The Dimethylsulfoniopropionate (DMSP) Lyase and Lyase-Like Cupin

Mar 21, 2018 - Here, we attempted to define and map this superfamily dubbed cupin-DLL (DMSP lyases and lyase-like). To this end, we have pursued the c...
0 downloads 4 Views 6MB Size
Subscriber access provided by Université de Strasbourg - Service Commun de la Documentation

The DMSP Lyase and Lyase-Like Cupin family consists of bona fide DMSP lyases as well as other enzymes with unknown function Lei Lei, Kesava Phaneendra Cherukuri, Uria Alcolombri, Diana Meltzer, and Dan S. Tawfik Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.8b00097 • Publication Date (Web): 21 Mar 2018 Downloaded from http://pubs.acs.org on March 22, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

1 2

The DMSP Lyase and Lyase-Like Cupin family consists of bona fide DMSP lyases as well as other enzymes with unknown function

3 4

Lei Lei, Kesava Phaneendra Cherukuri, Uria Alcolombri, Diana Meltzer & Dan S. Tawfik

5 6

Department of Biomolecular Sciences, Weizmann Institute of Science, Rehovot 76100, Israel.

7

*Correspondence to: [email protected]

8 9

Abstract

10

Marine

11

dimethylsulfoniopropionate (DMSP). Different genes encoding proteins with DMSP lyase

12

activity are known, yet these exhibit highly variable levels of activity. Most assigned

13

bacterial DMSP lyases, including DddK, DddL, DddQ, DddW and DddY, appear to belong

14

to one, cupin-like superfamily. Here, we attempted to define and map this superfamily

15

dubbed Cupin-DLL: DMSP lyases and lyase-like. To this end, we have pursued the

16

characterization of various recombinant DMSP lyases belonging to this superfamily of

17

metallo-enzymes, and especially of DddY and DddL that seem to be the most active

18

DMSP lyases in this superfamily. We identified two conserved sequence motifs that

19

characterize this superfamily. These motifs include the metal-ligating residues that are

20

absolutely essential, and other residues including an active site tyrosine (Y131 in DddQ)

21

that seems to play a relatively minor role in DMSP lysis. We also identified a transition

22

metal-chelator, TPEN, that selectively inhibits all known members of the Cupin-DLL

23

superfamily that exhibit DMSP lyase activity. A phylogenetic analysis indicated that the

24

known DMSP lyase families are sporadically distributed suggesting that DMSP lyases

25

evolved within this superfamily multiple times. However, unusually low specific DMSP

26

lyase activity, and genome context analysis, suggest that DMSP lyase is not the native

27

function of most cupin-DLL families. Indeed, a systematic profiling of substrate

28

selectivity with a series of DMSP analogues indicated that some members, most

organisms

release

dimethylsulfide

(DMS)

ACS Paragon Plus Environment

via

cleavage

of

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

29

distinctly DddY and DddL, are bona fide DMSP lyases, while others, foremost DddQ, may

30

only exhibit promiscuous DMSP lyase activity.

31 32

Key words: Dimethylsulfide, enzyme superfamily, enzyme evolution, enzyme

33

promiscuity.

34 35

Abbreviations: DMSP, dimethylsulfoniopropionate; DMS, dimethylsulfide; Cupin-DLL,

36

cupin DMSP lyase and lyase-like. TPEN, N,N,N',N'-tetrakis(2-pyridylmethyl)ethane-1,2-

37

diamine; NMR, Nuclear magnetic resonance; HMRS, High-resolution mass spectrometry.

38 39

Funding: Financial support by the Estate of Mark Scher, and the Sasson & Marjorie

40

Peress Philanthropic Fund, are gratefully acknowledged. D.S.T. is the Nella and Leon

41

Benoziyo Professor of Biochemistry.

