Genomic-based taxonomic classification of the family Erythrobacteraceae

The family Erythrobacteraceae , belonging to the order Sphingomonadales , class Alphaproteobacteria , is globally distributed in various environments. Currently, this family consist of seven genera: Altererythrobacter , Croceibacterium , Croceicoccus , Erythrobacter , Erythromicrobium , Porphyrobacter and Qipengyuania . As more species are identified, the taxonomic status of the family Erythrobacteraceae should be revised at the genomic level because of its polyphyletic nature evident from 16S rRNA gene sequence analysis. Phylogenomic reconstruction based on 288 single-copy orthologous clusters led to the identification of three separate clades. Pairwise comparisons of average nucleotide identity, average amino acid identity (AAI), percentage of conserved protein and evolutionary distance indicated that AAI and evolutionary distance had the highest correlation. Thresholds for genera boundaries were proposed as 70 % and 0.4 for AAI and evolutionary distance, respectively. Based on the phylo-genomic and genomic similarity analysis, the three clades were classified into 16 genera, including 11 novel ones, for which the names Alteraurantiacibacter, Altericroceibacterium, Alteriqipengyuania, Alteripontixanthobacter, Aurantiacibacter, Paraurantiacibacter, Parerythrobacter, Parapontixanthobacter, Pelagerythrobacter, Tsuneonella and Pontixanthobacter are proposed. We reclassified all species of Erythromicrobium and Porphyrobacter as species of Erythrobacter . This study is the first genomic-based study of the family Erythrobacteraceae , and will contribute to further insights into the evolution of this family.

Development of genome sequencing technologies has made bacterial genomic data more and more accessible, resulting in a revolution in bacterial taxonomy [52][53][54]. Phylogenomic reconstruction can provide a higher-resolution phylogeny than that based on 16S rRNA gene or several housekeeping genes [55][56][57][58]. In addition, genomic similarity calculations including average nucleotide identity (ANI), average amino acid identity (AAI) and percentage of conserved protein (POCP) provide numerical thresholds for delineation of each taxon [59][60][61]. Therefore, a genome-wide investigation of the taxonomy of the family Erythrobacteraceae was performed to revise the taxonomic status of this family.

Collection of Erythrobacteraceae type strains
In addition to the 47 Erythrobacteraceae type strains for which genome sequences were available, 27 type strains were obtained from culture collections including the China General Microbiological Culture Collection (CGMCC), the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ), the Japan Collection of Microorganisms (JCM), the Korean Collection for Type Cultures (KCTC), the Korea Environmental Microorganism Bank (KEMB), the Collection of the Laboratorium voor Microbiologie en Microbiele Genetica (LMG) and the Marine Culture Collection of China (MCCC) or received as gifts from other scholars ( Table 1 and Acknowledgements). These type strains were cultivated under appropriate conditions proposed previously [3, 12, 14, 20, 28, 31-33, 36, 44, 62-77] for subsequent genomic sequencing.

Genomic annotation and comparative genomic analysis
Genomes for annotation and comparative analysis were selected following assessment of genomic completeness (>95 %) and contamination (<5 %) using CheckM software version 1.0.7 [80] with the command 'checkm lineage_wf -x fasta bins/ checkm/' . rRNA and tRNA genes were searched by the command RNAmmer 1.2 package [81] and the tRNAscan-SE web server (http:// lowelab. ucsc. edu/ tRNAscan-SE/) [82], respectively. Annotated 16S rRNA genes were used to compare sequence identities on the EzBioCloud web server ( www. ezbiocloud. net/ identify) [83] to confirm that a genome represented its corresponding type strain. Coding sequences (CDSs) were predicted and annotated by using Rapid Annotation using Subsystem Technology (RAST) web server version 2.0 (http:// rast. nmpdr. org/ rast. cgi) [84]. The DNA G+C contents were also calculated on the RAST web server version 2.0.
Comparative genomic analysis was performed as previously described [85,86]. Orthologous clusters (OCs) were identified by comparing whole protein sequences translated from CDSs pairwise with the execution of Proteinortho version 5.16b [87] with command '-e 1e-5 -cov=50 -identity=50' , which is accordance with the threshold values for a group of OCs sharing identities more than 50 % and coverage longer than half of their sequence lengths. Subsequently, single-copy OCs were filtered by an in-house Perl script.

