2.1 Inoculum source

The Anammox inoculum used in this study was sourced from a continuously stirred tank reactor (CSTR) that had been operated for one year. The initial biomass concentration was recorded as 14,000 mg-MLSS/L. To prevent oxygen from inhibiting Anammox activity, the reactor was maintained under anaerobic conditions, ensuring that the dissolved oxygen (DO) level remained below 0.2 mg/L. The operating temperature was maintained at 30 ± 2°C. Ammonium chloride (NH₄Cl) and sodium nitrite (NaNO₂) were supplied as nitrogen sources, resulting in a total nitrogen (TN) concentration of 264 mg N/L in the feed water. The reactor consistently achieved a nitrogen removal efficiency (NRE) of over 95% while maintaining the pH at 7.5 ± 0.3. Microbial community analysis revealed that Candidatus Kuenenia, Candidatus Brocadia, and Candidatus Brocadiaceae_unclassified were the dominant genera, collectively accounting for > 23% of the total microbial population.

2.2 Reactivation strategy

To evaluate the impact of Freeze-Thaw Cycles (FTC) on the recovery of Anammox bacteria, three cycles were conducted, each followed by a reactivation phase(Fig. 1). The samples were frozen at -80°C or -200°C and subsequently reactivated at 15°C for five days. These freezing temperatures were chosen based on previous studies which demonstrated superior reactivation rates and structural preservation of Anammox cells at cryogenic conditions^{(Ali et al., 2014; Heylen et al., 2012; Rothrock et al., 2011)}. The thawing process was performed gradually by immersing the frozen samples in ice water maintained at approximately 15°C. Biomass was obtained from the parent culture and standardized to 5,000 mg-MLSS/L to ensure uniformity across all test groups. All experimental steps were performed in an anaerobic chamber to limit exposure to dissolved oxygen. The FTC process spanned nine days, with each cycle lasting for three days. Prior to freezing, the residual substrates were removed by rinsing the Anammox biomass three times with 0.1 M phosphate buffer (pH 7.2). The prepared samples were maintained at -80°C in a deep freezer and at -200°C in liquid nitrogen. After completing one, two, or three cycles, the biomass was transferred to 160 mL serum bottles with an operational volume of 100 mL and incubated at 15°C for reactivation. The reactivation phase was conducted using a Sequencing Batch Reactor (SBR) system. The SBR operation cycle included a 30-minute filling stage, a 22-hour reaction stage, a one-hour settling stage, and a 30-minute drawing stage, yielding a Hydraulic Retention Time (HRT) of 24 h. On the fifth day of reactivation, biological and chemical analyses were conducted to assess the adaptation and recovery of Anammox bacteria.

Fig. 1. Reactivation protocol.

2.3 Synthetic wastewater

The synthetic wastewater used in this study was prepared based on the composition of the mineral medium, as listed in Table 1. Ammonium chloride (NH₄Cl) and sodium nitrite (NaNO₂) were used as nitrogen sources, with final concentrations of 60 mg N/L and 72 mg N/L, respectively. To stabilize the alkalinity, 5 mM sodium bicarbonate (NaHCO₃) was added as a buffering agent. Essential nutrients, including KH₂PO4 (27 mg/L), MgSO⋅7H₂O (300 mg/L), and CaCl₂⋅2H₂O (180 mg/L), were incorporated to support the bacterial growth. The medium was supplemented with two distinct trace element solutions. Trace element solution I provided essential iron through 5000 mg/L of EDTA and 5000 mg/L of FeSO₄⋅7H₂O. Trace element solution II contained EDTA (15,000 mg/L) and other trace elements critical for metabolic activity and bacterial proliferation: ZnSO₄⋅7H₂O (430 mg/L), CoCl₂⋅6H₂O (240 mg/L), MnCl₂⋅4H₂O (990 mg/L), CuSO₄⋅5H₂O (250 mg/L), Na₂MoO₄⋅2H₂O (220 mg/L), NiCl₂⋅6H₂O (190 mg/L), Na₂SeO₄⋅10H₂O (210 mg/L), and H₃BO₃ (14 mg/L). This composition ensured that the medium provided all the necessary nutrients and trace elements to maintain bacterial activity and support their growth.

