Introduction

For decades, nitrification has been thought as a two-step process involving the oxidation of ammonia to nitrite by ammonia oxidizers (ammonia-oxidizing bacteria (AOB) and ammonia oxidizing archaea (AOA)), and followed by the oxidation of nitrite to nitrate by nitrite oxidizers (nitrite oxidizing bacteria (NOB)) (Prosser 1990). In the context of wastewater treatment, nitrification plays a crucial role in the removal of inorganic nitrogen compounds, and it is typically conducted under neutral to moderately alkaline conditions (with a pH range of 7.5 to 8) (Gieseke et al. 2006). However, the oxidation of ammonia consistently leads to a decrease in pH in the reactor, due to the generation of H+ ions from the oxidized NH4+. Without adequate buffer addition, incomplete nitrification may occur, resulting in nitrite accumulation and subsequently impacting the effectiveness of the downstream treatment process (Li et al. 2021). Consequently, the cultivation of acid-tolerant nitrifiers with a high nitrifying capacity would be advantageous in terms of cost reduction by eliminating the need for buffer solution supplementation.

This traditional understanding of the two-step nitrification process has now been largely revised. Recently, a distinct subgroup affiliated to the sub-lineage II of the genus Nitrospira, previously recognized as strict NOB, has been identified. This particular subgroup possesses the ability to perform complete ammonia oxidation (comammox reaction) in a single cell (Daims et al. 2015; Kits et al. 2017), and termed as comammox Nitrospira. Prior studies have indicated that comammox Nitrospira may be categorized as oligotrophic bacteria due to their high affinity for ammonia (Kits et al. 2017; Koch et al. 2019), which is supported by subsequent studies demonstrating their prevalence in environments with limited nutrient availability, such as drinking water treatment plants (DWTPs) (Fowler et al. 2018). However, comammox Nitrospira can also exhibit high abundance in environments with relatively high ammonia supply, such as wastewater treatment plants (WWTPs) (Maddela et al. 2021). Moreover, comammox Nitrospira seems to be widely distributed in natural aquatic and terrestrial ecosystems (Pjevac et al. 2017; Hu et al. 2021), suggesting that typical niches for comammox Nitrospira may not be restricted to low nutrient environment. Despite the uncertainties on the ecological characteristics of comammox Nitrospira, their presence may bring potential implications in the field of water treatment: the predominance of comammox Nitrospira may contribute to the reduction of N2O emission (a strong greenhouse gas with a global warming potential 265-fold greater than carbon dioxide) (Han et al. 2021; Li et al. 2021), and its high affinity of oxygen (Roots et al. 2019) may contribute to reduce the aeration cost. Since the enrichment of comammox Nitrospira still remains as a challenging task, the key factors determining the niche segregation of comammox Nitrospira and canonical ammonia oxidizers need to be further understood.

Interestingly, comammox Nitrospira was also reported to be dominant in weekly acidic membrane bioreactors (Li et al. 2021), acidic soils (Hu and He 2017), and acidic mine lakes (Li et al. 2023), indicating a potential adaptation of comammox Nitrospira to acidic environments. It should also be noted that, in previous studies (before the discovery of comammox Nitrospira), Nitrospira was reported to be the dominant NOB in other acidic soils (Wertz et al. 2012; Stempfhuber et al. 2017), acidic cultures (Tarre and Green 2004; Cytryn et al. 2012), and acidic biofilms (Gieseke et al. 2006), but was presumed as only nitrite oxidizers in these systems. Taken together with recent reports, the involvement of Nitrospira in nitrification in acidic environments may be more extensive than previously thought. In addition, the adaption of comammox Nitrospira to acidic environment may be a common physio-ecological trait of this bacterium. Therefore, the relative abundance as well as the potential competition strategy of comammox Nitrospira and canonical ammonia oxidizers under acidic conditions need to be further investigated.

