Strains H1-1-2A^T and ZN11-R3-1 were isolated from a coastal sediment sample and from an enrichment culture inoculated with coastal styrofoam plastics, respectively. The sediment sample was collected from a Spartina alterniflora growing area in a wetland (24°86' N, 118°68' E) in Quanzhou Bay, Quanzhou, PR China, on September 05, 2019. A water-extracted medium (WEM) prepared using the nutrients extracted from the sediment with pure water (w/v=1:1) was used to isolate strain H1-1-2A^T (Huang et al., 2020b). The 0.2g sediment sample was subjected to 10-fold serial dilutions and spread on the WEM plates and incubated for 2weeks at 28°C. Strain H1-1-2A^T was picked and then streaked onto Marine Broth 2216 (MB, BD) agar plates to obtain a pure culture. For the isolation of strain ZN11-R3-1, styrofoam plastic was collected from a mangrove preservation area (24^o27' N, 117^o53' E) in Longhai, Zhangzhou, PR China, on November 23, 2019. The plastics were placed into an enrichment medium of 300ml sterile MMC (NaCl 24g/L; MgSO₄·7H₂O 7.0g/L; NH₄NO₃ 1g/L; KCl 0.7g/L; KH₂PO₄ 2.0g/L; and Na₂HPO₄·12H₂O 3.0g/L, pH=7.4) and maintained at 150rpm shaking at 28°C for 2months. An aliquot (2ml) of enriched culture was then transferred to another 100ml fresh MMC medium containing sterile plastics and cultured for another 2months. Then, the enrichment was repeated as above. The biomass in the third enrichment culture was collected using centrifugation at 6,000rpm for 15min and plated on an MB agar plate and maintained at 30°C. Strains H1-1-2A^T and ZN11-R3-1 grew well on MB agar plates and MB medium and were stored at −80°C with 20% glycerol (v/v) in the laboratory.

The nearly complete 16S rRNA gene sequences of strain H1-1-2A^T and strain ZN11-R3-1 were obtained using Sanger sequencing performed as described in a previous study (Huang et al., 2020b). The sequences were also compared with rRNA genes extracted from the genome sequences.

Sequences of the closely related relatives of the two strains were obtained from the EzBioCloud database (Yoon et al., 2017a) and the NCBI nucleotide database.³ Burkholderia cepacia ATCC 25416^T was selected as an outgroup. Then, the 16S rRNA gene sequences were aligned and subjected to phylogenetic analysis using two algorithms, neighbor-joining (NJ) and maximum likelihood (ML) methods with 1,000 bootstraps using MEGA 7.0 (Huang et al., 2019). The best model (T92+G+I) with the lowest Bayesian information criterion (BIC) scores was selected.

The draft genome sequences of strain H1-1-2A^T and strain ZN11-R3-1 were determined using the Illumina HiSeq X-Ten platform (Shanghai Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China). A library of ~400bp fragments was constructed, and paired-end (PE) short reads of ~1 Gb were obtained. The PE reads were firstly trimmed to remove the low base of quality <20 and length <50bp using sickle.⁴ Then, clean reads were assembled into contigs using SPAdes v 3.8.0 with a serial of k values of 21, 33, 55, 77, 99, 127 and –careful flag (Huang et al., 2019). Then, contigs shorter than 1kb were removed from the assembled contigs. The genome quality was evaluated using QUAST (Gurevich et al., 2013).

The complete genome of type strain H1-1-2A^T was obtained using PacBio sequencing with one SMART cell. The 10-kb fragment library was constructed followed the manufacturer’s instructions. The long reads were assembled using the SMRT Link (V6.0.0.47841) of PacBio.

The complete 16S rRNA gene sequence was extracted from the genome sequence using RNAmmer (Lagesen et al., 2007). Genome completeness was evaluated using CheckM v1.0.1 (Parks et al., 2015). Gene annotation was carried out using the RAST server (Aziz et al., 2008) and the KAAS system.⁵ Functional genes with high similarity to close relatives were searched using the blast+ program with e-value cutoff of 1e-5 (Camacho et al., 2009).

