bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1 4-hydroxyphenylpyruvate dioxygenase thermolability is responsible for

2 temperature-dependent melanogenesis in Aeromonas salmonicida subsp. salmonicida

3

4 Yunqian Qiao,a Jiao Wang,a He Wang,a Baozhong Chai,a Chufeng Rao,a Xiangdong

5 Chen,a,b,# Shishen Dua,#

6

7 aState Key Laboratory of Virology, College of Life Sciences, Wuhan University, Wuhan,

8 China

9 bChina Center for Type Culture Collection, Wuhan, China

10

11 Running Head: Thermosensitive melanization in Aeromonas salmonicida

12

13 #Address correspondence to Xiangdong Chen, [email protected], or Shishen Du,

14 [email protected]

15

16 Keywords: Aeromonas salmonicida subsp. salmonicida, melanin, temperature-dependent,

17 4-hydroxyphenylpyruvate dioxygenase, thermolability, evolution

18

19

20

21

1 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

22 Abstract

23 Aeromonas salmonicida subsp. salmonicida (A.s.s) is a major pathogen affecting fisheries

24 worldwide. It is a well-known member of the pigmented Aeromonas species, which

25 produces melanin at  22 °C. However, melanogenesis decreases as the culture

26 temperature increases and is completely suppressed at 30–35 °C while bacterial growth is

27 not affected. The mechanism and biological significance of this temperature-dependent

28 melanogenesis are not clear. Heterologous expression of an A.s.s.

29 4-hydroxyphenylpyruvate dioxygenase (HppD), the most crucial enzyme in the

30 HGA-melanin synthesis pathway, results in thermosensitive pigmentation in Escherichia

31 coli, suggesting that HppD plays a key role in this process. In the current study, we

32 demonstrated that the extreme thermolability of HppD is responsible for the

33 temperature-dependent melanization of A.s.s. Substitutions in three residues, Ser18,

34 Pro103, or Leu119 of HppD from A.s.s increases the thermolability of this enzyme and

35 results in temperature-independent melanogenesis. Moreover, replacing the

36 corresponding residues of HppD from Aeromonas media strain WS, which forms pigment

37 independent of temperature, with those of A.s.s HppD leads to thermosensitive

38 melanogenesis. Structural analysis suggested that mutations at these sites, especially at

39 position P103, can strengthen the secondary structure of HppD and greatly improve its

40 thermal stability. In addition, we found that HppD sequences of all A.s.s isolates are

41 identical and that two of the three residues are completely conserved within A.s.s isolates,

42 which clearly distinguishes these from other Aeromonas strains. We suggest that this

2 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

43 property represents an adaptive strategy to the psychrophilic lifestyle of A.s.s.

44 Importance

45 Aeromonas salmonicida subsp. salmonicida (A.s.s) is the causative agent of furunculosis,

46 a bacterial septicemia of cold water fish of the Salmonidae family. As it has a

47 well-defined host range, A.s.s has become an ideal model to investigate the co-evolution

48 of host and pathogen. For many pathogens, melanin production is associated with

49 virulence. Although other species of Aeromonas can produce melanin, A.s.s is the only

50 member of this genus that has been reported to exhibit temperature-dependent

51 melanization. Here we demonstrate that thermosensitive melanogenesis in A.s.s strains is

52 due to the thermolability of 4-hydroxyphenylpyruvate dioxygenase (HppD). The strictly

53 conserved hppD sequences among A.s.s and the exclusive thermosensitive pigmentation

54 of these strains might provide insight into the role of melanin in the adaptation to a

55 particular host, and offer a novel molecular marker to readily differentiate A.s.s strains

56 from other A. salmonicida subspecies and Aeromonas species.

57

58 Introduction

59 Melanins are structurally complex high molecular weight pigments formed by the

60 oxidative polymerization of phenolic and/or indolic compounds (1). They represent the

61 most ubiquitous natural pigment and are produced by organisms from every kingdom (2).

62 Melanin has various functions in diverse organisms, and it is believed that it confers a

63 survival advantage for microorganisms, especially under stressful environmental

3 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

64 conditions such as radiation, oxidative stress, the presence of digestive enzymes, and

65 heavy metal toxicity (2). These molecules have also been associated with the virulence

66 and pathogenicity of a variety of microbes such as Cryptococcus neoformans, Wangiella

67 dermatitidis, and Vibrio cholerae (3-5), by protecting pathogens against host defenses and

68 interfering with the host immune response (2). Melanins are classified into several

69 categories based on their synthesis pathway, and bacteria usually synthesize two types,

70 namely eumelanin or pyomelanin. Eumelanin is catalyzed by tyrosinase or other

71 polyphenol oxidases to hydroxylate tyrosine, via the intermediate

72 L-3,4-dihydroxyphenylalanine (L-DOPA), into dopaquinone, which then auto-oxidizes

73 and polymerizes to form melanin (6). Pyomelanin is synthesized by the homogentisic

74 acid (HGA) pathway, in which 4-hydroxyphenylpyruvate is converted to HGA by

75 4-hydroxyphenylpyruvate dioxygenase (HppD), which is followed by the oxidative

76 polymerization of HGA to produce the pigment (6). Our previous study showed that

77 Aeromonas media WS, a strain that produces melanin at high levels, mainly synthesizes

78 pyomelanin through the HGA pathway (7).

79 The Aeromonas genus is a collection of gram-negative, rod-shaped, facultatively

80 anaerobic bacteria that include motile, non-motile, mesophilic, and psychrophilic species.

81 They are widely distributed in aquatic environments and have been frequently isolated

82 from aquatic animal species. Some Aeromonas strains are principal or opportunistic

83 pathogens of fish and other cold-blooded animals, as well as the causative agent of

84 infectious complications in humans (8). Melanogenesis has been described in some

4 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

85 species of this genus, such as A. media, Aeromonas salmonicida, Aeromonas bestiarum,

86 and Aeromonas eucrenophila (9). However, a number of Aeromonas species are not

87 considered to produce melanin, including Aeromonas hydrophila, Aeromonas

88 allosaccharophila, and Aeromonas encheleia (10). In fact, melanin synthesis is one of the

89 key criteria for the taxonomy of this genus.

90 Among melanogenic Aeromonas species, A. salmonicida subsp. salmonicida (A.s.s),

91 an important salmonid fish pathogen and the etiologic agent of furunculosis, is the only

92 member that has been described to exhibit temperature-dependent melanogenesis. It has

93 been demonstrated that A.s.s can initiate pigmentation at lower temperatures such as

94 22 °C. However, this process is thermosensitive. The level of melanin production reduces

95 as the incubation temperature increases, and this process is completely suppressed at

96 temperatures > 30 °C, at which the bacteria still grow normally (11-13). Nevertheless, the

97 reason for this phenomenon is not yet clear. Increasing the temperature can also alter

98 some other characteristics of this bacterium, including the loss of virulence (> 25 °C)

99 (14-16) and the auto-agglutination trait (>25 °C) (17). Since A.s.s is psychrophilic

100 (optimal growth at 22–25 °C) , these temperature-dependent phenotypes probably

101 represent adaptations to psychrophilic environments.

102 In the recent study, we suggested that Aeromonas species mainly produce

103 pyomelanin through the HGA pathway (7) and showed that heterologous expression of

104 hppD from an A.s.s strain in Escherichia coli BL21 leads to similar

105 temperature-dependent pigmentation. This result suggests that the abolished pigmentation

5 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

106 of A.s.s at elevated temperatures is likely due to the thermolabile activity of HppD. In this

107 study, we confirmed this hypothesis and identified the critical residues of HppD,

108 specifically, S18, P103, and L119, that are responsible for its thermolability. Substitutions

109 at these residues resulted in temperature-independent melanogenesis in A.s.s. Moreover,

110 replacing the corresponding residues of HppD from other Aeromonas species with those

111 from A.s.s resulted in thermal sensitive melanization. Taken together, our results

112 demonstrate that the thermosensitive melanogenesis of A.s.s is dependent on individual

113 residues in the HppD protein.

