bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Title:

2 Catabolism of 3-hydroxypyridine by Ensifer adhaerens HP1: a novel four-component

3 gene encoding 3-hydroxypyridine dehydrogenase HpdA catalyzes the first step of

4 biodegradation

5

6 Running title:

7 Microbial 3-hydroxypyridine degradation

8

9 Authors:

10 Haixia Wang a, Xiaoyu Wang a, Hao Ren a, Xuejun Wang a, Zhenmei Lu a#

11

12 a MOE Laboratory of Biosystem Homeostasis and Protection, College of Life Sciences,

13 Zhejiang University, Hangzhou, China

14

15 #Address correspondence to Zhenmei Lu, [email protected]

16

17

18

19

20

21

22

23

1 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

24 Abstract

25 3-Hydroxypyridine (3HP) is an important natural pyridine derivative. Ensifer

26 adhaerens HP1 can utilize 3HP as the sole source of carbon, nitrogen and energy to

27 grow. However, the genes responsible for the degradation of 3HP remain unknown. In

28 this study, we predicted that a gene cluster, designated 3hpd, may be responsible for the

29 degradation of 3HP. The initial hydroxylation of 3HP is catalyzed by a four-component

30 dehydrogenase (HpdA1A2A3A4), leading to the formation of 2,5-dihydroxypyridine

31 (2,5-DHP) in E. adhaerens HP1. In addition, the SRPBCC component in HpdA existed

32 as a separate subunit, which is different from other SRPBCC-containing

33 molybdohydroxylases acting on N-heterocyclic aromatic compounds. Our findings

34 provide a better understanding of the microbial degradation of pyridine derivatives in

35 nature. Additionally, research on the origin of the discovered four-component

36 dehydrogenase with a separate SRPBCC domain may be of great significance.

37

38 Importance

39 3-Hydroxypyridine is an important building block for synthesizing drugs, herbicides

40 and antibiotics. Although the microbial degradation of 3-hydroxypyridine has been

41 studied for many years, the molecular mechanisms remain unclear. Here, we show that

42 3hpd is responsible for the catabolism of 3-hydroxypyridine. The 3hpd gene cluster was

43 found to be widespread in Actinobacteria, Rubrobacteria, Thermoleophilia, and Alpha-,

44 Beta-, and Gammaproteobacteria, and the genetic organization of the 3hpd gene

45 clusters in these bacteria showed high diversity. Our findings provide new insight into

2 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

46 the catabolism of 3-hydroxypyridine in bacteria.

47

48 Keywords

49 3-hydroxypyridine, Ensifer adhaerens HP1, 3-hydroxypyridine catabolism, 3-

50 hydroxypyridine dehydrogenase, 3hpd

51

52 Introduction

53 The pyridine ring is a major constituent of natural compounds such as plant alkaloids,

54 coenzymes and antibiotics. 3-Hydroxypyridine (3HP), a useful and valuable pyridine

55 derivative, is a monohydroxypyridine in which the hydrogen at position 3 of the

56 pyridine has been replaced by a hydroxyl group. 3HP has been detected as a thermal

57 degradation product in the smoke from burning Salvia divinorum leaves and as a

58 significant constituent of tobacco smoke[1, 2]. Many bioactive compounds contain 3HP

59 as an important structural unit, and 3HP is widely used as a building block to synthesize

60 drugs, herbicides, insecticides and antibiotics[3-5]. Large amounts of 3HP are

61 synthesized each year, and there are many synthetic methods for 3HP, such as

62 ruthenium-catalyzed ring-closing olefin metathesis[6]. The widespread use of 3HP has

63 made its release to the environment inevitable, which may have serious implications

64 for human health. While 3HP can be eliminated by physical and chemical methods,

65 microbial biodegradation has been considered one of the most economical and effective

66 approaches to remediating 3HP pollution.

67 Catabolism of pyridine, particularly the initial steps of the hydroxylation of

3 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

68 monohydroxylated pyridines such as 2-hydroxypyridine (2HP), 3HP and 4-

69 hydroxypyridine (4HP), has received much attention. In particular, numerous bacteria

70 have been reported to use 2HP as the sole carbon and energy source to grow. Many

71 intermediates have been identified, and metabolic pathways have been proposed for

72 2HP biodegradation. One pathway of 2HP biodegradation involves the formation of

73 2,5-dihydroxypyridine (2,5-DHP), which then proceeds through the maleamate

74 pathway. The other pathway, involving the formation of 2,3,6-trixydroxypyridine first,

75 produces a blue pigment (nicotine blue) in the medium. Rhodococcus rhodochrous

76 PY11 was reported to use 2HP as the sole source of carbon and energy through the

77 nicotine blue-production pathway. A gene cluster (hpo) has been characterized as being

78 responsible for the catabolism of 2HP in strain PY11, and the initial hydroxylation of

79 2HP is catalyzed by a four-component dioxygenase (HpoBCDF)[7]. Burkholderia sp.

80 MAK1 was also reported to degrade 2HP, but through the maleamate pathway. A gene

81 cluster (hpd) was responsible for the degradation of 2HP in strain MAK1, and the 2-

82 hydroxypyridine 5-monooxygenase is a soluble di-iron monooxygenase (SDIMO)

83 encoded by a five-gene cluster hpdABCED[8]. Several strains have been reported to

84 partially or completely degrade 3HP. Achromobacter sp. (G2 and 2L)[9], Pusillimonas

85 sp. 5HP[10] and Agrobacterium sp. DW-1[11] could use 3HP as a carbon and nitrogen

86 source to grow. Nocardia Z1 was reported to slowly oxidize 3HP to pyridine-2,3-diol

87 and pyridine-3,4-diol[12], while Achromobacter 7N could convert 3HP to only 2,5-DHP

88 and could hardly further metabolize the diol[9].

89 Moreover, microbial degradation of 3HP, which has been proposed for several different

4 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

90 strains, was thought to proceed via the maleamate pathway[9, 12-14],

91 3HP →2,5-DHP → formate + maleamate → NH3 + maleate ↔ fumarate,

92 and was recently confirmed in strain DW-1[11]. However, the genes and enzymes

93 responsible for 3HP biodegradation have seldom been reported. Pyridine-2,5-diol

94 dioxygenase has been partially purified and characterized in only strains G2 and 2L[9].

95 The pyridine-2,5-diol dioxygenase in these two strains required Fe2+ to restore full

96 activity after purification, and the hydroxylases of strains G2 and 2L showed clear

97 specificity because they produced only the para-substituted 2,5-DHP from 3HP. No

98 enzymes responsible for the initial hydroxylation step of 3HP leading to the formation

99 of 2,5-DHP have been reported to date.

100 In this study, the bacterial strain Ensifer adhaerens HP1 was isolated from soil and

101 showed effective degradation and utilization of 3HP. We report the isolation and

102 characterization of the 3HP catabolic pathway in E. adhaerens HP1. A gene cluster

103 (3hpd) encoding the putative proteins required for 3HP biodegradation in this bacterium

104 was discovered and characterized. The results of bioinformatics analysis, gene

105 knockout and complementation of hpdA, and heterologous expression of HpdA suggest

106 that multicomponent HpdA is involved in the transformation of 3HP to 2,5-DHP.

107

108 Results and discussion

109 Isolation of a bacterium capable of degrading 3HP

110 A bacterium was isolated from enrichments of soil with 3HP and was selected for

111 detailed study. This strain could utilize 3HP as the sole source of carbon, nitrogen and

5 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

112 energy to grow (Figure 1) and was designated HP1. The 16S rRNA sequence of strain

113 HP1 exhibited the highest similarity (99%) to E. adhaerens strain Casida A (accession

114 number: CP015882.1). Therefore, strain HP1 was identified as E. adhaerens HP1.

115 During the growth of strain HP1 on 3HP-containing media, a green pigment

116 accumulated and gradually turned dark brown. This phenomenon has been reported in

117 strains Pusillimonas sp. 5HP[10], P. putida S16[15], and Agrobacterium sp. S33[16]. All

118 these strains converted the corresponding substrate to 2,5-DHP, and this color change

119 indicated the formation of 2,5-DHP. 2,5-DHP was detected in the supernatant of strain

120 HP1 by LC-MS analysis (Figure S1), suggesting that the degradation of 3HP in strain

121 HP1 occurs through the maleamate pathway (Figure 2B). Moreover, strain HP1 utilized

122 only 3HP but not the other two hydroxypyridine isomers, 2HP and 4HP (data not

123 shown).

124

125 A putative 3HP metabolism gene cluster is present in the genome of strain HP1

126 After performing a BLAST analysis against the genome of strain HP1 using previously

127 known metabolic genes involved in the maleamate pathway, we found a putative

128 metabolic gene cluster, hpdDBCE, in the genome of strain HP1 (Figure 2A). Analysis

129 of the pyridine ring hydroxylase showed that there were four sets of putative genes that

130 may be responsible for the initial metabolism of 3HP. Among them, a four-component

131 gene, hpdA1A2A3A4, was contiguous with the putative maleamate pathway metabolic

132 gene cluster. We predicted that the putative maleamate pathway metabolic gene and the

133 four-component gene were the most likely components of the 3HP-degrading operon.

