**4. Discussion**

Bacteria evolve through a variety of genetic events, such as mutations, intergenomic shuffling, and horizontal gene transfer. ICEs are being identified in increasing numbers via bacterial genomic analysis. Genomic islands, including ICEs, are discrete DNA segments and play an important role in bacterial evolution [42,45,46]

In this study, we revealed that *bph* genes differentially existed in ten biphenyl/PCB degrading strains isolated from biphenyl-contaminated soil. The types I and II *bph* genes of KF strains belonging to groups I and II were very similar in terms of gene organization and nucleotide sequences, except that the *bphX* region was missing in the type II *bph* gene cluster. The ICE structures carrying the *bph* genes were identified in several strains. Typical ICE*bph-sal* of approximately 120 kb was observed in the group I strains and KF701 of group II, and they were accompanied by the *sal* genes, approximately 6-kb downstream of the *bph* genes. The KF715 strain of group II carried *bph* and *sal* gene clusters in the chromosome and also on a plasmid (pKF715A, 483 kb) [16]. ICEs are thought to reintegrate into the recipient's chromosome immediately after transfer. However, a recent study suggested that certain ICEs, such as ICE*Bs1* from *Bacillus subtilis* [47] and SXT/R391 ICEs from *Vibrio cholerae* [48], are capable of autonomous replication. Transconjugants of *P. putida* AC30Bph+ and *P. putida* KT2440Bph+ carry the *bph-sal* cluster as an extrachromosomal circular form [16]. The mobile element carrying the *bph-sal* cluster replicates autonomously like plasmid, maintained stably, and consists of genes sharing homologies to components of the DNA replication and stabilization machinery.

The first example of ICE-harboring genes for the degradation of xenobiotic compounds is the *clc* element (ICE*clc*, 105 kb) from *P. knackmussii* B13. ICE*clc* encodes for the catabolic pathway involved in 3- and 4-chlorocatechol degradation [43,49–51]. Besides this catabolic gene cluster, ICE*clc* has the core genes, such as type IV secretion system-encoding genes, relaxase, and integrase. ICE*clc* can excise through recombination between short direct repeats at either end (*attL* and *attR*). The excised ICE*clc* can transfer to a new recipient cell through the conjugation apparatus and it integrates into the recipient's chromosome between the 18-bp sequence at the 3' end of the tRNA-Gly gene on the chromosome (*attB*) and the identical sequence on the excised ICE*clc* (*attP*), thereby restoring the tRNA-Gly gene. Both excision and integration are mediated by the IntB13 integrase. The ICE*bph-sal*<sup>s</sup> of group I (strains KF702, KF703, KF707, KF710, and KF716) were related to ICE*clc* in terms of the conjugation apparatus and the core gene set. However, the genetic organizations of ICE*bph-sal*<sup>s</sup> are different from those of ICE*clc*. The integrases in the group I ICE*bph-sal* were almost identical (~99% in amino acid sequences), but it was as low as 59% between that of ICE*bph-sal* and ICE*clc*. Such low identity of the integrases between ICE*bph-sal* and ICE*clc* reflected the differences of the insertion sites. The ICE*bph-sal*<sup>s</sup> of group I were inserted at the 3- end of the tRNA-Gly gene (76 bp in length), which carries the CCC anticodon. ICE*clc* inserts into a number of tRNA-Gly genes, but only the genes which carry the GCC anticodon [51]. It should also be noted that the ICE*bph-sal* in strains belonging to groups I and II are present in *Pseudomonas* spp., whereas other strains than *Pseudomonas* possess different ICE*bph* as seen in group IV strains, indicating that ICE*bph-sal* have restricted host ranges.

In this study, we found that various types of *bph* genes are present in ten different strains isolated from the same soil sample. Several lines of evidence suggested that many *bph* genes in these strains were present on the chromosome as an integrated form. However, it is also true that certain ICE*bph-sal* is present stably as a plasmid. ICE*bph-sal*KF715 was stably maintained as a circular form in the two transconjugants, *P. putida* AC30Bph+ and *P. putida* KT2440Bph<sup>+</sup>. On the other hand, ICE*bph-sal*KF715 (circular) from *P. putida* AC30Bph+ and *P. putida* KT2440Bph+ seemed to integrate differently in another recipient *P. putida* F39/D. When *P. putida* AC30Bph+ was used as a donor strain, the largest *Spe*I DNA fragment of the F39/DBph+ transconjugant was hybridized with the KF715 *bphA1* probe (Figure 9). When *P. putida* KT2440Bph+ was used as a donor strain, two copies of the *bphA1* DNA were detected in the F39/DBph+ transconjugant strain at different positions (Figure 9). Investigations are currently underway to reveal how the ICE*bph-sal*KF715 is integrated in the genome of the F39/DBph+ transconjugants. The ICE*clc* of the *P. knackmussii* strain B13 was transferred by conjugation and integrated into two nonadjacent sites on the chromosome of toluene utilizing *P. putida* F1 [49]. Our repeated attempts to conjugally transfer the ICE*bph-sal* of other strains than KF715 have not been successful. This may be due to a lack of expression of the integrase genes or mutations in certain gene(s) involved in the excision or conjugal transfer.