The average nucleotide identity (ANI) values were estimated using OrthoANI computation on the EzBioCloud Database (Yoon et al., 2017b). Digital DNA-DNA hybridization (dDDH) estimates were calculated on the GGDC website.⁶ Average amino acids identity among genomes was calculated using CompareM v0.1.2.⁷ The percentage of conserved proteins (POCP), proposed as genus boundary values was also calculated for genomic comparison (Qin et al., 2014).

The genomes affiliated to the order Rhodocyclales were downloaded from the genome portal in NCBI.⁸ A total of 303 genomes were obtained (until Feb.19, 2021), and the genome quality was checked using CheckM v1.0.1 (Parks et al., 2015). Genomes of <50% completeness and >10% contamination were removed from the following study. In addition, 9 genomes, identified using GTDB-tk v. 0.3.2 (Chaumeil et al., 2019), did not belong to the order Rhodocyclales, and these were removed from the study. Then, the phylogenomic tree of the genomes was inferred using a concatenated alignment of 120 bacterial single-copy genes with GTDB-tk v. 0.3.2 by using FastTree (Parks et al., 2018). The tree was edited using the Interactive Tree of Life (iTOL) online (Letunic and Bork, 2007). In addition, a phylogenomic tree based on the genomes of type strains belonging to the order Rhodocyclales was also constructed using GTDB-Tk.

Gram staining was carried out using a Gram staining kit (Hangzhou Tianhe Microorganism Reagent, Co., Ltd.). Colony morphology was recorded on a MB agar plate after incubation at 30°C for 3days. Catalase activity was tested by using 3% H₂O₂ solution. Oxidase activity was tested using the oxidase reagent (1% aqueous solution of N,N,N',N'-tetramethyl-p-phenylenediamine dihydrochloride, bioMérieux, France). Motility was observed by puncturing the cells into 0.5% agar. Growth under the anaerobic condition was tested by inoculating the cells into an anaerobic MB medium for 7days. The growth temperature range, NaCl tolerance range, pH range of the strains, and hydrolysis of substrates were determined as described in our previous study (Huang et al., 2020b). Growth on nitrogen-free medium was tested following the method of Huang et al. with 5g/l NaCl and 5g/l glucose (Huang et al., 2014). N. mangrovi M9-3-2^T (=MCCC 1K03313^T), obtained from the Marine Culture Collection Center (MCCC), was used as a reference strain.

Physiological and biochemical characterization was carried out using API ZYM, API 20NE, and API 20E kits according to the manufacturer’s instructions (bioMérieux, France). The tested strains and the reference strain were maintained under identical laboratory conditions. Test strips were maintained at 35°C for determining the physiological and biochemical properties.

For the analysis of fatty acids composition, the strains and reference strain were cultured in MB at 35°C for 3days and cells were collected by centrifugation at 8,000rpm for 10min. The cellular fatty acids were saponified, methylated and extracted, and then identified following the standard MIDI protocol (Sherlock Microbial Identification System, version 6B).

For the polar lipids analysis, strain H1-1-2A^T was cultured in MB medium at 35°C for 3days, and cells were harvested by using centrifugation as above. Polar lipids were extracted using a chloroform/methanol system and analyzed using one- and two-dimensional TLC using Merck silica gel 60F254 aluminum-backed thin-layer plates. Lipids were detected and identified by spraying the specific reagents (Huang et al., 2020a).

The 16S rRNA gene sequences of strains H1-1-2A^T and ZN11-R3-1, obtained by Sanger sequencing or extracted from the genome sequences, had 100% identity, indicating the two strains belonged to same species. The BOX-PCR genotypic fingerprinting profiles of two strains were similar but distinctive (Supplementary Figure 1), which confirmed that they were not clonal. Also, the fingerprinting of the two strains were totally different from N. mangrovi M9-3-2^T, indicating they may belong to a novel species different from N. mangrovi.