6 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

134 The cluster consists of 24 genes, hpdA1, hpdA2, hpdA3, hpdA4, hpdB, hpdC, hpdD,

135 hpdE, and orf1-orf16 (Figure 2A), and we designated this cluster 3hpd. The functional

136 annotations of hypothetical Hpd proteins are listed in Table 1. orf1 and orf2 encode

137 transposases. The amino acid sequences of orf3-orf5 showed similarity to the bacterial

138 ABC transporter. Gene orf7 in 3hpd encodes a potential GntR family DNA-binding

139 transcriptional regulator. Moreover, it showed that the transcription of orf3-orf5 was

140 induced by 3HP (Figure 3). This allowed us to presume that orf3-orf5 may be involved

141 in the transport of 3HP, but this requires further research. BLAST analysis based on

142 3hpd showed that hpdA had few homologous genes, which showed relatively low

143 nucleotide sequence similarities (<75%), except for one uncharacterized gene cluster in

144 Rhizobiales bacterium isolate AFS066724 (sequence accession number:

145 UCDB01000016.1), with nucleotide sequence similarity >95.92%. It was reported that

146 bacterial species display a wide range of variation in their total G+C content, and the

147 genes in a certain species’ genome share similar nucleotide sequence patterns[17, 18]. The

148 G+C content of the whole genome of strain HP1 (7.18 Mb) was 62.07%, whereas the

149 G+C content of the plasmid (334.2 kb) containing the 3hpd gene cluster and the 3hpd

150 gene cluster alone (calculated for genes between hpdD and hpdA4, 13.2 kb) were 60.40%

151 and 60.30%, respectively. We observed that the G+C contents of the plasmid and the

152 3hpd gene cluster were obviously lower than that of the whole genome. In addition,

153 considering the mobile element remnants flanking the 3hpd gene cluster (Figure 2A),

154 which is a direct indication of a foreign origin of the sequence flanked by mobile

155 elements, the 3hpd gene cluster was probably introduced through horizontal gene

7 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

156 transfer (HGT) to the genome of strain HP1.

157

158 Transcription of putative 3HP-degrading genes in 3hpd is induced by 3HP

159 To elucidate the correlation between 3HP degradation and putative 3HP-degrading

160 genes in 3hpd and other candidate 3HP dehydrogenase genes, the expression levels of

161 putative target genes involved in the 3HP degradation of strain HP1 were estimated

162 using reverse transcription quantitative PCR (RT-qPCR) and the 2∆∆CT method with or

163 without 3HP supplementation in minimal salt medium (MSM). As shown in Figure 3,

164 the expression levels of hpdA1, hpdA2, hpdA3, hpdA4, hpdB, hpdC, hpdD, and hpdE in

165 3hpd in experimental group were more than 10-times higher than the levels of the

166 corresponding genes in the control group, suggesting that the expression of these genes

167 was induced by 3HP or other 3HP degradation intermediates, while other candidate 3HP

168 dehydrogenase genes exhibited opposite results or even no transcription in the tested

169 growth conditions (Figure S2).

170

171 The hpdA genes encode the 3HP dehydrogenase

172 The first step in the metabolism of 3HP is the hydroxylation of C6 on the pyridine ring

173 to generate 2,5-DHP. The hydroxylation of N-heterocyclic aromatic compounds

174 consisting of a pyridine ring, such as nicotine to 6-hydroxynicotine and nicotinic acid

175 to 6-hydroxynicotinic acid is usually catalyzed by multicomponent molybdopterin-

176 containing dehydrogenases[19, 20]. A BLAST homology search of genes in 3hpd against

177 database sequences revealed that the genes hpdA1, hpdA2, hpdA3, and hpdA4 shared

8 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

178 the highest amino acid sequence identity (33%-52%) with the respective subunits of the

179 three-component molybdopterin-containing dehydrogenases, such as nicotine

180 dehydrogenase (NdhLSM)[19] and quinoline 2-oxidoreductase (QorLSM)[21]. hpdA1

181 was predicted to contain binding domains for the molybdopterin cytosine dinucleotide

182 (MCD), hpdA4 contained a flavin adenine dinucleotide (FAD) binding domain and

183 specific FAD-binding site of members of the xanthine dehydrogenase/oxidase family

184 (CODH), hpdA2 contained two predicted [Fe-S] clusters, and hpdA3 contained an

185 SRPBCC ligand-binding domain (Figure 4). Therefore, hpdA1, hpdA2, hpdA3, and

186 hpdA4 were predicted to encode a four-component molybdopterin-containing

187 dehydrogenase that catalyzes the initial step in the 3HP metabolic pathway.

188 To investigate the role of hdpA in the initial oxidation of 3HP, hpdA1A2 was

189 disrupted. Then, the ability of the mutant strain HP1∆hpdA to metabolize 3HP was

190 analyzed. The disruption of hpdA1A2 genes abolished the growth of the mutant

191 HP1∆hpdA in 3HP-containing medium (Figure 1) but did not affect its growth in 2,5-

192 DHP-containing medium (data not shown). After trans complementation of hdpA1A2

193 to HP1∆hpdA, the ability to utilize 3HP to grow was restored. The complementation of

194 hpdA1 alone could not recover the degradation ability (data not shown). To confirm the

195 role of hpdA genes, they were cloned and expressed in plasmid pRK415-hpdA-3. When

196 plasmid pRK415-hpdA-3 was transferred to E. adhaerens ZM04, which is not able to

197 utilize 3HP, the culture media turned green and then gradually dark brown, suggesting

198 the formation of 2,5-DHP. Results showed that the recombinant strain ZM04-hpdA-3

199 acquired the ability to efficiently convert 3HP into a stoichiometric amount of 2,5-DHP

9 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

200 (Figure 5). UV spectral analysis showed that 3HP decreased over time (absorbance

201 peaks at 244, 275, and 311 nm, pH 7.5), with the formation of a 2,5-DHP peak

202 (absorbance peaks at 228 and 321 nm, pH 7.5) (Figure 6). LC-MS analysis also

203 confirmed that the product was 2,5-DHP [molecular ion peak (M+H)+ 112.3680]

204 (Figure S3). The recombinant ZM04-hpdA could convert 1 mM 3HP into equivalent

205 amounts of 2,5-DHP (Figure 5), suggesting that 3HP was completely transformed into

206 2,5-DHP and that 2,5-DHP was the sole product of 3HP.

207 When truncated HpdA lacking hpdA3 or hpdA4 was constructed and expressed

208 from plasmids pRK415-hpdA-4 and pRK415-hpdA-5, respectively, these two truncated

209 enzymes did not show 3HP dehydrogenase activity, and no 2,5-DHP or color change in

210 the culture media was detected during incubation (data not shown). These results

211 showed that all of the components encoded by hpdA1, hpdA2, hpdA3 and hpdA4 are

212 essential for the function of HpdA. In summary, our data demonstrate that hpdA encodes

213 the dehydrogenase that catalyzes the oxidation of 3HP.

214

215 Prediction and analysis of transmembrane domain and signal peptide in hpdA3

216 HpdA3 contained a SRPBCC ligand-binding domain. It was reported that SRPBCC

217 domains have a deep hydrophobic ligand-binding pocket that can bind diverse ligands

218 and spans all three kingdoms of life, but its function remains unclear. Moreover,

219 SRPBCC-containing genes play a wide range of roles, including strawberry fruit

220 ripening[22], regulation of anthocyanin accumulation[23], encoding type II secretion

221 chaperones[24], functioning in the stabilization of mRNA[25], catalyzing the free radical

10 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

222 reactions of polyunsaturated fatty acids[26], and -pyridone ring formation[27].

223 Hydroxylases that hydroxylate the C6 of pyridine, such as 3-succinoylpyridine

224 dehydrogenase (SpmABC)[28] and picolinic acid dehydrogenase (PicA)[29], include

225 SRPBCC-containing components and are fused with [2Fe-2S] clusters, whereas most

226 pyridine hydroxylases, such as NdhLSM and KdhLMS, have no SRPBCC domain

227 (Figure 4A). In the case presented here, the SRPBCC domain (encoded by hpdA3) was

228 not fused to the HpdA2 protein (containing the [2Fe-2S] cluster) but appeared in a

229 separate protein, with an adjacent [2Fe-2S]-binding protein and FAD-binding subunit

230 (Figures 2A and 4A). In addition, HpdA1, HpdA2 and HpdA4 showed relatively low

231 evolutionary relationships with other molybdenum-containing hydroxylases (Figure

232 4B). The origin of this interesting and unique feature observed in HpdA is mysterious,

233 and the function of this SRPBCC domain in HpdA deserves further study.

234 Analysis of the amino acid sequence of the hpdA3 gene, encoding a carbon

235 monoxide dehydrogenase, from strain HP1 using TMHMM, HMMTOP, TMpred and

236 SignalP-5.0 predicted the presence of a TAT signal peptide (Sp1) (amino acids [aa] 138-

237 176; likelihood, 0.913), followed by a transmembrane domain (Tmd) (aa 216-233). In

238 addition, a different Sec signal peptide (Sp2) was also predicted by SignalP-5.0 (aa 155-

239 176; likelihood, 0.8819). The predicted SRPBCC conserved domain of HpdA3 was

240 located at aa 14-159. The 22 amino acid residues at the C-terminus of the SRPBCC

241 domain were included in the Sp1 sequence. A previous study showed that some of the

242 bacterial hydroxylases, such as nicotine dehydrogenase NDH in A. nicotinovorans

243 pAO1[19], CO dehydrogenase in P. carboxydovorans strain OM5[30], and carbon

11 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

244 monoxide dehydrogenase (CODH) in the eubacterium Oligotropha carboxidovorans[31],

245 were membrane associated. Bioinformatics analysis of the full-length HpdA3 predicted

246 the presence of a signal peptide and transmembrane segment, which led us to presume

247 that the protein encoded by hpdA3 may function in the localization of HpdA. To confirm

248 the roles of HpdA3 in the localization of HpdA, the full-length HpdA3 was fused to the

249 N terminus of green fluorescent protein (GFP) and expressed in Escherichia coli.

250 Confocal microscopy analysis of E. coli cells expressing GFP showed that the green

251 fluorescent signal was uniformly dispersed in the cytoplasm, whereas the cells

252 expressing HpdA3-GFP showed distinct uneven green fluorescence throughout the cell

253 (Figure 7). Although the results for infusion proteins did not show typical membrane-

254 associated localization, HpdA3 had a significant impact on the localization of infusion

255 proteins. Based on these observations, we speculate that HpdA3 is probably essential

256 for the localization of HpdA. SRPBCC-containing PicA3 and SpmC did not contain a

257 signal peptide or transmembrane sequence, whereas HpdA3 contained both a signal

258 peptide and Tmd domain. Therefore, in addition to the separate SRPBCC domain in