16s rRNA gene phylogenetic and phylogenomic reconstructions
In accordance with previous polyphasic taxonomic studies of the members in the family Erythrobacteraceae [19,22,35,40,88], Rhodospirillum rubrum ATCC 11170 T was chosen as an outgroup, with its 16S rRNA gene sequence and genomic sequences obtained from the NCBI GenBank database under the accession numbers D30778 and CP000230-CP000231, respectively. 16S rRNA gene phylogeny was   [89]. Gene sequences of 80 Erythrobacteraceae type strains and an outgroup were aligned with clustal_w [90] built in in the mega7 software [91]. Then, aligned sequences were processed into maximumlikelihood phylogenetic analysis [92], using mega7 software with the substitution model and the bootstrap value set as Kimura's two-parameter model [93] and 1000 replicates, respectively.
Protein and gene sequences of filtered single-copy OCs were both performed in the phylogenomic analyses. Protein sequences were aligned by using mafft version 7 [94] with the parameter '-auto' , while gene sequences were aligned by mapping nucleotides on amino acids based on aligned protein sequences through PAL2NAL program version 14 [95]. Aligned sequences were refined to select the most reliable positions through trimAL version 1.4.1 [96] with the parameter '-automated1' and concatenated through our in-house perl script. Concatenation and partition methods were both applied in this study. The best substitution models for were proposed by IQ-Tree 1.6.1 software [97] with the command '-m MFP' . Subsequently, LG+F+R9 and GTR+F+R8 were estimated as the best substitution models for concatenations of amino acid and nucleotide sequences, respectively, and the best substitution models for partition methods are listed in Table S2. Maximum-likelihood phylogenomic trees were reconstructed by using IQ-Tree 1.6.1 software [97] with the bootstrap value set to 100 replicates.

Genomic similarity analysis
ANI, AAI and POCP values were used to calculate genomic similarities. ANI values were calculated by the orthologous average nucleotide identity tool (OrthoANI version 0.93.1) [98] implemented with the blast algorithm [99]. AAI values were obtained using the Kostas lab AAI calculator web server (http:// enve-omics. ce. gatech. edu/ aai/) [100]. POCP values were obtained according to the formula 'POCP=(C1 +C2)/ (T1 +T2)×100 %' where C1 and C2 indicated the conserved number of predicted proteins in the two pairwise compared genomes, respectively, as well as T1 and T2 stands for the total number of predicted proteins in the two pairwise compared genomes, respectively [59], following comparative genomic analysis by using Proteinortho version 5.16b with the command '-e=1e-5 -cov=50 -identity=40' . In addition, the t-tests of AAI, ANI and POCP values of inter-and intra-group were calculated by using the function 't.text' within R version 3.4.2 [101].

Characteristics of Erythrobacteraceae genomes
All obtained genomes were of high quality with genomic completeness of 97.  (Table S4) were included in our phylogenomic analyses.

16s rRNA gene phylogeny
As stated before, several genera within the family Erythrobacteraceae did not form an independent clade in the 16S rRNA gene phylogenetic tree ( Fig. S1) Croceibacterium and Croceicoccus formed two independent clades, and they did not belong to monophyletic clades which could be separated from other genera. Thus, 16S rRNA gene sequences did not confirm monophyletic relationships within the genera of the family [2,4,30,37,38]. Only 19 nodes accounting for 26.0 % exhibited bootstrap values higher than 70 %, indicating that this phylogenetic tree was not reliable enough to correctly reveal the taxonomic status of the genera of the family.

Phylogenomic and genomic similarity analyses proposing three clades
Four phylogenomic trees, based on 288 single-copy OC, amino acid or nucleotide sequences, with annotations and substitution models are shown in Table S4  Based on these phylogenomic trees, the family Erythrobacteraceae can be divided into three separate clades, Clades I, II and III, consisting of 47, 23 and four species, respectively (Fig. 2). Genomic similarity analyses by AAI, ANI and POCP calculations also supported that the three clades were significantly separated with p value<2.2×10 −16 (Fig. 3). Clades I and II contained most species. Clade III only contained four Croceicoccus species, indicating that the taxonomic status of this genus should not be changed.

