Four coal-based ACs (AC-1, AC-2, AC-3, AC-4) and four wood-based ACs (AC-5, AC-6, AC-7, AC-8) were purchased from the market. These ACs are steam-activated ones and the main difference regarding their physiochemical properties is PSDs. The PSD of each AC was measured by high-resolution Micromeritics 3Flex instrument (Micromeritics 3Flex, USA) based on nitrogen adsorption. The pore volume and pore surface values of the ACs are shown in Table 1 and Table 2, respectively. The pore size region of AC was categorized based on previous studies that investigated the effect of PSD of AC on the adsorption capacity of dissolved organic matter in drinking water sources [25] . The majority of pores are concentrated in the size range below 20 Å for all eight ACs. For AC-2, the total pore volume and surface area are obviously lower than other types of AC, and this AC has more pores in the size region below 10 Å both its pore volume and pore surface.

Further, the functional groups on the surface of the AC were examined by Fourier transform infrared spectroscopy (FTIR). The FTIR spectra of all eight ACs are displayed in Figure 1. The existence of a strong peak at 3500 cm⁻¹ indicates the presence of hydroxyl groups. In addition, amide groups and carboxylic acid groups were detected at wavenumbers of 1650 and 1400 cm⁻¹, respectively [26] . All eight ACs had similar functional groups, suggesting these ACs had similar chemical structure. The different properties of all types of ACs were clearly seen in the total surface area, total pore volume and PSD of ACs, demonstrating these properties might contribute more to the uptake of tetG.

The tetracycline resistance gene (tetG) was prepared by referring to the TA cloning method (TaKaRa, Japan). The non-quantitative PCR instrument was used to amplify the target gene. Further, the Gel and PCR Clean-up kit (Macherey-Nagel, Germany) was utilized to purified PCR product. The ligation reaction of the target gene was performed with plasmid pMD20-T (TaKaRa, Japan). The resulting product was transformed into E. coli DH5α competent cells and cultured overnight in LB medium with tetracycline. Further, white colonies were picked for PCR amplification to confirm the insert of the target gene. The confirmed colonies containing the desired plasmid were cultured overnight in LB medium with tetracycline. Then, the plasmid was extracted from E. coli DH5 using a plasmid extraction kit (Macherey-Nagel, Germany), and its concentration was measured using a Quantus Fluorometer. Further, the qPCR analysis was performed to amplify tetG at 133 bp with designed primers (forward primer: TTA TCG CCG CCG CCC TTC T, reverse primer: TCA TCC AGC CGT AAC AGA AC) and utilizing SYBR^® Premix Ex TaqTM (TaKaRa, Japan) based on the manufacturer’s protocol [27] . The amplification product was isolated by gel electrophoresis and purified using a Gel and PCR Clean-up kit. The purified amplification product was diluted with sterile deionized water to obtain the amplicon solution and used for adsorption experiments.

Adsorption experiments were carried out using the batch equilibrium technique in 5 mL vials at 20˚C. Adsorption kinetics experiments were conducted for 180 minutes until equilibrium was attained using an AC dose of 1 g·L⁻¹ and an initial tetG concentration of 6.86 × 10¹³ copies·L⁻¹. Adsorption isotherm experiments were conducted at AC doses within the range of 0.01 - 1 g·L⁻¹, and the initial concentration of tetG was 6.86 × 10¹³ copies·L⁻¹. After shaking for 180 minutes, centrifugation was performed for 10 minutes at 12,000 g to remove the AC particles. The obtained supernatant was subjected to tetG quantification. The concentration of tetG before and after adsorption was determined by qPCR.

The changes of the uptake rate of tetG with time are shown in Figure 2. The uptake was rapid at the beginning and gradually slowed as the contact time increased until equilibrium was reached in 150 minutes. For the coal-based ACs, as displayed in Figure 2(a), compared to AC-2, the uptake rate for AC-1, AC-3, and AC-4 was higher, being about 4.04, 3.80, and 3.90 log at 180 min, respectively. AC-2 showed a lower uptake rate, which only reached 0.75 log after 180 min. It was found that uptake rate by AC-2 was approximately 5 times lower than other types of coal-based AC. Further, the uptake rate of tetG for wood-based ACs is presented in Figure 2(b). AC-7 had the highest uptake rate of tetG (4.50 log) compared to the other wood-based ACs at 180 min. Meanwhile, the uptake rates for AC-5, AC-6, and AC-8 were 3.20, 3.44, and 3.46 logs, respectively. The differences in the uptake rate of tetG on AC were attributed to the different surface properties of the ACs, such as surface chemistry, total pore volume, total surface area, etc.

Pseudo-first-order and pseudo-second-order kinetic models were used in this study to determine the uptake rate and possible mechanisms involved in the uptake of tetG. The pseudo-first-order kinetic model is given as [28] :

d q t d t = k 1 ( q e − q t ) (1)

where q_t (copies·mg⁻¹) and q_e (copies·mg⁻¹) are the uptake capacity of tetG onto AC at time t and equilibrium, respectively. The k₁ (min⁻¹) and t (min) are the pseudo-first-order rate constant and time of reaction, respectively. The k₁ and q_e were obtained from linear plot of ln (q_e – q_t) with time as displayed in Figure 3. Further, the pseudo-second-order kinetic model is exhibited by the following equation [28] :

d q t d t = k 2 ( q e − q t ) 2 (2)

where k₂ (copies·mg⁻¹·min⁻¹) is the pseudo-second-order rate constant. The k₂ and q_e were computed from the intercept and slope of the linear plot of t/q_t against t as demonstrated in Figure 4.

