganisms (Aratani et al., 2000; Klebanoff, 2005; Nauseef, 2007). Invading microorganisms are ingested by neutrophiles, which then fuse and release their content into an adjacent phagosome. The process consists of a series of events. First the neutrophil exhibits an abrupt increase in oxygen consumption, referred to as the respiratory burst, leading to rapid production of the superoxide anion, which then reacts to produce H2O2 (Nauseef, 2007). Myeloperoxidase then catalyses the formation of HOCl from H2O2 and Cl, followed by subsequent formation of biocidal agents such as chlorine, chloramines, hydroxyl radicals and singlet oxygen (Klebanoff, 2005; Pattison and Davies, 2006; Nauseef, 2007).

There is evidence that reactive chlorine is involved in direct antagonistic interactions between various types of organisms as well. For example, the Japanese lily (Lilium maximowiczii) reportedly produce chlorinated orcinols in response to attack by a pathogenic fungus (Monde et al., 1998). Furthermore, the presence of hydrogen peroxide-producing bacteria in the vagina and urinary tract protects against vaginosis and urinary tract infections, caused by bacteria such as Gardnerella vaginalis and E. coli (Klebanoff and Smith 1970; Eschenbach et al., 1989; Gupta et al., 1998; Klebanoff, 2005). The presence of peroxide and a halide greatly enhances the effectiveness of hydrogen peroxide-mediated antagonistic interactions (Hamon and Klebanoff, 1973), and peroxidase and chlorine are present in vitro in the cervical mucus in sufficient amounts to have a toxic effect (Blain et al., 1975; Klebanoff et al., 1991). Thus, the production of reactive chlorine such as HOCl seems to play a major part in these interactions (Klebanoff, 2005; Merk et al., 2005). The same system also seems to have a viricidal effect and may reduce the likelihood of transmission of the HIV virus and other STDs (Klebanoff and Coombs, 1991; Ab et al., 2000),

Just as mammals react with a H2O2 producing an oxidative burst in defence against invading microorganisms (Nauseef, 2007), plants produce H2O2 in response to microbial attack (Doke et al., 1991; Jacks and Davidonis, 1996). Most organisms react to oxygen stress by producing oxygen-scavenging enzymes, such as catalase, superoxide dismutase and melanin, a compound that can act as a sponge for free radicals (Duffy et al., 2003). Thus reactive oxygen species such as hydrogen peroxide are only moderately toxic. The ability of some plant-associated microorganisms to use the H2O2 to produce reactive chlorine greatly enhances the efficiency of this defence system, as demonstrated by the finding that the lethality of hydrogen peroxide at low concentrations on germinating conidia of the phytopathogen A. flavus was increased 30-fold by the addition of chloroperoxidase (Jacks et al., 1999).

Taken together, experimental evidence from various disciplines suggests that the ability to produce haloperoxidases and to convert moderately toxic hydrogen peroxide to highly toxic HOCl may well represent a form of chemical attack/defence in order to gain a competitive advantage against other microorganisms. Hence the extensive natural formation of chlorinated organic matter may, at least partly, be a side product of reactions with HOCl that is produced by microorganisms to suppress competitors. The ability to produce haloperoxidases is also recorded, for example, for plant endosymbiots and parasites such as Bradyrhizobium, Rhizobium and Agrobacterium. There is to our knowledge little or no empirical evidence that suggests that these organisms are antagonistic. It has been hypothesized that these microorganisms excrete haloperoxidase and produce HOCl in order to form hydroxyl radicals to penetrate the lignocellulose barrier of their host plant (Barnett et al., 1997; Öberg et al., 1997b).

Concluding remarks

The ecological role of the extensive chlorination of organic matter chlorination that occurs in soil and on leaves has long been an enigma. In this paper we have discussed the possibility that the chlorination occurs as a result of antagonistic interactions among microorganisms. The literature survey shows that there are both theoretical support and empirical evidence from various research fields that support the hypothesis. However, most of the evidence is circumstantial as, to our knowledge, no studies have specifically addressed the role of the chloroperoxidase system in microbial antagonistic interactions in terrestrial ecosystems. There are also other theories regarding the ecological role of organic matter chlorination, for example, involvement in the decomposition of recalcitrant organic compounds such as lignin (deJong and Field, 1997; Ortiz-Bermudez et al., 2003; van Pee and Unversucht, 2003; Macias et al., 2007; Ortiz-Bermudez et al., 2007), penetration of the lignocellulose barrier of host plants (Barnett et al., 1997) or detoxification of reactive oxygen species, for example, hydrogen peroxide (Barloy-Hubler et al., 2004). In order to increase our understanding of chlorination reactions in various environments, experiments designed to resolve those questions need to be performed. For example, the role of the chloroperoxidase system in antagonistic interactions among microorganisms could be resolved by comparing the antagonistic effects of mutants lacking the ability to produce chloroperoxidases with those of wild-type strains. The role of chloroperoxidase in detoxifying reactive oxygen species could be determined in more detail by studying the conditions that result in gene expression. Until such studies are performed, the primary role of the chloroperoxidase system and the ecological role of organic matter chlorination will remain an enigma.