AAI value and evolutionary distance classifying genera
A robust core-genome phylogeny of the family was obtained.
Although there is no generally recognized genus boundary, recent studies suggested AAI (60-80 %) and POCP (50 %) could be thresholds for distinguishing genera [59,61]. Evolutionary distance is also a relatively conserved criterion for inferring evolutionary relationships [102]. Pairwise comparisons of ANI, AAI, POCP and evolutionary distance indicated that the pair of AAI and evolutionary distance had a much higher correlation coefficient (r cc =0.85) than other pairs (Fig. 4). Type strains shared pairwise >50 % of POCP values, which is similar to the result of a phylogenomic study of the Roseobacter group [58], suggesting that POCP values could not be applied for delineating genera within the family Erythrobacteraceae. AAI was more suitable to distinguish each taxon in the family than ANI and POCP (Figs. S5-S7). Thus, AAI and evolutionary distance were selected to classify genera of Erythrobacteraceae.
Since phylogenetic tree topology is a major criterion for classifying genera, we propose that one genus should be clustered into one group only. Based on this criterion, Clade I is then composed of 10 genera (Fig. 5), including Genus I-I (Aeb. flavus  We therefore propose that phylogenomic topology supplemented with AAI values and evolutionary distance values could replace the phylogeny based on 16S rRNA gene sequences in the taxonomy of the family Erythrobacteraceae.

Genotype and phenotype support the proposal of new genera
Comparison of genomic contents within the family Erythrobacteraceae revealed that 12 443 OCs could be indicators for distinguishing newly proposed genera. While considerable metabolic diversity is found within the family, metabolic pathways involving carbon, nitrogen, phosphorus and sulfur could not be applied to refine the taxonomic status of this family. Therefore, the pathways of aerobic anoxygenic photosynthesis and flagella biosynthesis, which contain multiple genes and reactions [47,103], were selected to investigate their value as indicators for their taxonomic status.
Aerobic anoxygenic photosynthesis is encoded by a series of genes that were found in all Genus I-VI species, Aeb. ishigakiensis NBRC 107699 T (Genus I-V), Erb. marinus KCTC 23554 T (Genus II-III) and Erb. odishensis KCTC 239891 T (Genus II-III), as shown in Cells are Gram-stain-negative, ovoid to rod, non-sporeforming and non-motile. Aerobic or facultatively aerobic. Contains carotenoid pigments but not bacteriochlorophyll a. The predominant ubiquinone is Q-10. The major fatty acid (>10 %) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipids are diphosphatidylglycerol and phosphatidylethanolamine. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 60.5-67.0 % (by genome). The type species is Tsuneonella dongtanensis.
The description is the same as for Aeb. aerius [28]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The description is the same as for Aeb. dongtanensis [111]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The type strain, JM27 T (=KCTC 22672 T =CCTCC AB 209199 T ), was isolated from a tidal flat (Dongtan Wetland, Chongming Island, Shanghai, PR China). The DNA G+C content of the type strain is 65.8 % (by genome). Aerobic anoxygenic photosynthesis The description is the same as for Aeb. flavus [44]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The type strain, MS1-4 T (=MCCC 1K02683 T =NBRC 112977 T ), was isolated from mangrove sediment of the Jiulong River Estuary, Fujian Province, PR China. The DNA G+C content of the type strain is 60.5 % (by genome).
The description is the same as for Aeb. mangrovi [6]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The type strain, C9-11 T (=MCCC 1K03311 T =JCM 32056 T ), was isolated from a mangrove sediment sample collected from Yunxiao Mangrove National Nature Reverse in Zhangzhou, Fujian Province, PR China. The DNA G+C content of the type strain is 63.5 % (by genome).
The description is the same as for Aeb. rigui [112]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The description is the same as for Aeb. troitsensis [34]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Tsuneonella. The type strain, KMM 6042 T (=KCTC 12303 T =JCM 17037 T ), was isolated from the sea urchin Strongylocentrotus intermedius. The DNA G+C content of the type strain is 64.7 % (by genome).

EMENDED DESCRIPTION Of ThE GENuS QipEngyuania fENG ET al. 2015
The description is as given by Feng et al. [2] with the following amendment. Cells are aerobic or facultatively aerobic. Contains carotenoid pigments but not bacteriochlorophyll a. Positive or negative for oxidase. The major fatty acid (>10%) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipids are phosphatidylcholine, phosphatidylethanolamine and phosphatidylglycerol. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 60.6-66.7 % (by genome). The type species for the genus is Qipengyuania sediminis.
The description is the same as for Erb. aquimaris [73]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Qipengyuania. The description is the same as for Erb. gaetbuli [3]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Qipengyuania. The type strain, SW-161 T (=KCTC 12227 T =DSM 16225 T ), was isolated from a tidal flat of the Yellow Sea in the Republic of Korea. The DNA G+C content of the type strain is 64.1 % (by genome). The description is the same as for Erb. marisflavi [22]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Qipengyuania. The description is the same as for Aeb. oceanensis [70]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Qipengyuania.