259 HpdA, the signal peptide and the transmembrane domain in HpdA3 also distinguish

260 HpdA from SpmABC and PicA.

261

262 A phosphoenolpyruvate (PEP)-utilizing protein mobile subunit and pyruvate-

263 phosphate dikinase are essential for 3HP conversion

264 To confirm the role of HpdA genes, we constructed several heterologous expression

265 plasmids, including pRK415-hpdA-1, pRK415-hpdA-2, pRK415-hpdA-3, and

12 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

266 pRK415-hpdA-6 (Table 2). Plasmid pRK415-hpdA-6 contains the hpdA1A2A3A4

267 cluster, but when this plasmid was transferred to strain ZM04, the recombinant strain

268 ZM04-hpdA-6 did not acquire the ability to convert 3HP. To determine whether the

269 genes contiguous with hpdA are also required for the function of HpdA, we constructed

270 plasmid pRK415-hpdA-1, which contains hpdA1A2A3A4, as well as orf7 and orf8 at

271 the 3-terminus and orf9 and orf10 at the 5-terminus of hpdA. As expected, when the

272 plasmid was transferred to strain ZM04, the recombinant strain ZM04-hpdA-1 obtained

273 the ability to convert 3HP. To further confirm whether orf7-orf8 or orf9-orf10 are

274 essential for HpdA function, we constructed plasmids pRK415-hpdA-2 and pRK415-

275 hpdA-3, with the former containing hpdA1A2A3A4 and orf7-orf8 and the latter

276 containing hpdA1A2A3A4 and orf9-orf10. Surprisingly, the latter imparted the

277 recombinant strain ZM04-hpdA-3 with the ability to transform 3HP as aforementioned,

278 but the former did not. As described in Table 1, orf10 encodes the PEP-utilizing protein

279 mobile subunit, and orf9 encodes pyruvate-phosphate dikinase. To further confirm the

280 function of orf9 and orf10 in vivo, orf9 and orf10 were disrupted. The two mutants lost

281 the ability to grow with 3HP as expected (data not shown), which showed that orf9 and

282 orf10 are essential for HpdA function. What role do they play in the conversion of 3HP?

283 In a previous study, CreH was found to encode a PEP-utilizing enzyme, and CreI

284 encodes pyruvate-phosphate dikinase. The complex CreHI is a novel 4-methylbenzyl

285 phosphate synthase that is responsible for the phosphorylation of the hydroxyl group of

286 4-cresol[32]. Another similar phosphorylation reaction is catalyzed by the

287 phenylphosphate synthase reported in anaerobic phenol metabolism in Thauera

13 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

288 aromatica, and it reported that the molecular and catalytic features of phenylphosphate

289 synthase resemble those of phosphoenolpyruvate synthase[33]. According to these two

290 previous studies, we speculated that orf9 and orf10 may be responsible for the

291 conversion of 3HP to an intermediate compound, but during the experiment, we did not

292 detect any product other than 2,5-DHP. The functions of orf9 and orf10 remain to be

293 clarified.

294

295 Distribution and diversity of 3hpd genes in other bacteria

296 Eleven genome sequences of E. adhaerens strains (Casida A, OV14, RP12G, L18, ST2,

297 WJB133 25_10, M78, AG1206, YX1, SD006, and X097) were available in NCBI,

298 among which the genome sequence of strain Casida A was complete. The 3hpd gene

299 cluster was not identified in these genomes or other Ensifer species. Orthologous 3hpd

300 gene clusters were found in Actinobacteria, Rubrobacteria, Thermoleophilia, and

301 Alpha-, Beta-, and Gammaproteobacteria (21 genera and 38 strains) (Tables S1 and S2).

302 Most of the strains belong to the order Burkholderiales of the class Betaproteobacteria

303 and the orders Corynebacteriales, Geodermatophilales, and Pseudonocardiales of the

304 class Actinobacteria, including the following genera: Pusillimonas, Acidovorax,

305 Comamonas, Variovorax, Mycobacterium, Mycolicibacterium, Blastococcus,

306 Geodermatophilus, Amycolatopsis, and Pseudonocardia. Interestingly, some of the

307 strains are degraders (e.g., Acidovorax sp. KKS102 is a polychlorinated-biphenyl-

308 degrading strain[34], Pusillimonas noertemannii BS8 can cooperate with another strain

309 to degrade poly-γ-d-glutamic acid[35], and Mycobacterium sp. MS1601 can oxidize

14 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

310 branched polyols[36]), whereas some of the strains are unique (e.g., Geodermatophilus

311 sabuli strain DSM 46844 is a -radiation-resistant actinobacterium[37], Amycolatopsis

312 acidiphila strain JCM 30562 is a member of the genus Amycolatopsis inhabiting acidic

313 environments[38], and Salinicola peritrichatus strain JCM 18795 was isolated from

314 deep-sea sediment[39]).

315 The genetic organization of the 3hpd gene clusters in these bacteria was highly

316 diverse (Figure 8). The hpdA1A2A3A4 genes were generally contiguous, except for

317 orthologous genes in the class Actinobacteria and two orthologous genes in the class

318 Alphaproteobacteria. In most genera, orf9 and orf10 were also contiguous and adjoined

319 hpdA4, except in Gaiella occulta strain F2-233 and Solirubrobacterales bacterium 70-

320 9 SCNpilot. Interestingly, the 3hpd genes in the class Alphaproteobacteria, class

321 Gammaproteobacteria, class Actinobacteria and family Alcaligenaceae (belonging to

322 class Betaproteobacteria) were diverse, while the 3hpd genes in the family

323 Comamonadaceae (belonging to the class Betaproteobacteria) were highly conserved.

324

325 Conclusions

326 3HP, generated through tobacco burning or industrially synthesized, is released into the

327 environment and is potentially toxic or otherwise unacceptable. The degradation and

328 utilization of 3HP by microbes have been studied for decades. This study isolated a

329 3HP-degrading bacterium, E. adhaerens HP1, that can utilize 3HP as the sole source of

330 carbon and nitrogen to grow via the maleamate pathway. We also revealed that the 3hpd

331 gene cluster may be responsible for the degradation of 3HP in strain HP1 and that a

15 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

332 four-component dehydrogenase, hpdA, catalyzes the initial hydroxylation of 3HP to

333 2,5-DHP. The genetic pathway of 3HP to 2,5-DHP is reported here for the first time. In

334 addition, the separate SRPBCC component with predicted Sp and Tmd domains

335 suggested the unusual evolution of HpdA, which needs further study. Moreover, the

336 participation of a PEP-utilizing protein and pyruvate-phosphate dikinase in the

337 conversion of 3HP by HpdA makes the catalytic mechanism of HpdA unique but

338 requires further research. In summary, this work shows that the four-component HpdA

339 protein constitutes a previously uncharacterized and unique dehydrogenase.

340

341 Materials and methods

342 Chemicals and reagents

343 3HP (>98%) and 2,5-DHP were obtained from Aladdin (Shanghai, China). TransStart®

344 FastPfu DNA Polymerase and 2×T5 Super PCR mix for fragment amplification were

345 purchased from TransGen Biotech (Beijing, China) and Beijing Tsingke Biotech Co.,

346 Ltd. (Beijing, China). Restriction enzymes used for plasmid construction were

347 purchased from Takara Biotechnology Co., Ltd. (Dalian, China). Antibiotics, 2,6-

348 diaminopimelic acid (2,6-DAP), -D-1-thiogalactopyranoside (IPTG), and other

349 reagents were purchased from Shanghai Sangon Biotech Co., Ltd. (Shanghai, China).

350 All reagents and solvents were of analytical or chromatographic grade. A plasmid

351 extraction kit, gel extraction kit and DNA purification kit were obtained from Omega

352 Bio-tek, Inc. (Norcross, GA, USA). Bacterial genomic DNA was extracted using the

353 TIANamp Bacterial DNA Kit from Tiangen Biotech Co., Ltd. (Beijing, China).

16 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

354

355 Bacterial strains, plasmids, media, and growth conditions

356 Microorganisms with the ability to utilize pyridine as the sole source of carbon, nitrogen

357 and energy were obtained by selective culture in shaken flasks, as described by C.

358 Houghton and R. B. Cain[12]. Soil (1 g) was added to 250 mL Erlenmeyer flasks

359 containing 100 mL of MSM containing the following (per liter): K2HPO4, 1 g; KCl,

360 0.25 g; MgSO4•7H2O, 0.25 g (sterilized separately and added aseptically to the cooled

361 medium before use); and trace-element solution, 1 mL. The pH was adjusted to 7.5

362 using HCl (1 M). The trace-element solution contained the following (per liter):

363 FeSO4•7H2O, 40 mg; MnSO4•H2O, 40 mg; ZnSO4•7H2O, 20 mg; CuSO4•5H2O, 5 mg;

364 CoCl2•6H2O, 4 mg; Na2MoO4•2H2O, 5 mg; CaCl2•2H2O, 0.5 mg; and NaCl, 1 g. The

365 16S rRNA gene of the isolated 3HP-degrading strain was amplified with the universal

366 primer set 27F and 1492R. Strain HP1 (collection number, CGMCC 1.13748) and its

367 plasmid transformants were grown aerobically in MSM supplemented aseptically with

368 0.1% (wt/vol) 3HP in 250 mL Erlenmeyer flasks at 30 ℃ and 200 rpm. E. coli strains

369 DH5 and WM3064 used in this study were grown in LB medium at 37 ℃ and 200

370 rpm. When necessary, erythromycin, bleomycin and tetracycline were used at final

371 concentrations of 100, 25 and 10 μg/mL, respectively. 2,6-DAP was used at a final

372 concentration of 0.3 mM for E. coli WM3064 and its transformants. The strains and

373 plasmids used in this study are summarized in Table 2, and the primers are listed in

374 Table 3.

375

17 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

376 Genome sequence analysis and prediction of 3HP metabolism-related genes in

377 strain HP1

378 To identify putative genes involved in 3HP degradation in strain HP1, we conducted a

379 BLAST analysis against the genome sequence of strain HP1 using known metabolic

380 genes involved in the maleamate pathway, such as ndpHFEG in Sphingomonas melonis

381 TY[40] and vppHFEG in Ochrobactrum rhizosphaerae SJY1[41]. We also used some of

382 the hydroxylases acting on the pyridine ring, such as hpdABCED[8], ndhLSM[19],

383 kdhLMS[42], spmABC[28], and nicAB[20], to predict the unknown dehydrogenase in initial

384 metabolism of 3HP.

