Effects of selected herbicides on soil microbial populations in oil palm plantation of Malaysia : A microcosm experiment

Herbicides are commonly used in Malaysia to control weeds in oil palm plantation. In addition to their impact on weeds, these herbicides are also affecting soil microorganisms which are responsible for numerous biological processes essential for crop production. In the present study, we assessed the impact of four commonly used herbicides (paraquat, glyphosate, glufosinate-ammonium and metsulfuron-methyl) on soil microbial populations in oil palm plantation. Our study showed that the herbicide treatments significantly inhibited the development of microbial populations in the soil, and the degree of inhibition closely related to the rates of their applications and varied with the types of herbicide. Paraquat caused the highest inhibitory effect to bacteria and actinomycetes, whereas fungi were most affected by glyphosate. Metsulfuron-methyl had least inhibitory effects to all the microbial populations. The highest inhibition (59.3%) for fungal population was observed at 6 DAT (days after treatment), whereas for the bacteria and actinomycete s (82.0 and 70.6%, respectively) were at 4 DAT. Increasing trend of inhibition on growth of microbial populations was observed from the initial effect until 6 DAT, followed by a drastic decrease of the inhibition at 10 DAT. No inhibition was observed at 20 DAT. The study suggests that the herbicide application to soil of oil palm plantation cause transient impacts on microbial population growth, when applied at recommended or even as high as double (2x) of the recommended field application rate.


INTRODUCTION
Soil, an important component of the ecosystem, serves as a medium for plant growth through the activity of microbial c ommunities.This soil microbial communities (like bacteria, fungi and actinomycetes) play critical role in litter decomposition and nutrient cycling, which in turn, affect soil fertility and plant growth (Singh et al., 1999;Chauhan et al., 2006;Tripathi et al., 2006;Pandey et al., *Corresponding authors. E-mail: rosli@agri.upm.edu.my;mmorshed_bd@yahoo.com. Tel: 603-89474831. Fax: 603-89474918. 2007).However, soil micro-organisms are greatly influenced by factors including the application of herbicides (Pampulha et al., 2007), which are applied in modern agricultural practices to attain optimum crop yields (Zabaloy et al., 2008).If, microorganisms are sensitive to particular herbicide, its application will interfere wit h vital met abolic activities of microbes (Oliveira and P ampulha, 2006), t hus affect the availability of nutrients in the soil (Nautiyal, 2006).Numerous studies have shown the effect of herbicides on soil microorganism populations that ultimately affect the rates of decomposing labile, celluloses and recalcitrant like lignin, respectively, in a variety of ecosystems (Taylor and Parkinson, 1988;Tripathi and Singh, 1992a,b;Pandey et al., 2007;Osono et al., 2003;Osono and Takeda, 2007;Osono et al., 2008).Although, their accurate numbers are still not very clear mainly because of rapid changes in t he populations (Chauhan et al., 2006;Das et al., 2006), but a healthy population of mic roorganisms can stabilize t he ecological system in soil (Chauhan et al., 2006).Thus, the changes in the population of these micro-organisms will affect the ability of the soil to regenerate nutrients to support plant growth.
Malaysia is the world's largest producer and exporter of palm oil that covers over 5 million hectares of land (MPOB, 2011).Weed management is a major problem in the oil palm plantation during the immature phase to avoid suppression of growth and late yield of the oil palm (Chee et al., 1992), so the herbicides are frequently used to manage weeds.Most commonly used herbicides are paraquat, glufosinat e-ammonium, glyphosate and metsulfuron-methyl (Chuah et al., 2005;Kunt om et al., 2007).The presenc e of herbicide residues in soil could have direct impacts on soil microorganisms is matter of great concern.At normal field recommended rat es, herbicides are considered to have no major or long-term effect on microbial populations (Audus, 1964;Bollen, 1961;Fletcher, 1960).It has been reported that some microorganisms were able t o degrade the herbicide, while some others were adversely affected depending on the application rat es and the type of herbicide us ed (Wilkinson and Lucas, 1969;Sebiomo et al., 201 1).Therefore, effects of herbicides on microbial growth, either stimulating or depressive, depend on the chemicals (type and concentration), microbial species and environmental conditions (Bollen, 1961;Hattori, 1973).
Studies on pesticide residual effects on soil microorganisms are often done in soil microcosm smallscale experiment which can be interpreted accurately at larger scales (Benton et al., 2007).Microc osms containing soil microfauna of field communities offer higher resolution of ecotoxicological effects of chemicals in soil environments (Parmelee et al., 1993).As the precise assessment of t he potential non-target effects of herbicides on soil microorganisms in oil palm plantation are of growing interest, therefore, soil microcosm can provide better understanding of possible respons e of soil microbes t o herbicides.The study was aimed to evaluate the effect of commonly used herbicide on bacterial, fungal and actinomycetes populations in soil microcosms from oil palm plantation.

