J Biosci (2020)45:58 Ó Indian Academy of Sciences

DOI: 10.1007/s12038-020-00030-9 (0123456789().,-volV)(0123456789().,-volV)

A soil bacterial catabolic pathway on the move: Transfer of nicotine catabolic genes between Arthrobacter genus megaplasmids and invasion by mobile elements

1 2 RODERICH BRANDSCH and MARIUS MIHASAN * 1Institute of Biochemistry and Molecular Biology, Faculty of Medicine, University of Freiburg, Freiburg, Germany 2Biochemistry Laboratory, Biology Department, Alexandru Ioan Cuza University of Iasi, Iasi, Romania

*Corresponding author (Email, [email protected]) MS received 16 November 2019; accepted 20 March 2020

The 165,137 bp plasmid pAO1 of Paenarthrobacter nicotinovorans carries the genes of a nicotine catabolic pathway. The genes are organized into several gene modules responsible for the catabolism of L- and D-nicotine to nicotine blue, a-ketoglutarate and succinate. Various modules of these genes have been shown to be present in gram-positive (Gram?) soil bacteria. The presence of the identical pAO1 nic-genes on the 288,370 bp plasmid pZXY21 of Arthrobacter sp. ZXY2 (96% to 100% at the nucleotide level) permitted the identification of the limits of this DNA fragment. At the 50 end of the nic-genes are located the ORFs of two predicted integrases of the tyrosine recombinase family with conserved R, H, R and Y catalytic residues and that of a small transposase with a predicted leucine zipper motive. They are related to Tn554A, Tn554B and Tn554Cof Staphylococcus aureus and suggest that the entire nic-genes DNA fragment represents a large catabolic transposon. Surprisingly the nic-genes on pZXY21 were found to be interspersed by mobile elements encoding transposases of various IS families. Insertion of these IS elements disrupts nicotine degradation and divide the nic-genes DNA into potentially new transposons. This finding may illustrate how nicotine catabolic genes can be mobilized and spread by horizontal gene transfer to other soil bacteria.

Keywords. Arthrobacter; DNA transposable elements; horizontal gene transfer; nicotine metabolism; plasmids