DESCRIPTION Of QipEngyuania marisflavi
The type strain, Y2 T (=CGMCC 1.12752 T =LMG 28109 T ), was isolated from a deep-sea sediment of the western Pacific Ocean. The DNA G+C content of the type strain is 63.9 % (by genome).
Qipengyuania pelagi (pe' la. gi. L. gen. n. pelagi of/from the sea, reflecting isolation of the type strain from seawater of the Red Sea). The description is the same as for Erb. pelagi [77]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Qipengyuania. The type strain, UST081027-248 T (=JCM 17468 T =NRRL 59511 T ), was isolated from shallow seawater collected from the middle of the Red Sea. The DNA G+C content of the type strain is 64.2 % (by genome).

EMENDED DESCRIPTION Of QipEngyuania sEdiminis fENG ET al. 2015
Qipengyuania sediminis ( se. di' mi. nis. L. gen. n. sediminis of sediment) The description is identical to that given for Qpy. sediminis [2], except for the DNA G+C content. The description is the same as for Aeb. halimionae [32]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteriqipengyuania.
The type strain, CPA5 T (=CECT 9130 T =LMG 29519 T ), was isolated from the surface-sterilized aboveground tissues of the halophyte Halimione portulacoides. The DNA G+C content of the type strain is 65.5 % (by genome).

Basonym: Erythrobacter lutimaris Jung et al. 2014.
The description is the same as for Erb. lutimaris [114]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteriqipengyuania. The description is the same as for Erb. jejuensis [76]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Parerythrobacter. The type strain, CNU001 T (=KCTC 23090 T =JCM 16677 T ), was isolated from seawater collected off Jeju Island, Republic of Korea. The DNA G+C content of the type strain is 60.2 % (by genome).
The description is the same as for Aeb. lutipelagi [42]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Parerythrobacter. The type strain, GH1-16 T (=KCTC 52845 T =NBRC 113275 T ), was isolated from a tidal mudflat sample collected at the seashore of Gangwha Island, Republic of Korea. The DNA G+C content of the type strain is 60.6 % (by genome).
The description is identical to that given for Aeb. ishigakiensis [115], except for the DNA G+C content. The type strain, JPCCMB0017 T (=NITE-AP48 T =ATCC BAA-2084 T = NBRC 107699 T ), was isolated from the coastal area of Okinawa, Japan. The DNA G+C content of the type strain is 56.9 % (by genome).
The description is identical to that given for Aeb. xiamenensis [18], except for the DNA G+C content. The type strain, LY02 T (=CGMCC 1.12494 T =KCTC 32398 T =NBRC 109638 T ), was isolated from red tide seawater in Xiamen, Fujian Province, PR China. The DNA G+C content of the type strain is 61.8 % (by genome).

EMENDED DESCRIPTION Of ThE GENuS EryThrobacTEr ShIbA ET al. 1982
The description is as given by Shiba et al. 1982 [29] with the following amendment. Cells are motile or non-motile. Positive or negative for oxidase. Requires NaCl for growth. The major fatty acids (>10%) are C 18 : 1 ω7c and C 17 : 1 ω6c. The major polar lipids include a sphingoglycolipid. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 57.4-67.9 % (by genome). The type species for the genus is Erythrobacter longus.
Erythrobacter colymbi ( co. lym'bi. L. gen. n. colymbi, of a swimming pool, thus indicating the site of isolation of the type strain).
The description is the same as for Por. colymbi [24]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, TPW-24 T (=JCM 18338 T = KCTC 32078 T ), was isolated from swimming pool water in Tokyo, Japan. The DNA G+C content of the type strain is 66.5 % (by genome).
The description is the same as for Por. cryptus [26]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, ALC-2 T (=DSM 12079 T =ATCC BAA-386 T ), was isolated from the hot spring at Alcafache in Portugal. The DNA G+C content of the type strain is 67.9 % (by genome). The description is the same as for Por. donghaensis [105]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, SW-132 T (=KCTC 12229 T =DSM 16220 T ), was isolated from sea water from the East Sea in the Republic of Korea. The DNA G+C content of the type strain is 66.2 % (by genome).
The description is identical to that given for Erb. litoralis [31], except for the DNA G+C content. The type strain, T4 T (=ATCC 700002 T =CIP 106926 T =DSM 8509 T =JCM 10281 T =NBRC 102620 T ), was isolated from a marine cyanobacterial mat in a supralitoral zone. The DNA G+C content of the type strain is 65.2 % (by genome).
The description is identical to that given for Erb. longus [29], except for the DNA G+C content. The description is the same as for Por. neustonensis [23]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, ACM 2844 T (=CIP 104070 T =DSM 9434 T ), was isolated from air-water interface of freshwater subtropical pond in Brisbane, Australia. The DNA G+C content of the type strain is 65.3 % (by genome).
The description is the same as for Erm. ramosum [31]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, E5 T (=ATCC 700003 T =CIP 106927 T =DSM 8510 T =JCM 10282 T =NBRC 102621 T ), was isolated from a cyanobacterial mat from an alkaline spring. The DNA G+C content of the type strain is 64.3 % (by genome).
The description is the same as for Por. sanguineus [104]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter.
The type strain, A91 T (=ATCC 25659 T =DSM 11302 T =IAM 12620 T =ICPB 4167 T =NBRC 15763 T =JCM 20691 T ), was isolated from sea water collected in Baltic Sea. The DNA G+C content of the type strain is 63.6 % (by genome).
The description is the same as for Por. tepidarius [27]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Erythrobacter. The type strain, OT3 T (=DSM 10594 T ), was isolated from a cyanobacterial mat in brackish water of a hot spring in Shidzuoka Prefecture, Japan. The DNA G+C content of the type strain is 65.9 % (by genome).