Soil sam pling and preparation of m icrocosm
Soils w ere collected from a young oil palm (3 years old) area at Universiti Putra Malaysia ( UPM), Serdang, Selangor, Malaysia.The site has a history of herbic ide application at 6-months interval, and the herbicide used is glyphosate ( Roundup®).Eighty soil cores (approximately 40 kg) w ere sampled to a depth of 15 cm using auger, collected randomly from under neath the surrounding palms and betw een the palm row s.The samples w ere mixed thoroughly to form a composite sample and taken back to Microbiology Laboratory, Department of Plant Pr otection, Faculty of Agriculture, UPM, and processed accordingly.The pH value of sampled soil was deter mined as 4.1 ± 0.01.Soil chemical properties w ere deter mined w hich w ere as follows: 1.94% C, 0.32% N; 219 ppm P, 104 ppm K, 119 ppm Ca and 32 ppm Mg, and the soil w as classified as sandy c lay (40% clay, 10% silt and 50% sand).The microcosms w ere prepared according to Oliveira and Pampulha (2006) w ith minor modifications.The soils w ere air-dried slow ly in laboratory env ironment (25°C; 50% RH) for 24 h before sieving through a 2 mm mesh.The sieved soils w ere then analyzed to estimate the moisture content and the Water Holding Capacity (WHC).The laboratory deter mination of the moisture content of soil samples w as done by placing 10 g of soil sample in a w eighing glass beaker w as initially w eighed, follow ed by oven drying at 70°C for 24 h.Glass beaker containing the dried soil w as then w eighed again to get the final w eight of the soil.The moisture content w as calculated as percentage using the for mula: Water Holding Capacity (WHC) of the soil w as deter mined by placing 3 g of soil sample on a piece of Whatman filter paper w hich had been initially w eighed, follow ed by oven drying at 70 °C for 24 h.Oven-dried soil on the w eighed Whatman filter paper w as weighed before dipping into w ater until the soil w as saturated.The soil w as then placed in humid enclosure to drain off the w ater before weighing again, and calculated using the for mula (ASTM, 2010): The bulked soils w ith deter mined moisture content of 13% w ere then mixed together, and 56 ml ster ile distilled w ater w as added to achieve the moisture level of 18.5%, w hich w as 50% of its maximum MHC.The soil w as then placed in 39 ster ile glass bottles, each containing 1 kg of soil.Each bottle w as loosely fit w ith cap to allow gas exchange.The soil-containing glass bottles w ere then incubated in dark, in a 25 °C incubator, for 10 days to allow time for adaptation of microorganis ms before treatment w ith the herbicides.The herbicide treatments w ere applied w ith the follow ing procedures, conducted aseptically under laminar flow unless stated otherw ise.50 ml of each herbicide treatments w ere sprayed to 36 out of 39 glass bottle accordingly, using hand sprayer.The herbicide w as mixed thoroughly by constant shaking for 5 min.The remaining 3 glass bottle soils w ere served as control, and sprayed w ith 50 ml sterile distilled w ater.The soil microcos ms w ere then formed by transferring the treated soils into each sterile square plastic container (15 c m x 15 c m x 7.4 c m) w ith lids loosely fitted.The soil microcos ms w ere then incubated in dar kness at 25 °C.Sterile distilled w ater was added on w eekly basis to restore the initial w eight of each microcos m, maintaining the constant moisture content.