42 43

Introduction

44

Dimethyl sulfide (DMS) is a key regulator of marine life and possibly also a climate

45

regulator. DMSP (dimethylsulfoniopropionate) is the primary precursor of DMS, and is

46

synthesized mostly by marine phytoplankton (single-cell algae) but also by macroalgae,

47

corals, some angiosperms and bacteria.1-5 DMSP probably acts as osmolyte and

48

antioxidant, although its precise physiological role remains unknown as is the role of

49

DMS release. 2, 4 Also unknown is the relative contribution of different marine species to

50

the global DMS release, or even of bacterial relative to algal mediated release.6 Several

51

putative bacterial DMSP lyase families were identified that catalyze the elimination of

52

DMSP to generate DMS and acrylate (Scheme 1). These were assigned as ‘Ddd’ (DMSP

53

dependence DMS releasing) genes,7 as well as one eukaryote family, dubbed Alma,

54

found in organisms such as algae and corals.8

ACS Paragon Plus Environment

Page 2 of 41

Page 3 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

55 56

We aimed at better understanding the enzymology and phylogenetics of the

57

most abundant bacterial DMSP lyases. The known bacterial DMSP lyases belong to

58

several different superfamilies: DddD, a DMSP CoA-transferase-lyase;7,

59

belongs to the M24 proteinase family;10 DddQ, DddL, DddW and DddK that constitute

60

the largest DMSP lyases class, and possess a cupin motif.11-14 Most recently, DddY, has

61

also been classified as belonging to the cupin family.4,

62

evolutionary origins, these bacterial enzymes exhibit highly variable levels of DMSP

63

activity, especially the cupin DMSP lyases. The reported specific activities seem to range

64

from 0.002 up to 0.028 Units (µM DMS min-1mg-1 enzyme) for DddQ 19-21 up to 675 Units

65

for DddY.15, 16 Kinetic parameters are known for only some of these enzymes, but the

66

specific activities relate to kcat/KM values that may be as low as or even lower than 1 M-

67

1 -1

68

been systematically characterized. Foremost, the low DMSP lyases activity of some

69

members seems insufficient to support their assignment as DMSP lyases.

s

15-18

9

DddP that

In addition to variable

for DddQ,19-21 up to 106 M-1s-1 for DddY.15, 16 However, some Ddd families have not

70

Within enzyme superfamilies, promiscuity, i.e., the ability to transform

71

substrates other than the native substrate, or even to catalyze a different reaction, is

72

widespread. 22-24 Promiscuity is also postulated to be a major driving force in the natural

73

evolution of new enzyme functions. 22, 23, 25 A common key chemical step allows a range

74

of reactions to be catalyzed by the same configuration of catalytic active-site residues.

75

Owing to the shared chemistry, the native function of one superfamily member is often

76

manifested as a promiscuous activity in related members and vice versa. The prevalence

77

of a given promiscuous activity in different members of the same superfamily may result

78

in multiple independent emergences of new families for which this promiscuous activity

79

has become the native function. Promiscuity has become instrumental in unraveling the

80

evolutionary origins of enzymes,26 but it has also led to mis-annotation of enzymes

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

81

whereby a promiscuous activity has been defined as the native function (for example,

82

see Ref. 27 ).

83

To deepen our understanding of the enzymology of DMS release, we have

84

systematically characterized DddY, DddQ, DddL, DddW and DddK – the five gene

85

families that have been assigned as cupin DMSP lyases. Together, these five families

86

define a new superfamily of enzymes that share the conserved cupin metal active site

87

and the jellyroll fold. However, analysis of the genome context, and a systematic

88

substrate profiling with a series of DMSP analogues, have both indicated that some of

89

these families most likely exhibit promiscuous rather than native DMSP lyases. Overall,

90

our results indicate that DddL and DddY are clearly bona fide DMSP lyases. However,

91

other families, most distinctly DddQ, appear to exhibit DMSP lyase activity as a

92

promiscuous, side-activity. This promiscuous activity stems from the shared active site

93

architecture, and the overlapping substrate or/and reaction patterns of members of this

94

newly identified superfamily dubbed the Cupin-DLL (Cupin DMSP lyase and lyase-like)

95

superfamily.

96 97

Material and Methods

98

Phylogenetic analysis. Hypothetical cupin-like DMSP lyases sequences were collected by

99

using a Hidden Markov Model based search via the HMMsearch of the HMMer

100

package.28 Briefly, the existing HMM model of DMSP lyase, PF16867, was used to search

101

the NCBI non-redundant protein database and the Tara metagenomics database.29 All

102

hits were collected, aligned, and manually filtered by the presence of the two conserved

103

motifs and a minimal length of 90 amino acids. The sequence redundancy in the

104

remaining set was minimized using CD-hit with 70% cutoff. The resulting 361 sequences

105

were aligned by MUSCLE,30 followed by minimization of gaps (manual removal of

106

sporadic insertions) to obtain a core alignment of 115 amino acids length (the smallest