Sequence similarity search showed that the 16S rRNA gene sequence of strain H1-1-2A^T had the maximum similarity (99.6%) with an uncultured bacterium clone IWNB003 (accession number: FR744543), followed by N. mangrovi M9-3-2^T (98.4%), and had sequence similarities of ≤94.5% with other species affiliated to the family Zoogloeaceae. The clone IWNB003 was found in nitrate-amended injection seawater from an oil field (Gittel et al., 2012), and N. mangrovi M9-3-2^T has the ability of denitrification (Liao et al., 2021), which may indicate that Nitrogeniibacter members play valuable roles in nitrogen cycle in the environment.

Phylogeny of 16S rRNA gene sequence inferred from the ML and NJ methods placed strains H1-1-2A^T and ZN11-R3-1 within the genus Nitrogeniibacter as a novel monophyletic line, distinct from N. mangrovi M9-3-2^T. This indicated that the two strains could be considered as a novel species of the genus Nitrogeniibacter (Figure 1; Supplementary Figure 2).

Phylogeny of 16S rRNA gene sequences indicated that the members of Azoarcus and the members of Thauera were separated into different clades, which were clearly separated from the type species, A. indigens and T. selenatis. Firstly, A. pumilus SY39^T and “A. taiwanensis” NSC3^T formed a separate cluster, which did not cluster with the type species A. indigens. Here, we named this cluster as a novel genus Pseudazoarcus, which was equal to the group name “Azoarcus_D” of the Genome Taxonomy Database (GTDB; Chaumeil et al., 2019). Thus, A. pumilus should be transferred into the genus Pseudazoarcus. Azoarcus pumilus was designated the type species of this genus and was renamed as Pseudazoarcus pumilus comb. nov. “A. taiwanensis” (a name effectively but not validly published; Lee et al., 2014) was also affiliated to this genus. Secondly, in the phylogenetic clade of the genus Thauera, there were four species, including T. lacus, T. hydrothermalis, A. nasutitermitis, and A. rhizosphaerae that formed a monophyletic cluster. Though this cluster formed a node with other Thauera members, bootstrap support was low (<70% of both ML and NJ; Figure 1; Supplementary Figure 2). The four species may be assigned to a new genus named Pseudothauera, which is equivalent to “Thauera_A” in the GTDB taxonomy. Thus, the four species, T. lacus, T. hydrothermalis, A. nasutitermitis, and A. rhizosphaerae, should be transferred to a novel genus Pseudothauera and renamed as Pseudothauera lacus comb. nov., Pseudothauera hydrothermalis comb. nov., Pseudothauera nasutitermitis comb. nov., and Pseudothauera rhizosphaerae comb. nov., respectively. Thirdly, A. halotolerans HKLI-1^T formed an independent line on the phylogenomic tree, which clearly branched with Azoarcus. This species should be reclassified into a novel genus; Cognatazoarcus halotolerans gen. nov., comb. nov. was therefore proposed. Fourthly, A. communis SWub3^T did not cluster together with the type species A. indigens and should be transferred into a novel genus. Here, we named this cluster as a novel genus Parazoarcus, which was equal to the genus name “Azoarcus_C” of the Genome Taxonomy Database (GTDB).

The complete genome of strain H1-1-2A^T included one chromosome (4,678,511bp) and one plasmid (66,515bp). The draft genome size of strain ZN11-R3-2 was 4,656,485bp on 42 contigs (>1kb) with N50 value of 343,918bp (Table 1). Three copies of rrn operon (16S-