Enum eration of m icrobial population
Enumeration of the microbial populations w as done using spec ific media for each microorganis m.Thr ee different grow th media supplemented w ith inhibitors w ere prepared: Potatoe Dextrose Agar (PDA, Difco) supplemented w ith 30 mg/L streptomycin sulphate (Sigma-Aldrich) for enumer ation of fungi; Nutrient Agar (NA, Oxoid) supplemented w ith 0.1g/L cyclohexamide ( Merck) for enumeration of bacteria; and Actinomycetes Isolation Agar (AIA, Difco) supplemented w ith 0.5 g/L cyclohexamide ( Merck) for enumeration of actinomycetes (Araujo et al., 2003).The inhibitors w ere added into ster ilized media (121°C, 15 min) accordingly, and mixed thoroughly on hotplate and stirrer (Jenw ay) before pouring into each Petri dish, mar ked at the bottom div iding it into three sections.
Soil w as collected from each microcosm at 2, 4, 6 , 10 and 20 DA T (days after treatment) to assess the herbicidal effect on the microbial populations present in the soil.Five sub-samples w ere collected randomly from each microcos m treatment using sterile cork borer ( 10 mm diameter).Sub-samples from each microcos m were mixed together, and 1 g of the soil w as taken to make a serial dilution.Serial dilutions w ere made aseptically under laminar flow by suspending the soil in 9 ml of sterile distilled w ater in a test tube and vortexed us ing vortex mixer (Vision Scientific) for 30 s to thoroughly mix them.This pr ocess w as repeated until the dilutions were made up to 10 -5 to complete the serial dilutions.
The drop plate method, conducted under sterile condition, w as used for enumeration of the colonies.The test tubes of the serial dilutions w ere vortexed before five drops (10 µL drop -1 ) of the suspension w ere pipetted out onto each particular section of the media ( marked by div iding lines) according to dilution value of the suspensions.Dilutions selected for plating on PDA w ere 10 -2 to 10 -4 (for culturing fungi), w hereas, NA and AIA w ere plated w ith the dilutions of the 10 -3 to 10 -5 (for cultur ing bacter ia and actinomycetes, respectively).The plates w ere prepared in triplicates, covered and allow ed to dry.After 1 h, the plates w ere inverted, sealed w ith parafilm to avoid contamination and incubated in dar kness at 25°C.
Enumeration of colonies for bacteria, fungi and actinomycetes were done using the Colony Counter ( Rocker) after 24 h, 7 and 10 days, respectively.The total up of the colonies w as used to calculate the Colony-forming unit ( CFU) /g dry w eight of soil.Dry weight of soil w as deter mined after oven drying at 70 °C for 24 h using the for mula: Dry w eight of soil = (w eight of moist soil) X (1% moistur e soil sample/100), and the CFU w as calculated using the for mula:

Dat a analysis
The exper iment w as conducted by Complete Randomized Design Zain et al. 369 (CRD) w ith three replicates.Data w ere expressed as inhibition percentages relative to the control, and analyzed follow ing 2-w ay Analysis of Variance (ANOVA) betw een her bicides and each exposure dates.Means w ere compared using Duncan's Multiple Range Test ( DMRT) at P<0.05 using Statistical Analysis System (SAS).