1. Introduction pyridine ring of nicotine, opening of the pyrrolidine ring and cleavage of N-methyl-c-amino-butyrate with The molecular evolution of catabolic pathways carried the formation of 4,40,5,50-tetrahydroxy-3,30-di- on soil bacterial plasmids and their spread by hori- azadiphenoquinone-(2,20) or nicotine blue (NB) zontal gene transfer was the focus of our study. As a (Brandsch 2006). Recently it was shown that a putative model we made use of the gene cluster of nicotine trihydroxypyridine hydrolase (thpH) can hydrolytically degradation carried on the Paenarthrobacter nicoti- cleave trihydroxylated pyridines with the formation of novorans (formerly Arthrobacter nicotinovorans) 2-oxo-glutaramat (Vaitekunas et al. 2016) which can be plasmid pAO1 (GenBank accession number: deamidated by a kgA encoded a a-ketoglutaramat AJ507836) (Igloi and Brandsch 2003; Busse 2016) amidase (Cobzaru et al. 2011). Both the thpH and kgA which is organized into several gene modules. The genes encoded by pAO1 were shown to have a nico- upper pathway genes encode enzymes, transporters and tine-dependent expression (Miha˘s¸an et al. 2018) and regulatory proteins required for hydroxylation of the might connect the nicotine pathway to the citric acid Electronic supplementary material: The online version of this article (https://doi.org/10.1007/s12038-020-00030-9) contains supplementary material, which is available to authorized users. http://www.ias.ac.in/jbiosci 58 Page 2 of 12 Roderich Brandsch and Marius Mihasan cycle via a glutamate dehydrogenase (Cobzaru et al. degradation (nic-genes DNA) was different from the 2011). rest of the plasmid DNA (Igloi and Brandsch 2003). The lower pathway genes encode enzymes and Here we report the finding that the entire pAO1 nic- transcriptional regulators required for the degradation genes cluster, upper and lower pathway, including a of N-methyl-c-amino-butyrate generated in the upper 6-hydroxy-D-nicotine oxidase gene and its regulator, pathway. The end products are methylamine which is are present in a virtually identical copy on the pZXY21 excreted in the medium and succinic acid which is plasmid of Arthrobacter sp. ZXY-2 (GenBank acces- integrated in the main metabolic pathways of the cell. sion number: NZ_CP017422). The nic-genes on the Brandsch (2006) and Gurusamy and Natarajan (2013) two plasmids showed high identity at the DNA level as provide reviews on the molecular organization of well as at the deduced translation products. The high nicotine degradation gene clusters and the chemical identity of the nic-gene DNA segments of the two intermediates of the pathway, while updated informa- plasmids provided the unique opportunity to determine tion on nicotine metabolism enzymes and their struc- the DNA limits of the transposed segment containing tures and mechanisms of actions is provided by the nic-genes. The entire DNA segment containing nic- Fitzpatrick (2018). genes showed the characteristics of large catabolic With the increasing number of bacterial genomes transposons (Rauch and De Vos 1992; Rauch and de sequenced, a more detailed insight into the relation- Vos 1994; Wright and Grossman 2016) with ORFs of ships of particular catabolic pathways of soil bacteria, integrases of the tyrosine family of recombinases sit- into the mechanisms of their evolution and spread uated at one end of the 69.8 kbp DNA fragment. In among different members of the bacterial soil com- addition, the analysis of the two DNA segments con- munity becomes possible. taining the nic-genes of pAO1 and pZXY21 revealed We have shown that the Paenarthrobacter pAO1 the surprising finding that the DNA of the nicotine- genes of the upper pathway of nicotine degradation genes cluster on pZXY21 was invaded in the host including those of nicotine dehydrogenase (ndhMSL), Arthrobacter sp. ZXY2 by mobile elements. These 6-hydroxy-L-nicotine oxidase (6hlno), ketone dehy- mobile elements were present on pZXY21 outside the drogenase (kdhMSL), 2,6-dihydroxy-pseudooxyni- nic-genes cluster as well as on the bacterial chromo- cotine hydrolase (ponH), dihydroxypyridine some. Our findings present a snapshot view on the hydroxylase (dhpH), trihydroxypyridine hydrolase breakdown of a large catabolic gene cluster into (thpH), a-ketoglutaramat amidase (kga) and some potentially new transposons. This leads to loss of accessory genes (coxG,D,E,F and mobA) are present in function of nicotine degradation. The newly generated similar clusters in Nocardioides sp. JS614 and transposons, with their cargo of catabolic genes, may Rhodococcus opacus B4 (Mihasan and Brandsch disseminate these genes by horizontal gene transfer and 2013). These gene clusters were apparently spread thus eventually produce new catabolic pathways among the bacteria by horizontal gene transfer aided by (Koonin et al. 2001; Toussaint and Merlin 2002; pAO1-related plasmids (Ganas et al. 2008; Cobzaru Wiedenbeck and Cohan 2011; Harrison and Brockhurst et al. 2011). A particular feature of the pAO1-mediated 2012). nicotine degradation by Paenarthrobacter nicotinovo- rans is the presence of a 6-hydroxy-D-nicotine specific oxidase (6hdno) responsible for the catabolism of the 2. Methods small amounts of D-nicotine present in tobacco plants (Brandsch, 2006). This gene, its regulator hdnoR as All nucleotide and protein sequences were downloaded well as the entire lower pathway of c-N-methy- from NCBI GenBank (Clark et al. 2016) using the laminobutyrate degradation were absent from the following IDs: pAO1 – AJ507836.1; pZXY21 – gene clusters in R. opacus B4 and Nocardioides sp. NZ_CP017422; pZXY22 - GI: NZ_CP017423; JS614. pZXY23 – NZ_CP017424; pZXY24 – NZ_CP017425; We have also shown that pAO1 is member of a pZXY25 – NZ_CP017426. Biological sequence simi- family of related Arthrobacter catabolic plasmids car- larity search was performed by using Basic Local rying a predicted type IV secretion system (T4SS) Alignment Search Tool (BLAST) (Altschul et al. 1990) involved in plasmid conjugation and similar replication using a minimum value of e-2 for assessing signifi- and partitioning genes (Miha˘s¸an and Brandsch 2016). cance. The amino acid sequence of proteins was ana- The 165,137-nucleotide sequence of pAO1 indicated lyzed by the InterProScan protein signature database that the GC content of the gene cluster for nicotine using InterProScan tools (Finn et al. 2017). Protein Nicotine catabolic genes in Arthrobacter Page 3 of 12 58 motifs were identified using Motif finder (Combet et al. sites and translational start sites were not taken into 2000). SignalP was employed for the prediction of consideration. The difference in GC content of the nic- signal peptide cleavage sites in proteins (Petersen et al. gene cluster and the rest of the plasmid DNA observed 2011) and prediction of transmembrane regions was for pAO1 (Igloi and Brandsch 2003) was found to performed by DAS-TM-filter according to (Cserzo apply also for pZXY21 (mean GC content of the nic- et al. 2002). MAFFT version 7 program (Yamada et al. gens DNA was 58.53% vs 61.92% of the plasmid, see 2016) and Cobalt (Papadopoulos and Agarwala 2007) supplementary figure 1). were used for multiple alignments of amino acids and pZXY21 is not pAO1 because outside the nic-genes nucleotides. Gene synteny was analyzed and visualized clusters the percent identity of homologous genes drops with ACT (Carver et al. 2005) and progressive Mauve significantly (supplementary figure 2). pAO1 and (Darling et al. 2010). GC content analysis were per- pZXY21 belong to a related family of plasmids carrying formed and visualized with Artemis (Carver et al. ORFs of a predicted conserved type IV secretion system 2012) (T4SS, figure 1B) (Miha˘s¸an and Brandsch 2016). The pAO1 ORFs Q8GAF0 to Q8GAD9 are homologs of pZXY21 ORFs A0A1D9FFC0 to A0A1D9FF11. The 3. Results average identity of the amino acid sequences of these 12 ORFs was 76.58% (table 2). Both plasmids contain a 3.1 Two identical gene clusters for nicotine kfrA similar ORF, a gene found typically in Gram neg- degradation on two different Arthrobacter plasmids ative bacteria where it is involved in plasmid mainte- nance (Adamczyk et al. 2013). The increase in the available bacterial genome This data showed the presence of two related but sequences prompted us to reevaluate by BLAST anal- distinct Arthrobacter plasmids carrying the same set of ysis the presence of related pAO1 nic-catabolic genes nic-genes and ORFs. in other soil bacteria. The analysis indicated the pres- ence of virtually identical genes on pZXY21 of Arthrobacter sp. ZXY-2 (GenBank accession number: 3.2 The ends of similarity of the DNA segments NZ_CP017422, Zhao X., Wang L., Ma F., Yang J.; carrying the nic-gene cluster Submitted on 27-Sep-2016, School of Municipal and Environmental Engineering and State Key Laboratory The DNA segments containing the nic-genes were in of Urban Water Resource and Environment, Harbin inverted orientation on pAO1 and pZXY21. One end of Institute of Technology, Harbin City, Heilongjiang identity was situated 30 of the 6hdnoR regulator gene 150090, China). (ORF Q8GAF9) at pAO1nucleotide position G118165. The DNA identity at the nucleotide level of the two On pZXY21 this corresponded to ORF A0A1D9FFI3 DNA segments of pAO1 and pZXY21, including the (6hdnoR regulator) at nucleotide position G149715. The coding sequences of the nic-catabolic genes and the G nucleotides represent the last nucleotide in the palin- intergenic sequences, was 96% to 100%. On pAO1 this drome sequence 50CTGAC.GTCAG 30 present at this DNA segment started with nucleotide 48,358 and end of the DNA segments (figure 1B). The other end of ended with nucleotide 118,165. On pZXY21 it started identity of the DNA segments of the two plasmids was with nucleotide 223,517 and ended with nucleotide situated 113 nt 50 to the pAO1 ORFs Q8GAJ9 to 149,715 (figure 1A). That is a 69,807 bp pAO1 DNA Q8GAJ5 of transposases related to those of S. aureus fragment and a 73,802 bp pZXY21 DNA fragment. Tn554. These ORFs are part of the DNA fragment car- According to the nucleotide numbering, these DNA rying the nic-genes on pAO1, but not on pZXY21. segments were in inverted orientation on the plasmids The DNA identity between pAO1 and pZXY21 (figure 1B). The difference of 3995 nucleotides restarts at pAO1 nucleotide 48292 (pZXY21 nt 223582) between the two DNA segments was accounted for by with the AT rich sequence 50CTTTACATAA 30 con- the nucleotides of the ORFs of the transposases present taining a short palindrome (underlined) and ended after on the pZXY21 fragment, but not present on the pAO1 65 nucleotides at pAO1 nt 48358 (pZXY21 nt 223517) fragment (see further down). with the palindrome sequence 50CGGGA(AC)TC The 50 genes of the nic-gene clusters of the two CCG30 (figure 2). Between pZXY21 nt 223517 and nt plasmids showed an average identity of over 99% 223524 an 1894 nt long sequence of pAO1 similar DNA (table 1). Small ORFs of uncertain reality resulting was missing including the ORFs Q8GAJ9, Q8GAJ8 and from differences in assignment of ribosomal binding 39 N-terminal amino acids of Q8GAJ7. The 7nt between 58 Page 4 of 12 Roderich Brandsch and Marius Mihasan A