2.4 Transmission electron microscope (TEM)

The ultrastructure of freeze-thawed (FT) Anammox bacteria was examined by transmission electron microscopy (TEM). To prepare the samples, an initial fixation was performed for at least one hour in a solution containing 2% paraformaldehyde, 2% glutaraldehyde, and 0.1 M sodium cacodylate, all dissolved in 0.1 M phosphate buffer (pH 7.4). Post-fixation was carried out for 2 h with 1% osmium tetroxide (OsO₄) prepared in 0.1 M phosphate buffer. Each dehydration step was performed using an ethanol gradient (50%, 60%, 70%, 80%, 90%, 95%, and 100%) for 10 min. Subsequently, the samples were exposed to propylene oxide for an additional 10 min. Infiltration and polymerization were conducted at 65°C for 12 h using a Poly/Bed 812 kit (Polysciences) in a polymerization oven(TD-700; DOSAKA, Japan). The polymerized sample blocks were sectioned into semi-thin slices (200 nm thick) using an ultramicrotome with a diamond knife. The sections were stained with toluidine blue and visualized using an optical microscope. Ultrathin sections (80 nm thick) were prepared from the semi-thin sections, placed on copper grids, and stained sequentially with 3% uranyl acetate for 30 min and 3% lead citrate for 7 min. Imaging was performed using a TEM(TEM-1011, JEOL, Tokyo, Japan) operating at 80 kV and equipped with a MegaView III CCD camera(Soft Imaging System, Germany).

2.5 Molecular biological analysis

3.1 Morphology analysis

In the pre-treatment phase(Fig. 2), physical changes in the cell structure of the Anammox bacteria were observed. Structural damage in the Anammox bacterial cells was observed due to Freeze–Thaw Cycles (FTCs), even though no substrates such as ammonium (NH₄⁺) and nitrite (NO₂⁻) were supplied. While substrate limitation may have temporarily suppressed metabolic activity, the physical damage to cellular membranes and intracellular structures likely compromised the viability and subsequent metabolic potential of the cells. After reactivation, when ammonium and nitrite were supplied as substrates(Fig. 2), these structural changes appeared to have a more pronounced effect on the bacterial activity. In particular, cells subjected to more freeze–thaw cycles showed more significant morphological damage. For example, in the samples exposed to three freeze–thaw cycles at -80°C and -200°C (3FT_80 and 3FT_200), the structural damage was more evident, suggesting that a higher frequency of freeze-thaw treatment leads to more severe cell disruption. This indicates that the presence of ammonium and nitrite during the reactivation phase may have revealed the extent of structural damage sustained during pre-treatment, potentially influencing the metabolic readiness of the cells. As the degree of structural impairment increased, the cells exhibited reduced physiological resilience.

In conclusion, the observed morphological changes reflect the cellular damage induced by FTCs, and these structural alterations may have contributed to diminished functional stability in Anammox bacteria. As cell damage increased, the structural integrity essential for maintaining metabolic function may have been increasingly compromised, potentially affecting nitrogen transformation processes.

Fig. 2. Morphological changes in anammox bacteria during FTC pre-treatment and on day 5 of reactivation. TEM images of anammox bacterial cells exposed to different experimental conditions, highlighting structural changes in cell membranes and intracellular components. (a) The images include control samples maintained at 30°C, (b) NFT samples reactivated directly at 15°C without FTC pre-treatment and samples treated with freeze-thaw cycles (FTC) at -80°C or -200°C. Treatments during the FTC pre-treatment phase: (c) Samples were treated with a single FTC at -80°C, (d) two FTCs at -80°C, (e) three FTCs at -80°C, (f) a single FTC at -200°C, (g) two FTCs at -200°C, and (h) three FTCs at -200°C. In the reactivation day 5 phase, images include samples initially treated with (i) a single FTC at -80°C, (j) two FTCs at -80°C, (k) and three FTCs at -80°C, (l) a single FTC at -200°C, (m) two FTCs at -200°C, and (n) three FTCs at -200°C.