23S-5S rRNA genes) were found in the complete genome, and 16S rRNA gene copies were 1,528bp in length and had 100% identity. The genome size of the strains was a little larger than N. mangrovi M9-3-2^T (a chromosome of 4,236,644bp; accession number: CP048836). The DNA G+C content of strains H1-1-2A^T and ZN11-R3-1 were 62.67 and 62.71%, respectively, which were a little lower than N. mangrovi M9-3-2^T (67.13%). Gene prediction showed that there were 4,364 and 3,884 predicted genes in strains H1-1-2A^T and M9-3-2^T, respectively. dDDH and ANI values between strain H1-1-2A^T and ZN11-R3-1 were estimated to be 85.6 and 98.4%, respectively. These values exceeded the threshold of species delineation, which strongly supported that the two strains belonged to the same species (Kim et al., 2014). dDDH and ANI values between strains H1-1-2A^T and ZN11-R3-1 and N. mangrovi M9-3-2^T were 21.4–21.6% and 78.50–78.72%, indicating that strains H1-1-2A^T and ZN11-R3-1 represented a novel species.

Strains H1-1-2A^T and ZN11-R3-1 contained nitrogen fixation gene clusters encoding nitrogenase reductase (nifH, KO list: K02588), nitrogenase molybdenum-iron protein (nifD, K02586; nifK, K02591), and related proteins (modABCD; Supplementary Table 1), which were also found in N. mangrovi M9-3-2^T (Liao et al., 2021). The nitrogen fixation genes were assumed to enable their growth on nitrogen-free medium. Interestingly, denitrification genes, nirBD (K000362 and K000363), norBC (K004561 and K002305), and narGHIJ (K000370, K000371, K000373, and K000374), were not found in strains H1-1-2A^T and ZN11-R3-1, but they were present in N. mangrovi M9-3-2^T (Supplementary Table 2), in which denitrification was confirmed in a laboratory experiment (Liao et al., 2021). In addition, the sox system (soxABCDXYZ gene cluster) was present in strain H1-1-2A^T (Supplementary Table 3) and in N. mangrovi M9-3-2^T (Liao et al., 2021), suggesting that Nitrogeniibacter members may have the ability of sulfur oxidation.

The development of MAG binning and single-cell genomes contributed large numbers of genome sequences of uncultivated bacteria, including members of family Zoogloeaceae and the order Rhodocyclales, which could expand knowledge on the phylogenetic diversity based on core genome analysis. Here, the genomes of the order Rhodocyclales with ≥50% completeness and≤10% contamination were used, of which the genome quality was verified to perform accurate phylogenetic analysis by GTDB-Tk (Bowers et al., 2017). A total of 277 genomes affiliated to the order Rhodocyclales that meet the above standards were used in the phylogenomic analysis. Compared to the 78 and 92 genomes analyzed in phylogenomic studies of the order Rhodocyclales by Wang et al. (2020) and Liao et al. (2021), respectively, our study further expanded the known phylogenetic groups within the order Rhodocyclales. The described species account for a minor part of the phylogenomic tree, indicating that majority of the members of Rhodocyclales are still waiting to be cultivated (Figure 2).

Phylogenomic analysis based on 120 bacterial conserved single-copy genes strongly placed strains H1-1-2A^T and ZN11-R3-1 in a sister group of the genus Nitrogeniibacter, which was neighbored by “Denitromonas.” This agreed with the phylogeny based on concatenated core genome sequences (Liao et al., 2021). “Denitromonas” should be transferred into the family Zoogloeaceae and did not belong to the family Rhodocyclaceae.⁹ In the lineages of the family Zoogloeaceae, the relationship between A. pumilus SY39^T and “A. taiwanensis” NSC3^T showed congruent topology with 16S rRNA gene phylogeny (Figure 1), which strongly supported the two species should be reclassified into a novel genus, for which we propose the name Pseudazoarcus. Also, in the phylogenomic tree, T. lacus D20^T, T. hydrothermalis GD-2^T, A. nasutitermitis CC-YHH838^T, and A. rhizosphaerae CC-YHH848^T formed a monophyletic cluster, which also supported the phylogeny of the 16S rRNA gene. The four species should be assigned to a new genus, for which we propose the name Pseudothauera. In addition, phylogenomic analysis of A. olearius DQS-4^T, A. indigens VB32^T, and A. communis SWub3^T showed topology incongruent with the 16S rRNA gene, possibly due to the small number of sequences used. Thus, it is proposed that A. communis SWub3^T be reclassified into a novel genus named Parazoarcus gen. nov. Azoarcus halotolerans HKLI-1^T, which is only distantly related to the type species A. indigens, should also be reclassified into a novel genus. Thus, Cognatazoarcus gen. nov. was proposed. Niveibacterium firstly proposed in the family Rhodocyclaceae (Chun et al., 2016) should be transferred to the family Zoogloeaceae based on the phylogenetic analysis. Finally, a family-level lineage including the genus Rugosibacter was clearly separated from the family Zoogloeaceae, indicating that Rugosibacter may represent a novel family.