B

Figure 1. The nic-genes are present in a virtually identical copy on the pAO1 plasmid of Paenarthrobacter nicotinovorans and pZXY21 plasmid of Arthrobacter sp. ZXY-2. (A) Schematic drawing depicting the nic-genes location and their order on the two plasmids. Known ORFs of the upper nicotine pathway: NdhMSL – nicotine dehydrogenase; 6HlnO – 6-hydroxy-L- nicotine oxidase; 6HdnO – 6-hydroxy-D-nicotine oxidase; KdhMSL – ketone dehydrogenase; PonH – 2,6-dihydroxy- pseudooxy-nicotine hydrolase; ThpH – trihydroxypyridine hydrolase; DhpH – dihydroxypyridine-3-hydroxylase; KgA – a- ketoglutaramat amidase. Known ORFs of the lower nicotine pathway: pmfR – transcriptional regulator; MabO – N-methyl- gamma-aminobutyrtae oxidase; SsaDH – succinate semialdehyde dehydrogenase; MaO – gamma-aminobutyrate oxidase; The UniProtKB IDs for each ORFs in the corresponding plasmids are available in table 1.(B) Gene co-linearity and nucleotide sequence identity between pAO1 and pZXY21; green colored arrows – nic-gene cluster, orange colored arrows – type IV secretion system, yellow colored arrows – kfrA related gene cluster involved in plasmid maintenance; UniProtKB IDs are indicated for ORFs that are not found on both plasmids, magenta text and arrow – transposases, cyan text and arrows – unknown function (A0A1D9FFL8, predicted transposase related to Q8GAH5/IS1473; Q8GAG5 predicted major facilitator superfamily permease). Nucleic acid similarities (BLASTN comparisons showing C95% identity) are depicted as solid lines between sequences, where red indicates regions of identity in the same orientation and blue indicates regions of identity with the opposite orientation. The figure was generated with the Artemis Comparison Tool (ACT) provided by the Sanger Centre (Carver et al. 2005). Nicotine catabolic genes in Arthrobacter Page 5 of 12 58 Table 1. Identity between genes and ORFs of the nic clusters on pAO1 and pZXY21

UniProtKB IDs

Protein pAO1 pZXY21 Aminoacid Identities (%) Nucleotide Identities (%) HdnoR Q8GAF9 A0A1D9FFI3 140/140 (100%) 422/423 (99%) aaTrsp Q8GAG0 A0A1D9FF88 461/465 (99.1%) 1375/1395 (99%) 6hdno Q8GAG1 A0A1D9FF21 452/458 (98.7%) 1363/1377 (99%) aaTrsp Q8GAG2 A0A1D9FFQ8 493/497 (99.2%) 1485/1494 (99%) Hypothetical protein (101 aac) Q8GAG3 A0A1D9FF84 100/101 (99%) 302/306 (99%) MalR/LacI Q8GAG4 A0A1D9FFD6 273/276 (98.9%) 834/840 (99%) ß-glucosidase Q8GAG6 A0A1D9FFE3 499/506 (98.9%) 1624/1632 (99%) modC Q8GAG7 A0A1D9FF98 346/349 (99%) 1045/1050 (99%) modB Q8GAG8 A0A1D9FF30 237/239 (99.2) 714/721 (99%) ModA Q8GAG9 A0A1D9FF66 258/262 (98.9%) 779/789 (99%) MoaE Q8G970 A0A1D9FF69 152/155 (98.1%) 462/468 (99%) MoaC Q8GAH0 A0A1D9FFR9 167/168 (99.4%) 504/507 (99%) MoeA Q8GAH1 A0A1D9FF94 420/429 (98%) 1271/1290 (99%) moaA Q8GAH2 A0A1D9FFE6 361/372 (97%) 1103/1119 (99%) moaD Q8GAH3 A0A1D9FFF4 87/88 (98.9%) 264/267 (99%) ndhM Q59127 A0A1D9FF74 283/283 (100%) 850/852 (99%) ndhS Q59128 A0A1D9FF75 50/51 (98%) 154/156 (99%) ndhL Q93NH5 A0A1D9FFS9 815/816 (99.1%) 2441/2451 (99%) 6hlno Q93NH4 A0A1D9FFA5 425/425 (100%) 1275/1278 (99%) ORF92 (124 aac) Q8GAH6 * 98.5% – CAD47950 kdhS O87682 ** 100% – CAD47946 kdhM O87681 A0A1D9FFM2 296/296 (100%) 889/891 (99%) TscReg Q8GAH7 A0A1D9FF57 393/394 (99.7%) 1182/1185 (99%) TscReg Q8GAH8 A0A1D9FFC7 325/326 (99.7%) 979/981 (99%) ORF116 Q93NG7 A0A1D9FF91 115/116 (99.1%) 348/350 (99%) ponH Q93NG6 A0A1D9FF93 367/367 (100%) 1103/1104 (99%) kdhL Q933N0 A0A1D9FFU9 793/794 (99.9%) 2381/2385 (99%) cupin Q93NG5 A0A1D9FFC4 305/306 (99%) 100/101 (99%) thpH Q93NG4 A0A1D9FFH2 310/310 (100%) 933/933 (100%) dhpH Q93NG3 A0A1D9FFH4 397/397 (100%) 1194/1194 (100%) coxG Q93NG2 A0A1D9FFQ1 234/235 (99.6%) 707/708 (99%) kgA Q93NG1 A0A1D9FFN0 291/291 (100%) 876/876 (100%) coxD Q93NG0 A0A1D9FFD8 297/297 (100%) 893/894 (99%) coxE Q93NF9 A0A1D9FF65 361/363 (99.4%) 1089/1092 (99%) coxF Q93NF8 A0A1D9FFA2 377/377 (100%) 1132/1134 (99%) ORF73 Q93NF7 A0A1D9FFA1 116/117 (99.1%) 353/354 (99%) mobA Q93NF6 A0A1D9FFV9 203/204 (99.5%) 613/615 (99%) part. Transpos. Q8GAH8 A0A1D9FFR5 85/85 (100%) 258/258 (100%) pmfR Q8GAH9 A0A1D9FFB2 470/470 (100%) 1411/1413 (99%) aa perm Q8GAI0 A0A1D9FFB1 480/481 (99.6%) 1443/1446 (99%) purU Q8GAI2 A0A1D9FFE4 287/287 (100%) 864/864 (100%) mabO Q8GAI3 A0A1D9FFI4 824/824 (100%) 2474/2475 (99%) folD Q8GAI4 A0A1D9FFJ0 308/308 (100%) 926/927 (99%) nepA Q8GAI5 A0A1D9FFS5 73/74 (98.6%) 225/225 (100%) nepB Q8GAI6 A0A1D9FFP8 111/112 (99.1%) 339/339 (100%) faadh Q8GAI7 A0A1D9FFF6 145/146 (99.3%) 441/441 (100%) ssadh Q8GAI8 A0A1D9FF85 448/450 (99.6%) 1352/1353 (99%) mao Q8GAJ0 A0A1D9FFY2 420/421 (99.8%) 1265/1266 (99%) cupin Q8GAJ1 A0A1D9FFF3 48/49 (98%) 144/147 (99%) nbor Q8GAJ2 A0A1D9FFK3 203/204 (99.5%) 612/615 (99%) Average identity*** 99,38% 99.19%