114

115 Results

116 hppD genes of A.s.s isolates are identical

117 To investigate the molecular mechanism responsible for the temperature-dependent

118 pigmentation of A.s.s, we first ensured that this phenotype is ubiquitous in various A.s.s

119 isolates. As shown in Fig. S1, all four A.s.s strains maintained in our laboratory (listed in

120 Table 1; AB98041, KACC14791, 2013-8, and 2014-235) exhibited similar

121 thermosensitive pigmentation as reported in the literature. We reasoned that if HppD

122 actually plays an important role in this phenotype, the amino acid sequences of HppD

123 from various A.s.s isolates should be highly conserved. Thus, we cloned and sequenced

124 hppD genes from the four isolates and discovered that the sequences were 100% identical

125 at the nucleotide level. Furthermore, alignment of the sequences of hppD from all 26 A.s.s

126 isolates available in GenBank revealed that hppD was 100% conserved at the nucleotide

6 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

127 level (data not shown), even though these strains were isolated in different years, from

128 diverse areas, and from a variety of hosts (see detailed strain information in Table S1).

129 Therefore, the thermal sensitive melanization observed in A.s.s isolates is likely due to the

130 different activities of HppD at different temperatures. For simplicity, we refer to the hppD

131 gene from A.s.s strains as hppD-AS in the rest of this paper.

132 Identification of residues in HppD-AS for site-directed mutagenesis

133 HppD is the most important enzyme in the HGA-melanin synthesis pathway. Almost all

134 Aeromonas species for which sequence data has been submitted to GenBank have

135 homologous hppD genes, regardless of pigment production. Since thermosensitive

136 pigmentation does not occur in other Aeromonas species and the expression of HppD

137 alone is sufficient for this phenotype, we hypothesized that comparing HppD-AS with

138 HppD from other Aeromonas species would reveal the crucial residues that are important

139 for HppD-AS thermolability. We used the “consensus approach” to identify potential

140 amino acids for mutagenesis. This approach is based on replacing a specific amino acid

141 with the most frequent amino acid at a given position based on the alignment of

142 homologous proteins, which has been applied successfully for many protein engineering

143 studies (18-20). Fifteen homologous HppDs from different representative Aeromonas

144 species were selected for sequence alignments. These all shared greater than 80% amino

145 acid identities with that of HppD-AS. As shown in Fig. 1, sequence analysis revealed a

146 total of eleven positions in HppD-AS that differed from the most conserved amino acids

147 in homologous proteins. These residues were found to be S5, S10, S18, A96, P103, T107,

7 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

148 L119, T160, S298, D301, and Q303. In addition, we found five other residues that are

149 present in nearly half of the homologues, specifically S8, Q180, S232, L255, and Q262

150 (Fig. 1). Therefore, we replaced all sixteen of these residues in HppD-AS with the most

151 prevalent amino acid at that position. Subsequently, single variants (S5T, S8A, S10T,

152 S18T, A96S, P103Q, T107A, L119P, T160A, Q180H, S232A, L255M, Q262H, S298A,

153 D301E, and Q303R) were generated and termed Mut1 to Mut16, respectively.

154 Identification of residues in HppD-AS that are critical for temperature-dependent

155 pigmentation in recombinant E. coli strains

156 As heterologous expression of hppD-AS in E. coli BL21 results in pigmentation at 22 °C

157 but not at ≥ 30 °C, we tested whether any of our HppD-AS mutants resulted in melanin

158 production at ≥ 30 °C in recombinant E. coli strains. hppD-WS (hppD from A. media

159 strain WS) and empty plasmid were used as controls. As presented in Fig. 2A, compared

160 to that in the strain harboring an empty plasmid, all the recombinant E. coli strains

161 expressing HppD could synthesize pigment at 22 °C, indicating that all hppD variants

162 were functional. As expected, BL21(pET26-hppD-AS) lost melanogenic ability when the

163 induction temperature was increased to 30 °C or 37 °C, whereas BL21(pET26-hppD-WS),

164 which expressed HppD-WS, still formed pigment. Interestingly, when analyzing the

165 sixteen HppD-AS mutants, E. coli expressing three variants (Mut4 [S18T], Mut6 [P103Q],

166 and Mut8 [L119P]) produced pigment at 30 °C (Fig. 2A, which only shows partial results

167 of mutants), and Mut6 was even found to result in pigment accumulation at 37 °C.

168 To further confirm the role of these three residues in the thermosensitive

8 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

169 melanogenesis phenomenon, double mutants Mut46 (S18T/P103Q), Mut48

170 (S18T/L119P), and Mut68 (P103Q/L119P), as well as a triple mutant Mut468

171 (S18T/P103Q/L119P) were generated and studied. As shown in Fig. 2B, any combination

172 of these three mutations enabled recombinant E. coli strains to form melanin at 30 °C.

173 Notably, any mutant containing the P103Q substitution (Mut46, Mut68, and Mut468)

174 could generate a significant amount of pigment at 37 °C, comparable to that with

175 hppD-WS. The double mutant Mut48 also resulted in a slight sign of melanization at

176 37 °C, whereas the single mutant Mut4 or Mut8 did not. These data demonstrate that the

177 effect of each mutation is additive. Taken together, these results indicate that the three

178 substitutions (S18T, P103Q, and L119P) in HppD-AS significantly increase heat

179 resistance of melanogenesis in recombinant expression strains, of which Mut6 (P103Q)

180 has the strongest effect.

181 The aforementioned results clearly demonstrate a key role for three residues in

182 HppD-AS in temperature-dependent pigmentation. We suggested that in theory, replacing

183 the corresponding residues of HppD from other Aeromonas species that undergo

184 temperature-independent pigmentation with those of HppD-AS should result in

185 temperature-sensitive melanization. We thus substituted the corresponding residues of

186 HppD-WS with the amino acids of HppD-AS at positions S18, P103, and L119. As a

187 result, single point mutants W4(T16S), W6(Q101P) W8(P117L), and double/triple

188 mutants W46(T16S/Q101P), W48(T16S/P117L), W68 (Q101P/P117L), and

189 W468(T16S/Q101P/P117L) of HppD-WS were constructed and tested. As shown in Fig.

9 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

190 3, cells expressing the W4 or W8 mutant formed pigment at all tested temperatures,

191 suggesting that these mutations do not dramatically affect the function of HppD-WS.

192 However, cells expressing the W6 mutant failed to produce pigment at 37 °C, consistent

193 with its important role in temperature-dependent pigmentation. Similar to the observed

194 additive effect of these mutations (mentioned previously herein), strains expressing any

195 of the double mutants failed to synthesize pigment at 37 °C. Importantly, cells expressing

196 the triple mutant W468 produced less pigment at 22 °C, compared to that in other strains,

197 and completely lost melanin formation ability at 30 °C and 37 °C. This indicated that we

198 had constructed an HppD-AS-like HppD-WS protein by altering only three residues.

199 Verification of the critical role of HppD in temperature-dependent pigmentation in

200 A.s.s

201 Using the heterogeneous E. coli system, we established that temperature-dependent

202 melanogenesis in A.s.s is due to the function of HppD-AS at different temperatures and

203 identified critical residues of HppD-AS that are involved in this phenotype. However,

204 drawing conclusions from this artificial system is dangerous unless we could prove it in

205 the natural biological system. To verify that our findings in E. coli are relevant for A.s.s,

206 we modified chromosomal hppD in A.s.s strains. We first knocked out the hppD gene in

207 wild-type A.s.s strains to determine if this would abolish pigmentation. As expected, the

208 three hppD deletion mutants (AB98041ΔhppD, 2013-8ΔhppD, and 2014-235ΔhppD) all

209 completely lost their ability to generate pigment regardless of the culture temperature (Fig.

210 4). This confirmed that pigmented A.s.s strains mainly synthesize pyomelanin through the

10 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

211 HGA pathway. We next constructed AB98041 mutant strains carrying the hppD mutant

212 alleles that eliminated temperature-dependent pigmentation in E. coli using the

213 AB98041ΔhppD strain as the background. Three mutant strains, AB98041(hppD-6

214 [P103Q]), AB98041(hppD-46 [S18T/P103Q]), and AB98041(hppD-68 [P103Q/L119P]),

215 were generated and tested for pigmentation at different temperatures. As shown in Fig.

216 4A, unlike the wild-type strain, AB98041(hppD-6), AB98041(hppD-46), and

217 AB98041(hppD-68) were able to produce pigment at 30 °C. Furthermore, when the hppD

218 gene of AB98041 was replaced with hppD-WS, the resultant strain, designated

219 AB98041(hppD-WS), synthesized melanin at 30 °C (Fig. 4A). Similar phenotypes were

220 observed with two other A.s.s isolates when their hppD genes were replaced with

221 hppD-WS, respectively (2013-8[hppD-WS] and 2014-235[hppD-WS]) (Fig. 4B and 4C).

222 Based on these results, we ascertained that the inhibition of pigmentation in A.s.s strains

223 at elevated temperatures is caused by the diminished function of HppD.