Figure 3 presents a small phylogenomic tree reconstructed using GTDB-tk, only including the type strains. Two genomes, Thauera selenatis AX^T (type species) with high genome contamination and “Zoogloea ramigera” ATCC 19544, possibly incorrectly named, were excluded (Supplementary Table 4). The topology of the small tree was congruent with that of the large tree, which supported the above analysis. AAI values calculated among the 40 members of the family Zoogloeaceae ranged from 60.34 to 94.53% (Figure 4), which exceeded the family boundary of >45% (Konstantinidis et al., 2017). Thus, the members should be considered to belong to the family Zoogloeaceae. Our analysis did not support the proposal of Uliginosibacterium as an independent family (Wang et al., 2020). Compared to POCP values, AAI values demonstrated certain advantages to delineate the genus boundary of the members of the Zoogloeaceae (Figure 4). The calculation of POCP values depends on the similarity of the protein contents of genomes, which had similar genome size (Qin et al., 2014). It is reported that POCP values are also not effective and appropriate for delineating the genera of the families Acetobacteraceae (Rai et al., 2021), Rhodobacteraceae (Suresh et al., 2019), and Methylococcaceae (Orata et al., 2018). For instance, the four species, T. lacus, T. hydrothermalis, A. nasutitermitis, and A. rhizosphaerae, clearly grouped together, ranging from 78.37 to 80.80% of the AAI values for the type strains, which were below the recommended genus cutoff of <80% (Luo et al., 2014). The four species were distinctly separated from Thauera members and other genera (Figure 4). The AAI values of Nitrogeniibacter compared to the genera Thauera, Parazoarcus, Azoarcus, Pseudothauera, Pseudazoarcus, and Cognatazoarcus were 65.5–67.4%, 65.9–66.8%, 65.6–67.0%, 66.2–68.5%, 64.1–66.1%, and 66.4–67.5%, respectively, which were below the genus cutoff of <80% (Luo et al., 2014). Thus, our study expanded the family Zoogloeaceae into 11 genera, including Zoogloea, Azoarcus, Aromatoleum, Thauera, Niveibacterium, Uliginosibacterium, Nitrogeniibacter, Parazoarcus, Cognatazoarcus, Pseudazoarcus, and Pseudothauera.

Colonies of strains H1-1-2A^T and ZN11-R3-1 on MB agar plates at 30°C were round, transparent, convex, and~1mm in diameter. The cells were rod-shaped, motile, and stained Gram-negative. Catalase activity and oxidase activity were found to be positive, similar to N. mangrovi M9-3-2^T. The tested strains did not degrade soluble starch, skim milk, carboxymethyl cellulose, and Tweens 40, 60, and 80. Anaerobic growth was not observed for strains H1-1-2A^T and ZN11-R3-1. The two strains can grow at 15–40°C with the optimum at 35°C and a pH range of 7.0–8.0 (Table 1). NaCl tolerance was observed at 0–4% (w/v) with the optimum of 0.5%, similar to N. mangrovi M9-3-2^T (Table 1). Strains H1-1-2A^T and ZN11-R3-1 can grow on nitrogen-free medium, similar to N. mangrovi M9-3-2^T. Nitrate cannot be reduced by strains H1-1-2A^T and ZN11-R3-1, in contrast to N. mangrovi M9-3-2^T which has denitrification ability. Additional biochemical and physiological properties of strains H1-1-2A^T and ZN11-R3-1 are listed in the species description.

