Soil sample was collected from the top 20 cm from an arable field located at Mymensingh (Upazila: Bhaluka), Bangladesh. It is a terrace soil belonging to the Noadda-1 soil series. According to the USDA [18] soil taxonomy the soil is Ultic Ustochrept and the land type is high land. The bulk of soil samples were collected by composite soil sampling method and processed. To date most of the studies regarding microplastic were focused on the soil contaminated with industrial wastes. Arable soil was often neglected. Thus, we selected an arable field to observe the impacts of microplastics on arable soil.

The soils were air-dried, visible roots and plant debris were discarded and the soils were ground gently to break up larger soil aggregates. After that the soils were sieved at 2 mm, thoroughly homogenized and finally characterized. Physico-chemical properties of the soil (pH, texture and organic matter content) were determined with the standard method described by Rowell [19] . The soil had a loam texture with a pH of 5.24 and organic matter content of 1.63%. The moisture content of the soil was 13.2%.

Microplastics were generated using a cryogenic grinder and liquid nitrogen [20] [21] . Low density polyethylene (LDPE) sheets were purchased from Hebei Qiudu Technology Co., Ltd. located in Hebei, China. These plastic sheets were cut into pieces (1 × 1 cm) using scissors. A 100 ml grinding vial was filled half with the plastic pieces, followed by placing the vial in the cryogenic grinder. The lid of the cryogenic grinder was closed. After that, the vial was submerged in liquid nitrogen for 15 minutes for precooling. The plastic was then agitated using the grinder. The grinding procedure consisted of eight cycles of 2 min grinding and 2 min cooling, with a grinding rate of 12 cps (cycles per second). Finally, the vial was opened using the vial extractor and microplastics were collected for further experimental use.

An incubation experiment was conducted following a randomized block design to investigate the impacts of LDPE microplastic on soil microbiological properties. The incubation experiment was based on four groups of microplastic treatments viz., control, 0.05%, 0.50% and 5.00% (w/w) which were selected on the basis of values of microplastics found in arable soil [6] . All treatments were replicated four times (n = 4). Each container contained 400 g air-dried soil. LDPE microplastic was added to the soil according to the treatments. After adding LDPE microplastics to the soil, the soils were thoroughly mixed and homogenized with a glass rod to distribute the microplastics as evenly as possible. 100 ml of deionised water was added to each container and thoroughly mixed into the soil to establish a soil water content of 25% w/w. The containers were wrapped within plastic film to prevent evaporation. Finally, all containers were kept at 30˚C for an incubation period of 30 days. Deionised water was added (0.5 - 1.0 g) on a weekly basis to maintain a constant water content. After 30 days the soil was collected to determine bacterial isolates, bacterial population, β-glucosidase, urease and dehydrogenase enzymes activities.

Various physical, chemical and physico-chemical properties of the soil sample were analyzed [19] . Total viable count (TVC) of bacteria was enumerated manually by the number of colonies forming units. TVC was determined following the pour-plate technique (Obenauf and Finazzo). Following streaking and incubating, single bacterial colony was isolated. Inocula from the colonies were sub cultured in slants and bacteria were tested for purity for morphological parameters; and when homogeneity of a single isolate was ensured, a number of morphological and biochemical tests were done for identifying the bacteria. Morphological tests involve the observation of bacterial colony (color, shape, size, elevation and transparency). Biochemical tests include catalase, oxidase, nitrate reduction, TSIA (Triple Sugar Iron Agar), LIA (Lysine Iron Agar) and KIA (Kligler Iron Agar). Finally bacterial isolates were identified using “Bergey’s Manual of Determinative Bacteriology” [22] .