3.2 Microbial community composition analysis

In the pre-treatment phase (Fig. 3), the microbial community was generally similar to that of the control sample (C, activated at 30°C). Notably, Candidatus_Brocadiaceae_unclassified was the dominant genus in most samples. Specifically, it accounted for 16.49% of 1FT_80, 14.67% of 2FT_80, 15.23% of 3FT_80, and 19.09% of 1FT_200, maintaining relatively high proportions. This indicates that even before the freeze–thaw cycles, Anammox bacteria remained dominant, with over 70% of their abundance compared to the parent sample. Additionally, Aggregatilinea and Hyphomicrobium played significant roles during the pre-treatment phase. The proportion of Aggregatilinea ranged between 10.97% and 9.72%, whereas that of Hyphomicrobium ranged from 4.00% to 4.67%(Fig. 2). This suggests that these two microorganisms held an important position within the microbial community, even before the freeze–thaw cycles.

After day 5 of reactivation(Fig. 3), the microbial community composition shifted considerably. The proportion of Candidatus_Brocadiaceae_unclassified generally decreased in all reactivated samples. For instance, in 1FT_80, the proportion decreased from 16.49% to 9.26%, and in 3FT_200, it decreased sharply from 12.86% to 1.62%. This significant decrease suggests that during reactivation following freeze–thaw cycles, the survival rate of Anammox bacteria was reduced, or they became less competitive in utilizing available resources. Conversely, Pseudomonas exhibited a substantial increase on day five of reactivation(Fig. 3). In 1FT_80, the proportion increased from 1.61% to 7.61%, and in 3FT_200, it increased from 0.0048% to 15.66%, respectively. This sharp increase indicated that Pseudomonas gained a competitive advantage, likely by utilizing nitrite as an electron acceptor during denitrification^{(Arai, 2011)}. This metabolic strategy may have contributed significantly to enhancing nitrogen removal efficiency in the reactivated samples.

Moreover, the freezing temperature itself appeared to significantly influenced the microbial community composition, in addition to the number of freeze–thaw cycles. For example, in the samples subjected to three FTCs, the relative abundance of Candidatus_Brocadiaceae_unclassified decreased from 15.23% to 4.10% at –80°C, whereas a more pronounced decline from 12.86% to 1.62% was observed at –200°C. Similarly, Pseudomonas exhibited a substantial increase under more extreme freezing conditions, rising from 5.96% to 15.66% in 3FT_200, compared with a more moderate increase from 4.29% to 7.61% in 3FT_80. These results suggest that cryogenic freezing temperatures may impose greater physiological stress on Anammox bacteria, thereby facilitating a shift in community structure that favors opportunistic taxa, such as Pseudomonas, during the reactivation phase.

In summary, the microbial community composition underwent substantial changes following the Freeze-Thaw Cycles during the reactivation process(Fig. 3). The abundance of Anammox bacteria experienced a notable decline, whereas that of Pseudomonas and other microorganisms became dominant. This shift in community structure highlights the impact of Freeze-Thaw Cycles, with important implications for microbial competition and nitrogen removal performance in the reactivated system.

The changes in the microbial community during the pre-treatment process and on day 5 of reactivation were examined in detail. Fig. 4(a) shows the relative abundance of Anammox bacteria at the genus level, primarily composed of Candidatus_Brocadiaceae_unclassified, Candidatus_Kuenenia, and Candidatus_Brocadia, with Candidatus_Brocadiaceae_unclassified being the most dominant.

During the pre-treatment phase, Candidatus_Brocadiaceae_unclassified, was the dominant genus across most samples, with particularly high proportions in 1FT_80 (16.49%) and 1FT_200 (19.09%). This suggests that the Anammox bacterial community was stable before the freeze-thaw cycle, demonstrating its initial resilience to low-temperature conditions.