Thus, the aim of this research was to evaluate the community composition of ammonia oxidizers under acidic conditions in a biofilm reactor. The reactor was supplied with 40 mg NH4+-N/L, initiated at slightly acidic pH (pH 6), then gradually changed to acidic conditions (as low as 5). qPCR approach combined with amplicon next-generation sequencing was applied to monitor the microbial community change during the continuous operation.

Materials and methods

Bioreactor operation

A 1.1-L fluidized-bed biofilm reactor (FBBR) (Fig. 1) fed with artificial wastewater (containing 0.158 mg/L of NH4Cl, 0.469 mg/L of NaHCO3, and 0.046 mg/L of Na2HPO3·12H2O) was used in this research. The reactor was continuously aerated at 1 L/min and dissolved oxygen (DO) was maintained in the range of 7.5 to 9.4 mg/L. The water temperature was monitored and maintained in the range of 24.0 to 26.4°C by a thermal heater (DP-3A, As One Co., Ltd., Tokyo, Japan). The pH was monitored and controlled by a pH controller (FP-01, Tokyo Garasu Kikai Co., Ltd., Tokyo, Japan), which can simultaneously adjust the pH by adding 5% NaHCO3 solution or 0.5N HCl solution.

Fig. 1
Fig. 1
Full size image

The schematic image of the biofilm reactor used in this research

In the initial setup stage, the reactor was seeded with 10% (vol/vol) of PVA carriers (mature carriers: new carriers = 1:9) (4 mm square, KURARAY CO., Ltd., Ehime, Japan). Mature carriers used as inoculum were obtained from a biofilm reactor in a previous study, which was initially set up at a neutral pH using activated sludge from a WWTP as the inoculum, and subsequently operated at various acidic conditions (ranging from pH 3 to 6.5) for a duration over 1 year. In the present study, the initial start-up stage was carried out for 51 days at pH 6, and the hydraulic retention time (HRT) was 24h for the first 28 days, and 12h for the next 23 days, and kept at 6 ~ 8h since day 52. After the initial setup, the reactor was operated in the following three periods: first, at pH 6 for another 49 days (day 52 to day 149); second, at pH 5.5 for the next 19 days (day 150 to day 168); third, at pH 5 from day 169 to day 305.

During all the periods, influent and effluent water samples were collected every 2 ~ 3 days, filtered through a 0.45-μm pore-size nylon membrane filter (5190–5272, Agilent Co., Ltd., CA, US), and then subjected to water chemistry analysis (NH4+–N, NO2−–N, and NO3−–N) using methods as previously described (Li et al. 2022). Ammonia removal rates and nitrification rates were defined as follows.

$$\begin{array}{c}Ammonia\;removal\;rate\;\left(AMR\right)\;\left(\%\right)=\left[{\left(NH_4^+-N\right)}_{In.}-{\left(NH_4^+-N\right)}_{Eff}\right]/{\left(NH_4^+-N\right)}_{In.}\times100\%\\Nitrification\;rate\;\left(NR\right)\;\left(kg-N/m^3/day\right)={\left(NH_4^+-N\right)}_{In.}\times\left(24/HRT\right)/1000\times AMR/100\end{array}$$

DNA extraction and qPCR

Carrier samples were collected in triplicate on days 9, 48, 62, 76, 150, 155, 162, 176, 195, 199, 207, 213, 220, 253, and 281 (samples on days 100 and 125 were also collected, but inadvertently exposed to inappropriate conditions and were therefore not included in subsequent analysis). Each sample was centrifuged in centricut ultramini tubes (W-MO-045, KURABO Industries Ltd., Osaka, Japan) to remove excess water and then stored at − 30°C in a freezer until further analysis. DNA of each sample was extracted with the DNeasy PowerBiofilm Kit (Qiagen Co., Ltd., Hilden, Germany) according to the manufacturer’s instructions. Concentrations of extracted DNA were then measured by NanoDropTM 2000 spectrophotometers (Thermo Fisher Scientific, MA, USA). DNA samples were stored at − 30°C until further analysis. The qPCR assays were performed on a StepOnePlus™ Real-time PCR System (Applied Biosystems, Foster City, CA, USA) equipped with StepOne™ and StepOnePlus™ v2.3 software. The copy numbers of bacterial amoA, archaeal amoA, and comammox amoA in all samples were determined in triplicate on undiluted samples using the SYBR Green method by absolute quantitation strategy as previously described (Li et al. 2022). Primers used in the qPCR analysis are listed in Table 1. Prior to conducting the qPCR assays, validation was performed on each primer set to determine the optimal primer concentration and thermal condition, to ensure optimum assay efficiency (within the range of 90 to 110%) and specificity (as indicated by a single peak in the melt curve analysis). Additionally, the qPCR amplification product was checked on 1.5% agarose gel by electrophoresis to detect any instances of unspecific amplification.