385

386 RT-qPCR analysis

387 RT-qPCR analyses of strain HP1 cultures grown under different conditions for select

388 genes were performed using specific primers (Table 3). For the experimental group,

389 3HP was used as the sole carbon and nitrogen source in MSM, while for the control

390 group, glucose and ammonium sulfate were used as the carbon and nitrogen source in

391 MSM, respectively. Total RNA was extracted from mid-exponential cultures of strain

392 HP1 (both the control and experimental groups) using the RNAprep Pure Bacteria Kit

393 (Tiangen Biotech, Beijing, China) according to the manufacturer’s protocol. All RNA

394 samples were stored at -80 °C until further processing. To obtain the cDNA and remove

395 potential traces of genomic DNA, RNA was reverse-transcribed into cDNA using

396 random hexamer primers and the PrimeScript RT Reagent Kit with gDNA Eraser

397 (Perfect Real Time; Takara, Dalian, China). For the target genes, qPCR experiments

18 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

398 were performed in 72-well PCR plates using SYBR® Premix Ex TaqTM II-Tli RNaseH

399 Plus (Takara, Dalian, China) in a 20-μL final reaction volume. PCR mixtures were

400 carried out in molecular biology-grade water in the presence of 1×SYBR® Premix, 0.2

401 μM of the forward and reverse primers, and one to ten nanograms of cDNA template.

402 Amplifications were performed using a Rotor-Gene Q real-time PCR detection system

403 (Qiagen, Germany) with the following cycling conditions: 95 °C for 30 s, followed by

404 [95 °C for 5 s; 60 °C for 20 s; 72 °C for 12 s] × 40 cycles and a melting curve analysis.

405 16S rRNA was selected as the reference gene, and the primers used were 338F/518R

406 (Table 3). Untranscribed RNA was used as a control to ensure that no remaining

407 genomic DNA could be detected. Reaction mixtures with water as the template were

408 set as the negative control. Melting curve analysis with temperatures ranging from 72 ℃

409 to 95 ℃ at the end of the amplification cycles and agarose gel analyses of the

410 amplification products were used to confirm the specificity of the qPCR products. All

411 experiments were performed with four biological replicates, and each sample had four

412 technical replicates. In addition, the mean estimated PCR efficiency for each amplicon

413 group was calculated by the LinRegPCR program (version 2013.0) and used instead of

414 the theoretical efficiency to increase the accuracy of the results[43-45].

415

416 Gene disruption and complementation

417 pEX18Tc-hpdA was constructed for gene knockout by fusing the erythromycin

418 resistance gene and two upstream and downstream fragments of the target gene

419 amplified with the primers shown in Table 3 to Sac I/Hind III-digested pEX18Tc with

19 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

420 the In-Fusion® HD Cloning Kit (TaKaRa, Dalian, China). The resulting plasmid

421 pEX18Tc-hpdA was transformed into E. coli WM3064 (2,6-DAP auxotroph)-

422 competent cells before conjugation with strain HP1 as described previously[46]. The

423 double-crossover recombinants of HP1∆hpdA were screened on LB plates containing

424 10% sucrose (w/v) and erythromycin. The obtained mutant strain HP1∆hpdA was

425 verified by PCR amplification and sequencing analysis (data not shown) using specific

426 primers. Sequencing analysis was performed by YKang Biological Technology Co., Ltd.

427 (Hangzhou, China). orf9 and orf10 were disrupted with the same procedure as hpdA.

428 pBBRBleo was constructed by replacing the kanamycin resistance gene in

429 pBBR1MCS-2 with the bleomycin resistance gene. For gene complementation,

430 pBBRBleo-hpdA was constructed by fusing the PCR products of hpdA amplified with

431 the primers shown in Table 3 to Kpn I/Xba I-digested pBBRBleo. pBBRBleo-hpdA was

432 used to transform E. coli WM3064 and then mated into the mutant strain HP1∆hpdA

433 by conjugation to obtain the complementary strain HP1∆hpdA(pBBRBleo-hpdA).

434 Moreover, pBBRBleo was also introduced into the mutant strain HP1∆hpdA through

435 conjugation to obtain the control strain HP1∆hpdA(pBBRBleo).

436

437 Growth and 3HP degradation analysis

438 The growth of and 3HP degradation by the wild-type strain HP1 and strains HP1∆hpdA,

439 HP1∆hpdA(pBBRBleo-hpdA), and HP1∆hpdA(pBBRBleo) were investigated. All four

440 strains were cultivated in MSM with 1 g/L 3HP. Small samples of the medium were

441 removed regularly for measurement of the strain growth and substrate concentration.

20 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

442 The growth of strains was detected by measuring the optical density of the medium

443 sample at 600 nm (OD600 nm) using a spectrophotometer. To determine the 3HP

444 concentration in the medium, the cells were removed by centrifugation at 14,000×g for

445 2 min, and the supernatants were carefully collected after filtering through a 0.22 m

446 filter and used for high-performance liquid chromatography (HPLC) analysis. The

447 growth of and 3HP degradation by other mutants were analyzed following the same

448 methods as those for the mutant strain HP1∆hpdA.

449

450 Bioinformatics analysis

451 The domain annotation of hpdA was analyzed by BLAST analysis of NCBI using the

452 Conserved Domain Database (CDD). The amino acid sequence encoded by hpdA3 was

453 analyzed by TMHMM (http://www.cbs.dtu.dk/services/TMHMM/), HMMTOP

454 (http://www.enzim.hu/hmmtop/), TMpred (https://embnet.vital-

455 it.ch/software/TMPRED_form.html) and SignalP-5.0

456 (http://www.cbs.dtu.dk/services/SignalP/) to predict the transmembrane domain (Tmd)

457 and signal peptide (Sp).

458

459 Heterologous expression of hpdA

460 To verify the function of hpdA, HpdA was heterologously expressed. S. melonis TY, E.

461 adhaerens ZM04 (collection number, CCTCC AB 2019220) and Alcaligenes faecalis

462 JQ135 were chosen as the expression hosts. The pRK415-hpdA series expression vector

463 was generated with a fragment of hpdA just downstream of the promoter of the vector

21 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

464 pRK415 (Hind III/EcoR I-digested). After sequencing and obtaining the desired

465 construction, the plasmid construct was transformed into E. coli WM3064 and then

466 mated into S. melonis TY, E. adhaerens ZM04 and A. faecalis JQ135 to generate S.

467 melonis TY-hpdA, E. adhaerens ZM04-hpdA and A. faecalis JQ135-hpdA, respectively.

468 The expression of HpdA was induced by adding 0.1 mM IPTG. The cells were

469 harvested by centrifugation at 6,000×g for 5 min and washed twice with 12 mM

470 phosphate‐buffered saline (PBS), pH 7.4, and the cell pellets were resuspended in MSM

471 until an OD600 nm value of 1.0 was reached (resting cells). Then, the 3HP transformation

472 ability of resting cells was determined. 3HP was added at a final concentration of 0.1

473 mg/mL in the resting cell suspension, which was then shaken at 200 rpm at 30 °C for 8

474 h. Then, samples were taken aseptically for UV and LC-MS analysis to detect the

475 formation of the product.

476

477 Construction, expression and cellular localization of GFP fusion protein

478 The full-length amino acid sequence of HpdA3 was fused with GFP at the N terminus.

479 pGFPehpdA3 was constructed by fusing the PCR fragment of hdpA3 to pGFPe

480 (digested with EcoR I and BamH I, yielding an 11-amino acid linker region

481 (GSENLYFQGQF) followed by GFP). The primers used to construct the GFP fusion

482 protein are listed in Table 3. The construct was transformed into E. coli BL21 (DE3)

483 and grown in LB medium containing kanamycin. Transformant colonies verified by

484 sequencing were grown in LB medium containing kanamycin at 37 ℃ to an OD600 nm

485 of 0.5. Cultures were induced with 0.5 mM IPTG for 4 h at 30 ℃ on a rotary shaker

22 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

486 (200 rpm). The cells were harvested at 10,000×g for 1 min at 4 ℃, washed twice with

487 PBS (pH 7.4) and resuspended in PBS (pH 7.4). Samples were imaged under a confocal

488 microscope with a 488-nm excitation filter and a 520-nm emission filter.

489

490 Analytical methods

491 The concentration of 3HP was determined by HPLC with diode array detection. The

492 column used was an Eclipse XDB-C18 reverse-phase column (5 μm; 4.6 × 250 mm;

493 Agilent, USA) at 35 °C. The mobile phase was 20% (v/v) methanol and 80% (v/v) 0.1%

494 triethylamine at a flow rate of 1.0 mL min−1, and the detection wavelength was set at

495 254 nm. LC-MS analysis to determine the heterologous expression product of hpdA

496 was performed on a liquid chromatograph (Agilent 1200, USA) equipped with an

497 Eclipse XDB-C18 reverse-phase column (5 μm; 4.6 × 250 mm; Agilent, USA) and an

498 LCQ Deca XP Max MS instrument (Thermo Finnigan) with an electrospray interface

499 (Turbo Ion Spray). The iron spray voltage was set at 3,000 V. Nitrogen was used as the

500 sheath gas (60 arb) and auxiliary gas (15 arb). The capillary temperature was set at

501 350 ℃, and the capillary voltage was set at 10 V. The mobile phase was 95% (v/v) of

502 0.1% (v/v) formic acid and 5% (v/v) of methanol at a flow rate of 1 mL/min, and 10 μL

503 of sample was injected. The column temperature was set at 35 ℃, and the detection

504 wavelength was 310 nm. Positive electrospray ionization with continuous full scanning

505 from m/z 50 to 350 was performed. Samples for HPLC and LC-MS analysis were

506 centrifuged at 14,000×g for 2 min, and the supernatant was filtered through a 0.22 μm

507 filter prior to injection.