224 HppD-AS is thermolabile in vitro

225 The previously mentioned results suggest that thermosensitive melanization in A.s.s is

226 likely due to the thermolability of HppD-AS, and that mutations that stabilize HppD-AS

227 result in a temperature-independent phenotype. To confirm this, wild-type HppD-AS,

228 HppD-WS, and seven HppD-AS variants (Mut4, Mut6, Mut8, Mut46, Mut48, Mut68, and

229 Mut468) that resulted in the melanization of E. coli at 30 °C were homogenously purified

230 and characterized as described in the Materials and Methods. Based on sodium dodecyl

231 sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), all purified mutant enzymes

11 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

232 appeared as a single band with the same molecular mass (~42 kDa) as HppD-AS, in

233 agreement with the predicted molecular mass of 41227.47 Da (Fig. S2).

234 The effect of temperature on the activities of wild-type and the seven HppD-AS

235 mutant enzymes was determined at temperatures ranging from 16 to 53 °C. To our

236 surprise, purified HppD-AS displayed no activity at any temperature. Moreover, we

237 found that the HppD-AS crude extract lost enzyme activity at 4 °C within 8–10 h

238 probably due to its excessive thermolability. Therefore, the enzymatic activity and

239 thermal stability of HppD-AS was assayed using its fresh crude extract, whereas purified

240 enzymes were used for all other mutants. As shown in Fig. 5A and Fig. S3, HppD-AS

241 showed maximum activity at 22 °C. However, mutant enzymes such as Mut6 and Mut468

242 exhibited higher optimum temperatures of 30 °C and 37 °C, respectively. HppD-WS

243 displayed the highest activity at 37 °C. And at 45 °C, wild-type HppD-AS exhibited only

244 43% of its maximal activity, whereas Mut8, Mut6, and Mut468 mutants retained around

245 51%, 94%, and 92% of their individual maximum activities, respectively. To evaluate the

246 thermostability of each enzyme, protein was incubated at 40 °C and after various

247 incubation times, samples were taken to determine residual activity. As shown in Fig. 5B,

248 HppD-AS activity was highly sensitive to thermal inactivation. The residual activity of

249 HppD-AS decreased exponentially to only 8% of initial activity within 10 min, whereas

250 the activity of HppD-WS was not significantly reduced even after incubation for 80 min.

251 However, Mut4, Mut8, and the double mutant Mut48 exhibited longer half-lives than

252 wild-type HppD-AS, indicating that these two mutations contributed to enhancing the

12 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

253 thermostability of HppD-AS. Remarkably, the substitution P103Q obviously improved

254 protein stability. The single mutant Mut6 retained approximately 75% of initial activity

255 after incubation at 40 °C for 80 min, and combining this variant with the other two

256 mutations further increased enzyme stability, as indicated by the activity of the triple

257 mutant Mut468. These results together demonstrate that wild-type HppD-AS is extremely

258 thermolabile and that substitutions at three residues (S18T, P103Q, L119P) can

259 substantially increase enzyme thermostability such that A.s.s strains containing these

260 mutated hppD genes can produce melanin at 30 °C. In other words, HppD thermolability

261 is suggested to mediate the abolished pyomelanin formation of A.s.s strains at higher

262 temperatures.

263 Structural analysis of HppD-AS mutation sites

264 To gain insight into the mechanisms of enhanced thermostability mediated by the three

265 mutations (S18T, P103Q, L119P) in HppD-AS, we constructed a structural model of

266 wild-type HppD-AS based on the crystal structure of HppD from Pseudomonas

267 fluorescens (PDB ID, 1CJX) (Fig. 6). For this, the three substituted residues were mapped

268 to the generated model structure. All three residues were found to be located in the

269 N-terminal domain of the enzyme and the effect of these substitutions on HppD

270 thermostability was then evaluated (Fig. 6).

271 The S18 residue resides in the first β strand (β1) of the enzyme, which forms a β-sheet

272 with the parallel β4 strand and antiparallel β5 strand (Fig. 7A). The S18T substitution was

273 determined to be a very conservative change because both Ser and Thr are uncharged

13 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

274 polar amino acids that differ only by one methyl group. It was predicted that this mutation

275 would not cause significant changes to the local structure or the number of hydrogen

276 bonds (Fig. 7B). However, it has been reported that Thr has a greater intrinsic propensity

277 to form β-sheets than Ser, which was verified by both statistical analysis of proteins with

278 known structures (21, 22) and experimental measurement using different host-guest

279 model systems (23). The distinct propensities to form β-sheets were considered due to

280 different mean curvature parameters of each amino acid (24). As Thr has a larger

281 accessible area in its side chain relative to that of Ser, which results in a smaller mean

282 curvature value in the backbone surface, it can readily adopt a β-sheet pattern and

283 stabilize β-sheets (24). Thus, the enhanced stability of the S18T mutant of HppD-AS is

284 likely due to a slightly more stable structure. Nonetheless, the effect of the mutation was

285 predicted to be moderate, consistent with the moderate increase in pigmentation at higher

286 temperatures conferred by the mutation in vivo.

287 The substitution P103Q resulted in temperature-independent pigmentation in A.s.s in

288 vivo and had the greatest effect on enzyme stability in vitro. Based on the modeled

289 structure, P103 is located in the internal α3 helix (Fig. 7C). It is well known that Pro is a

290 strong breaker of α helical structure and has a low tendency to form α-helices (21, 25, 26).

291 Due to the pyrrolidine ring in its side chain, Pro has no amide hydrogen to form hydrogen

292 bonds while it is in the middle of a polypeptide chain. Thus, Pro103 cannot participate in

293 hydrogen bonding to the carbonyl oxygen atom of the preceding fourth residue Gly99 in

294 the α-helix (Fig. 7C). Furthermore, it has been widely proven that the presence of a

14 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

295 proline residue in α-helices can cause a pronounced kink between the segments preceding

296 and following proline (27, 28). This kink results from the reduced number of helix

297 backbone hydrogen bonds and helps to avoid steric clash between the pyrrolidine ring

298 and the local backbone (28). As a consequence, the presence of proline would greatly

299 increase the flexibility of helices. However, substituting Pro103 with Gln would eliminate

300 the distortion in the helical structure caused by the pyrrolidine ring of proline.

301 Furthermore, glutamine contains two amide protons that would introduce two new

302 hydrogen bonds with the carbonyl oxygen of Gly99 (Fig. 7D). These changes are

303 suggested to significantly increase the rigidity of the α3 helix, thus stabilizing the overall

304 structure of the protein. This is consistent with the potent effect of this mutation on the

305 activity of HppD-AS in vivo and in vitro.

306 L119 was found to be situated at the second residue of a β-turn between strands β6

307 and β7 (Fig. 7E). The reason for increased thermostability of the L119P mutant enzyme

308 appears to lie in a reduction in the flexibility of the β-turn local backbone. As proline

309 adopts fewer conformations in its backbone than other amino acids because of the rigid

310 pyrrolidine ring, it can constrain the conformational freedom of Cα-N rotation (29) and

311 stabilize the protein by reducing the entropy of the denatured state (30). The introduction

312 of a proline into flexible regions to improve the thermostability of proteins has been well

313 documented (31-33). Certain key positions for protein thermostabilization by proline

314 have also been declared, namely, flexible loops, the N-caps of α-helices, and the first or

315 second site of β-turns (33). In addition, except for the hydrogen bonds formed by its

15 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

316 oxygen atom with K53, L119 is not involved in other H-bonds via its backbone amide

317 (Fig. 7E). Therefore, the L119P mutation would make the β-turn more rigid without

318 negative side effects. According to the above analysis, we concluded that all three

319 substitutions (S18T, P103Q, and L119P) stabilize the structure of the protein by

320 strengthening secondary structural elements of HppD-AS.

321

322 Discussion

323 The temperature-sensitive melanization of A.s.s strains has been observed for many years.

324 However, since very little is understood about the associated melanin synthesis pathway,

325 the mechanism underlying this phenomenon remains unclear. We previously showed that

326 A. media WS produces pyomelanin through the HGA pathway. Moreover, we found that

327 heterologous expression of HppD from an A.s.s strain in E. coli resulted in

328 temperature-sensitive pigmentation, indicating that A.s.s strains likely produce melanin

329 through the HGA pathway and that HppD plays a critical role in the thermal sensitive

330 melanization process. Here, we confirmed that HppD is critical for melanization in A.s.s

331 and demonstrated that the extreme thermolability of HppD-AS is responsible for the

332 temperature-dependent melanogenesis. Here, we also identified the critical residues of

333 HppD-AS that are important for its thermolability and showed that replacing the

334 corresponding residues in HppD from A. media WS results in thermolabile enzymes.