*Q8GAH6 homolog from pAO1 is cut in two on pXYZ2 by transposases A0A1D9FF51 (locus ARZXY2_4548) and A0A1D9FFB7 (locus ARZXY2_4547), the resulting fragments are not annotated; **O87682 homolog from pAO1 is present on pXYZ2 but on a different reading frame and not properly annotated; ***small ORFs encoding hypothetical proteins of uncertain relevance were not considered. 58 Page 6 of 12 Roderich Brandsch and Marius Mihasan Table 2. Identity between genes and ORFs of a predicted T4 secretion system of pAO1 and pZXY21

UniProtKB IDs pAO1 pZXY21 Aminoacid Identities (%) Nucleotide Identities (%) Q8GAF0 A0A1D9FFC0 101/133 (76%) 283/346 (82%) Q8GAE9 A0A1D9FEZ2 521/588 (89%) 1451/1720 (84%) Q8GAE8 A0A1D9FF02 469/530 (88%) 1251/1450 (86%) Q8GAE7 A0A1D9FF50 453/502 (90%) 1274/1506 (85%) Q8GAE6 A0A1D9FF73 339/434 (78%) 934/1249 (75%) Q8GAE5 A0A1D9FF96 122/251 (49%) 272/385 (71%) Q8GAE4 A0A1D9FEW1 77/89 (87%) 230/271 (85%) Q8GAE3 A0A1D9FES9 147/192 (77%) 470/596 (79%) Q8GAE2 A0A1D9FEW3 417/507 (82%) 836/947 (88%) Q8GAE1 A0A1D9FEW4 135/170 (79%) 413/511 (81%) Q8GAE0 A0A1D9FFC9 288/357 (81%) 814/982 (83%) Q8GAD9 A0A1D9FF11 48/111 (43%) 37/54 (69%) Average identity 76.58% 80.66%

A

B

Figure 2. Insertion of the A0A1D9FFG5 integrase from pZXY21 into pAO1. (A) Gene collinearity and nucleotides similarities at the beginning of the nic-genes on the pAO1 plasmid of Paenarthrobacter nicotinovorans and the pZXY21 plasmid of Arthrobacter sp. ZXY-2 indicating the deletion of Q8GAJ9 and Q8GAJ8 ORFs; Nucleic acid similarities (BLASTN comparisons showing C95% identity) are depicted as solid lines blue area between sequences. (B) Detail of the A0A1D9FFG5 insertion site. Underlined with red is the repetition 50CTCCC(G)TC 30 at the two ends of the pAO1 DNA; boxed withe dashed blue line is the palidrome CGGGA(AC)TCCCG. Arrows denote ORF’s, UniProtKB IDs are indicated with red text for transposases, and blue text for unknown function.