23 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

508

509 Data availability

510 The 16S rRNA sequence of E. adhaerens HP1 has been deposited in the GenBank

511 database under the accession number MN083303. The Whole Genome Shotgun project

512 of E. adhaerens HP has been deposited at DDBJ/ENA/GenBank under accession

513 VHKK00000000. The version described in this paper is version VHKK01000000.

514

515

516 Acknowledgments

517 This work was financially supported by grants from the National Natural Science

518 Foundation of China (Nos. 41721001, 41630637 and 31800089).

519

520 References

521 1. Kim, S.G. and R.F. Novak, Role of P450IIE1 in the metabolism of 3-

522 hydroxypyridine, a constituent of tobacco smoke: redox cycling and DNA strand

523 scission by the metabolite 2, 5-dihydroxypyridine. Cancer research, 1990.

524 50(17): p. 5333-5339.

525 2. Ma, Z., et al., Thermal degradation products derived from the smoke of Salvia

526 divinorum leaves. Tetrahedron Letters, 2010. 51(41): p. 5480-5482.

527 3. Huq, F., et al., Studies on the activity of three palladium (II) compounds of the

528 form: Trans-PdL2Cl2 where L= 2-hydroxypyridine, 3-hydroxypyridine, and 4-

529 hydroxypyridine. Journal of inorganic biochemistry, 2007. 101(1): p. 30-35.

24 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

530 4. Schmidt, A., Biologically active mesomeric betaines and alkaloids, derived

531 from 3-hydroxypyridine, pyridin-N-oxide, nicotinic acid and picolinic acid:

532 Three types of conjugation and their consequences. Current Organic Chemistry,

533 2004. 8(8): p. 653-670.

534 5. Koch, V. and S. Schnatterer, Chemistry of 3-hydroxypyridine Part 1:

535 Bromination and iodination of 3-hydroxypyridine. Synthesis, 1990. 1990(06): p.

536 497-498.

537 6. Yoshida, K., et al., Synthesis of 3-hydroxypyridines using ruthenium-catalyzed

538 ring-closing olefin metathesis. Organic letters, 2008. 11(3): p. 515-518.

539 7. Vaitekūnas, J., et al., A 2-hydroxypyridine catabolism pathway in Rhodococcus

540 rhodochrous strain PY11. Appl. Environ. Microbiol., 2016. 82(4): p. 1264-1273.

541 8. Petkevičius, V., et al., Catabolism of 2-hydroxypyridine by Burkholderia sp.

542 MAK1: a five-gene cluster encoded 2-hydroxypyridine 5-monooxygenase

543 HpdABCDE catalyses the first step of biodegradation. Appl. Environ.

544 Microbiol., 2018: p. AEM. 00387-18.

545 9. Cain, R.B., C. Houghton, and K.A. Wright, Microbial metabolism of the

546 pyridine ring. Metabolism of 2-and 3-hydroxypyridines by the maleamate

547 pathway in Achromobacter sp. Biochemical Journal, 1974. 140(2): p. 293-300.

548 10. Karvelis, L., et al., Pusillimonas sp. 5HP degrading 5-hydroxypicolinic acid.

549 Biodegradation, 2014. 25(1): p. 11-19.

550 11. Zhao, S., et al., Isolation of a 3-hydroxypyridine degrading bacterium,

551 Agrobacterium sp. DW-1, and its proposed degradation pathway. AMB Express,

25 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

552 2019. 9(1): p. 65.

553 12. Houghton, C. and R. Cain, Microbial metabolism of the pyridine ring.

554 Formation of pyridinediols (dihydroxypyridines) as intermediates in the

555 degradation of pyridine compounds by micro-organisms. Biochemical Journal,

556 1972. 130(3): p. 879-893.

557 13. Kaiser, J.-P., Y. Feng, and J.-M. Bollag, Microbial metabolism of pyridine,

558 quinoline, acridine, and their derivatives under aerobic and anaerobic

559 conditions. Microbiol. Mol. Biol. Rev., 1996. 60(3): p. 483-498.

560 14. Shukla, O. Microbial transformation of pyridine compounds. in Proceedings of

561 the Indian Academy of Sciences-Chemical Sciences. 1984. Springer.

562 15. Wang, S., et al., Identification of nicotine biotransformation intermediates by

563 Agrobacterium tumefaciens strain S33 suggests a novel nicotine degradation

564 pathway. Appl Microbiol Biotechnol, 2012. 95(6): p. 1567-78.

565 16. Wang, S.N., Z. Liu, and P. Xu, Biodegradation of nicotine by a newly isolated

566 Agrobacterium sp. strain S33. Journal of Applied Microbiology, 2009. 107(3):

567 p. 838-847.

568 17. Karlin, S., A.M. Campbell, and J. Mrazek, Comparative DNA analysis across

569 diverse genomes. Annual review of genetics, 1998. 32(1): p. 185-225.

570 18. Muto, A. and S. Osawa, The guanine and cytosine content of genomic DNA and

571 bacterial evolution. Proceedings of the National Academy of Sciences, 1987.

572 84(1): p. 166-169.

573 19. Grether-Beck, S., et al., Structural analysis and molybdenum-dependent

26 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

574 expression of the pAO1-encoded nicotine dehydrogenase genes of Arthrobacter

575 nicotinovorans. Molecular Microbiology, 1994. 13(5): p. 929-936.

576 20. Jiménez, J.I., et al., Deciphering the genetic determinants for aerobic nicotinic

577 acid degradation: the nic cluster from Pseudomonas putida KT2440.

578 Proceedings of the National Academy of Sciences, 2008. 105(32): p. 11329-

579 11334.

580 21. Bläse, M., et al., Cloning, expression, and sequence analysis of the three genes

581 encoding quinoline 2-oxidoreductase, a molybdenum-containing hydroxylase

582 from Pseudomonas putida 86. Journal of Biological Chemistry, 1996. 271(38):

583 p. 23068-23079.

584 22. Chai, Y.-m., et al., FaPYR1 is involved in strawberry fruit ripening. Journal of

585 experimental botany, 2011. 62(14): p. 5079-5089.

586 23. Gao, Z., et al., Characterization of the ABA receptor VlPYL1 that regulates

587 anthocyanin accumulation in grape berry skin. Frontiers in plant science, 2018.

588 9.

589 24. Kinsella, R.L., et al., Defining the interaction of the protease CpaA with its type

590 II secretion chaperone CpaB and its contribution to virulence in Acinetobacter

591 species. Journal of Biological Chemistry, 2017. 292(48): p. 19628-19638.

592 25. Manavski, N., L.-M. Schmid, and J. Meurer, RNA-stabilization factors in

593 chloroplasts of vascular plants. Essays in biochemistry, 2018. 62(1): p. 51-64.

594 26. Zhu, Z., et al., A lipoxygenase from red alga Pyropia haitanensis, a unique

595 enzyme catalyzing the free radical reactions of polyunsaturated fatty acids with

27 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

596 triple ethylenic bonds. PloS one, 2015. 10(2): p. e0117351.

597 27. Liu, Q., et al., Elucidation of piericidin A1 biosynthetic locus revealed a

598 thioesterase-dependent mechanism of α-pyridone ring formation. Chemistry &

599 biology, 2012. 19(2): p. 243-253.

600 28. Tang, H., et al., Systematic unraveling of the unsolved pathway of nicotine

601 degradation in Pseudomonas. PLoS Genet, 2013. 9(10): p. e1003923.

602 29. Qiu, J., et al., Identification and characterization of a novel pic gene cluster

603 responsible for picolinic acid degradation in Alcaligenes faecalis JQ135.

604 bioRxiv, 2019: p. 530550.

605 30. Rohde, M., et al., Attachment of CO dehydrogenase to the cytoplasmic

606 membrane is limiting the respiratory rate of Pseudomonas carboxydovorans.

607 FEMS microbiology letters, 1985. 28(2): p. 141-144.

608 31. Spreitler, F., et al., Interaction of CO dehydrogenase with the cytoplasmic

609 membrane monitored by fluorescence correlation spectroscopy. ChemBioChem,

610 2010. 11(17): p. 2419-2423.

611 32. Du, L., et al., Characterization of a Unique Pathway for 4-Cresol Catabolism

612 Initiated by Phosphorylation in Corynebacterium glutamicum. J Biol Chem,

613 2016. 291(12): p. 6583-94.

614 33. Schmeling, S., et al., Phenylphosphate synthase: a new phosphotransferase

615 catalyzing the first step in anaerobic phenol metabolism in Thauera aromatica.

616 J Bacteriol, 2004. 186(23): p. 8044-57.

617 34. Ohtsubo, Y., et al., Complete Genome Sequence of Acidovorax sp. Strain

28 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

618 KKS102, a Polychlorinated-Biphenyl Degrader. Journal of Bacteriology, 2012.

619 194(24): p. 6970-6971.

620 35. Stabler, R.A., et al., Draft Genome Sequences of Pseudomonas fluorescens BS2

621 and Pusillimonas noertemannii BS8, Soil Bacteria That Cooperate To Degrade

622 the Poly-γ-d-Glutamic Acid Anthrax Capsule. Genome Announcements, 2013.

623 1(1): p. e00057-12.

624 36. Sayed, M., et al., Complete Genome Sequence of Mycobacterium sp. MS1601,

625 a Bacterium Performing Selective Oxidation of Polyols. Genome

626 Announcements, 2017. 5(15): p. e00156-17.

627 37. Hezbri, K., et al., Geodermatophilus sabuli sp. nov., a γ-radiation-resistant

628 actinobacterium isolated from desert limestone. International Journal of

629 Systematic and Evolutionary Microbiology, 2015. 65(10): p. 3365-3372.

630 38. Oyuntsetseg, B., et al., Amycolatopsis acidiphila sp. nov., a moderately

631 acidophilic species isolated from coal mine soil. International Journal of

632 Systematic and Evolutionary Microbiology, 2017. 67(9): p. 3387-3392.

633 39. Huo, Y.-Y., et al., Salinicola peritrichatus sp. nov., isolated from deep-sea

634 sediment. Antonie van Leeuwenhoek, 2013. 104(1): p. 55-62.

635 40. Wang, H., et al., Characterization of a novel nicotine degradation gene cluster

636 ndp in Sphingomonas melonis TY and its evolutionary analysis. Frontiers in

637 Microbiology, 2017. 8: p. 337.

638 41. Yu, H., et al., Molybdenum-containing nicotine hydroxylase genes in a nicotine

639 degradation pathway that is a variant of the pyridine and pyrrolidine pathways.