335 Among the three identified residues, S18 and L119 were shown to have a moderate

336 effect, whereas P103 had the strongest impact on thermal lability. A single P103Q

16 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

337 substitution greatly increased the thermostability of HppD-AS and was sufficient to

338 eliminate the thermo-sensitive characteristic in both recombinant E. coli and A.s.s strains.

339 This indicates that mutation in only one site can substantially effect protein conformation,

340 thus changing the property of the protein dramatically. Through homology modeling, we

341 revealed that this site resides in the sixth position of an α-helix in HppD-AS. The proline

342 residue is rarely found after the fourth residue in α-helices of globular proteins because of

343 its ability to disrupt α-helical structure (25). This should be the main reason for the

344 thermolability of HppD-AS. However, why would such a seemingly unfavorable residue

345 has been reserved in HppD-AS?

346 We aligned the HppD-AS sequence with its homologs from all other A. salmonicida

347 subspecies and Aeromonas species published in GenBank to examine the degree of

348 conservation of this site. Intraspecific sequence alignment was conducted among isolates

349 including A. salmonicida subsp. masoucida RFAS1, subsp. pectinolytica 34melT, and

350 nine un-subspeciated strains (see Table S1 for detailed information). Alignment results

351 showed that there are some residue variations in the HppD of other A. salmonicida

352 subspecies compared to the sequence of HppD-AS (Fig. S4). Notably, among the variant

353 positions, Pro103 and Leu119 of HppD-AS were found to be absolutely conserved and

354 clearly distinct from the Gln and Pro in other subspecies, respectively. More strikingly,

355 interspecific sequence comparison with totally 206 recognized and published Aeromonas

356 spp. isolates found that Pro and Leu at these two sites were only conserved within A.s.s

357 isolates and that in other bacterial species, these were substituted with different residues

17 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

358 including Gln, Leu, Arg, and Pro (data not shown). As A.s.s are strict psychrophilic

359 bacteria, we surmise that the consistent preservation of Pro103 and Leu119 in HppD from

360 all A.s.s strains is a protein adaptation that occurred during evolution to adopt a

361 psychrophilic lifestyle. The wide distribution of such substitutions might imply their

362 advantage. The assessment of the activities of each enzyme at various temperatures

363 showed that the optimal temperature of HppD-AS was 22 °C, whereas that of other

364 mutants gradually shifted from 22 to 37 °C, and HppD-WS had maximum activity at

365 37 °C (Fig. 5A). This agrees with a previous study suggesting that psychrophilic enzymes

366 usually possess high catalytic activity at the optimal growth temperatures of bacterium

367 (34). It indicates that although Pro113 and Leu119 residues decrease thermostability, they

368 substantially improve the catalytic activity of HppD-AS at low temperatures. This is in

369 accordance with the common strategy of cold adaptation for proteins; specifically,

370 psychrophilic enzymes always have amino acid substitution preferences that avoid

371 forming helical structures and participating in local interactions to confer higher protein

372 flexibility to offset the slow catalytic reaction rate inherent to low temperatures (34, 35).

373 Furthermore, the high conservation of hppD sequences among A.s.s strains might be

374 attributed to the clonal population structure of this subspecies (36). Such high clonality is

375 probably an adaptation to the particular host, salmonid fish, in which clonal outbreaks are

376 prone to occur (37).

377 Although A. salmonicida species are classified into five subspecies including

378 salmonicida, achromogenes, smithia, masoucida, and pectinolytica, as reported in

18 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

379 Bergey’s Manual of Systematic Bacteriology (10), many laboratories currently designate

380 subspecies salmonicida as “typical” and any isolate diverging phenotypically and

381 molecularly as “atypical” strains. The typical isolates comprise an extremely compact

382 group and are considered clonal (38), whereas atypical isolates are more heterogeneous

383 and show a wide range of hosts including non-salmonid fish as well as salmonids (39).

384 The convoluted history of taxonomy and nomenclature for A. salmonicida species makes

385 the clear assignment of strains difficult. Concerning the particular possession of Pro103

386 and Leu119 in HppD-AS, these residues could be used as potential molecular markers for

387 identifying A.s.s isolates. The conventional and readily observable feature to differentiate

388 subspecies salmonicida from atypical strains is the production of a brown diffusible

389 pigment. However, this rule was not always applicable; for example, some A.s.s isolates

390 lack pigment production and some atypical isolates synthesize melanin (40-42). Other

391 biochemical tests such as indole production and gluconate oxidation have similar caveats

392 (43). A variety of molecular genetic techniques have also been applied for the

393 characterization of typical and atypical A. salmonicida strains, including DNA:DNA

394 reassociation (44), ribotyping (45), PCR-based typing analysis (46), plasmid profiling

395 (47), pulse-field gel electrophoresis (48), amplified fragment length polymorphism

396 analysis (49), and IS-based RFLP (50), among others. Nonetheless, these methods are

397 often time-consuming and occasionally controversial. DNA sequencing analysis based on

398 the 16S rRNA gene also makes the identification challenging at the species level because

399 of the extreme intraspecies conservation of this gene (51). Recently, vapA, a gene

19 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

400 encoding the A-layer, was used to establish sequence schemes and to distinguish typical A.

401 salmonicida from atypical isolates, with partial nucleotide sequence homogeneity

402 between the former but heterogeneity among the latter (52). However, it has been shown

403 that there can be a high frequency of mutations in vapA when the strains are cultured in

404 unfavorable conditions. The identical hppD nucleotide sequence described here, and the

405 distinct presence of Pro103 and Leu119 residues compared to its homologs from other

406 Aeromonas species, provide a new pragmatic parameter to easily discriminate A.s.s from

407 atypical strains and other Aeromonas species. Furthermore, this could complement the

408 A-layer typing method, which cannot classify strains lacking the A-layer (e.g. subspecies

409 pectinolytica) (52).

410 It has been generally reported that inactivation of homogentisate 1,2-dioxygenase

411 (HmgA), an enzyme that degrades HGA into Kreb's cycle, can result in the accumulation

412 of pyomelanin (53-55). By examining the hmgA genes from laboratorial and published

413 A.s.s strains, we observed that they all have frameshift mutations and form pseudogenes.

414 This suggests that all of these A.s.s isolates are pyomelanin-producing. Moreover, the

415 presence of the conserved Pro103 residue in HppD protein implies that the temperature

416 susceptibility of melanization is exclusive and ubiquitous for A.s.s. As reported previously,

417 the genome of A.s.s A449 revealed a significant number of pseudogenes, including many

418 virulence genes, genes encoding transcription factors, and genes encoding basic

419 metabolic enzymes, among others (56). Since A.s.s is a specific pathogen of Salmonidae,

420 the accumulation of pseudogenes has been suggested to be a hallmark of adaptation to the

20 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

421 specific host (56). Considering that melanin has been associated with virulence in many

422 pathogenic microorganisms, it is tempting to speculate that both disruption of the hmgA

423 genes and temperature-sensitive pigmentation represent pathoadaptation towards

424 particular salmonid hosts. However, the correlation between melanin and host specificity

425 has not been reported. Subsequently, further pathogenicity experiments are required to

426 validate this contention.

427

428 Materials and Methods

429 Bacterial strains, plasmids, and growth conditions

430 All the parental and derivative strains and plasmids used in this study are listed in Table 1.

431 A.s.s strain AB98041 was obtained from the China Center for Type Culture Collection

432 (CCTCC) and A.s.s strain KACC14791 was purchased from the Korean Agricultural

433 Culture Collection (KACC). The other two A.s.s isolates, 2013-8 and 2014-235, were

434 kindly donated by Dr. Tongyan Lu from Heilongjiang River Fisheries Research Institute,

435 Chinese Academy of Fishery Sciences. E. coli strains were routinely cultured in

436 Luria-Bertani (LB) broth or agar at 37 °C unless otherwise indicated. Heterologous

437 expression strains were also grown at 22 or 30 °C to compare melanin production. The

438 wild-type and mutant A.s.s strains were grown in LB medium at 22 or 30 °C to observe

439 the melanin synthesis phenotype. Antibiotics were supplemented in the medium at the

440 following concentrations when necessary: ampicillin (100 μg/ml), kanamycin (50 μg/ml),

441 chloramphenicol (34 μg/ml). For cultivation of E. coli strain β2155, 100 μg/ml

21 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

442 diaminopimelic acid (DAP) was added.