pZXY21 223517 and pZXY21 223524 represent a rep- which ends the DNA identity between the two plasmids etition 50CTCCC(G)TC 30 present at the two ends of the on this 50 side. DNA (underlined in red in figure 2B). The sequence Thus, the identity of the DNA segments of the two TCCCG is part of the palindrome CGGGA(AC)TCCCG plasmids, virtually identical at the nucleotide level, Nicotine catabolic genes in Arthrobacter Page 7 of 12 58 ended each at both ends 30and 50 with a palindromic XerD site specific recombinase were conserved in the sequence. proteins (supplementary figure 3B). The pAO1 ORF Q8GAJ5 and its similar pZXY21 ORF A0A1D9FFQ7 encode small proteins related to 3.3 The predicted integrases of pAO1 and pZXY21 Tn554C and to many Mycobacterial ORFs (e. g. M. gordonae A9W98_15650, 134 aa with 62% identity to The pAO1 nic-genes DNA segment starts at its 50end Q8GAJ5). Characteristic for these proteins is the with several ORFs encoding transposases related to presence of a predicted leucine zipper motif in the those of the S. aureus transposon Tn554. These pAO1 C-terminal half of the proteins (supplementary ORFs Q8GAJ8, Q8GAJ7 and Q8GAJ5 are transla- figure 3C). tionally coupled as indicated by the overlap of their The next ORF on the plasmids was predicted as STOP and START codons. Q8GAJ7 resulted by fusion encoding a HTH transcriptional regulator (Q8GAJ3 of the reading frames Q8GAJ7 and Q8GAJ6 originally and A0A1D9FFT5, respectively) and was followed by annotated on pAO1 (Igloi and Brandsch 2003) but ORFs related to nicotine catabolism. shown to form in fact one reading frame (see further These data revealed the presence of ORFs related to down). the phage integrase family of site-specific recombi- Q8GAJ8 is predicted as belonging to the phage nases at one of the ends of the nic-genes DNA seg- integrase family of site-specific recombinases of the ments of pAO1 and pZXY21. tyrosine family (Cornet et al. 1997). Phage integrase proteins cleave DNA substrates with the formation of a covalent link between the 50 phosphate of one strand 3.4 Invasion by mobile elements of the 69,807 bp and a catalytic tyrosine residue at the carboxyl end of nic-genes DNA fragment on pXYZ21 the integrase (Kwon et al. 1997; Guo et al. 1997). An alignment of the Q8GAJ8, Tn554A, and the protein Transfer of the nic-gene cluster to pZXY21 resulted in sequence of the well characterized phage recombinase the invasion of this DNA by mobile elements of the CRE (RECR_BPP1) revealed that the catalytic site new host. Copies of transposable elements present on residues R-173, H-289, R-292 and Y-324 of the CRE pZXY21 or the host chromosome were inserted at recombinase (P06956) were conserved in the proteins various sites of the acquired DNA segment as shown of the alignment (supplementary figure 3A), supporting schematically in figure 3. Apparently in this host there the assumption that Q8GAJ8 protein is a site specific was a relaxed control of transposition of its mobile recombinase. elements. pAO1 Q8GAJ7 is 93.4% similar to ORF The complementary transcribed ORFs A0A1D9FF38 A0A1D9FFG5 of plasmid pZXY21. This pZXY21 and A0A1D9F9U6 form the IstA transposase and the ORF is the result of the DNA-fusion following the IstB AAA-ATPase transposase helper protein of an deletion from pZXY21 of the DNA corresponding to IS21 family mobile element (PS50994; IPR028350; pAO1 Q8GAJ9 and Q8GAJ8 (figure 2). ORF IPR003593; IPR002611). Related transposase proteins A0A1D9FFG5 and pAO1 ORF Q8GAJ7 show a high are found on pZXY21 outside the nic-gene DNA level of amino acid identity with many predicted My- (ORFs A0A1D9FFY0 and A0A1D9FFR1). The DNA cobacterial integrases (e.g. M. gordonae insertion site of this mobile element is close to the start A9W98_15655, 771 aa of 63% identity to Q8GAJ7). of the ORF of the nicotine dehydrogenase middle sized These deduced proteins of similar size were all pre- subunit gene (ndhM), which forms a transcriptional dicted as phage integrases and XerD-related site- unit with the ORFs of the subunits genes ndhS and specific recombinases (IPR011010 – DNA_brk_joi- ndhL and the ORF of the 6hlno gene (one operon n_enz.; IPR013762 – Integrase-like_cat; IPR002104 – ndhMSL6hlno). The trimeric nicotine dehydrogenase Integrase_catalytic; IPR023109 – Integrase_recombi- enzyme encoded by these genes hydroxylates the pyr- nase_N; IPR004107 – Integrase_SAM-like_N). The idine ring of nicotine in position 6, starting nicotine predicted Q8GAJ7 and A0A1D9FFG5 proteins and the degradation. 6-Hydroxy-L-nicotine oxidase catalysis related Mycobacterial proteins were about 90 aa longer the second step of nicotine degradation, which consists at the N-terminus than Tn554B and about twice the in the oxidation of the pyrrolidine ring with formation size of XerD. Sequence alignments however revealed of 6-hydroxy-pseudo-oxynicotine. The entire operon that the C-terminal half of the proteins were similar and has been cloned and the enzymes synthesized recom- the position of the catalytic residues R, H, R and Y of binantly (Sachelaru et al. 2006). Insertion at this site 58 Page 8 of 12 Roderich Brandsch and Marius Mihasan pAO1 Reg mao modC modB gluc kga modA moaE coxD moaC ponH malR moeA kdhS coxE dhph coxG thph kdhM kdhL coxF moaA nepB pmfR nepA nbor moaD folD mobA ndhS ndhM ndhL mabo purU aatsp aaper hdnoR cupin 6hdno aatsp cupin Q8GAG3 Q8GAG5 Q93NF3 IS1473 ssadh Q93NH3 Q93NH2 faadh 6hlno Q8GAJ6 Q8GAJ7

Score: 34750 Score: 35440 Score: 19800 % id: 100% % id: 100% % id: 99%

pZXY21 A0A1D9FFB1 A0A1D9FFB2 A0A1D9FFE4 A0A1D9FFS5 A0A1D9FFJ0 A0A1D9FFI4 A0A1D9FFY2 A0A1D9FF85 A0A1D9FFF6 A0A1D9FFP8 A0A1D9FF93 A0A1D9FFQ7 A0A1D9FFF3 A0A1D9FFU9 A0A1D9FFQ1 A0A1D9FFH4 A0A1D9FFH2 A0A1D9FFN8 A0A1D9FFM2 A0A1D9FFK3 A0A1D9FFC4 A0A1D9FFT5 A0A1D9FF94 A0A1D9FFE6 A0A1D9FF98 A0A1D9FFF4 A0A1D9FFD6 A0A1D9FF21 A0A1D9FF30 A0A1D9FFD4 A0A1D9FF69 A0A1D9FFI3 A0A1D9FFE3 A0A1D9FF65 A0A1D9FF66 A0A1D9FF88 A0A1D9FFA5 A0A1D9FFD8 A0A1D9FFB4 A0A1D9FFV9 A0A1D9FFS9 A0A1D9FFH7 A0A1D9FF84 A0A1D9FF83 A0A1D9FF75 A0A1D9FFN0 A0A1D9FF74 A0A1D9FF80 A0A1D9FFQ8 A0A1D9FFR9 A0A1D9FFL2