RESULTS
The effect of herbicide treatments on soil microbial population was determined based on the inhibition percentages of the growth of fungal, bacterial and actinomycetes colonies in eac h treatment media.The growth inhibition showed an increasing trend with increased herbicide concentrations, and the microbial population showed different degree of sensitivity to t he herbicide compounds at different sampling dates (exposure periods).The inhibition percentages of fungal colony development by the herbicides relative to the control (without herbicide treatment) were shown in Table 1.The inhibition percentage of fungi increased with higher application rates of each herbicid e. Highest inhibitions of 63. 1 to 81.4% were obs erved at 2x the recommended field application rate.At 0.5x the recommended field application rate, the herbicides inhibited fungal development by 42.2 to 54.1%.At recommended field application rate, these herbicides could be considered as only moderately toxic to t he fungal colony development, causing moderate inhibition of 54 to 59. 3%.This indicated that applications of the herbicides even at lower than the recommended field rates could be moderately detrimental to the fungal development in soil.
Moderately high inhibition percent ages of the fungal colony development of more than 44% were observed within 2 DA T for the herbicides, except for paraquat.Paraquat, however, caus ed significantly lower inhibition (25.8% ) at 2 DA T. The highest inhibition for paraquat of 54.3% was observed at 6 DA T, but was statistically insignificant compared with the inhibition rate of glyphosat e and glufosinate-ammonium. Inhibition observed for glyphosate and glufosinate-ammonium were comparable at specific rates of application and times of sampling.S ubsequently, the inhibition percentages of t he fungal colony development at recommended field rate were insignificant among the herbicides from 6 DA T onwards.Inhibition of the fungal colony development was abruptly low for all the treatments at 10 DA T, ranging from 2.3 to 10.6%.The fungal colonies, therefore, showed their ability to recover from the toxic effect by 10 DA T, and at 20 DA T, no furt her inhibition or full colony recovery was observed.
Bacterial population development in soil was also affected significantly until 10 DA T by paraquat, glyphosat e, glufosinat e-ammonium and metsulfuronmethyl.The percentages of inhibition of the bacterial colony development relative to the control are shown in Table 2.The herbicides caused higher inhibition to bacterial population development compared wit h that of the fungi.At all sampling times and treatment rates of t he herbicides, the inhibition percentages of bacterial colonies were higher than thos e observed for the fungal colony development, except for t he glufosinat eammonium treatment at 4 DA T and 6 DA T.
The highest inhibitions of the bacterial population were from 77.9 to 87.9%.These highest inhibitions, however, were observed from t he 2 times recommended fi eld rate for all herbicides.Treatment of herbicides at 0.5, 1 and 2 times their recommended field rate also indicat ed increased inhibition percentages with the increased in t he herbicide rates, when sampled at 2 days after treatment until 10 DA T. However, the lowest treatment at 0.5 times the field recommended rat e had also caused significantly high inhibition of the colony development compared with the control, and comparable wit h those of treatments at recommended field rate.
At the recommended field rat e, the herbicides could be considered as moderately to highly toxic to bacterial population.Highest inhibition of bacterial growth was recorded at 68.7, 74 and 82% at 4 DA T for metsulfuronmethyl, glyphosate and paraquat, respectively, and 73% for glufosinate-ammonium at 2 DA T. However, glufosinat e-ammonium caused the maximum suppression through growth inhibition of the bacterial colony development (73%) at faster rate (2 DA T) than paraquat, glyphosate and metsulfuron-met hyl with 45.5, 55 and 67.3%, res pectively.The inhibition percentages of bacterial population for all treatments reduc ed significantly by 10 DA T with a range of 8 to 22.8%.The observations were comparable wit h that observed for the fungal colony development discussed earlier.No inhibition at 20 DA T indicates that the bacterial population recovers from the earlier effects, similar to the fungal population.
Paraquat, gly phosat e, glufosinate-ammonium and metsulfuron-methyl treatment to soil also affected t he development of actinomycetes population (Table 3).The growth inhibition of actinomycetes colonies caused by the herbicides was similar to those recorded for the fungi and bacteria, which increased with the inc reased of the herbicides application rates.However, treatments at 0.5x and 1x the recommended field rates were significantly Values in the same column follow ed by superscript similar letter(s) are not signi ficantly di fferent by DMRT ( P<0.05).Data are presented as mean values (standard error) of three replicates at each exposure period.RFR, Recommended field rate; the rate which i s recommended in the product label to apply in the field.
lower than that at 2x the recommended field rate.Herbicides, at rates recommended for use in the field, were considered as moderately toxic to actinomycetes population in soil.Highest inhibition at the recommended field rate for all herbicides were 70.6, 47.0, 64.3 and 59.4% for paraquat, glyphosate, glufosinat e-ammonium and metsulfuron-methyl, respectively.These inhibition percentages were observed by 4 DA T for paraquat, glyphosat e and metsulfuron-methyl, whereas it was slower for glufosinate-ammonium, observed at 6 DA T. By 10 DA T, however, the inhibition rate for actinomycetes were still relatively high, in comparison with that of t he earlier sampling period, and als o to that of the fungal and bacterial populations.This could indic ate slower recovery period of actinomycetes after the initial effect of the herbicides.However, by 20 DA T, no further inhibition to the actinomycetes population was observed for all treatments, which indicate full recovery from t he treatment.