107

known Cupin Ddd+ enzymes (DddK’s) are 130 amino acids length). A phylogenetic tree

108

was built from this core alignment using the Markov chain Monte Carlo (MCMC) method

109

of the MrBayes program.31

ACS Paragon Plus Environment

Page 4 of 41

Page 5 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

110

Enzyme cloning and mutagenesis. All tested dddY genes (including DadddY, AfdddY,

111

Shewanella dddY, Ferrimonas dddY, and Synechococcus dddY) were synthesized by Gen9

112

(USA). The Ruegeria pomeroyi DSS-3 dddQ and the Rhodobacter sphaeroides 2.4.1 dddL

113

genes were kindly provided by Professor Andrew Johnston, University of East Anglia.

114

The synthesized genes were amplified and the PCR product was digested with NcoI and

115

HindIII, and then cloned into the expression vector pET28a for expression with a C-

116

terminal His-tag. These plasmids were also used as the template for site-directed

117

mutagenesis. The TpdddL gene was synthesized by Gen9 (USA) and contained an N-

118

terminal Strep-tag and a stop codon before the HindIII site. The genes dddK, dddW and

119

genes selected from other Cupin-DLL families were synthesized by Gen9 (USA). The

120

synthesized gene fragments dddK and dddW were amplified and the PCR product was

121

digested with NcoI and HindIII, and then cloned into the expression vector pET28a. Point

122

mutations in DadddY, TpdddL, dddK and dddW were introduced with mutagenesis oligos

123

using the SOE-PCR-based approach. The PCR fragments were digested with NcoI and

124

HindIII and cloned into pET28a. All mutants were verified by DNA sequencing.

125

Enzyme purification. Enzymes were typically expressed using pET28a plasmids in E. coli

126

BL21 (DE3). Cells were grown for overnight in 5 mL LB medium at 37°C. These cultures (1

127

mL) were used to inoculate 1 liter LB cultures that were subsequently grown at 37°C to

128

OD600nm of 0.6-0.8. The growth temperature was reduced to 16°C, and enzyme

129

expression was induced with 0.1 mM IPTG. Following overnight growth at 16°C, the cells

130

were harvested by centrifugation at 4°C. DddY, DddK, DddW, DddQ, and other

131

representatives of cupin-DLL families were purified on Ni-NTA beads. Briefly, 2 g cells

132

were re-suspended in 50 mL lysis buffer (100 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM

133

CaCl2, 10 mg lysozyme and 10 µg bezonase). Cell suspensions were incubated in ambient

134

temperature for 30 min and sonicated. Lysates were clarified by centrifugation and

135

loaded on 2 mL Ni-NTA agarose beads (Millipore). Binding was performed at 4° C, beads

136

were washed with 50 mL lysis buffer followed by 100 mL lysis buffer with 35 mM

137

imidazole. The bound enzymes were eluted with 150 mM imidazole. Fractions

138

containing the expressed DMSP lyases were analyzed by SDS-PAGE and by enzymatic

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

139

activity, combined, and the purified enzyme was concentrated by ultrafiltration

140

(Amicon). Typical yields of purified enzymes were: 8 mg DddY, 10 mg DddQ, 20 mg DddK,

141

5 mg DddW, with ≥ 80% purity. For TpDddL with an N-terminal Strep-tag, clarified cell

142

lysates were loaded on 2 mL Strep-tactin beads (IBA), and the beads were rinsed with 50

143

mL lysis buffer. The protein was eluted with 2.5 mM desthiobiotin. Final enzyme

144

concentrations were determined by the BCA assay.32

145

DMSP lyase assays. DMS release was measured as previously described.10 Briefly,

146

freshly prepared 100 mM Tris-HCl buffer pH 8.0 with 100 mM NaCl was used

147

supplemented with 10 mM DMSP. Note that DMSP (and its analogues) are synthesized

148

as hydrochloride salts, and their cleavage releases protons. Thus, to maintain a constant

149

pH, high buffer capacity is critical. Reactions were performed at 30 °C and aliquots were

150

taken (typically for 5 min) and immediately diluted 1000-fold into 30 mL chilled 10 mM

151

glycine buffer pH 3.0 in sealed glass vials. Enzyme concentrations were typically as