CACGTA CACGTA CGTC CGTC GTAGGG A0A1D9FCQ4 A0A1D9F9U6; A0A1D9FF38 A0A1D9F9U6; A0A1D9FFG5

A0A1D9FCQ4 GTAGGG A0A1D9FDJ6 CGCTTAAGG CATTACT CGCTTAAGG CATTACT

A0A1D9FFB7;A0A1D9FF51 CAGGT CAGGT

Figure 3. Colinearity of the nic-genes and insertion sites of pZXY21 ORFs encoding predicted transposases. Direct DNA- repeats generated at the insertion sites are indicated, as well as the location of the insertion sites. Black arrows indicate genes experimentally shown to be involved in nicotine metabolism; white arrows – genes with no experimental function; red arrows – genes involved in transposition events; blue text – genes that were split by the insertion of a predicted transposases and the corresponding resulting ORFs. The correspondence between the pAO1 and pZXY21 encoded proteins are detailed in table 1. The colored blue blocks between sequences represent nucleic acid similarities [BLASTN comparisons with a minimum score cut-off level of 20 and a significance threshold (e-value) of 10] on different strands. would disrupt transcription of this operon from its Insertion of the DNA element carrying the ORFs of native promoter. the predicted transposases A0A1D9FFB7 and Insertion of this mobile element could have led to the A0A1D9FF51 did split the pAO1 ORF of the hypo- deletion of pAO1 IS1473 (Mene´ndez et al. 1997) from thetical protein Q93NH3 (figure 3), found in many the nic-genes DNA of pZXY21, leaving only the 58 Arthrobacter strains, into the two pZXY21 ORFs N-terminal and 54 C-terminal amino acids of the A0A1D9FFL2 and A0A1D9FF80. The transposase transposase encoded by this element (A0A1D9FFL8). A0A1D9FF51 is related to bacteriophage Mu trans- About 66 nt downstream from the end of the 6HLNO posase and A0A1D9FFB7 to the IS3 family. This small ORF was the insertion site of the DNA carrying the ORF protein contains a helix-turn-helix motif and a C-ter- of the predicted transposase A0A1D9FCQ4 (figure 3). minal leucine zipper and may act as helper protein This transposase is predicted as belonging to the IS204/ during transposition (IPR025948; IPR001584; IS1001/IS1096/IS116 family (IPR002560; IPR032877; IPR012337). There are additional copies of these pre- IPR029261). The transposase A0A1D9FCQ4 is identi- dicted transposases on pZXY21 (A0A1D9FF51 and cal to ORF A0A1D9F401 and there are 12 identical A0A1D9FFB7). ORFs of predicted transposases of this family on The DNA-element encoding the predicted trans- pZXY21 as well as on the additional plasmids pZXY22 posase A0A1D9FCQ4 was inserted at the C-terminus and pZXY23 and on the chromosome of this strain. of the 78 amino acids pAO1 ORF Q93NH2, which Nicotine catabolic genes in Arthrobacter Page 9 of 12 58 corresponds to pZXY21 ORF A0A1D9FF83. This gene of a 6-hydroxy-D-nicotine-specific enzyme, must ORF with an encoded protein of unknown function was have been transferred between the two Arthrobacter present in the genome of many Arthrobacter and other plasmids. Despite some differences in annotation of bacteria species. It is part of a predicted gene cluster ORFs between the pAO1 and the pZXY21 nic-genes implicated in recombination events (Mihasan and segment, at DNA level the nucleotide identity per- Brandsch 2013). Following ORF A0A1D9FCQ4 there sisted, indicating that the nic-genes DNA segment was are 4 ORFs of 41 to 45 aa of hypothetical proteins transferred as a unit. The markedly reduced identity at positioned between the ORF of the transposase and the the amino acid level of the rest of the plasmid ORFs next ORF of the nic-gene cluster - kdhM and the reduced DNA identity level of the rest of the (A0A1D9FFM2). Translation of the DNA of the small plasmid DNA outside the nic-gene DNA (supplemen- open reading frames revealed the ORF of the kdhS tary figure 2) further support this assumption. gene (compl 183863-183360, not annotated on The results of this investigation corroborate well pZXY21) with its start codon ATG overlapping with with the reports on the nic-genes clusters identified in the stop codon TGA of kdhM, suggesting the two the draft genomes of Arthrobacter AK-YN10 (Sagarkar ORFs are translationally coupled, as it is the case on et al. 2016) and M2012083 (Yao et al. 2015) sug- pAO1. gesting that the entire nic-genes DNA fragment carry- The pZXY21 ORF A0A1D9FFV9 is a homolog of ing the upper and lower nicotine degradation genes the pAO1 MobA ORF (Q93NF8) which marks the end represents a large catabolic transposon. This suggestion of the gene cluster encoding the upper pathway of is further supported by the similarity of the proteins nicotine catabolism. This cluster of ORFs is responsi- derived from the pAO1 ORFs Q8GAJ8, Q8GAJ7 and ble for molybdopterin cofactor (MoCo) biosynthesis Q8GAJ5 with the site-specific recombinases and and of gene products required for nicotine dehydro- transposases Tn544A, Tn554B and Tn554C of the S. genase (NdhMSL) and 6-hydroxy-pseudo-oxynicotine aureus transposon 544 and other predicted site-specific dehydrogenase (ketone-dehydrogenase KdhMSL) recombinases. Similar to Tn554 (Murphy and Lo¨fdahl assembly (Sachelaru et al. 2006). Following this MobA 1984), there are no direct or inverted repeats at the ends ORF was identified the insertion site of the predicted of the nic-fragment, but only the mentioned palin- transposase of ORF A0A1D9FDJ6 (figure 3). There are dromic sequences that may represent the attL and attR two more identical copies of this transposase on the recognition sites of the integrases. The pAO1 nic-genes plasmid. They belong to the mutator family of trans- DNA fragment may form a transposition competent posases characterized by a related winged helix-turn- transposon. This seems no longer to be the case with helix DNA binding motif (IPR001207). The ORF of the pZXY21 counterpart nic-fragment since Q8GAJ8, the transposase was followed by several ORFs of the ORF related to Tn544A and essential for transpo- unknown significance. The next functionally charac- sition, was deleted. terized ORF involved in nicotine catabolism was that The pAO1 ORFs Q8GAJ8, Q8GAJ7 and Q8GAJ5 of the transcriptional activator PmfR (A0A1D9FFB2 show overlapping translational start and stop codons, on pZXY21; Q8GAH9 on pAO1) which starts the set an indication that they are translationally coupled. of ORFs encoding the enzymes of the lower pathway Related ORFs encoding two similar site-specific of nicotine