Table 1 Primer information used in qPCR assays
Full size table

16S rRNA amplicon sequencing and comammox amoA amplicon sequencing

DNA samples on days 76, 176, and 253 were subjected to 16S rRNA gene amplicon sequencing and comammox amoA amplicon sequencing. For 16S rRNA gene, amplicon sequencing targeting v3 and v4 region was performed with primers Bakt_341F (TCGTCGGCAGCGTCAGATGTGTATAAG-AGACAGCCTACGGGNGGCWGCAG) and Bakt_805R (GTCTCGTGGGCTCGGAGATGTG-TATAAGAGACAGGACTACHVGGGTATCTAATCC) following the 16S rRNA gene Sequencing Library Preparation Illumina protocol (Part # 15044223 Rev. B) on a Miseq system (Illumina, San Diego, CA, USA). Raw reads were processed by DADA2 (v 1.18.0) in QIIME2 (ver 2020.2) for the removal of low-quality reads, the assembly of paired end reads, primer trimming and amplicon sequence variants (ASVs) picking, and chimera removal. SILVA v132 database was used to assign taxonomic classification to all ASVs by using the classify-sklearn algorithm in QIIME2. The numbers of high-quality paired end reads were 24,379 (generated from 106,633 raw reads), 16,685 (from 93,875 raw reads), and 24,962 (from 110,599 raw reads) in samples at pH 6, pH 5.5, and pH 5, respectively.

For comammox amoA amplicon sequencing, DNA samples were first amplified using primer set of comamoAF and comamoAR (Zhao et al. 2019), and TaKaRa Ex Taq (TakaraBio Inc., Shiga, Japan) according to the manufacturer’s instructions and previous report (Zhao et al. 2019). Thermal condition for the first PCR was 94°C for 2 min followed by 20 cycles of 94°C (30 s), 53 °C (30 s), and 72 °C (30 s), and a final extension step at 72°C for 5 min. After purification of those PCR products using AMPure XP (Beckman Coulter Life Sciences, Indianapolis, USA), the second PCR was performed using the primer 2ndF and 2ndR, and TaKaRa Ex Taq® Hot Start Version (TaKaRa Bio Inc., Shiga, Japan). The thermal condition of the 2nd PCR was 94 °C for 2 min followed by 8 cycles of 94 °C (30 s), 60°C (30 s), and 72°C (30 s), and a final extension step at 72°C for 5 min. Sequencing was performed using a Miseq system (Illumina, San Diego, CA, USA) by Bioengineering Lab. Co., Ltd. (Kanagawa, Japan). Raw reads containing the identical primer sequence were first extracted by FASTX-Toolkit (ver 0.0.14). Low-quality reads (slide windows, 40bp; quality threshold, 20) were then removed by sickle (ver. 1.33). Paired end reads were subsequentially assembled with an overlap of 10 bp using FLASH (ver.1.2.11). The assembled reads were then checked by Qiime2 (ver.2022.8) to remove chimera and noise, and to infer amplicon sequence variants (ASVs) using the pipeline for paired-end Illumina MiSeq data. The number of high-quality paired end reads were 26,133 (generated from 71,007 raw reads), 34,636 (from 45,438 raw reads), and 19,686 (from 52,367 raw reads) in samples at pH 6, pH 5.5, and pH 5, respectively.