DESCRIPTION Of ponTixanThobacTEr GEN. NOv.
Pontixanthobacter The description is the same as for Aeb. aestiaquae [62]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter. The description is the same as for Aeb. aquaemixtae [64]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter.
The type strain, JSSK-8 T (=KCTC 52763 T =NBRC 112764 T ), was isolated from the place where the ocean and a freshwater spring meet at Jeju Island, Republic of Korea. The DNA G+C content of the type strain is 58.5 % (by genome).
The description is the same as for Aeb. confluentis [20]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter.
The type strain, KEM-4 T (=KCTC 52259 T =NBRC 112305 T ), was isolated from water collected from an estuary environment where the ocean and a river meet at Seocheon, Republic of Korea. The DNA G+C content of the type strain is 59.1 % (by genome).
The description is the same as for Aeb. gangjinensis [67]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter.
The type strain, KJ7 T (=KACC 16190 T =JCM 17802 T ), was isolated from a tidal flat of the Gangjin bay in the Republic of Korea. The DNA G+C content of the type strain is 55.5 % (by genome).
Basonym: Altererythrobacter luteolus . emend. Kwon et al. 2007 The description is the same as for Aeb. luteolus [38,68]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter. The type strain, SW-109 T (=KCTC 12311 T =JCM 12599 T ), was isolated from a tidal flat of the Yellow Sea in the Republic of Korea. The DNA G+C content of the type strain is 59.3 % (by genome).
The description is the same as for Aeb. sediminis [72]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Pontixanthobacter. The type strain, CAU 1172 T (=KCTC 42453 T =NBRC 110917 T ), was isolated from a sample of lagoon sediment from along the east coast of the Republic of Korea. The DNA G+C content of the type strain is 61.5 % (by genome).
Cells are Gram-stain-negative, rod, non-spore-forming, non-motile and aerobic. Oxidase-and catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. Requires NaCl for growth. The predominant ubiquinone is Q-10. The major fatty acids (>10%) are summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c), summed feature 3 (C 16 : 1 ω7c and/or C 16 : 1 ω6c) and C 16 : 0 . The major polar lipids are diphosphatidylglycerol, phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol and sphingoglycolipid. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 60.8 % (by genome). The type species is Alteripontixanthobacter maritimus.
The description is the same as for Aeb. maritimus [43]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteripontixanthobacter. The type strain, HME9302 T (=KCTC 32463 T =KACC 17617 T =CECT 8417 T ), was isolated from seawater in the Republic of Korea. The DNA G+C content of the type strain is 60.8 % (by genome).
Cells are Gram-stain-negative, coccoid, non-sporeforming, non-motile and strictly aerobic. Oxidase-negative and catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. Requires NaCl for growth. Reduces nitrate to nitrite. The predominant ubiquinone is Q-10. The major fatty acids (>10 %) are summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c), summed feature 3 (C 16 : 1 ω7c and/or C 16 : 1 ω6c) and C 16 : 0 . The major polar lipids are diphosphatidylglycerol, phosphatidylethanolamine, phosphatidylglycerol and sphingoglycolipid. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 61.2 % (by genome). The type species is Parapontixanthobacter aurantiacus.
The description is the same as for Aeb. aurantiacus [65]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Parapontixanthobacter. The type strain, O30 T (=CGMCC 1.12762 T =JCM 19853 T =LMG 28110 T =MCCC 1A09962 T ), was isolated from a deep-sea sediment of the west Pacific Ocean. The DNA G+C content of the type strain is 61.2 % (by genome).
Cells are Gram-stain-negative, rod-shaped, non-sporeforming and aerobic. Motile or non-motile. Positive and negative for oxidase. Catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. Requires NaCl for growth. The predominant ubiquinone is Q-10. The major fatty acid (>10 %) is C 18 : 1 ω7c. The major polar lipids are diphosphatidylglycerol, phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol and sphingoglycolipid. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The description is the same as for Aeb. aerophilus [17]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Pelagerythrobacter. The description is the same as for Aeb. marensis [116]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Pelagerythrobacter. Croceibacterium, a genus name; N.L. neut. n. Altericroceibacterium, another or different Croceibacterium).
Cells are Gram-stain-negative, rod-shaped, non-sporeforming, aerobic and non-motile. Positive and negative for oxidase. Catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. Requires NaCl for growth. The predominant ubiquinone is Q-10. The major fatty acid (>10%) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipid is phosphatidylethanolamine. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 55.8-64.2 % (by genome). The type species is Altericroceibacterium indicum. The description is the same as for Aeb. endophyticus [32]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Altericroceibacterium. The type strain, BR75 T (=CECT 9129 T =LMG 29518 T ), was isolated from the surface-sterilized belowground tissues of the halophyte Halimione portulacoides. The DNA G+C content of the type strain is 58.6 % (by genome).