ICE*bph*KF708 and ICE*bph*KF712 were found in *Cupriavidus* and *Comamonas*, respectively. ICE*bph*KF708 was almost identical to the ICEKKS102*4677* from *Acidovorax* sp. KKS102, and they had several nucleotide differences, indicating that this type of ICE*bph* could be transferred between *Cupriavidus* and *Acidovorax*. The right wing corresponding to the core legion (33 kb) of ICE*bph*KF712 and Tn*4371* is highly conserved, but the left wing is diversified (Figure 8). This indicated that these two ICE*bph* were rearranged in the left wing. ICE*bph-sal* (from the groups I and II strain) and ICE*bph* (from the group IV strain) are typical ICEs, which possess type IV secretion machinery. However, there are few relationships between ICE*bph-sal* and ICE*bph* in terms of gene organization, nucleotide sequence, and size (Figure 6, Figure 8). No significant identity was detected between the integrases of ICE*bph* and ICE*bph-sal*. The region encoding two conjugative transfer components, the *tra*/*trb* genes adjacent to the *bph* gene cluster, was found in ICE*bph*. The corresponding gene clusters were not found in ICE*bph-sal*. Gene components downstream of the *sal* gene cluster are likely to be involved in conjugative transfer in ICE*bph-sal*; however, they have not been identified because of the lack of reliable homologous genes that are identified as conjugative transfer components in public databases. Although ICE*bph-sal* and ICE*bph* possess common genetic components involved in biphenyl catabolism, their platforms are different.

The biphenyl degrading bacteria are considered to be responsible for lignin degradation at the final stage. Lignin is a complex compound based on the phenylpropane structure and contains a variety of biphenyl related molecules. Thus, the *bph* genes could be very ancient and distributed across a wide range of soil bacteria. Mobilization of the *bph* genes in soil bacteria can be achieved through various mobile genetic elements, including ICEs, transposons, and plasmids. It is highly conceivable that these genes can be modified and rearranged in different ways in new host cells. The results in this study provide a better understanding as to how soil bacteria exchange genetic islands involved in the catabolism of aromatic compounds, as well as how such genes are rearranged and modified in the natural environment.

**Supplementary Materials:** The following tables and figure are available online at http://www.mdpi.com/2073- 4425/10/5/404/s1. Table S1: Distribution of the catabolic genes for aromatic compounds and the heavy metal resistance genes in the biphenyl/PCB degrading KF strains. Tables S2–S13: Identity (%) of nucleotide sequence of *bphR, bphA1, bphA2, bphA3, bphA4, bphB, bphC, bphX0, bphX1, bphX2, bphX3,* and *bphD.* Figure S1: Comparison of the *bphX3* (a) and *bphD* (b) genes belonging to types I and II.

**Author Contributions:** K.F. planned the genome sequencing project and performed Southern blot hybridization analysis of the transconjugants. H.F. performed the sequence analysis of KF701 and KF702. H.S. performed the sequence analysis of KF703, KF708, and KF715. N.K. performed the sequence analysis of KF707. T.W. performed the sequence analysis of KF709 and KF710. J.H. participated in the sequence analysis of KF712 and KF716 and annotation of ICE*bph-sal*/ICE*bph*. A.S., T.F., and M.G. were involved in the whole genome annotation of the KF strains and completed the manuscript. All authors read and approved the final manuscript.

**Funding:** This work was partly supported by a research gran<sup>t</sup> from the Institute for Fermentation, Osaka (G-2018-3-020 to T.W.). The APC was funded by a research gran<sup>t</sup> from Beppu University (to H.F.).

**Acknowledgments:** We thank Atsushi Yamazoe and Akito Nishi for their useful discussion and suggestions.

**Conflicts of Interest:** The authors declare no conflict of interest.