Nucleotide sequence accession numbers

The raw reads of the V3-V4 region in 16S rRNA gene sequence and comammox amoA sequence were deposited under accession number DRA017621 (BioProject PRJDB17236) in DNA Data Bank of Japan (DDBJ).

Results

Water treatment performance

As shown in Figs. 2a–c and 3, in general, the bioreactor demonstrated significant rates of ammonia removal throughout all three periods (pH 6, pH 5.5, pH 5) after the initial setup phase. The average rates of ammonia removal rates during each period were as follows: 54.4 ± 22.0% during the set-up stage, 83.2 ± 15.9% during the pH 6 period, 100 ± 0% during the pH 5.5 period, 87.2 ± 23.1% during the pH 5 period. From day 188 to day 199, the concentration of NH4+–N in the influent water experienced rapid decline due to inadvertent errors in reagent preparation, which unavoidably led to fluctuations in the NH4+–N and NO3−–N levels in the effluent water during the subsequent pH 5 period. With the exception of the setup stage, the ammonia removal rates were generally good, averaging at 86.5 ± 19.6% (and 80.8 ± 23.5% when considering the setup stage). This suggests a notable tolerance to acidic conditions. The nitrification rates further demonstrated this tolerance (Fig. 3), as they remained consistent at 0.12 ± 0.01 kg-N/m3/day and 0.12 ± 0.04 kg-N/m3/day in the pH 5.5 and the pH 5 periods respectively, surpassing the rate observed in the pH 6 period (0.07 ± 0.03 kg-N/m3/day). Additionally, the accumulation of nitrite was not observed under acidic conditions, with average effluent concentrations of 0.49 ± 2.00 mg N/L throughout all three periods, indicating complete nitrification that occurred in the bioreactor.

Fig. 2
Fig. 2
Full size image

Water chemical and microbial profiles during 305 days’ continuous operation period: a concentrations of nitrogen compounds, b ammonia removal rates, c nitrification rates, d functional gene abundances

Fig. 3
Fig. 3
Full size image

Boxplot of ammonia removal rates (left) and nitrification rates (right) in pre-operation (setup) period (D0 to D49), pH6-operation period (D50 to D147), pH5.5-operation period (D148 to D176), pH5-operation period (D177 to D305)

Quantification results of functional genes involved in ammonia oxidation

The copy number of functional genes involved in ammonia oxidation (amoA of AOB, comammox Nitrospira, and AOA) was quantified by qPCR assays, and the results are shown in Fig. 2d. During the initial setup stage and early stage of pH 6, AOB exhibited a competitive advantage over AOA by more than 1 order of magnitude, and outperformed comammox Nitrospira by 2 to 4 orders of magnitude. However, under the pH 5.5 condition, the population of comammox Nitrospira surpassed that of AOB by one order of magnitude, and the predominance of comammox Nitrospira became even more pronounced (by 1–3 orders of magnitude) when the pH was further reduced to 5. For AOA, although it was the second predominant ammonia oxidizers during the initial stages of the experiment, its abundance decreased by 2 ~ 3 orders of magnitude in pH 5.5 and pH 5, indicating a diminished role of AOA to ammonia oxidation under higher acidic conditions within this biofilm reactor. The dominance of comammox Nitrospira in pH 5.5 and pH 5 samples suggests that a persistent acidic environment may confer a competitive advantage to comammox Nitrospira over canonical AOB and AOA within the biofilm reactor.

Microbial communities of biofilm based on 16S rRNA amplicon sequencing

The community structure of biofilms at different pH levels (pH 6, pH 5.5, and pH 5) was analyzed using 16S rRNA amplicon sequencing, and the community structure was shown at class level and family level in Fig. 4. As a result, Gammaproteobacteria and Alphaproteobacteria exhibited dominance across all samples at the class level, with the former showing increased relative abundance in samples with higher acidic level. The relative abundance of Alphaproteobacteria remained consistent in a range of 13.93 to 19.34% in all samples. Betaproteobacteria and Nitrospira also displayed significant proportion, with the average relative abundances of 7.77% and 4.35%, respectively.