Basonym: Altererythrobacter indicus
The description is the same as for Aeb. indicus [33]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Altericroceibacterium. The type strain, MSSRF26 T (=LMG 23789 T =DSM 18604 T ), isolated from the rhizosphere of mangrove-associated wild rice (Porteresia coarctata Tateoka). The DNA G+C content of the type strain is 55.8 % (by genome).
The description is the same as for Aeb. xinjiangensis [15]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Altericroceibacterium.
The type strain, S3-63 T (=CCTCC AB 207166 T =CIP 110125 T ), was isolated from sand from the desert of Xinjiang, PR China. The DNA G+C content of the type strain is 64.2 % (by genome).

EMENDED DESCRIPTION Of ThE GENuS crocEibacTErium LIu ET al. 2019
The description is as given by Liu et al. [30]with the following amendment. Cells are pleomorphic. Some species can motile by means of polar flagella. Positive or negative for oxidase. The major fatty acid (>10 %) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipids are diphosphatidylglycerol, phosphatidylethanolamine and phosphatidylglycerol. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the coregenomic phylogeny. The DNA G+C content is 61.9-67.0 % (by genome). The type species for the genus is Croceibacterium ferulae.

DESCRIPTION Of crocEibacTErium aTlanTicum COMb. NOv.
Croceibacterium atlanticum ( at. lan' ti. cum. L. neut. adj. atlanticum of or pertaining to the Atlantic Ocean, where the type strain was isolated).
The description is the same as for Aeb. atlanticus [8]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Croceibacterium. The type strain, 26DY36 T (=CGMCC 1.12411 T =JCM 18865 T ), was isolated from a deep-sea sediment sample collected from the North Atlantic Rise. The DNA G+C content of the type strain is 61.9 % (by genome). The description is the same as for Aeb. salegens [71]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Croceibacterium. The type strain, XY-R17 T (=KCTC 52267 T =MCCC 1K01500 T ), was isolated from the surface sediment of Mai Po Inner Deep Bay Ramsar Site in Hong Kong. The DNA G+C content of the type strain is 64.6% (by genome). The description is the same as for Aeb. soli [14]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Croceibacterium. The type strain, MN-1 T (=KCTC 52135 T =MCCC 1K02066 T ), isolated from a desert sand sample collected from Tengger desert, north-western PR China. The DNA G+C content of the type strain is 67.0 % (by genome). The description is the same as for Aeb. xixiisoli [12]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Croceibacterium. The type strain, S36 T (=CGMCC 1.12804 T =NBRC 110413 T ), was isolated from soil of the Xixi wetland in Hangzhou, eastern PR China. The DNA G+C content of the type strain is 63.3 % (by genome).
Cells are Gram-stain-negative, pleomorphic, nonspore-forming, aerobic and non-motile. Oxidase-and catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. Requires NaCl for growth. The predominant ubiquinone is Q-10. The major fatty acid (>10%) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipids are phosphatidylethanolamine and phosphatidylglycerol. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 62.5-66.0 % (by genome). The type species is Alteraurantiacibacter aestuarii.