29 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

640 Applied and Environmental Microbiology, 2015. 81(24): p. 8330-8338.

641 42. Schenk, S., et al., Gene structures and properties of enzymes of the plasmid-

642 encoded nicotine catabolism of Arthrobacter nicotinovorans. Journal of

643 Molecular Biology, 1998. 284(5): p. 1323-1339.

644 43. Livak, K.J. and T.D. Schmittgen, Analysis of relative gene expression data using

645 real-time quantitative PCR and the 2−ΔΔCT method. Methods, 2001. 25(4): p.

646 402-408.

647 44. Ramakers, C., et al., Assumption-free analysis of quantitative real-time

648 polymerase chain reaction (PCR) data. Neuroscience Letters, 2003. 339(1): p.

649 62-66.

650 45. Tuomi, J.M., et al., Bias in the Cq value observed with hydrolysis probe based

651 quantitative PCR can be corrected with the estimated PCR efficiency value.

652 Methods, 2010. 50(4): p. 313-322.

653 46. Chen, Y.-F., H. Chao, and N.-Y. Zhou, The catabolism of 2,4-xylenol and p-

654 cresol share the enzymes for the oxidation of para-methyl group in

655 Pseudomonas putida NCIMB 9866. Applied Microbiology and Biotechnology,

656 2014. 98(3): p. 1349-1356.

657 47. Woodcock, D.M., et al., Quantitative evaluation of Escherichia coli host strains

658 for tolerance to cytosine methylation in plasmid and phage recombinants.

659 Nucleic Acids Research, 1989. 17(9 ): p. 3469-3478.

660 48. Dehio, C. and M. Meyer, Maintenance of broad-host-range incompatibility

661 group P and group Q plasmids and transposition of Tn5 in Bartonella henselae

30 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

662 following conjugal plasmid transfer from Escherichia coli. Journal of

663 Bacteriology, 1997. 179(2): p. 538-540.

664 49. Saltikov, C.W. and D.K. Newman, Genetic identification of a respiratory

665 arsenate reductase. Proceedings of the National Academy of Sciences of the

666 United States of America, 2003. 100(19): p. 10983-10988.

667 50. Wang, H., et al., Two Novel Sets of Genes Essential for Nicotine Degradation

668 by Sphingomonas melonis TY. Frontiers in Microbiology, 2017. 7(2060).

669 51. Qiu, J., et al., A Novel Degradation Mechanism for Pyridine Derivatives in

670 <span class="named-content genus-species" id="named-

671 content-1">Alcaligenes faecalis</span> JQ135. Applied and

672 Environmental Microbiology, 2018. 84(15): p. e00910-18.

673 52. Arnaud, M., A. Chastanet, and M. Debarbouille, New vector for efficient allelic

674 replacement in naturally nontransformable, low-GC-content, gram-positive

675 bacteria. Appl Environ Microbiol, 2004. 70(11): p. 6887-91.

676 53. Hoang, T.T., et al., A broad-host-range Flp-FRT recombination system for site-

677 specific excision of chromosomally-located DNA sequences: application for

678 isolation of unmarked Pseudomonas aeruginosa mutants. Gene, 1998. 212(1):

679 p. 77-86.

680 54. Kovach, M.E., et al., Four new derivatives of the broad-host-range cloning

681 vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene,

682 1995. 166(1): p. 175-176.

683 55. Yan, X., et al., Cre/lox system and PCR-based genome engineering in Bacillus

31 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

684 subtilis. Appl Environ Microbiol, 2008. 74(17): p. 5556-62.

685 56. Keen, N.T., et al., Improved broad-host-range plasmids for DNA cloning in

686 Gram-negative bacteria. Gene, 1988. 70(1): p. 191-197.

687 57. Rapp, M., et al., Experimentally based topology models for E. coli inner

688 membrane proteins. Protein Science, 2004. 13(4): p. 937-945.

689

690 Tables

691 Table 1 Functional annotations of hypothetical Hpd proteins Superfamily designation Protein Size (amino Putative function Region Superfamily (specific hit/conserved CDD accession no. E value acids) (positions) domain) Orf1 213 IS110 family transposase 83-156 Transposase_20 (Transposase pfam02371 1.25e-28 IS116/IS110/IS902 family) Orf2 115 IS6 family transposase 1-100 Rve (Transposase (or an inactivated COG3316 1.06e-18 derivative)) HpdD 208 Isochorismatase family 23-201 CSHase (N-carbamoylsarcosine cd01015 1.08e-68 protein amidohydrolase) HpdB 343 Leucyl aminopeptidase No putative conserved domains have been detected HpdC 277 Alpha/beta hydrolase 46-277 MhpC (Pimeloyl-ACP methyl ester COG0596 1.18e-19 carboxylesterase) HpdE 249 Asp/Glu racemase 5-247 COG3473 (Maleate cis-trans COG3473 4.98e-85 isomerase) Orf3 215 ABC transporter 1-203 TM_PBP1_transp_AraH_like cd06579 7.90e-48 permease Orf4 125 ABC transporter 17-120 AraH COG1172 1.88e-14 permease (Ribose/xylose/arabinose/galactosid e ABC-type transport system, permease component) Orf5 531 Sugar ABC transporter 32-521 MglA (ABC-type sugar transport COG1129 0e+00 ATP-binding protein system, ATPase component) Orf6 357 TMAO reductase system 48-356 PRK10936 (TMAO reductase PRK10936 7.43e-91 periplasmic protein TorT system periplasmic protein TorT) Orf7 233 GntR family 2-223 GntR (DNA-binding transcriptional COG1802 3.02e-33 transcriptional regulator regulator, GntR family) Orf8 224 Hypothetical protein 32-132 DUF4202 (Domain of unknown pfam13875 9.94e-05 function) HpdA1 819 Xanthine dehydrogenase 36-808 CoxL (CO or xanthine COG1529 0e+00 family protein dehydrogenase, Mo-binding subunit) molybdopterin-binding subunit HpdA2 147 (2Fe-2S)-binding protein 5-137 CoxS (Aerobic-type carbon COG2080 7.80e-56 monoxide dehydrogenase, small subunit, CoxS/CutS family) HpdA3 234 Carbon monoxide 14-159 SRPBCC_5 cd07823 2.26e-32 dehydrogenase HpdA4 292 Xanthine dehydrogenase 4-285 CoxM (CO or xanthine COG1319 1.26e-63 family protein subunit M dehydrogenase, FAD-binding subunit) Orf9 367 Pyruvate, phosphate 20-337 PPDK_N (Pyruvate phosphate pfam01326 2.85e- dikinase dikinase, PEP/pyruvate binding 106 domain) Orf10 612 PEP-utilizing protein 10-609 PRK08296 (hypothetical protein) PRK08296 0e+00 mobile subunit Orf11 42 Hypothetical protein No putative conserved domains have been detected Orf12 166 Flavin reductase 16-158 Flavin_Reduct (Flavin reductase like pfam01613 9.24e-24 domain) Orf13 304 LysR family 5-65 HTH_1 (Bacterial regulatory helix- pfam00126 7.41e-22 transcriptional regulator turn-helix protein, lysR family) 98-291 PBP2_PAO1_like cd08412 6.43e-53 Orf14 499 Hypothetical protein 27-487 YoaI (Aromatic ring hydroxylase) COG2368 7.24e-89 Orf15 313 Polysaccharide 28-283 CE4_HpPgdA_like (Catalytic cd10938 3.75e- deacetylase domain of Helicobacter pylori 110 peptidoglycan deacetylase 32 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Superfamily designation Protein Size (amino Putative function Region Superfamily (specific hit/conserved CDD accession no. E value acids) (positions) domain) (HpPgdA) and similar proteins) Orf16 416 M20 family peptidase 20-408 M20_ArgE_DapE-like cd08011 2.74e- 100 692 693 Table 2 Strains and plasmids used in this study Strain or plasmid Relevant characteristics Reference or source Strains Escherichia coli DH5 supE44 lacU169(80dlacZΔM15)hsdR17 recA1 endA1 [47] gyrA96Δthi relA1 - - - BL21(DE3) F ompT hsdSB (rB mB )gal dcm lacY1(DE3) Transgen BL21(DE3)pGFPe BL21(DE3) transformed with pGFPe, KanR This study BL21(DE3)pGFP-hpdA3 BL21(DE3) transformed with pGFPehpdA3, KanR This study BL21(DE3)pGFP-hpdA3-2 BL21(DE3) transformed with pGFPehpdA3-2, KanR This study BL21(DE3)pGFP-hpdA3-3 BL21(DE3) transformed with pGFPehpdA3-3, KanR This study WM3064 Donor strain for conjugation, 2,6-diaminopimelic [48, 49] acid auxotroph: thrB1004 pro thi rpsL hsdS lacZΔM15 RP4-1360 Δ(araBAD)567 ΔdapA1341::[erm pir(wt)] Ensifer species HP1 Wild type, 3HP-degrading strain, G-, ErmS, BlmS This study HP1ΔhpdA Ensifer adhaarens HP1 mutant with hpdA gene replaced This study by erythromycin resistance gene from pMAD, ErmR, BlmS HP1Δorf9 Ensifer adhaarens HP1 mutant with orf9 gene replaced This study by erythromycin resistance gene from pMAD, ErmR, BlmS HP1Δorf10 Ensifer adhaarens HP1 mutant with orf10 gene This study replaced by erythromycin resistance gene from pMAD, ErmR, BlmS HP1ΔhpdA(pBBRBleo- Complementary strain, in which hpdA gene was This study hpdA) complemented by pBBRBleo-hpdA in Ensifer adhaarens HP1ΔhpdA, ErmR, BlmR HP1ΔhpdA(pBBRBleo) HP1ΔhpdA containing pBBRBleo as a control for This study complementary strain, ErmR, BlmR Ensifer adhaerens ZM04 Wild type, non-3HP-degrading strain, G-, TcS This study ZM04-hpdA-1 Ensifer sp. ZM04 transformed with pRK415-hpdA-1, This study TcR ZM04-hpdA-2 Ensifer sp. ZM04 transformed with pRK415-hpdA-2, This study TcR ZM04-hpdA-3 Ensifer sp. ZM04 transformed with pRK415-hpdA-3, This study TcR ZM04-hpdA-4 Ensifer sp. ZM04 transformed with pRK415-hpdA-4, This study