Fig. 4
Fig. 4
Full size image

Bar plot of microbial community structure in biofilm carriers at class level (a) and at family level (b) revealed by 16S rRNA gene V3-V4 region amplicon sequencing. The top 29 classes or families were shown

At the family level, Xanthomonadaceae, Rhodospirillaceae, Nitrospiraceae, and Chitinophagaceae were found to be dominated in all samples. Nitrosomonadaceae, the only detected AOB, exhibited relative abundance of 5.21% in samples at pH 6 (day 76), and significantly declined to 0.36% (pH 5.5) and 0.91% (pH 5). In contrast to the low abundance of AOB in pH 5.5 and pH 5 samples as determined by 16S rRNA amplicon sequencing, qPCR analysis (Fig. 2d) indicated a significant presence of AOB amoA across all samples. To investigate potential shifts in the AOB community, the ASVs that assigned to the family Nitrosomonadaceae with relative abundance over 0.1% were extracted and subjected to phylogenetic analysis (Fig. S1). As a result, all these Nitrosomonadaceae-assigned ASVs formed two distinct clusters, one closely associated with Nitrosomonas, and the other with Nitrosospira. Notably, ASVs within the Nitrosomonas-like cluster were only detected in pH 6 samples, while ASVs within Nitrosospira-like cluster exclusively exist throughout all three samples. Moreover, in order to find out if there are any highly divergent sequences that cannot be accurately classified, ASVs assigned to the class Betaproteobacteria but unclassified at the family level (with a relative abundance > 0.1%) were extracted. These unclassified betaproteobacterial ASVs were then manually inspected and blasted to the NCBI rRNA database, the ones with Nitrosospira listed in the BLAST results (with a percent identity ranged from 89.48 to 91.65% to the know Nitrosospira species) were selected and subjected to the phylogenetic analysis too (Fig. S1, marked as italic, grey). These unclassified “Nitrosospira-like” ASVs were found to be extensively existed in all the samples, as the sum of which accounted for 1.83%, 1.86%, and 1.76% in samples at pH 6, pH 5.5, and pH 5, respectively. In fact, in the case of AOB, the utilization of 16S rRNA gene-based detection methods necessitates sequences that typically cover V2 to V7 regions (Voytek and Ward 1995). The utilization of relative short amplicon targeting 16S rRNA V3-V4 region may result in diminished resolution in the detection of AOB (such as the unclassified “Nitrosospira-like” ASVs).

As potential participant in either nitrite oxidation or comammox reaction, the family Nitrospiraceae maintained a relatively high abundance across all the samples (which is 5.16% in pH 6 sample, 5.06% in pH 5.5 sample, and 2.84% in pH 5 sample). The ASVs belonging to the family Nitrospiraceae were also extracted and subjected to further phylogenetic analysis. Figure 5 shows the phylogenetic tree generated from the top 12 Nitrospiraceae ASVs (ASVs with a proportion higher than 0.1%) and the normalized read number of each ASV in each sample. As a result, all ASVs were genetically close to Nitrospira sublineage I or sublineage II, while being distinct from Nitrospira sublineage III to VI. Among the 12 ASVs, 5 ASVs (ASV732, 785, 912, 955, 1053) were found to cluster with the known comammox Nitrospira (Candidatus Nitrospira nitrosa and Candidatus Nitrospira nitrificans), whereas the remaining ASVs showed more similarity to environmental sequences. Moreover, 5 ASVs show heterogeneously declines in abundance as pH decreased; 2 ASVs were exclusively present in all samples; and 4 ASVs (i.e., ASV732, ASV785, ASV918, ASV955) were more abundant in samples with lower pH levels, indicating a divergency in pH preference within Nitrospiraceae. However, it is important to note that distinguishing comammox Nitrospira from the phylum Nitrospira based on 16S rRNA analysis is not feasible. Therefore, these 16S rRNA V3-V4 amplicon sequencing results were insufficient to explain the ammonia removal observed during pH 5.5 and 5 operation periods. Thus, the aforementioned samples (days 76, 176, and 253) were subjected to high-throughput sequencing on the comammox amoA gene and subsequently analyzed in detail.