443 Site-directed mutagenesis and heterologous expression of the hppD gene

444 The hppD genes from A.s.s AB98041 and A. media WS were previously cloned using the

445 primer pairs hppD(AS)-S/hppD(AS)-A and hppD(WS)-S/hppD(WS)-A (7). The amplified

446 fragments were then digested with NdeI and HindIII (Thermo Fisher Scientific, USA) and

447 inserted into the pUC18 vector (TaKaRa, Japan) digested with the same enzymes to

448 construct plasmids pUC-hppDAS and pUC-hppDWS, respectively. The double-stranded,

449 dam-methylated recombinant plasmids were isolated from E. coli DH5α, confirmed by

450 DNA sequencing, and used as the PCR template to construct all mutants. Mutagenesis

451 was extended by KOD-Plus-Neo DNA polymerase (TOYOBO, Japan) using a pair of

452 mutagenic primers containing the appropriate base changes (Table S3). The products

453 were treated with DpnI (Thermo Fisher Scientific, USA) at 37 °C for 30 min, and nicked

454 plasmid DNA was transformed into E. coli DH5α for automatic repair. The plasmids

455 harboring expected mutations were identified by restriction analysis and DNA sequencing.

456 The mutated fragments were digested and ligated into expression vector pET26b

457 (Novagen, Germany) at NdeI and HindIII restriction sites. The ligation products were

458 transformed into E. coli BL21 (DE3), which were selected for transformants. Correct

459 transformants, confirmed by PCR, were cultivated at 37 °C overnight in 5 ml LB medium

460 containing 50 μg/ml kanamycin, and then transferred to the same fresh medium for

461 further culture until the optical density at 600 nm (OD600) reached ~0.6. Isopropyl

462 β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM to

22 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

463 induce enzyme expression. Subsequently, cultures were incubated at the indicated

464 temperatures (22, 30, and 37 °C, respectively) for another 36 h to examine the production

465 of melanin.

466 Construction of A.s.s strains with different hppD alleles

467 Deletion, site-directed mutation, and replacement of the hppD gene in A.s.s AB98041

468 were all performed by the double homologous recombination method. To construct the

469 AB98041 hppD deletion strain, DNA fragments carrying 500 bp regions upstream and

470 downstream of the hppD open reading frame were amplified from the chromosomal DNA

471 of AB98041 with the primer pairs hppD(up)-F/hppD(up)-R and hppD(dn)-F/hppD(dn)-R

472 (Table S3), respectively. Upstream and downstream fragments were spliced by

473 overlap-extension PCR via the overlapping homologous sequences of hppD(up)-R and

474 hppD(dn)-F. The assembled fragment was digested with XbaI and SmaI and cloned into

475 the suicide vector pDM4 containing the sacB gene. The resultant plasmid pDM4-hppDAS

476 was transformed into the E. coli β2155(λpir) strain, which can grow only in the presence

477 of DAP (57). Two-parental mating was performed to transfer the recombinant plasmid

478 pDM4-hppDAS into AB98041. Briefly, the AB98041 recipient strain and the β2155

479 (harboring pDM4-hppDAS) donor strain were grown overnight in LB medium and

480 LB-chloramphenicol medium containing 100 μg/ml DAP, respectively. Cultures were

481 mixed in a 200 μl volume at a ratio of 1:1, centrifuged at 8,000 × g for 5 min, washed

482 twice with 400 μl of 10% glycerol, resuspended in 50 μl of 10% glycerol, and then

483 spotted on LB-DAP agar plates. After incubation at 22 °C for 12 h, cells were harvested

23 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

484 and resuspended in 2 ml of LB. Fifty microliters of culture was then plated on

485 LB-chloramphenicol agar to select single recombinants. Colonies were transferred to LB

486 agar plates and grown for 24 h for a second allelic exchange, and then transferred to LB

487 agar plates containing 15% sucrose. Double recombinants that could grow in the presence

488 of sucrose were identified. Potential hppD deletion mutants were confirmed by PCR

489 analysis and DNA sequencing.

490 To construct hppD mutant strains of AB98041, the hppD mutant alleles containing

491 different base mutations were amplified from the corresponding pUC-hppDAS variant

492 plasmids. The amplified fragments were then linked to upstream and downstream arms

493 by overlapping PCR to generate continuous products that contain the mutant alleles

494 flanked by two arms. Similarly, to replace the hppD gene of AB98041 with the hppD-WS,

495 the latter was amplified and inserted into the two arms. After sequencing, these constructs

496 were inserted into pDM4 and transformed into E. coli β2155. Conjugation was performed

497 between the β2155 strain and AB98041ΔhppD. The following procedures were identical

498 to the construction of deletion mutants. The correct insertion of hppD mutant alleles and

499 replacement of the hppD gene with hppD-WS in the AB98041 chromosome were all

500 verified by PCR and sequenced for confirmation.

501 hppD deletion and replacement with hppD-WS in A.s.s strains 2013-8 and 2014-235

502 were also performed using the same approach.

503 Purification and enzymatic characterization of wild-type and mutant enzymes

504 Colonies of E. coli BL21(DE3) expressing wild-type and mutant HppD enzymes were

24 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

505 cultivated overnight in 5 ml LB medium containing kanamycin (50 μg/ml). The overnight

506 cultures were inoculated into 100 ml of the same fresh medium until the OD600 reached

507 ~0.6. IPTG was then added to a final concentration of 0.5 mM. After another 5 h of

508 induction at 22 °C (16 ℃ and 12 h for HppD-AS), the cells were harvested from culture

509 broth by centrifugation at 8,000 × g for 20 min at 4 °C, and then resuspended in 20 mM

510 Tris-HCl buffer (pH 7.9) containing 500 mM NaCl, 10% glycerol, and 20 mM imidazole.

511 The resuspended cells were disrupted on ice by sonication and the resulting lysate was

512 centrifuged at 13,000 × g for 20 min at 4 °C to pellet insoluble material. The supernatant

513 was filtered through a 0.22-μm filter (Merck Millipore, Germany) and passed over a

514 Ni2+-NTA matrix (GE Healthcare, USA) according to the manufacturer’s instructions to

515 purify His6-tagged enzymes. The column was washed with 20 mM Tris-HCl buffer (pH

516 7.9) containing 500 mM NaCl, 10% glycerol, and 40 mM imidazole. Proteins bound to

517 the column were eluted with an imidazole gradient (80 to 500 mM) in buffer (20 mM

518 Tris-HCl, 500 mM NaCl, and 10% glycerol, [pH 7.9]). All fractions were collected and

519 analyzed by SDS-PAGE. Homogeneous peak fractions were pooled and concentrated

520 using a centrifugal filter (Merck Millipore, Germany) equipped with a 10,000 MWCO

521 membrane against 100 mM Tris-HCl buffer (pH 7.5) to simultaneously remove imidazole.

522 Protein concentrations were determined by the Bradford assay, with bovine serum

523 albumin as the standard.

524 HppD activity was measured as described previously (58) with some modifications.

525 Briefly, the assay was performed in 100 μl of reaction buffer containing 100 mM

25 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

526 Tris-HCl (pH 7.5), 1 mM Na-ascorbate, 250 μM ferrous sulfate, 1 mM

527 4-hydroxyphenylpyruvate (Sigma, USA), and 100 μg purified protein (depending on

528 specific activity). The mixture was incubated for 30 min and then the reaction was

529 terminated by the addition of acetonitrile (15% [v/v] final concentration). The

530 precipitated proteins were removed by centrifugation at 16,000 × g at 4 °C for 5 min and

531 the supernatant was subjected to HPLC analysis. An injection volume of 20 μl was used

532 on an Agilent 1100 liquid chromatograph fitted with a C18 reverse phase column (4.6 ×

533 250 mm; 5-μm particle size) and a DAD-UV detector. The mobile phase consisted of

534 90% (v/v) 10 mM acetic acid and 10% (v/v) methanol with a flow rate of 1 ml/min. HGA

535 production was monitored at 290 nm and quantification was performed by measuring

536 peak areas. The areas were transformed to nmol based on a standard curve obtained with

537 commercially available HGA (Sigma, USA). One unit (U) of enzyme activity was defined

538 as the amount of HGA (μmol) generated by per g enzyme per min. The optimal

539 temperature of each enzyme was determined by measuring activity at temperatures

540 ranging from 16 and 53 °C. Thermostability was determined by incubating the enzyme at

541 40 °C, and aliquots were then taken at different time points to determine residual

542 activities at 22 °C as described previously herein.