33 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Strain or plasmid Relevant characteristics Reference or source TcR ZM04-hpdA-5 Ensifer sp. ZM04 transformed with pRK415-hpdA-5, This study TcR ZM04-hpdA-6 Ensifer sp. ZM04 transformed with pRK415-hpdA-6, This study TcR ZM04 pRK415 Ensifer sp. ZM04 transformed with pRK415, TcR This study Sphingomonas melonis TY Wild type, non-3HP-degrading strain, G-, TcS [50] Alcaligenes faecalis JQ135 Wild type, non-3HP-degrading strain, G-, TcS [51]

Plasmids pMAD Source of erythromycin resistance gene [52] pEX18Tc Gene knockout vector, oriT+, sacB+, TcR [53] pEX18Tc-hpdA hpdA gene knockout vector containing two DNA This study fragments homologous to the upstream and downstream regions of the hpdA and erythromycin resistance gene from pMAD pEX18Tc-orf9 orf9 gene knockout vector containing two DNA fragments homologous to the upstream and downstream regions of the orf9 and erythromycin resistance gene from pMAD pEX18Tc-orf10 orf10 gene knockout vector containing two DNA fragments homologous to the upstream and downstream regions of the orf10 and erythromycin resistance gene from pMAD pBBR1MCS-2 Broad-host-range cloning vector, GmR [54] p7Z6 pMD18-T containing lox71-zeo-lox66 cassette, source [55] of bleomycin resistance gene pBBRBleo Plasmid constructed by replacing the kanamycin This study resistance gene in p BBR1MCS-2 with bleomycin resistance gene pBBRBleo-hpdA hpdA gene complementation vector by fusing This study hpdA into the Kpn I/Xba I digested pBBRBleo pRK415 Broad host range vector, TcR [56] pRK415-hpdA-1 hdpA heterologous expression vector by fusing orf7- This study orf10 in 3hpd into the Hind III/EcoR I digested pRK415 pRK415-hpdA-2 hdpA heterologous expression vector by fusing orf7- This study hpdA4 in 3hpd into the Hind III/EcoR I digested pRK415 pRK415-hpdA-3 hdpA heterologous expression vector by fusing hpdA1- This study orf10 in 3hpd into the Hind III/EcoR I digested pRK415 pRK415-hpdA-4 hdpA heterologous expression vector by fusing hpdA1- This study orf10 (except hpdA3) in 3hpd into the Hind III/EcoR I

34 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Strain or plasmid Relevant characteristics Reference or source digested pRK415 pRK415-hpdA-5 hdpA heterologous expression vector by fusing hpdA1- This study orf10 (except hpdA4) in 3hpd into the Hind III/EcoR I digested pRK415 pRK415-hpdA-6 hdpA heterologous expression vector by fusing hpdA1A2A3A4 in 3hpd into the Hind III/EcoR I digested pRK415 pGFPe ColE1 replicon, derivative of pWaldo digested with [57] EcoR I and BamH I, yielding an 11-amino-acid linker sequence (GSENLYFQGQF) followed by GFP, KanR pGFPehpdA3 PCR-amplified fragment containing hpdA3 without the This study stop codon inserted into pGFPe at the EcoR I/BamH I restriction sites pGFPehpdA3-2 PCR-amplified fragment containing partial hpdA3 This study (Sp1+Tmd) without the stop codon inserted into pGFPe at the EcoR I/BamH I restriction sites pGFPehpdA3-3 PCR-amplified fragment containing partial hpdA3 This study (Sp2+Tmd) without the stop codon inserted into pGFPe at the EcoR I/BamH I restriction sites 694 695 Table 3 Primers used in this study Primer Sequence (5’-3’) Purpose HpdAupF ctggttgccgtcacttatc To amplify upstream fragment of hdpA HpdAupR tgcgcttctccatggg for gene knockout HpdAdownF aactcggcctctttccc To amplify downstream fragment of HpdAdownR tgccctgccccgtaatc hdpA for gene knockout Orf9-upF ttgctcagaaggctgttactc To amplify upstream fragment of orf9 Orf9-upR gatcgtcaagccaaaggat for gene knockout Orf9-downF gctttcagcagcaacaggg To amplify downstream fragment of Orf9-downR ggaccgaccttctgcgtaa orf9 for gene knockout Orf10-upF tacggcgtgcctatgac To amplify upstream fragment of orf10 Orf10-upR aaagcggttcgggatta for gene knockout Orf10-downF ggcgcatacggctattatct To amplify downstream fragment of Orf10-downR aggcccattcaatgtcctg orf10 for gene knockout ErmC-F aattgaatgagacatgcta To amplify bleomycin resistance gene ErmC-R atcgattcacaaaaaatagg from pMad for gene knockout pEX18Tc-VF gcacgacaggtttcccgactg For verification of the construction of pEX18Tc-VR ccgcttctgcgttctgattta knockout vector by PCR or sequencing HpdA-VF cgctgggagcattgacg For verification of strain HP1∆hpdA by HpdA-VR gcacctggccctcaaccaa PCR or sequencing Orf9-VF gacgatgggctcctatcact For verification of strain HP1∆orf9 by Orf9-VR ggatgcaacgtcactacgc PCR or sequencing

35 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Primer Sequence (5’-3’) Purpose Orf10-VF acccgccttaggtatcgtc For verification of strain HP1∆orf10 by Orf10-VR ttcatttgccttccttgct PCR or sequencing HpdA-CF tcagccgtttggtcgtcgtcag To amplify hdpA for gene HpdA-CR atcacagctgattttcctggt complementation pBBRBleo-VF ggcacgacaggtttcccgactg For verification of the construction of pBBRBleo-VR tgcgggcctcttcgctatt pBBRBleo-hpdA by PCR or sequencing hpdA-1-1F atgcacgtgagaatgctgc hpdA-1-1R ggtttatgggcgaggagat hpdA-1-2F atctcctcgcccataaacc To amplify orf6-orf10 in 3hpd to hpdA-1-2R gcccgtcaaaatctccac construct pRK415-hpdA-1 hpdA-1-3F gtggagattttgacgggc hpdA-1-3R atggcagatctgttttcc hpdA-2-1F atgaaagccgccgatttcct hpdA-2-1R gcccgtcaaaatctccac To amplify orf6-hpdA4 in 3hpd to hpdA-2-2F gtggagattttgacgggc construct pRK415-hpdA-2 hpdA-2-2R atggcagatctgttttcc hpdA-3-1F atgcacgtgagaatgctgc hpdA-3-1R ggtttatgggcgaggagat hpdA-3-2F atctcctcgcccataaacc To amplify hpdA1-orf10 in 3hpd to hpdA-3-2R gcccgtcaaaatctccac construct pRK415-hpdA-3 hpdA-3-3F gtggagattttgacgggc hpdA-3-3R tcacagctgattttcctggt hpdA-4-1F atgcacgtgagaatgctgc hpdA-4-1R ggtttatgggcgaggagat To amplify hpdA1-orf10 (except hpdA-4-2F atctcctcgcccataaacc hpdA3) in 3hpd to construct pRK415- hpdA-4-2R caacttggaattcctgaaca hpdA-4 hpdA-4-3F ttcagccgtttggtcgtcgt hpdA-4-3R tcacagctgattttcctggt hpdA-5-1F atgcacgtgagaatgctgc hpdA-5-1R tcatttgccttccttgctcac To amplify hpdA1-orf10 (except hpdA-5-2F atgctccgtcgcggacaaca hpdA4) in 3hpd to construct pRK415- hpdA-5-2R gcccgtcaaaatctccac hpdA-5 hpdA-5-3F gtggagattttgacgggc hpdA-5-3R tcacagctgattttcctggt hpdA-6-1F atgaaagccgccgatttcct To amplify hpdA1A2A3A4 in 3hpd to hpdA-6-1R tcacagctgattttcctggt construct pRK415-hpdA-6 GFPhpdA3-1F atgctccgtcgcggacaaca To amplify fragment containing hpdA3 GFPhpdA3-1R gttcagcgaaatcaagatcgc without the stop codon to construct pGFPehpdA3 338F cctacgggaggcagcagcag To amplify V3 region in 16S rRNA 518R attaccgcggctgctgg gene for qPCR RT-hpdA1F attgaggaacatctatcg To amplify 78-bp fragment of hpdA1