The β-glucosidase activity was estimated by using p-nitrophenyl-β-D-glucoside (PNG) as a substrate and incubating 1 g of soil with 0.25 ml toluene, 4 ml modified universal buffer (pH 6), and 1 ml PNG solution (25 mM) for one hour at 37˚C. After incubation at 37˚C, 1 ml of CaCl₂ solution and 4 ml Tris buffer (pH 12) were added, and absorbance was taken at 400 nm using a spectrophotometer [23] . The urease activity was determined by using urea as a substrate. Five grams of moist soil was incubated with 1 ml methylbenzene, 10 ml of 10% urea 20 ml citrate buffer (pH 6.7) for 24 hours at 37˚C. One milliliter of filtered soil solution, 1 ml of sodium phenolate, and 3 ml of sodium hypochlorite were added and diluted to 50 ml, and absorbance was determined at 578 nm using a spectrophotometer [11] [24] . Dehydrogenase activity was measured using triphenyl tetrazolium chloride (TTC) as a substrate where the TTC solution (0.3 - 0.4 g/100 ml) was mixed with 5 g of moist soil and incubated for 24 h at 30˚C. The triphenyl formazan (TPF) formed was extracted with acetone and measured spectrophotometrically at 546 nm. Finally the dehydrogenase activity was expressed as μg TPF g⁻¹ dry soil h⁻¹ [23] [24] .

All data were analyzed using SigmaPlot (version 14) software. Data were tested for normal distribution using the Shapiro-Wilk and Kolmogorov-Smirnov test and equal variance using Levene’s mean test [25] [26] . Data were normally distributed for all analyses except urease activity. Thus, within the urease activity dataset, data were square root transformed. A one-way analysis of variance (ANOVA; treatment) was used to detect the differences in bacterial population, β-glucosidase, urease and dehydrogenase enzymes activities across four microplastic treatments. Pearson correlations were performed to determine how bacterial population, β-glucosidase, urease and dehydrogenase enzymes activities related to each other.

After the end of incubation period, the bacteria present in the control and microplastic treatments were identified. Total 22 bacterial strains were identified in the soil control treatments (Table 1). However, bacterial strains were found to be reduced by microplastic treatments. Three bacterial strains were (Bacillus alvei, Bacillus thuringiensis, Bacillus krulwichiae) were disappeared in 0.05% treatments. Likewise eight (Bacillus subtilis, Bacillus bataviensis, Bacillus alvei, Bacillus thuringiensis, Bacillus krulwichiae, Azotobacter macrocytogenes, Azospirillum lipoferum, and Azotobacter armeniacus) and 16 (Bacillus subtilis, Bacillus

bataviensis, Bacillus thuringiensis, Bacillus krulwichiae, Bacillus sphaericus, Azotobacter macrocytogenes, Azotobacter indicum, Azotobacter chroococcum, Azotobacter bryophylli, Micrococcus luteus, Xanthomonas campestris, Paenibacillus apiaries, Bradyrhizobium japonicum, Bradyrhizobium elkanii, Azospirillum lipoferum, and Azospirillum amazonense) bacterial strains were disappeared in the 0.50% and 5.00% microplastic treatments respectively.

Bacterial colonies started to appear after 48 hours of incubation in soil and microplastic treated inocula indicating the presence of bacteria in these materials. Control soils had a total viable count (TVC) of 90 × 10⁶ CFU/ gm. Microplastic treatments significantly affected the TVC; TVC was significantly (p ≤ 0.05) reduced by 4%, 13% and 21% in 0.05%, 0.50% and 5.00% LDPE treatments respectively (Figure 1).

All three enzyme activities (β-glucosidase, urease and dehydrogenase) in the present experiment closely followed the pattern of the total viable count in various treatments. The AVOVA test indicates that the treatment had highly significant (p ≤ 0.05) effect on enzyme activities. The control soil had higher activities of β-glucosidase, urease and dehydrogenase compared to the treatments. β-glucosidase activity was significantly (p ≤ 0.05) reduced to 10%, 31% and 54% in 0.05%, 0.50% and 5.00% LDPE treatments respectively (Figure 2(a)). Urease activity was significantly (p ≤ 0.05) decreased to 6%, 14% and 30% in 0.05%, 0.50% and 5.00% LDPE treatments respectively (Figure 2(b)). Likewise dehydrogenase activity was significantly (p ≤ 0.05) reduced to 24%, 41% and 60% in 0.05%, 0.50% and 5.00% LDPE treatments respectively (Figure 2(c)). Strong positive correlations (p = 0.00) were observed between bacterial population and β-glucosidase, and between urease and bacterial population (Table 2). Negative correlations were observed between β-glucosidase and urease, between β-glucosidase and

dehydrogenase. However, no relationship was found between urease and dehydrogenase (Table 2).