543 Homology Modeling

544 The possible effects of substitutions of critical residues on protein structural stability were

545 evaluated by analyzing the three-dimensional structural model. A molecular model of the

546 HppD-AS enzyme was generated by homology modeling using SWISS-MODEL (59).

26 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

547 The HppD of Pseudomonas fluorescens A32 (60) (PDB accession number 1CJX) that

548 shared 51% identity with HppD-AS was used as the template. The PyMOL molecular

549 graphics system was used to visualize and analyze the structure of wild-type HppD-AS

550 and mutants.

551

552 Acknowledgments

553 This work was supported by the National Natural Science Foundation of China (No.

554 31770052), the National Fund for Fostering Talents of Basic Sciences (J1103513), and

555 Research (Innovative) Fund of Laboratory of Wuhan University. The funders had no role

556 in study design, data collection, and interpretation, or the decision to submit the work for

557 publication. We thank Dr. Tongyan Lu (Heilongjiang River Fisheries Research Institute,

558 Chinese Academy of Fishery Sciences) for kindly providing A.s.s strains 2013-8 and

559 2014-235. There are no conflicts of interest to declare.

560 Yunqian Qiao and Xiangdong Chen contributed to the conception and design of the

561 study. Yunqian Qiao, Jiao Wang, He Wang, Baozhong Chai, and Chufeng Rao performed

562 the experiments and generated the data. Yunqian Qiao, Xiangdong Chen, and Shishen Du

563 analyzed the data and wrote the paper.

564

565 References

566 1. Singh S, Malhotra AG, Pandey A, Pandey KM. 2013. Computational model for

567 pathway reconstruction to unravel the evolutionary significance of melanin synthesis.

27 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

568 Bioinformation 9:94-100.

569 2. Nosanchuk JD, Casadevall A. 2006. Impact of melanin on microbial virulence and

570 clinical resistance to antimicrobial compounds. Antimicrob Agents Chemother

571 50:3519-3528.

572 3. Kwon-Chung KJ, Rhodes JC. 1986. Encapsulation and melanin formation as indicators

573 of virulence in Cryptococcus neoformans. Infect Immun 51:218-223.

574 4. Dixon DM, Migliozzi J, Cooper CR, Jr., Solis O, Breslin B, Szaniszlo PJ. 1992.

575 Melanized and non-melanized multicellular form mutants of Wangiella dermatitidis in

576 mice: mortality and histopathology studies. Mycoses 35:17-21.

577 5. Valeru SP, Rompikuntal PK, Ishikawa T, Vaitkevicius K, Sjöling Å, Dolganov N, Zhu

578 J, Schoolnik G, Wai SN. 2009. Role of melanin pigment in expression of Vibrio

579 cholerae virulence factors. Infect Immun 77:935-942.

580 6. Chai B, Qiao Y, Wang H, Zhang X, Wang J, Wang C, Zhou P, Chen X. 2017.

581 Identification of YfiH and the catalase CatA as polyphenol oxidases of Aeromonas

582 media and CatA as a regulator of pigmentation by its peroxyl radical scavenging

583 capacity. Front Microbiol 8:1939.

584 7. Wang H, Qiao Y, Chai B, Qiu C, Chen X. 2015. Identification and molecular

585 characterization of the homogentisate pathway responsible for pyomelanin production,

586 the major melanin constituents in Aeromonas media WS. PLoS One 10:e0120923.

587 8. Janda JM, Abbott SL. 2010. The genus Aeromonas: taxonomy, pathogenicity, and

588 infection. Clin Microbiol Rev 23:35-73.

28 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

589 9. Abbott SL, Cheung WK, Janda JM. 2003. The genus Aeromonas: biochemical

590 characteristics, atypical reactions, and phenotypic identification schemes. J Clin

591 Microbiol 41:2348-2357.

592 10. Martin-Carnahan A, Joseph SW. 2005. Family I. Aeromonadaceae Colwell,

593 MacDonell and De Ley 1986, p 556-578. In Brenner DJ, Krieg NR, Staley JT, Garrity

594 GM (ed), Bergey’s Manual of Systematic Bacteriology, 2nd ed, vol 2. Springer-Verlag,

595 New York, USA.

596 11. McIntosh D, Austin B. 1991. Atypical characteristics of the salmonid pathogen

597 Aeromonas salmonicida. J Gen Microbiol 137:1341-1343.

598 12. Griffin PJ, Snieszko SF, Friddle SB. 1953. Pigment formation by Bacterium

599 salmonicida. J Bacteriol 65:652-659.

600 13. Griffin PJ, Snieszko SF, Friddle SB. 1953. A more comprehensive description of

601 Bacterium salmonicida. Trans Am Fish Soc 82:129-138.

602 14. Daher RK, Filion G, Tan SG, Dallaire-Dufresne S, Paquet VE, Charette SJ. 2011.

603 Alteration of virulence factors and rearrangement of pAsa5 plasmid caused by the

604 growth of Aeromonas salmonicida in stressful conditions. Vet Microbiol 152:353-360.

605 15. Stuber K, Burr SE, Braun M, Wahli T, Frey J. 2003. Type III secretion genes in

606 Aeromonas salmonicida subsp. salmonicida are located on a large thermolabile

607 virulence plasmid. J Clin Microbiol 41:3854-3856.

608 16. Ishiguro EE, Kay WW, Ainsworth T, Chamberlain JB, Austen RA, Buckley JT, Trust

609 TJ. 1981. Loss of virulence during culture of Aeromonas salmonicida at high

29 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

610 temperature. J Bacteriol 148:333-340.

611 17. Moki ST, Nomura T, Yoshimizu M. 1995. Effect of incubation temperature for

612 isolation on auto-agglutination of Aeromonas salmonicida. Fish Pathol 30:67-68.

613 18. Anbar M, Gul O, Lamed R, Sezerman UO, Bayer EA. 2012. Improved

614 thermostability of Clostridium thermocellum endoglucanase Cel8A by using

615 consensus-guided mutagenesis. Appl Environ Microbiol 78:3458-3464.

616 19. Vázquez-Figueroa E, Chaparro-Riggers J, Bommarius AS. 2007. Development of a

617 thermostable glucose dehydrogenase by a structure-guided consensus concept.

618 ChemBioChem 8:2295-2301.

619 20. Amin N, Liu AD, Ramer S, Aehle W, Meijer D, Metin M, Wong S, Gualfetti P,

620 Schellenberger V. 2004. Construction of stabilized proteins by combinatorial consensus

621 mutagenesis. Protein Eng Des Sel 17:787-793.

622 21. Chou PY, Fasman GD. 1974. Conformational parameters for amino acids in helical,

623 β-sheet, and random coil regions calculated from proteins. Biochemistry 13:211-222.

624 22. Munoz V, Serrano L. 1994. Intrinsic secondary structure propensities of the amino

625 acids, using statistical ϕ–ψ matrices: Comparison with experimental scales. Proteins:

626 Struct Funct Genet 20:301-311.

627 23. Kim CA, Berg JM. 1993. Thermodynamic β-sheet propensities measured using a

628 zinc-finger host peptide. Nature 362:267-270.

629 24. Koh E, Kim T, Cho HS. 2006. Mean curvature as a major determinant of β-sheet

630 propensity. Bioinformatics 22:297-302.

30 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

631 25. Robson B, Suzuki E. 1976. Conformational properties of amino acid residues in

632 globular proteins. J Mol Biol 107:327-356.

633 26. O'Neil KT, DeGrado WF. 1990. A thermodynamic scale for the helix-forming

634 tendencies of the commonly occurring amino acids. Science 250:646-651.

635 27. von Heijne G. 1991. Proline kinks in transmembrane α-helices. J Mol Biol

636 218:499-503.

637 28. Sankararamakrishnan R, Vishveshwara S. 1992. Geometry of proline-containing

638 alpha-helices in proteins. Int J Pept Protein Res 39:356-363.

639 29. Schimmel PR, Flory PJ. 1968. Conformational energies and configurational statistics

640 of copolypeptides containing L-proline. J Mol Biol 34:105-120.

641 30. Matthews BW, Nicholson H, Becktel WJ. 1987. Enhanced protein thermostability

642 from site-directed mutations that decrease the entropy of unfolding. Proc Natl Acad Sci

643 USA 84:6663-6667.

644 31. Zhou C, Xue Y, Ma Y. 2010. Enhancing the thermostability of α-glucosidase from

645 Thermoanaerobacter tengcongensis MB4 by single proline substitution. J Biosci

646 Bioeng 110:12-17.

647 32. Yu H, Zhao Y, Guo C, Gan Y, Huang H. 2015. The role of proline substitutions

648 within flexible regions on thermostability of luciferase. Biochim Biophys Acta

649 1854:65-72.