36 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Primer Sequence (5’-3’) Purpose RT-hpdA1R ctcagtctgtttatggat for qPCR RT-hpdA2F gactatcttcgccacaa To amplify 117-bp fragment of hpdA2 RT-hpdA2R aagcatcaagcacgatt for qPCR RT-hpdA3F attccaagttgccgagag To amplify 104-bp fragment of hpdA3 RT-hpdA3R tgtctattgcctggatgg for qPCR RT-hpdA4F tggttgccgtcacttatc To amplify 84-bp fragment of hpdA4 RT-hpdA4R aaaggattgcgaaatcattgt for qPCR RT-hpdBF ccgattatctatacgattgagaa To amplify 85-bp fragment of hpdB for RT-hpdBR atgccatgtaggacttga qPCR RT-hpdCF cagggatgtattgagtgg To amplify 77-bp fragment of hpdC for RT-hpdCR atagtgaagcaggtgttg qPCR RT-hpdDF gtgactgttgtgctgacc To amplify 103-bp fragment of hpdD RT-hpdDR cacgccgaaatgtcttca for qPCR RT-hpdEF ttgatcgacggattgaaa To amplify 92-bp fragment of hpdE for RT-hpdER atatagtccacaaccatctc qPCR RT-orf3F gttcttcatcacgctcttg To amplify 92-bp fragment of orf3 for RT-orf3R cgaggacgagtaccaatc qPCR RT-orf4F cgccgacaatatgtttaacc To amplify 80-bp fragment of orf4 for RT-orf4R tcagcaggatcaggaact qPCR RT-orf5F agagccttagcaagaatt To amplify 87-bp fragment of orf5 for RT-orf5R cgaagagaacatgaacct qPCR RT-orf6F gaagcaagaaggtcaagg To amplify 76-bp fragment of orf6 for RT-orf6R aaaccatcagcagaatcg qPCR RT-orf7F tttccaaatccgatacac To amplify 84-bp fragment of orf7 for RT-orf7R gagtgccagacaagatat qPCR RT-orf8F aatggtgattgatactgctt To amplify 80-bp fragment of orf8 for RT-orf8R acttcggttgagttgtct qPCR RT-orf9F gtgaaggcgtggtgaagg To amplify 83-bp fragment of orf9 for RT-orf9R aattctgcggcgaaggac qPCR RT-orf10F gtcggctccaagttcttc To amplify 77-bp fragment of orf10 for RT-orf10R tcgtagagtgcgtcataca qPCR RT-orf11F caggaagtcgccagcaat To amplify 75-bp fragment of orf11 for RT-orf11R aagtggtgacgccttctt qPCR RT-orf12F aacgacctatgccaagac To amplify 75-bp fragment of orf12 for RT-orf12R aacagggaaattaggatgct qPCR RT-orf13F atacgcagccatatctca To amplify 80-bp fragment of orf13 for RT-orf13R atcctccacctctttcaa qPCR RT-orf14F ttcgacaaggtcttcatc To amplify 86-bp fragment of orf14 for RT-orf14R aactgaccgagaatgttg qPCR RT-orf15F attcaccttcgacatggatg To amplify 78-bp fragment of orf15 for RT-orf15R atggttgcgacgagattg qPCR RT-orf16F cccttctcccaagaccat To amplify 77-bp fragment of orf16 for RT-orf16R aatgcaccttcacgacat qPCR

37 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

696 Figures and figure legends

697

698 Figure 1 Cell growth of and 3HP degradation by strain HP1 and its plasmid

699 transformants. The solid and broken lines represent cell growth and 3HP degradation,

700 respectively. The results presented in these histograms are the means of three

701 independent experiments, and the error bars indicate the standard error.

702

703

704 Figure 2 (A) Genetic organization of the putative 3HP degradation gene cluster 3hpd

705 in E. adhaerens HP1. The diagram shows the arrangement of genes that BLAST

706 analysis suggests are involved in 3HP catabolism. The arrows indicate the size and

707 direction of transcription of each gene. hpdA is a putative 3-hydroxypyridine

708 dehydrogenase; hpdBCDE are putative 2,5-dihydroxypyridine dioxygenase, N-formyl 38 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

709 maleamic acid deformylase, maleamate amidase, and maleate isomerase, respectively,

710 which are responsible for the maleate pathway of 2,5-dihydroxypyridine metabolism.

711 Orf1-orf16 are genes flanking the putative 3HP metabolic genes. The predicted

712 functions of these genes are summarized in Table 1. (B) Proposed catabolism of 3HP

713 by E. adhaerens HP1.

714

715

716 Figure 3 Transcriptional analysis of the 3hpd gene cluster. RT-qPCR analysis of target

717 gene transcripts produced in E. adhaerens HP1 grown with (black bars) or without (gray

718 bars) 3HP. The expression levels of the 3hpd genes were normalized to the 16S rRNA

719 expression level and are expressed as the fold change in expression in cells. The results

720 presented in these histograms are the means of four independent experiments, and error

721 bars indicate the standard error. 39 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

722

723

724 Figure 4 Bioinformatic analysis of HpdA. (A) Molecular architecture of several typical

725 multicomponent molybdenum-containing hydroxylases. Subunits Ⅰ, Ⅱ, and Ⅲ are

726 MCD-, FAD-, and two [Fe-S] cluster-containing components, respectively. HpdA, 3-

727 hydroxypyridine dehydrogenase from E. adhaerens HP1; SpmABC (GenBank

40 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

728 accession numbers AEJ14617 and AEJ14616), 3-succinoylpyridine dehydrogenase

729 from P. putida; PicA (KY264362), picolinic acid dehydrogenase from Alcaligenes

730 faecalis JQ135; NdhLSM (YP 007988778, YP 007988779, and YP 007988780),

731 nicotine dehydrogenase from Arthrobacter nicotinovorans; KdhLMS (YP 007988766,

732 YP 007988771, and YP 007988772), ketone dehydrogenase from A. nicotinovorans;

733 QorLSM (X98131), quinoline 2-oxidoreductase from P. putida 86; QoxLMS

734 (AJ537472), quinaldine 4-oxidase from Arthrobacter ilicis Ru61a; XdhABC

735 (BAE76932.1, BAE76933.1, and BAE76934.1), xanthine dehydrogenase from E. coli

736 W3110; CdhABC (ADH15879.1, ADH15880.1, and ADH15881.1), caffeine

737 dehydrogenase Pseudomonas sp. strain CBB1; CoxLSM (X82447), carbon monoxide

738 dehydrogenase from Oligotropha carboxidovorans; CutLSM (AAD00363.1,

739 AAD00362.1, and AAD00361.1), carbon monoxide dehydrogenase from

740 Hydrogenophaga pseudoflava; IorAB (CAA88753 and CAA88754), isoquinoline 1-

741 oxidoreductase from Brevundimonas diminuta 7; NicAB (PP3947 and PP3948),

742 nicotinate hydroxylase from P. putida KT2440. The letters below the proteins indicate

743 the subunit names of the corresponding proteins. The conserved domains are MPT1,

744 MPT2 and MPT3, which are the three motifs for binding to the MCD cofactor; FAD,

745 consensus FAD-binding site; CODH, specific FAD-binding site of members of the

746 xanthine dehydrogenase/oxidase family; [FeS]1, binding site for the ferredoxin-like

747 [2Fe-2S] cluster; [FeS]2, binding site for the second [2Fe-2S] cluster; SRPBCC,

748 SRPBCC ligand-binding domain; CytC1, CytC2 and CytC3, the three cytochrome c

749 binding motifs. (B) Phylogenetic analysis of HpdA and related molybdenum-containing

41 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

750 hydroxylases: HpdA1 and the Moco binding subunit of other enzymes (a); HpdA4 and

751 the FAD-binding subunit of other enzymes (b); HpdA2 and the [2Fe-2S]-binding

752 subunit of other enzymes (c). The phylogenetic trees were constructed using the

753 neighbor-joining method (with 1,000 bootstraps) with MEGA software, version 6.06.

754 The bar represents the number of amino acid substitutions per site.

755

756 757 Figure 5 Time course of the conversion of 3HP to 2,5-DHP by recombinant strain

758 ZM04-hpdA. The control group with strain ZM04 did not show transformation of 3HP,

759 and no product was formed.

760

42 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

761

762 Figure 6 Detection of the catalytic activity of hpdA in strain ZM04. UV spectrum for

763 the conversion of 3HP by transformant ZM04 pRK415-hpdA-3 with substrate 3HP (A),

764 ZM04 pRK415-hpdA-3 without substrate 3HP (B), ZM04 pRK415 with substrate 3HP

765 (C), ZM04 pRK415 without substrate 3HP (D), and substrate 3HP only (E) and UV

766 spectrum of 0.025 mg/mL of 2,5-DHP standard (F). Samples in A-E were diluted 5

767 times before detection. Each group was analyzed three times independently, and the 43 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

768 results of each group showed a similar trend. The best one is shown.

769

770

771 Figure 7 Localization of GFP by confocal microscopy. E. coli BL21 (DE3)-pGFPe (A)

772 and E. coli BL21 (DE3)-pGFPehpdA3 (B).

773

774

775 Figure 8 Predicted 3HP catabolism gene clusters in bacterial genomes. Ⅰ to XXII: the

776 22 different 3HP catabolism loci. Abbreviations and representative strains are as

44 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.08.898148; this version posted January 9, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

777 follows: Ead, Ensifer adhaerens HP1; Hyp, Hyphomicrobium sp. 99; Rhi, Rhizobiales

778 bacterium isolate AFS066724; Sno, Starkeya novella isolate S2; Con,

779 Confluentimicrobium sp. EMB200-NS6; Pca, Pusillimonas caeni strain KCTC 42353;

780 Pno, Pusillimonas noertemannii BS8; Pus, Pusillimonas sp. 17-4A; Pus2, Pusillimonas

781 sp. L52-1-41; Pus3, Pusillimonas sp. isolate EAC49; Pus4, Pusillimonas sp. isolate

782 SAT20; Aci, Acidovorax sp. KKS102; Com, Comamonas sp. A23 and Comamonas sp.

783 Z1; Cte, Comamonas testosteroni I2 and Comamonas testosteroni NBRC 100989; Cth,

784 Comamonas thiooxydans strain S44; Vpa, Variovorax paradoxus isolate S2; Hal,

785 Halomonas sp. MES3-P3E; Hba, Halomonadaceae bacterium R4HLG17; Pse,

786 Pseudoxanthomonas sp. SGD-5-1; Spe, Salinicola peritrichatus strain JCM 18795; Bla,

787 Blastococcus sp. DSM 46838; Gru, Geodermatophilus ruber strain DSM 45317; Myc,

788 Mycobacterium sp. MS1601; Pau, Pseudonocardia autotrophica strain NRRL B-16064;

789 Por, Pseudonocardia oroxyli strain CGMCC 4.3143; Myc2, Mycobacterium sp. GA-

790 2829; Msm, Mycolicibacterium smegmatis MKD8; Pam, Pseudonocardia

791 ammonioxydans strain CGMCC 4.1877; Pku, Pseudonocardia kunmingensis strain

792 DSM 45301; Pse2, Pseudonocardia sp. MH-G8; Gsa, Geodermatophilus sabuli strain

793 DSM 46844; Aac, Amycolatopsis acidiphila strain JCM 30562; Goc, Gaiella occulta

794 strain F2-233; Sol, Solirubrobacterales bacterium 70-9 SCN. The detailed genomic

795 accession numbers and the gene locus tags are listed in Table S1 in the supplemental

796 material. Identities (percent) and similarity (percent) of amino acid sequences between

797 3hpd proteins of strain E. adhaerens HP1 and representative homologs are listed in

798 Table S2.

45