DESCRIPTION Of alTErauranTiacibacTEr aEsTuarii COMb. NOv.
Alteraurantiacibacter aestuarii ( aes. tu.a'ri.i. L. gen. n. aestuarii of a tidal flat, from where the type strain was isolated). The description is the same as for Aeb. aestuarii [63]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteraurantiacibacter. The type strain, KYW147 T (=KCTC 22735 T =JCM 16339 T ), was isolated from a seawater sample collected from the South Sea, Republic of Korea. The DNA G+C content of the type strain is 62.5 % (by genome).

DESCRIPTION Of alTErauranTiacibacTEr aQuimixTicola COMb. NOv.
Alteraurantiacibacter aquimixticola ( a. qui. mix. ti' co. la. L. fem. n. aqua water; L. masc. perf. part. mixtus mixed; L. suff. -cola inhabitant; N.L. masc. n. aquimixticola an inhabitant of mixed waters). The description is the same as for Aeb. aquimixticola [40]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteraurantiacibacter. The type strain, SSKS-13 T (=KACC 19863 T =KCTC 62900 T =NBRC 113545 T ), was isolated from sediment sampled at the junction between the ocean and a freshwater spring at Jeju island of the South Sea, Republic of Korea. The DNA G+C content of the type strain is 63.9 % (by genome).
The description is the same as for Aeb. buctensis [66]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Alteraurantiacibacter. The type strain, M0322 T (=CGMCC 1.12871 T =JCM 30112 T ), was isolated from the Mohe Basin, PR China. The DNA G+C content of the type strain is 66.0 % (by genome).
Cells are Gram-stain-negative, pleomorphic and nonspore-forming. Aerobic or facultative anaerobic. Positive or negative for oxidase. Catalase-positive. Contains carotenoid pigments but not bacteriochlorophyll a. The predominant ubiquinone is Q-10. The major fatty acid (>10 %) is summed feature 8 (C 18 : 1 ω7c and/or C 18 : 1 ω6c). The major polar lipid is phosphatidylethanolamine. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 58.3-67.2 % (by genome). The type species is Aurantiacibacter gangjinensis. The description is the same as for Erb. aquimixticola [88]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, JSSK-14 T (=KCTC 52764 T =NBRC 112765 T ), was isolated from water from the place where the ocean and a freshwater spring meet at Jeju island, Republic of Korea. The DNA G+C content of the type strain is 63.0 % (by genome). The description is the same as for Erb. arachoides [74]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, RC4-10-4 T (=CGMCC 1.15507 T =JCM 31277 T ), was isolated from an ice core in the East Rongbuk Glacier, Tibetan Plateau. The DNA G+C content of the type strain is 65.4 % (by genome).
The description is the same as for Erb. atlanticus [109]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, s21-N3 T (=MCCC 1A00519 T =KCTC 42697 T ) was isolated from deep-sea sediment of the Atlantic Ocean. The DNA G+C content of the type strain is 58.3 % (by genome).

DESCRIPTION Of auranTiacibacTEr gangjinEnsis COMb. NOv.
Aurantiacibacter gangjinensis ( gang. jin. en'sis. N.L. masc. adj. gangjinensis referring to Gangjin, the name of the bay in Korea from which the type strain was isolated). The description is the same as for Erb. gangjinensis [117]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, K7-2 T (=KCTC 22330 T =JCM 15420 T ), was isolated from seawater of Gangjin Bay, Republic of Korea. The DNA G+C content of the type strain is 62.7 % (by genome).
The description is the same as for Erb. luteus [118]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, KA37 T (=MCCC 1F01227 T =KCTC 42179 T ), was isolated from a mangrove sediment sample collected from Yunxiao mangrove National Nature Reserve, Fujian Province, PR China. The DNA G+C content of the type strain is 67.2 % (by genome).