catabolism. recombinases of the tyrosine family and a transposase Five insertion sites of four predicted mobile elements with leucine zipper motif are found in many My- were identified on the nic-genes DNA segment of cobacterial strains in the GenBank (M. parafinicum pZXY21. Copies of these mobile elements were pre- BRW65_19530, BRW65_19535, BRW65_19540; M. sent on the chromosome and on additional plasmids of gordonae A9W98_15660, A9W98_15655, the Arthrobacter sp. ZXY2. A9W98_15650 and many more). They may be part of a transposon commonly found in Mycobacterial strains. Regarding the size of the suggested nic-transposon, 4. Discussion there are other large transposons known. Thus, nisin biosynthetic genes and sucrose catabolic genes are The gene order in the nic-genes clusters on pAO1 and encoded on Tn5301 and on Tn5276 of 70 kb in size pZXY21, as well as the ORFs amino acid sequence and (Horn et al. 1991; Rauch and De Vos 1992; Rauch and DNA sequence identity levels (99.1% and 96% to de Vos 1994) which is of the same order of magnitude 100%, respectively) indicated that the entire DNA as the nic-genes DNA. As in the case of these trans- segment carrying the nic-genes cluster, including the posons the genes of the site-specific integrases and 58 Page 10 of 12 Roderich Brandsch and Marius Mihasan transposases are located at one extreme end of the nic- association of these genes in Nocardioides sp. JS614 fragment. However, different from these conjugative with predicted genes of integrases (Ganas et al. 2008; transposons (Scott and Churchward 1995; Burrus et al. Cobzaru et al. 2011). Thus, the Nocardioides sp. JS614 2002) the nic-genes cluster lacks ORFs encoding pro- transposase IstA (AISE85) and IstB (AISE86) are teins involved in conjugal transfer of DNA. homologs of the pZXY21 predicted IstA We suppose that the nic-genes DNA was first trans- (A0A1D9FF38) and IstB (A0A1D9F9U6) (65% and ferred to plasmid pZXY21 and after its integration into 80% identity, respectively). The predicted integrases the plasmid DNA it was invaded by transposable ele- A0A1D9FFG5 of pZXY21 and Q8GAJ7 of pAO1 are ments present in the genome of Arthrobacter sp. homologs of the Nocardioides integrase A1SEA1-3 ZXY2. With exception of the phage related integrase and they are flanking the nic-gene clusters in these A0A1D9FFG5 all transposable elements in the nic- bacteria. gene DNA of pZXY21 were present in additional copies in the genome of the strain. Integration of these transposable elements has func- 5. Conclusions tional consequences regarding expression of the genes of nicotine catabolism. Although no essential enzyme A virtually identical nic-gene cluster, upper and lower gene of the pathway was affected by the insertion of an pathway, was present on the pAO1 plasmid of Pae- IS element, transposition of A0A1D9F9U6 and narthrobacter nicotinocorans and the pZXY21 plasmid A0A1D9FF38 close to the translational start of the of Arthrobacter sp. ZXY-2. The nic-genes DNA seg- ndhMSL operon disrupts the promoter from its genes ments showed the characteristics of large catabolic and thus abolish the initial enzymatic steps of nicotine transposons with ORFs of integrases of the tyrosine degradation. The pAO1 ORFs Q93NH3 and Q93NH2 family of recombinases situated at the 50 end of the are highly conserved in a large number of Arthrobacter DNA fragment. The DNA of the nic-genes cluster on and other soil bacteria. They are located on the chro- pZXY21 was invaded in the host Arthrobacter mosomes associated with ORFs encoding enzymes of ARZXY2 by mobile elements present on pZXY21 side specific recombination (Mihasan and Brandsch outside the nic-genes cluster as well as on the bacterial 2013). The function of these proteins in pAO1 biology chromosome. These findings represent a unique and nicotine catabolism is unknown but insertion of the opportunity to determine the DNA limits of the nic- IS elements would abolish their function. The identical genes DNA fragment. They offer a snapshot view on A0A1D9FCQ4 and A0A1D9FCQ4 transposases the breakdown of a large catabolic gene cluster and the showed inverted orientations. They were positioned generation of putative new transposons. These trans- like the terminal IS elements of a transposon. posons may illustrate how modules of nicotine cata- The features of the pZXY21 nic-genes DNA frag- bolic genes can be mobilized to other soil bacteria by ment may reflect steps in the decay of this large horizontal gene transfer contributing to the generation transposon and the generation of new composite of new assemblies of catabolic pathways. transposons carrying nic-gene modules (Toussaint and Merlin 2002). These transposons may aid in the transfer of individual parts of the nic-genes DNA to Acknowledgements other replicons (related Arthrobacter plasmids, chro- mosomes or plasmids of other soil bacteria) by hori- RB was supported by the Faculty of Medicine of the zontal gene transfer (Koonin et al. 2001; Wiedenbeck University of Freiburg. MM was partially supported by and Cohan 2011; Harrison and Brockhurst 2012). a grant of the Romanian National Authority for Sci- Indeed, the nic-pathway genes of Rhodococcus opacus entific Research and Innovation, CNCS – UEFISCDI B4 and Nocardioides sp. JS614 include only those of [Grant No. PN-III-P1-1.1-TE-2016-0367, 2016]. the upper pathway. The genes of 6hdno, most of the MoCo biosynthesis and the lower pathway genes for the metabolism of N-methyl-gamma-amino-butyrate to succinate are not included (Mihasan and Brandsch References 2013). In Rhodococcus sp. B4 no transposable ele- ments were identified in its nic-gene cluster. However, Adamczyk Z, Kujda M, Nattich-Rak M, Ludwiczak M, the possible distribution of nic-gene modules with the Jagura-Burdzy G and Adamczyk M 2013 Revealing aid of transposable elements was illustrated by the properties of the KfrA plasmid protein via combined Nicotine catabolic genes in Arthrobacter Page 11 of 12 58 DLS, AFM and electrokinetic measurements. Colloids Gurusamy R and Natarajan S 2013 Current status on Surfaces B Biointerfaces 103 635–641 biochemistry and molecular biology of microbial degra- Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ dation of nicotine. Sci. World J. 2013 125385 1990 Basic local alignment search tool. J. Mol. Biol. 215 Harrison E and Brockhurst MA 2012 Plasmid-mediated 403–410 horizontal gene transfer is a coevolutionary process. Brandsch R 2006 Microbiology and biochemistry of nicotine Trends Microbiol. 