Herbicide
treatments of paraquat, glyphosate, glufosinat e-ammonium and metsulfuron-methyl showed significant effects on microbial growt h and development in soil environment.Significant increased of fungal, bacterial and actinomycetes growth inhibition were observed from 0.5x to 2x their recommended field application rates, indicating a positive correlation between growth inhibition and treatment rates.Bacterial and actinomycetes populations were severely affected by Paraquat which inhibited their population growth by 70 to 82% at recommended field rate.However, the fungal population in soil was moderately inhibited (54.3%).Paraquat has also been reported to inhibit several microorganisms in soil by Smith and Mayfield (1977).They report ed that paraquat could inhibit a great number of cellulolytic microflora and that might cause injurious effects to symbiotic, anaerobic and nitrogen fixing microorganisms.Paraquat is also known to be bounded strongly and coherently to soil components, including clay minerals and organic matter, therefore limits the access of microorganisms to paraquat in soil water (B romilow, 2003;Isenring, 2006).Thus, adsorption of paraquat to soil rapidly decreases the bioavailability of the herbicide in the soil environment and demonstrated the capability of ads orption process to deactivate hundreds or even thousands of paraquat application over many soil types (Roberts et al., 2002).Th e sandy clay classification of the experimental soils might have reduced the binding of paraquat to soil components and thus increasing t he availability of paraquat in soil water, and henc e affecting the soil microorganisms significantly.
Glyphosate was observed to be less toxic than paraquat to bacterial and actinomycetes populations.At recommended field rate, it inhibited the bacterial population by 74%.The inhibition of actinomycetes and fungal populations were moderat e with 47 to 59.3%.Findings from this study were supported by other studies (Anderson and Kolmer, 2005;Franz et al., 1997;Mekwatanakarn and Sivasithamparam, 1987;Toubia-Rahme et al., 1995;Turkingt on et al., 2001;Wong et al., 1993;Wyss and Muller-Scharer, 2001).However, few studies contradict this result (Busse et al., 2001;Muller et al., 1981;Stratton and Stewart, 1992;Wardle and Parkinson, 1990;Weaver et al., 2007).As a weed killer, glyphosat e targets a single enzyme called 5enolpyruvylshikimate-3-phosphate synthase (EPSPS) (Franz et al., 1997) which plays important role in t he shikimic acid pathway responsible for biosynthesis of aromatic amino acids, and this enzyme is widely present in plants and microorganisms, including bacteria and fungi (Kishore and Shah, 1998;CaJacob et al., 2004).The presence of EPSPS prot eins in bacteria and fungi, therefore, made the microorganisms vulnerable to glyphosat e. CaJacob et al. ( 2004) also reported that EPSPS proteins have been isolated and characteriz ed from microorganisms, which some can tolerate glyphosat e while others were sensitive to the herbicide.
Glufosinate-ammonium was considered to be more toxic than glyphosate to the actinomycetes population.The inhibition of bacterial population by glufosinat eammonium (73%) was considered as being equally toxic compared with glyphosate (74% ).The growth-inhibition by glufosinate-ammonium c ould be due to negative effects on the dehydrogenase activity of soil microorganisms as explained by Pampulha et al. (2007), and s ubsequent decline of growth-inhibition likely due to the compound's rapid degradation process in soil (Ismail and Ahmed, 1994).A study done by Ahmad and Malloch (1995) reported that bacterial growth was reduced only about 40% by glufosinate-ammonium herbicide in agricultural soils.