On day 5 of reactivation, Candidatus_Kuenenia showed a significant increase in several samples. In particular, in the 2FT_200 and 3FT_200 samples, Candidatus_Kuenenia reached 8.31% and 5.41%, respectively, representing a significant increase compared to that in the pre-treatment phase. This indicates that Candidatus_Kuenenia adapted to and grew better under low-temperature conditions following freeze–thaw cycles. In contrast, Candidatus_Brocadiaceae remained unclassified but experienced a substantial decrease, with its proportion dropping to 9.26% in 1FT_80 and as low as 1.62% in 3FT_200. This suggests that as the number of freeze–thaw cycles increased, the survival and activity of Anammox bacteria were likely reduced.

Fig. 4(b) illustrates the relative abundance of Anammox bacteria and Pseudomonas. In the pre-treatment phase, Pseudomonas was almost absent, and Anammox bacteria remained dominant in all samples. For instance, in 1FT_80, Pseudomonas was only 0.0051%, whereas the Anammox bacteria were stable.

However, by day 5 of reactivation, the proportion of Pseudomonas had significantly increased. In 1FT_80, the proportion of Pseudomonas increased to 7.61%, and in 3FT_200, it surged to 15.66%, indicating that Pseudomonas outcompeted Anammox bacteria for the substrate under low-temperature conditions. This suggests that Pseudomonas may play a more critical role in nitrite removal during reactivation, compensating for the reduced activity of Anammox bacteria. Even as Anammox activity declined, Pseudomonas became a key player in nitrogen removal.

The influence of freezing temperature on microbial community shifts was further evident at the genus level(Fig. 4). Notably, Candidatus_Kuenenia exhibited greater proliferation at –200°C, particularly in the 2FT_200 and 3FT_200 samples, where its relative abundance reached 8.31% and 5.41% respectively. In contrast, Candidatus_Brocadiaceae_unclassified declined more sharply at –200°C than at –80°C. Additionally, Pseudomonas consistently dominated reactivated samples frozen at –200°C, suggesting that ultra-low freezing conditions may not only suppress specific Anammox populations but also selectively enrich psychrotolerant or opportunistic taxa. These results highlight the critical role of freezing temperature, in addition to FTC frequency, in shaping microbial community dynamics under extreme environmental stress.

In conclusion, although Candidatus_Kuenenia showed some adaptation to low-temperature conditions in certain samples, the overall proportion of Anammox bacteria decreased. Conversely, Pseudomonas became the dominant microorganism during the reactivation process, showing a higher competitive advantage for substrates than that of Anammox bacteria. These findings suggest that after the freeze-thaw cycle, Pseudomonas likely plays a more prominent role in nitrogen removal, compensating for the reduced activity of Anammox bacteria in low-temperature environments.

Fig. 3. Comparison of microbial community composition based on NGS analysis between freeze-thaw cycle (FTC) pre-treatment and day 5 reactivation. The relative microbial community composition (%) was analyzed using next-generation sequencing (NGS) for samples subjected to different FTCs during pre-treatment and reactivation on day 5. The bars represent the microbial genera identified in samples treated with one (1FT), two (2FT), and three freeze-thaw cycles (3FT) at freezing temperatures of -80°C (1FT_80, 2FT_80, 3FT_80) and -200°C (1FT_200, 2FT_200, 3FT_200). NFT represents the sample reactivated directly at 15°C without FTC pre-treatment, and C represents the parent culture maintained at 30°C.

Fig. 4. Relative abundance of anammox bacteria and Pseudomonas at the genus level during FTC pre-treatment and reactivation on day 5. The relative abundance of anammox bacteria (Candidatus_Brocadiaceae_unclassified, Candidatus_Brocadia, and Candidatus_Kuenenia) and Pseudomonas (genus level) was analyzed based on NGS data from samples collected after FTC pre-treatment and reactivation on day 5. (a) Distribution of anammox bacterial genera and (b) comparison of anammox bacteria and Pseudomonas.