Fig. 5
Fig. 5
Full size image

Neighbor-joining phylogenetic tree were generated from an alignment of the representative Nitrospirae ASVs (ASVs that assigned to the family Nitrospiraceae with proportion higher than 0.1% were shown) and reference sequences based on the 470 bp fragment for 16S V3-V4 amplicon sequencing. Bootstrap higher than 0.8 were marked with light purple circles. Scale bar represents the 1% nucleotide divergency. The bar plot on the right indicates the normalized read number in each sample (sampling dates were as follows: pH6: 22nd Nov, 2021 (day 76); pH5.5: 2nd Mar, 2022 (day 176); pH5: 18th May, 2022 (day 253))

Phylogenetic analysis of comammox Nitrospira based on amoA amplicon sequencing

Figure 6 shows the phylogenetic tree based on the 396-bp fragment for comammox amoA amplicon sequencing. ASVs with a proportion higher than 0.1% and with long reads (396 bp) were selected and analyzed. Similar to the results shown in Fig. 5, amoA ASVs also reflected the pH-dependent divergency. Notably, these ASVs formed two distinct clusters, with one cluster (the upper cluster marked in yellow in Fig. 6) demonstrating a higher tolerance towards lower pH levels (specially, pH 5), while the other cluster (the lower cluster marked in green in Fig. 6) displayed a higher presence in samples under less acidic conditions (pH 6). Environmental sequences derived from acidic mine lakes (Li et al. 2023) and sequences of comammox Nitrospira enriched from acidic conditions (Takahashi et al. 2020) also fell into the lower pH tolerant cluster (the upper cluster). Intriguingly, this acid-tolerant cluster appeared to be distinct from previously identified comammox Nitrospira isolates, enrichments, and the majority of reported environmental sequences, suggesting the existence of a potential acidophilic cluster within comammox Nitrospira. In the less-acidic favored cluster (the lower cluster), the majority of ASVs showed a significant decline in relative abundance as pH decreased. However, ASV001, which represented 46.9% of all reads on average, demonstrated a remarkable tolerance to lower pH levels, with its relative abundance only declined from 28.5% at pH 6 to 16.0% at pH 5. These findings suggest the existence of potential acidophilic (or acid-tolerant) species within comammox Nitrospira, which may account for the observed the ammonia removal without nitrite accumulation during operation periods at pH 5.5 and 5 in the studied biofilm reactor.

Fig. 6
Fig. 6
Full size image

Neighbor-joining phylogenetic tree were generated from an alignment of the representative ASVs (ASVs with proportion higher than 0.1% were inspected manually to remove short reads) and reference sequences based on the 396 bp fragment for comammox Nitrospira amoA amplicon sequencing. Bootstrap higher than 0.8 were marked with light purple circles. Scale bar represents the 5% nucleotide divergency. The bar plot on the right indicates the normalized read number in each sample (sampling dates were as follows: pH6: 22nd Nov, 2021 (day 76); pH5.5: 2nd Mar, 2022 (day 176); pH5: 18th May, 2022 (day 253))