As shown in Figure 3 and Figure 4, pseudo-second-order fitted well with the experimental data compared to the pseudo-first-order kinetic model, suggesting the experimental data follows the pseudo-second-order reaction. The R² value of the pseudo-second-order was very high (R² > 0.975), and the calculated q_e value was in good agreement with the experimental ones as shown in Table 3, indicating the applicability of the pseudo-second-order kinetic model to describe the uptake rate of tetG on AC. The highest uptake rate constant was found in AC-1 (4.64 × 10⁻⁹ copies·mg⁻¹·min⁻¹) for coal-based AC and AC-7 (1.59 × 10⁻⁸ copies·mg⁻¹·min⁻¹) for wood-based one. While AC-2 had the smallest uptake rate constant compared to all types of AC (coal-based and wood-based), which was (5.97 × 10⁻¹² copies·mg⁻¹·min⁻¹). Moreover, as displayed in Table 3, the uptake capacity (q_e) of tetG for coal-based and wood-based ACs was in the range 6.40 × 10⁹ - 1.12 × 10¹⁰ copies·mg⁻¹ and 8.91 × 10⁹ - 1.82 × 10¹⁰ copies·mg⁻¹, respectively.

The pseudo-second-order reaction assumes that the adsorption process is controlled by chemisorption, involving valence forces by sharing electrons between the adsorbent and the adsorbate [29] . As mentioned earlier, hydroxyl groups and carboxylic acid groups are present in eight ACs. These functional groups may allow for H-bonding interactions between the H atom of the AC and the N atom of the tetG DNA base. Moreover, π- π interactions may occur between the aromatic structure of AC and the DNA base of tetG. All DNA bases are aromatic and can easily interact with AC through this interaction. AC also consists of aromatic sheets that are flat and broken in places by slit-shaped pores. Heteroatoms (e.g. oxygen, nitrogen) exist in individual particles inserted between the aromatic sheets or incorporated as functional groups [30] . Similar findings were also obtained in other studies [31] .

The capability of AC to adsorb tetG was evaluated with the Freundlich isotherm model. The equation of this model is described as follows [32] .

q e = K F C e 1 / n (3)

where K_F is the Freundlich constant representing the adsorption strength, and 1/n reflects the affinity of tetG with AC.

The Freundlich isotherm data for the uptake of tetG on coal-based and wood-based ACs are shown in Figure 5. The Freundlich isotherm parameters and R² values are summarized in Table 4. The experimental data for all eight ACs were fitted well with this model, except AC-2. The 1/n value was found below 1, indicating the uptake of tetG on AC is favorable [33] . AC-2 had the highest 1/n value compared to the other types of ACs. This implies that weak affinity between tetG and AC-2 during the adsorption process. Meanwhile, other types of ACs exhibit small 1/n values compared to AC-2. A smaller 1/n value indicates a higher level of affinity or a stronger interaction between tetG and AC [33] . The K_F values for coal-based and wood-based ACs varied in the range of 1.71 × 10³ - 8.00 × 10⁹ (copies·g⁻¹)^1−1/n and 7.00 × 10⁸ - 3.00 × 10¹⁰ (copies·g⁻¹)^1−1/n, respectively. The highest K_F value was found from AC-1 and AC-7 for coal based and wood-based AC, respectively. Compared to all types of ACs, AC-2 possessed the smallest K_F value. The K_F value follows the order AC-7 > AC-1 > AC-5 > AC-6 > AC-4 > AC-8 > AC-3 > AC-2. While the 1/n value follows the order AC-2 > AC-3 > AC-4 > AC-8 > AC-5 > AC-6 > AC-1 > AC-7. Generally, wood-based ACs were found to have a higher K_F value than coal-based ACs, indicating wood-based ACs have a higher ability to accept ARGs compared to coal-based ones.

Table 5 shows the Pearson correlation analysis between the estimated K_F and the pore volume as well as the pore surface in different pore size regions of eight ACs. The highest positive correlation was noticed between K_F and the pore size region of 50 - 100 Å in both pore volume and pore surface, suggesting pores within this size region might be effective for the uptake of tetG. Pores in the range of 50 - 100 Å probably provide ideal accessibility for tetG, allowing the gene to enter and be fixed inside the pore. Pore sizes of 50 - 240 Å also generated a good positive relation with K_F but the correlation coefficient was slightly smaller compared to pores of 50 - 100 Å. It indicated that tetG is preferentially adsorbed on pores slightly larger than its size due to the greater attractive force and more contact points between tetG and the active site of AC. This result is attributed to the potential force overlap formed when the opposite pore wall is slightly larger than the size of the adsorbed molecule [34] . Pore sizes of 50 - 240 Å may be able to adsorb tetG, but the available active sites for uptake are smaller due to the limited surface area at larger pore sizes.

Pore sizes below 20 Å showed a negative correlation with K_F as displayed in Table 5, indicating these pore sizes are not accessible by tetG. Hence, it is obvious that AC-2 has the lowest K_F value compared to other types of ACs due to more pore sizes in the size region below 20 Å, with the percentage of pore volume and pore surface being 98.1% and 99.9%, respectively (Table 1 and Table 2). The size exclusion effect might prevent tetG from entering the pores of AC-2. Thus, the accessibility of tetG to the AC pores is crucial since the appropriate pore size can allow the penetration of tetG into the pore, resulting in more effective gene uptake. This result showed that micropores (<20 Å) were less effective for the uptake of tetG [35] . The micropore may have a larger surface area but it has no entry point and insufficient transport pathways for the gene.