650 33. Watanabe K, Hata Y, Kizaki H, Katsube Y, Suzuki Y. 1997. The refined crystal

651 structure of Bacillus cereus oligo-1,6-glucosidase at 2.0 Å resolution: structural

31 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

652 characterization of proline-substitution sites for protein thermostabilization. J Mol Biol

653 269:142-153.

654 34. De Vendittis E, Castellano I, Cotugno R, Ruocco MR, Raimo G, Masullo M. 2008.

655 Adaptation of model proteins from cold to hot environments involves continuous and

656 small adjustments of average parameters related to amino acid composition. J Theor

657 Biol 250:156-171.

658 35. Metpally RP, Reddy BV. 2009. Comparative proteome analysis of psychrophilic

659 versus mesophilic bacterial species: Insights into the molecular basis of cold adaptation

660 of proteins. BMC Genomics 10:11.

661 36. O'HIci B, Olivier G, Powell R. 2000. Genetic diversity of the fish pathogen

662 Aeromonas salmonicida demonstrated by random amplified polymorphic DNA and

663 pulsed-field gel electrophoresis analyses. Dis Aquat Organ 39:109-119.

664 37. Colston SM, Fullmer MS, Beka L, Lamy B, Gogarten JP, Graf J. 2014. Bioinformatic

665 genome comparisons for taxonomic and phylogenetic assignments using Aeromonas as

666 a test case. MBio 5:e02136.

667 38. Garcia JA, Larsen JL, Dalsgaard I, Pedersen K. 2000. Pulsed-field gel

668 electrophoriesis analyis of Aeromonas salmonicida ssp. salmonicida. FEMS Microbiol

669 Lett 190:163-166.

670 39. Wiklund T, Dalsgaard I. 1998. Occurrence and significance of atypical Aeromonas

671 salmonicida in non-salmonid and salmonid fish species: a review. Dis Aquat Organ

672 32:49-69.

32 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

673 40. Wiklund T, Lönnström L, Niiranen H. 1993. Aeromonas salmonicida ssp.

674 salmonicida lacking pigment production, isolated from farmed salmonids in Finland.

675 Dis Aquat Org 15:219-223.

676 41. Koppang EO, Fjølstad M, Melgård B, Vigerust M, Sørum H. 2000.

677 Non-pigment-producing isolates of Aeromonas salmonicida subspecies salmonicida:

678 isolation, identification, transmission and pathogenicity in Atlantic salmon, Salmo

679 salar L. J Fish Dis 23:39-48.

680 42. Pavan ME, Pavan EE, López NI, Levin L, Pettinari MJ. 2015. Living in an extremely

681 polluted environment: clues from the genome of melanin-producing Aeromonas

682 salmonicida subsp. pectinolytica 34melT. Appl Environ Microbiol 81:5235-5248.

683 43. Martínez-Murcia AJ, Soler L, Saavedra MJ, Chacón MR, Guarro J, Stackebrandt E,

684 Figueras MJ. 2005. Phenotypic, genotypic, and phylogenetic discrepancies to

685 differentiate Aeromonas salmonicida from Aeromonas bestiarum. Int Microbiol

686 8:259-269.

687 44. Belland RJ, Trust TJ. 1988. DNA:DNA reassociation analysis of Aeromonas

688 salmonicida. J Gen Microbiol 134:307-315.

689 45. Hänninen ML, Hirvelä-Koski V. 1997. Molecular and phenotypic methods for the

690 characterization of atypical Aeromonas salmonicida. Vet Microbiol 56:147-158.

691 46. Høie S, Dalsgaard I, Aase IL, Heum M, Thornton JM, Powell R. 1999. Polymerase

692 chain reaction (PCR)-based typing analysis of atypical isolates of the fish pathogen

693 Aeromonas salmonicida. Syst Appl Microbiol 22:403-411.

33 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

694 47. Sørum H, Holstad G, Lunder T, Håstein T. 2000. Grouping by plasmid profiles of

695 atypical Aeromonas salmonicida isolated from fish, with special reference to salmonid

696 fish. Dis Aquat Org 41:159-171.

697 48. Umelo E, Trust TJ. 1998. Physical map of the chromosome of Aeromonas

698 salmonicida and genomic comparisons between Aeromonas strains. Microbiology

699 144:2141-2149.

700 49. Lund V, Jenssen LM, Wesmajervi MS. 2002. Assessment of genetic variability and

701 relatedness among atypical Aeromonas salmonicida from marine fishes, using

702 AFLP-fingerprinting. Dis Aquat Org 50:119-126.

703 50. Studer N, Frey J, Vanden Bergh P. 2013. Clustering subspecies of Aeromonas

704 salmonicida using IS630 typing. BMC Microbiol 13:36.

705 51. Martinez-Murcia AJ, Benlloch S, Collins MD. 1992. Phylogenetic interrelationships

706 of members of the genera Aeromonas and Plesiomonas as determined by 16S

707 ribosomal DNA sequencing: lack of congruence with results of DNA-DNA

708 hybridizations. Int J Syst Bacteriol 42:412-421.

709 52. Gulla S, Lund V, Kristoffersen AB, Sorum H, Colquhoun DJ. 2016. vapA (A-layer)

710 typing differentiates Aeromonas salmonicida subspecies and identifies a number of

711 previously undescribed subtypes. J Fish Dis 39:329-342.

712 53. Rodriguez-Rojas A, Mena A, Martin S, Borrell N, Oliver A, Blazquez J. 2009.

713 Inactivation of the hmgA gene of Pseudomonas aeruginosa leads to pyomelanin

714 hyperproduction, stress resistance and increased persistence in chronic lung infection.

34 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

715 Microbiology 155:1050-1057.

716 54. Wang R, Wang H, Zhou H, Wang Y, Yue J, Diao B, Kan B. 2011. Characters of

717 homogentisate oxygenase gene mutation and high clonality of the natural

718 pigment-producing Vibrio cholerae strains. BMC Microbiol 11:109.

719 55. Han H, Iakovenko L, Wilson AC. 2015. Loss of homogentisate 1,2-dioxygenase

720 activity in Bacillus anthracis results in accumulation of protective pigment. PLoS One

721 10:e0128967.

722 56. Reith ME, Singh RK, Curtis B, Boyd JM, Bouevitch A, Kimball J, Munholland J,

723 Murphy C, Sarty D, Williams J, Nash JH, Johnson SC, Brown LL. 2008. The genome

724 of Aeromonas salmonicida subsp. salmonicida A449: insights into the evolution of a

725 fish pathogen. BMC Genomics 9:427.

726 57. Dehio C, Meyer M. 1997. Maintenance of broad-host-range incompatibility group P

727 and group Q plasmids and transposition of Tn5 in Bartonella henselae following

728 conjugal plasmid transfer from Escherichia coli. J Bacteriol 179:538-540.

729 58. Gunsior M, Ravel J, Challis GL, Townsend CA. 2004. Engineering

730 p-hydroxyphenylpyruvate dioxygenase to a p-hydroxymandelate synthase and

731 evidence for the proposed benzene oxide intermediate in homogentisate formation.

732 Biochemistry 43:663-674.

733 59. Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, Kiefer F, Gallo

734 Cassarino T, Bertoni M, Bordoli L, Schwede T. 2014. SWISS-MODEL: modelling

735 protein tertiary and quaternary structure using evolutionary information. Nucleic Acids

35 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

736 Res 42:W252-W258.

737 60. Serre L, Sailland A, Sy D, Boudec P, Rolland A, Pebay-Peyroula E, Cohen-Addad C.

738 1999. Crystal structure of Pseudomonas fluorescens 4-hydroxyphenylpyruvate

739 dioxygenase: an enzyme involved in the tyrosine degradation pathway. Structure

740 7:977-988.

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

36 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

757 Figure Legends

758 Figure 1. Amino acid sequence alignment of HppD from A.s.s and other fifteen

759 Aeromonas species. The sequences were aligned using the ClustalW algorithm. The

760 identical residues are plotted as dots. The mutant positions selected are highlighted in

761 gray. Asterisks mark the catalytic triad residues. The numbers above the sequences refer

762 to the amino acid positions in HppD-AS. The full strain names and accession numbers of

763 HppD are listed in Table S2.

764 Figure 2. The melanin production of E. coli strains expressing hppD-AS variants.

765 Recombinant E. coli strains expressing the single mutants (A) and double/triple mutants

766 (B) were cultured at 22, 30, and 37 ℃ after the induction of IPTG. E. coli strains

767 containing the empty vector or expressing wild type hppD-AS and hppD-WS were served

768 as controls.