The respiratory quinone of strain H1-1-2A^T was ubiquinone-8 (Q-8), as in the related N. mangrovi M9-3-2^T and other members of family Zoogloeaceae (Liao et al., 2021). The polar lipids consisted of phosphatidylethanolamine (PE), diphosphatidylglycerol (DPG), and phosphatidylglycerol (PG), two unidentified aminophospholipid (APL), one other phospholipid (PL), and one unidentified lipid (L; Supplementary Figure 3). The predominant fatty acids (>5%) of strain H1-1-2A^T consisted of summed feature 3 (43.2%), C_16:0 (23.0%), C_12:0 (9.5%), and C_10:0 3-OH (7.6%), similar to strain ZN11-R3-1 (37.9, 26.1, 6.3, and 5.1%, respectively; Table 2). Although the major fatty acids of strain H1-1-2A^T and strain ZN11-R3-1 were similar to N. mangrovi M9-3-2^T, the presence of minor fatty acids such as C_18:0 showed characteristic differences. The major isoprenoid quinone and major fatty acids of Nitrogeniibacter were similar to the closely related genera, Cognatazoarcus, Pseudazoarcus, Pseudothauera, and Parazoarcus, and their polar lipids composition showed somewhat different profiles (Supplementary Table 5).

Nitrogeniibacter aestuarii (aes.tu.a’ri.i. L. gen. n. aestuarii, of a coastal wetland, the source of the type strain isolated from wetland cordgrass and mangrove in estuary).

Colonies on MB agar plates cultured for 3days at 30°C are ~1mm, round, transparent, and convex. Cells are Gram stain-negative and rod-shaped. Growth occurs between 15 and 40°C with an optimum at 35°C, at 0–4% NaCl (w/v) with an optimum of 0.5% and a pH range of 7.0–8.0. Catalase-positive and oxidase-positive. Strains can grow on nitrogen-free medium. Nitrate cannot be reduced to nitrite. Positive for alkaline phosphatase, leucine arylamidase; weakly positive for esterase (C4), valine arylamidase, acid phosphatase, and naphtholAS-BI-phosphohydrolase. Hydrolysis of aesculin is weak positive. Malic acid and trisodium citrate can be used as sole carbon sources.

The quinone system is ubiquinone-8. The major fatty acids are summed feature 3 (C_16:1ω7c and C_16:1ω6c), C_16:0, C_12:0, and C_10:0 3-OH. The major polar lipids include phosphatidylethanolamine (PE), diphosphatidylglycerol (DPG), and phosphatidylglycerol (PG). The genome size is 4.7Mbp with DNA G+C content of 62.7%.

The type strain is H1-1-2A^T (=MCCC 1K04284^T=KCTC 82672^T), isolated from Spartina alterniflora wetland sediment. Another strain is ZN11-R3-1 (=MCCC 1A17971=KCTC 82671), isolated from the enrichment culture inoculated with plastics collected from a wetland mangrove.

The GenBank/EMBL/DDBJ accession numbers of 16S rRNA gene sequence of strains H1-1-2A^T and ZN11-R3-1 are MW644766 and MW644767, respectively. The whole-genome sequences of strains H1-1-2A^T and ZN11-R3-1 have been deposited at GenBank under the accession numbers CP071321-CP071322 and JAFKAB000000000, respectively.

In addition to the properties listed in the original description (Boden et al., 2017), the family Zoogloeaceae includes the genera Niveibacterium, Parazoarcus, Pseudothauera, Pseudazoarcus, and Cognatazoarcus. The AAI values among the members range from 60.34 to 94.53%. DNA G+C content is 56.6–68.7%.