Similarly, the herbicide at recommended field rat e reduced the bacterial population temporarily, as they recovered after 7 days (Ismail et al., 1995).Pampulha et al. (2007) report ed significant inhibition in growth of actinomycetes, Streptomyces spp., within six days after application of the herbicide t o soil microcosms.In contrary, Ahmad and Malloch (1995) obtained insignificant res ult for the effects of glufosinat eammonium towards soil actinomycetes.
Metsulfuron-methyl was observed to be the least toxic to fungal and bacterial populations compared to paraquat, glyphosat e and glufosinate-ammonium.However, the toxicity of metsulfuron-methyl to actinomycetes population was higher than glyphosate, but similar to paraquat and glufosinate-ammonium. Metsulfuron-methyl could be considered as being moderately t oxic to bacterial, actinomycetes and fungal populations at recommended field rate.Ismail et al. (1996) showed that bacterial population decreased when the concentrations of metsulfuron-methyl increas ed during the first 3 to 9 days after application, depending on soil types.However, Ismail et al. (1996) also demonstrated increase in fungal population with increasing metsulfuron-methyl concentrations, whic h may be influenced by the soil type.El-Ghamry et al. (2000) reflected the toxicity effect of metsulfuron -met hyl when the soil mic robial biomass significantly dec reased with increasing concent rations of the herbicide, which could either be due t o toxicity effect and t he adsorption of t he herbicide in soil or bec ause the soil microorganisms were not adapted to the herbicide itself.
In this study, the herbicide treatments to soil indicat ed short term growth-inhibitory effects on soil microbial population.The treatment effects on soil microbial population growth over the five exposure periods exhibited rapid dec reasing trends after 6 DA T, and t he effects were zero at 20 DA T which indicat e full recovery of the microbial populations from the initial herbicidal effects.This was to be expected becaus e the amounts of herbicides molecules present in the soil were negligible to have any influence on fungal population that ultimately lead to zero inhibition of fungal growth.

CONCLUSION
Paraquat, gly phosat glufosinate-ammonium and metsulfuron-methyl caused significant inhibitory effects on growth of fungal, bacterial and actinomycetes populations in soil microcosms.However, the exposures of the microorganisms upon herbicide applications cause short term changes on the growth and development of the microbial community in oil palm plantation soil.

Table 1 .
Effect of herbic ide treatments on soil fungal population at five exposure periods in soil microcos m.
Values in the same column follow ed by superscript similar letter(s) are not signi ficantl y different by DMRT (P<0.05).Data are presented as mean values (standard error) of three replicates at each exposure period.RFR, Recommended field rate; the rate which is recommended in the product label to apply in the field.

Table 2 .
Effect of herbicide tr eatments on soil bacterial population at five exposure per iods in soil microcos ms.

Table 3 .
Effect of herbicide treatments on soil actinomycete population at five exposure periods in soil microcos ms.Values in the same column follow ed by superscript similar letter(s) are not signifi cantly di fferent by DMRT ( P<0.05).Data are presented as mean values (standard error) of three replicates at each exposure period.RFR, Recommended field rate; the rate which i s recommended in the product label to apply in the field.