769 Figure 3. The melanin synthesis phenotype of E. coli strains expressing hppD-WS

770 variants. The corresponding residues of HppD-WS were replaced with the amino acids

771 of HppD-AS. Strains were induced by IPTG at different temperatures (22, 30, and 37 ℃,

772 respectively). E. coli strains carrying an empty vector or expressing wild type hppD-AS

773 and hppD-WS were employed as controls.

774 Figure 4. hppD-AS mutants result in temperature-independent melanogenesis in

775 A.s.s strains. Shown are pigmentation phenotypes of wild-type strain AB98041 and its

776 hppD mutants (A), wild-type strain 2013-8 and its hppD mutants (B), wild-type strain

777 2014-235 and its hppD mutants (C), which were cultivated in LB medium for 48 h at 22

37 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

778 and 30 ℃, respectively.

779 Figure 5. Effect of temperature on the activities and stabilities of wild-type and

780 mutant HppD enzymes. (A) Optimum temperatures of several representative enzymes.

781 Activity was measured in 100 mM Tris-HCl (pH 7.5) at 16, 22, 30, 37, 45, and 53 ℃ for

782 30 min, respectively. The total activity at the optimum temperature is set as 100%. (B)

783 Thermostabilities of HppD enzymes. The proteins were incubated at 40 °C, and aliquots

784 were then taken at different time points to determine residual activities at 22 °C. The

785 activity of each enzyme is normalized to its activity prior to incubation, which is defined

786 as 100%. The error bars show the standard deviations of triplicate assays.

787 Figure 6. The structural model of wild-type HppD-AS. The model structure was

788 generated based on the HppD structure of Pseudomonas fluorescens (PDB ID, 1CJX).

789 The two barrel-like domains, N-terminal domain and C-terminal domain, are depicted in

790 blue and beige, respectively. The catalytic triad residues (H174, H252, and E331) are

791 highlighted by red sticks. The three substituted residues (S18, P103, and L119) are

792 marked in yellow.

793 Figure 7. Potential impact of the substitutions on the structure of HppD-AS. Shown

794 are close-up views of local structures around the positions 18, 103, and 119 before (A, C,

795 E) and after (B, D, F) mutations, respectively. The original and substituted residues are

796 indicated by yellow sticks. The green dashed lines represent hydrogen bonds.

797

798

38 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

799 Table 1. Bacterial strains and plasmids used in this study. Strains or plasmids Characteristics Reference or source A.s.s AB98041 Wild type strain, producing melanin at CCTCC lower temperatures AB98041ΔhppD AB98041 derivative strain, hppD this study was deleted AB98041(hppD-6)/ AB98041 derivative strains, hppD this study AB98041(hppD-46)/ were mutated at respective sites AB98041(hppD-68) AB98041(hppD-WS) AB98041 derivative strain, hppD this study was replaced by hppD from A. media WS KACC 14791 Wild type strain, producing melanin at KACC lower temperatures 2013-8 Wild type strain, producing melanin at Kindly donated by lower temperatures Dr. Tongyan Lu from Heilongjiang River Fisheries Research Institute, Chinese 2014-235 Wild type strain, producing melanin at Academy of Fishery lower temperatures Sciences. Isolated from the liver of Rainbow trout. 2013-8ΔhppD 2013-8 derivative strain, hppD was this study deleted 2013-8(hppD-WS) 2013-8 derivative strain, hppD was this study replaced by hppD from A. media WS 2014-235ΔhppD 2014-235 derivative strain, hppD was this study deleted 2014-235(hppD-WS) 2014-235 derivative strain, hppD was this study replaced by hppD from A. media WS E. coli DH5α Cloning host; Commercially recA1 endA1 gyrA96 thi-1 hsdR17 available supE44 relA1 ΔlacU169 (φ80lacZΔM15) β2155(λpir) thrB1004 pro thi hsdS lacZΔM15 (57) (F′lacZΔM15 lacIq traD36 proA+proB+)

39 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

dap::erm recA::RP4-2-tet::Mu-km λpir; Ermr Tetr Kanr BL21(DE3) Expression host; F-ompT hsdSB Commercially - (rB-mB ) gal dcm (DE3) available Plasmids pUC18 Cloning vecor; Ampr Commercially available pUC-hppDAS pUC18 with the hppD cloning from this study AB98041, the template for mutation pUC-hppDWS pUC18 with the hppD cloning from this study A. media WS, the template for mutation pET26b Expression vector; Kmr Commercially available pET-Mut1~Mut16 pET-hppDAS derivatives with this study hppD carrying different mutations pET-W4/W6/W8/W46/ pET-hppDWS derivatives with this study W48/W68/W468 hppD carrying different mutations pDM4 Suicide plasmid, pir dependent with Kindly donated by sacAB genes, oriR6K; Cmr Dr. Shiyun Chen from Wuhan Institute of Virology, Chinese Academy of Sciences pDM4-hppD-AS pDM4 based plasmid for deletion of this study hppD in A.s.s strains pDM4-hppD-A6/A46/ pDM4 based plasmid for site-directed this study A68 mutations of hppD in A.s.s strains pDM4-hppD-WS pDM4 based plasmid for replacement this study of hppD with hppD-WS in A.s.s strains 800

801

802

803

804

805

40 bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 1. Amino acid sequence alignment of HppD from A.s.s and other fifteen

Aeromonas species. The sequences were aligned using the ClustalW algorithm. The

identical residues are plotted as dots. The mutant positions selected are highlighted in

gray. Asterisks mark the catalytic triad residues. The numbers above the sequences refer

to the amino acid positions in HppD-AS. The full strain names and accession numbers of

HppD are listed in Table S2. bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 2. The melanin production of E. coli strains expressing hppD-AS variants.

Recombinant E. coli strains expressing the single mutants (A) and double/triple mutants

(B) were cultured at 22, 30, and 37 ℃ after the induction of IPTG. E. coli strains

containing the empty vector or expressing wild type hppD-AS and hppD-WS were served

as controls.

bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 3. The melanin synthesis phenotype of E. coli strains expressing hppD-WS

variants. The corresponding residues of HppD-WS were replaced with the amino acids

of HppD-AS. Strains were induced by IPTG at different temperatures (22, 30, and 37 ℃,

respectively). E. coli strains carrying an empty vector or expressing wild type hppD-AS

and hppD-WS were employed as controls.

bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 4. hppD-AS mutants result in temperature-independent melanogenesis in

A.s.s strains. Shown are pigmentation phenotypes of wild-type strain AB98041 and its

hppD mutants (A), wild-type strain 2013-8 and its hppD mutants (B), wild-type strain

2014-235 and its hppD mutants (C), which were cultivated in LB medium for 48 h at 22

and 30 ℃, respectively.

bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 5. Effect of temperature on the activities and stabilities of wild-type and

mutant HppD enzymes. (A) Optimum temperatures of several representative enzymes.

Activity was measured in 100 mM Tris-HCl (pH 7.5) at 16, 22, 30, 37, 45, and 53 ℃ for

30 min, respectively. The total activity at the optimum temperature is set as 100%. (B)

Thermostabilities of HppD enzymes. The proteins were incubated at 40 ° C , and aliquots

were then taken at different time points to determine residual activities at 22 °C. The

activity of each enzyme is normalized to its activity prior to incubation, which is defined

as 100%. The error bars show the standard deviations of triplicate assays.

bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 6. The structural model of wild-type HppD-AS. The model structure was

generated based on the HppD structure of Pseudomonas fluorescens (PDB ID, 1CJX).

The two barrel-like domains, N-terminal domain and C-terminal domain, are depicted in

blue and beige, respectively. The catalytic triad residues (H174, H252, and E331) are

highlighted by red sticks. The three substituted residues (S18, P103, and L119) are

marked in yellow.

bioRxiv preprint doi: https://doi.org/10.1101/324731; this version posted May 17, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 7. Potential impact of the substitutions on the structure of HppD-AS. Shown

are close-up views of local structures around the positions 18, 103, and 119 before (A, C,

E) and after (B, D, F) mutations, respectively. The original and substituted residues are

indicated by yellow sticks. The green dashed lines represent hydrogen bonds.