DESCRIPTION Of auranTiacibacTEr marinus COMb. NOv.
Aurantiacibacter marinus ( ma. ri'nus. L. masc. adj. marinus of the sea, marine). The description is the same as for Erb. marinus [119]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, HWDM-33 T (=KCTC 23554 T =CCUG 60528 T ), was isolated from seawater of Hwang-do, an island of the Yellow Sea, Republic of Korea. The DNA G+C content of the type strain is 59.1 % (by genome).
The description is the same as for Erb. odishensis [13]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, JA747 T (=KCTC 23981 T =NBRC 108930 T ), was isolated from a soil sample of a solar saltern at Humma, Odisha, India. The DNA G+C content of the type strain is 63.7 % (by genome).

DESCRIPTION Of auranTiacibacTEr spongiaE COMb. NOv.
Aurantiacibacter spongiae (spon' gi. ae. L. gen. n. spongiae of a sponge, the source of the type strain). The description is the same as for Erb. spongiae [35]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, HN-E23 T (=MCCC 1K03331 T =LMG 30457 T ), was isolated from a sponge sample. The DNA G+C content of the type strain is 65.5 % (by genome).
The description is the same as for Erb. xanthus [19]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, SM1501 T (=KCTC 42669 T =CCTCC AB 2015396 T ), was isolated from surface seawater of the South China Sea. The DNA G+C content of the type strain is 64.5 % (by genome). The description is the same as for Erb. zhengii [9]. Coregenomic phylogenetic analysis strongly supported the placement of this species in the genus Aurantiacibacter. The type strain, V18 T (=KCTC 62389 T =MCCC 1K03475 T ), was isolated from deep-sea sediment of the Pacific Ocean. The DNA G+C content of the type strain is 62.7 % (by genome).
Cells are Gram-stain-negative, ovoid-to rod-shaped, nonspore-forming and aerobic. Oxidase-and catalase-positive. Requires NaCl for growth. Contains carotenoid pigments but not bacteriochlorophyll a. The predominant ubiquinone is Q-10. The major fatty acids (>10 %) are C 18 : 1 ω7c and summed feature 3 (C 16 : 1 ω7c and/or C 16 : 1 ω6c). The major polar lipids are diphosphatidylglycerol, phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol and sphingoglycolipid. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 65.0 % (by genome). The type species is Paraurantiacibacter namhicola.

DESCRIPTION Of parauranTiacibacTEr namhicola COMb. NOv.
Paraurantiacibacter namhicola ( nam. hi' co. la. N.L. n. namhae Namhae, the Korean name of the South Sea; L. suff. -cola from L. n. incola a dweller, inhabitant; N.L. masc. n. namhicola a dweller of the South Sea, referring to the isolation of the type strain).
The description is the same as for Aeb. namhicola [63]. Core-genomic phylogenetic analysis strongly supported the placement of this species in the genus Paraurantiacibacter. The type strain, KYW48 T (=KCTC 22736 T =JCM 16345 T ), was isolated from a seawater sample collected from the South Sea, Republic of Korea. The DNA G+C content of the type strain is 65.0 % (by genome).

EMENDED DESCRIPTION Of ThE GENuS crocEicoccus Xu ET al. 2009 EMEND. huANG ET al. 2015
The description is as given by Xu et al. [4] and Huang et al. [110] with the following amendments. Cells are coccid or rods. The major fatty acid (>10 %) is C 18 : 1 ω7c. The genus represents a distinct branch in the family Erythrobacteraceae of the class Alphaproteobacteria based on the core-genomic phylogeny. The DNA G+C content is 62.5-64.5 % (by genome). The type species for the genus is Croceicoccus marinus.

EMENDED DESCRIPTION Of crocEicoccus marinus Xu ET al. 2009 EMEND. huANG ET al. 2015
Croceicoccus marinus ( ma. ri'nus. L. masc. adj. marinus of or belonging to the sea, marine) The description is identical to that given for Ccc. marinus [4,110], except for the DNA G+C content. The type strain, E4A9 T (=CGMCC 1.6776 T =JCM 14846 T ), was isolated from a deep-sea sediment sample collected from a polymetallic nodule region in the East Pacific Ocean. The DNA G+C content of the type strain is 64.5% (by genome).
The description is identical to that given for Ccc. naphthovorans [110], except for the DNA G+C content. The type strain, PQ-2 T (=CGMCC 1.12805 T =NBRC 110381 T ) was isolated from marine biofilm collected from a boat shell at a harbour of Zhoushan island in Zhejiang Province, PR China. The DNA G+C content of the type strain is 62.6 % (by genome).

Funding information
This study was funded by the National Natural Science Foundation of China (No. 31770004 and No. 91851114), the Natural Science Foundation of Zhejiang Province (No. LR17D060001), the National Program for Support of Top-Notch Young Professionals.