Discussion

Low pH is known as a negative affecting factor on nitrification, as NH3, the actual substrate of ammonia-monooxygenase, is extensively protonated to NH4+ and becomes unavailable for most AOBs under acidic conditions. Moreover, NO2−, the intermediate product of ammonia oxidation, undergoes protonation to free nitric acid at low pH levels, further inhibiting the growth and activity of both AOB and NOB (Li et al. 2020; Wang et al. 2021). Despite these inhibitory effects, instances of nitrification under acidic conditions have been reported in both natural environments (such as acid soils and acidic mine lakes) (De Boer and Kowalchuk 2001; Li et al. 2023), and engineered systems (such as biofilms, chalk particles, and suspended biomass reactors) (Tarre and Green 2004; Gieseke et al. 2006; Han et al. 2021). However, the predominant ammonia oxidizers varied across different environments. In contrast to the ubiquitous predominance of AOA in acidic soils (as reviewed by (Li et al. 2018)), the prevalence of AOA over other ammonia oxidizers may be significantly reduced or reversed in water treatment bioreactors (Li et al. 2021; Meng et al. 2022; Pan et al. 2024) or acidic mine lake (Li et al. 2023). Consistent with prior research findings, this study observed low levels of AOA throughout the experiment, indicating a minor contribution of AOA to the ammonia removal in the acidic biofilm reactor. In this study, AOB was suggested to play a major role in the initial stages, and were the second important ammonia oxidizers when pH was 5.5 and 5. 16S V3-V4 amplicon sequencing identified the presence of a Nitrosomonas-like cluster in the ammonia oxidation at pH 6, while groups related to Nitrosospira potentially dominate in the bacterial ammonia oxidizers at more acidic levels. Consistent with our results, Nitrosomonas-associated species were prevalent in the AOB community in samples around pH 6, but their abundance largely declined in samples with lower pH in MBBR reactors (Fumasoli et al. 2017; Li et al. 2020). Nitrosospira-associated species were found to be predominant in acidic nitrification reactor (Gieseke et al. 2006) and isolation experiments that used samples from acidic terrestrial environments (De Boer et al. 1995; Jiang and Bakken 1999; De Boer and Kowalchuk 2001; Li et al. 2020), particularly at pH levels around 4 or lower. This suggests the significant role of Nitrosospira-associated species in the AOB guild under acidic conditions. Notably, previous studies have shown that Nitrospira, traditionally classified as strict NOB, existed exclusively in the acidic environments including reactors (Gieseke et al. 2006), cultures (Tarre and Green 2004; Cytryn et al. 2012) and terrestrial ecosystems (Wertz et al. 2012; Stempfhuber et al. 2017). This suggests a potentially overlooked role of Nitrospira in low pH conditions. With the discovery of comammox Nitrospira, its predominance over canonical ammonia oxidizers at acidic environment has recently been reported in acidic natural and artificial systems (Takahashi et al. 2020; Li et al. 2021, 2023). This study provides additional evidence for the significance of comammox Nitrospira in ammonia oxidation under acidic conditions, as well as underscoring the presence of a potential acidophilic subgroup within comammox Nitrospira in samples with higher acidic levels.

Competition for limited free ammonia under acidic conditions is a crucial factor for the survival of ammonia oxidizers in acidic environments, and potentially influencing their community composition. Acidophilic ammonia oxidizers typically exhibit a higher affinity for ammonia compared to their counterparts in neutral or slightly alkaline environments. For example, while Nitrosomonas europaea and Nitrosococcus mobilis (typical AOBs) demonstrate a Km (NH3) range of 18 to 70 μM (Knowles et al. 1965; Suzuki et al. 1974), Nitrosospira, which predominates the AOB community in acidic environments, displays a Km (NH3) range of 6 to 11 μM (Jiang and Bakken 1999). Additionally, the acidophilic AOB “Ca. Nitrosacidococcus” exhibit a Km (NH3) of 147 nM (Picone et al. 2021). Notably, the cultured comammox Nitrospira “Nitrospira inopinata” and “Ca. Nitrospira kreftii” demonstrate an even lower Km (NH3) of around 60 nM, and 40 nM, respectively, surpassing most AOA and all other cultured AOB (Kits et al. 2017; Sakoula et al. 2021; Jung et al. 2022). The high affinity of ammonia exhibited by comammox Nitrospira and the Nitrosospira-like AOB may confer an adaptive advantage for their thriving under acidic conditions.