20 262–267 degradation. Appl. Microbiol. Biotechnol. 69 493–498 Horn N, Swindell S, Dodd H and Gasson M 1991 Nisin Burrus V, Pavlovic G, Decaris B and Gue´don G 2002 biosynthesis genes are encoded by a novel conjugative Conjugative transposons: the tip of the iceberg. Mol. transposon. Mol. Gen. Genet. 228 129–135 Microbiol. 46 601–610 Igloi GL and Brandsch R 2003 Sequence of the 165-kilobase Busse H-J Review of the taxonomy of the genus Arthrobac- catabolic plasmid pAO1 from Arthrobacter nicotinovo- ter, emendation of the genus Arthrobacter sensu lato, rans and identification of a pAO1-dependent nicotine proposal to reclassify selected species of the genus uptake system. J Bacteriol 185 1976–1986 Arthrobacter in the novel genera Glutamicibacter gen. Koonin E V., Makarova KS and Aravind L 2001 Horizontal nov., Paeniglutamicibacter gen. nov., Pseudogluta. 2016 Gene Transfer in Prokaryotes: Quantification and Classi- Int. J. Syst. Evol. Microbiol. 66 9–37 fication. Annu. Rev. Microbiol. 55 709–742 Carver T, Harris SR, Berriman M, Parkhill J and McQuillan Kwon HJ, Tirumalai R, Landy A and Ellenberger T JA 2012 Artemis: an integrated platform for visualization Flexibility in DNA recombination: structure of the lambda and analysis of high-throughput sequence-based experi- integrase catalytic core. 1997 Science 276 126–131 mental data. Bioinformatics 28 464–469 Mene´ndez C, Igloi GL and Brandsch R 1997 IS1473, a Carver TJ, Rutherford KM, Berriman M, Rajandream M-A, putative insertion sequence identified in the plasmid Barrell BG and Parkhill J 2005 ACT: the Artemis pAO1 from Arthrobacter nicotinovorans: isolation, char- Comparison Tool. Bioinformatics 21 3422–3423 acterization and distribution among Arthrobacter species. Clark K, Karsch-Mizrachi I, Lipman DJ, Ostell J and Sayers Plasmid 37 35–41 EW 2016 GenBank. Nucleic Acids Res. 44 D67–D72 Miha˘s¸an M, Babii C, Aslebagh R, Channaveerappa D, Cobzaru C, Ganas P, Mihasan M, Schleberger P and Dupree E and Darie CC 2018 Proteomics based analysis Brandsch R 2011 Homologous gene clusters of nicotine of the nicotine catabolism in Paenarthrobacter nicoti- catabolism, including a new x-amidase for a-ketoglu- novorans pAO1. Sci. Rep. 8 16239 taramate, in species of three genera of Gram-positive Mihasan M and Brandsch R 2013 pAO1 of Arthrobacter bacteria. Res. Microbiol. 162 285–291 nicotinovorans and the spread of catabolic traits by Combet C, Blanchet C, Geourjon C and Dele´age G 2000 horizontal gene transfer in gram-positive soil bacteria. J. NPS@: network protein sequence analysis. Trends Mol. Evol. 77 22–30 Biochem. Sci. 25 147–150 Miha˘s¸an M and Brandsch R 2016 A predicted T4 secretion Cornet F, Hallet B and Sherratt DJ 1997 Xer recombination system and conserved DNA-repeats identified in a subset in Escherichia coli. Site-specific DNA topoisomerase of related Arthrobacter plasmids. Microbiol. Res. 191 activity of the XerC and XerD recombinases. J. Biol. 32–37 Chem. 272 21927–21931 Murphy E and Lo¨fdahl S 1984 Transposition of Tn554 does Cserzo¨ M, Eisenhaber F, Eisenhaber B and Simon I 2002 On not generate a target duplication. Nature 307 292–294 filtering false positive transmembrane protein predictions. Papadopoulos JS and Agarwala R 2007 COBALT: con- Protein Eng. 15 745–752 straint-based alignment tool for multiple protein Darling AE, Mau B and Perna NT 2010 progressiveMauve: sequences. Bioinformatics 23 1073–1079 multiple genome alignment with gene gain, loss and Petersen TN, Brunak S, Heijne G von and Nielsen H 2011 rearrangement. PLoS One 5 e11147 SignalP 4.0: discriminating signal peptides from trans- Finn RD et al. 2017 InterPro in 2017-beyond protein family membrane regions. Nat. Methods 8 785–786 and domain annotations. Nucleic Acids Res. 45 D190– Rauch PJ and Vos WM de 1994 Identification and charac- D199 terization of genes involved in excision of the Lactococ- Fitzpatrick PF 2018 The enzymes of microbial nicotine cus lactis conjugative transposon Tn5276. J. Bacteriol. metabolism. Beilstein J. Org. Chem. 14 2295–2307 176 2165–2171 Ganas P, Sachelaru P, Mihasan M, Igloi GL and Brandsch R Rauch PJ and Vos WM De 1992 Characterization of the 2008 Two closely related pathways of nicotine catabolism novel nisin-sucrose conjugative transposon Tn5276 and in Arthrobacter nicotinovorans and Nocardioides sp. its insertion in Lactococcus lactis. J. Bacteriol. 174 strain JS614. Arch. Microbiol. 189 511–517 1280–1287 Guo F, Gopaul DN and Duyne GD van 1997 Structure of Sachelaru P, Schiltz E and Brandsch R 2006 A functional Cre recombinase complexed with DNA in a site-specific mobA gene for molybdopterin cytosine dinucleotide recombination synapse. Nature 389 40–46 cofactor biosynthesis is required for activity and 58 Page 12 of 12 Roderich Brandsch and Marius Mihasan holoenzyme assembly of the heterotrimeric nicotine Wiedenbeck J and Cohan FM 2011 Origins of bacterial dehydrogenases of Arthrobacter nicotinovorans. Appl. diversity through horizontal genetic transfer and adapta- Environ. Microbiol. 72 5126–5131 tion to new ecological niches. FEMS Microbiol. Rev. 35 Sagarkar S, Bhardwaj P, Storck V, Devers-Lamrani M, 957–976 Martin-Laurent F and Kapley A 2016 s-triazine degrading Wright LD and Grossman AD 2016 Autonomous Replica- bacterial isolate Arthrobacter sp. AK-YN10, a candidate tion of the Conjugative Transposon Tn916. J. Bacteriol. for bioaugmentation of atrazine contaminated soil. Appl. 198 3355–3366 Microbiol. Biotechnol. 100 903–913 Yamada KD, Tomii K and Katoh K 2016 Application of the Scott JR and Churchward GG 1995 Conjugative Transpo- MAFFT sequence alignment program to large data— sition. Annu. Rev. Microbiol. 49 367–397 reexamination of the usefulness of chained guide trees. Toussaint A and Merlin C 2002 Mobile Elements as a Bioinformatics 32 3246–3251 Combination of Functional Modules. Plasmid 47 26–35 Yao Y, Tang H, Su F and Xu P 2015 Comparative genome Vaitekunas J, Gasparaviciute R, Rutkiene R, Tauraite D and analysis reveals the molecular basis of nicotine degrada- Meskys R 2016 A 2-Hydroxypyridine Catabolism Path- tion and survival capacities of Arthrobacter. Sci. Rep. 5 way in Rhodococcus rhodochrous Strain PY11. Appl. 8642 Environ. Microbiol. 82 1264–1273

Corresponding editor: SUDHA BHATTACHARYA