In addition to the importance of competition strategies for free ammonia, many studies have suggested that aggregates, such as biofilms, may also offer enhanced survival advantages in low pH environments compared to individual cells. This is attributed to the formation of slime-like extracellular polymeric substances (EPS) between microcolonies, which would potentially provide a relatively pH neutral shelter (microsites) for the nitrifiers (Keen and Prosser 1987; De Boer and Kowalchuk 2001; Gieseke et al. 2006). This would be particularly true for comammox Nitrospira, as it demonstrates a preference for dominance within attached growth system, such as biofilms, regardless of the acidity level of the environment (Pjevac et al. 2017; Fowler et al. 2018; Xia et al. 2018; Cotto et al. 2020; Hu et al. 2020; Li et al. 2021, 2023). In such systems, comammox Nitrospira is theoretically competitive against AOA and AOB due to its less energy-consuming carbon fixation pathway and higher ATP yield (Costa et al. 2006; Petruzzi et al. 2017). Furthermore, the presence of genes related to biofilm formation (Flagellar biosymthesis, EPS systhesis and transport, glycogen biosynthesis) in comammox Nitrospira bins identified through metagenomic analysis (Zhao et al. 2022) may further enhance its ability to actively create biofilm for better survival.

Moreover, the versatility of utilizing or producing various substances of comammox Nitrospira that revealed by metagenomic analysis may also contribute to its predominance under acidic conditions through interactions within the biofilm community. Substance exchanges (such as, cobalamin, polyamines) through microbial interactions may play an important role in stress response, cell growth regulation, and biofilm formation (Crespo et al. 2018; Krysenko and Wohlleben 2022). For example, comammox Nitrospira possesses a complete biosynthesis pathway of cobalamin, an important cofactor essential for primary and secondary metabolism, a feature that is uncommon among most microorganisms (Zhao et al. 2022). This feature would thereby establish comammox Nitrospira as hub microorganisms within the micro-community. In addition, comammox Nitrospira may utilize cytoplasmic buffering molecules such as polyamines, polyphosphate, and inorganic phosphate to consume protons and maintain neutral cytoplasmic pH (pH homeostasis), as revealed by the presence of large number of pstSCAB genes (Li et al. 2023). The presence of genes encoding carbonic anhydrase and urease (cah and ureABC), cation transporters including the Na+(or K+)/H+ antiport systems found in comammox Nitrospira bins, may also aid in acidic pH homeostasis by consuming protons (Li et al. 2023). The possession of these gene clusters in comammox Nitrospira demonstrates the versatility adaptability of this group, which may confer advantages over canonical ammonia oxidizers. Given the distinct phylogeny of potential acidophilic comammox Nitrospira cluster compared to the known isolates and enrichments, additional research is warranted to elucidate the physiological and metagenomic traits of this group.

In conclusion, our study elucidated the prevalence of comammox Nitrospira in an acidic biofilm reactor operating at pH 5.5 and 5. Shifts in the communities of Nitrospirae and comammox Nitrospira were observed, with a distinct acidophilic comammox Nitrospira cluster hypothesized to be responsible for the ammonia removal at these pH levels. The dominance of comammox Nitrospira under acidic conditions may be linked to their high ammonia affinity, propensity for biofilm formation, and versatility to produce and utilize various metabolites. The prevalence of comammox Nitrospira over canonical ammonia oxidizers under acidic pH may bring significant implications for the advancement of nitrification technology that obviates the necessity for neutralization procedures.

However, it is worth noting that the relative abundance of the potential comammox Nitrospira detected through amplicon sequencing does not necessarily correlate with the actual ammonia oxidation activity. Further investigation into the ecological characteristics of this distinct group of acidophilic comammox Nitrospira is warranted. Approaches such as Stable Isotope Probing (SIP) combined metagenomics analysis would be helpful to link the substrate consumption to the specific microbial groups. Additionally, efforts to isolate and enrich these organisms will be instrumental in elucidating their physiological characteristics.