152

follows: Ehux-Alma1, 0.3 ug/mL; Sym-Alma1, 1 μg/mL; Desufovibrio DddY 20 ng/mL;

153

Alcaligenes DddY 50 ng/mL; DddW, 15 μg/mL; DddQ, 100 μg/mL; DddK, 8 μg/mL. DddL

154

was assayed in crude lysate, at an estimated concentration of 8 μg/mL (Figure S1). DMS

155

levels were determined using an Eclipse 4660 Purge-and-Trap Sample Concentrator

156

system (OI Analytical) followed by separation and detection using GC-FPD (HP 5890)

157

equipped with RT-XL sulfur column (Restek). All measurements were calibrated using

158

DMS standards. To obtain the kinetic parameters, recombinant DddK, DddW and DddQ

159

enzymes (at 8, 15 and 200 µg/mL respectively were reacted with DMSP at different

160

concentrations up to 10 mM for 5 mins (at these enzyme concentrations, rates of

161

release were found to be linear up to 5 min). The amount of the released DMS was

162

measured to derive the initial rates (Figure S2). Activity measurements of DMSP

163

analogues were calibrated using the corresponding dialkylsulfides (ethyl methyl sulfide

164

for EMSP, diethyl sulfide for DESP, and tetrahydrothiopene for cDESP).

165

Metal chelation and reconstitution. Purified enzyme samples were incubated with

166

1mM EDTA or TPEN, at 30 °C for 1 hr. The chelation of the active-site metal was verified

167

by the complete loss of DMSP lyase activity. The chelators were subsequently removed

ACS Paragon Plus Environment

Page 6 of 41

Page 7 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

168

by dialysis at 4 °C overnight. Different metal ions were supplemented by adding the

169

corresponding chloride salts to the dialyzed apo-proteins at 2 mM concentration, and

170

incubating at 30 °C for 1 hr. The regain of ,DMSP lyase activity was subsequently

171

measured.

172

DMSP analogues synthesis. In general, all chemicals were purchased from Sigma-Aldrich

173

and other commercial suppliers in reagent grade and used without further purification.

174

Solvents are of AR grade and used without further purification. Deuterated solvents

175

were from Sigma and Cambridge Isotope Laboratories, Inc. 1H and

176

were recorded on BRUKER AVANCE III-400 (400 MHz) in D2O or CD3OD and all the signal

177

positions were recorded in δ ppm with abbreviations s, d, t, q, m and dd denoting singlet,

178

doublet, triplet, quartet, multiplet and doublet of doublet, respectively. All NMR

179

chemical shifts were referenced to residual solvent peaks (CD3OD δ =3.31 ppm and D2O

180

δ =4.79 ppm). Coupling constants J, were registered in Hz. HRMS was determined with

181

Xevo G2-XS QTOF mass spectrometer by electrospray ionization. The purity of all the

182

compounds was more than 90% by 1H NMR spectra.

13

C NMR spectra

183

The various DMSP analogues were synthesized by Michael addition of the

184

corresponding dialkylsulfide and acrylic acid, or acrylic acid derivative, essentially as

185

described.6 Typically, the corresponding acrylic acid (1 equivalent) was dissolved in 2 M

186

aqueous HCl (~10 mL for 1 gr), and the respective dialkylsulfide (3-6 molar equivalents)

187

was added portion-wise. The reaction mixture was refluxed at 80 °C for 3-12 hours.

188

After cooling to room temperature, the solvent and the excess of unreacted

189

dialkylsulfide were removed under reduced pressure. In the case of EMSP, DESP, cDESP,

190

the resulting crude product (hydrochloride salt; solid, or syrup) was purified by

191

recrystallization from isopropanol, and in the case of 2-methyl DMSP and 3-methyl

192

DMSP, the crude product hydrochloride salt (solid, or syrup) was triturated with diethyl

193

ether. The products were dried under vacuum to obtain the hydrochloride salts used for

194

the enzymatic assays. The products’ identity and purity were determined by 1H-NMR

195

(400 MHz), 13C-NMR (101 MHz) and HRMS analysis (NMR spectra are provided in Figure

196

S3).

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

197

DESP: obtained from diethyl sulfide and acrylic acid in 42% yield as a white solid.

198

1

199

Hz, 2H), 1.47 (t, J = 7.4 Hz, 6H); 13C NMR (101 MHz, D2O) δ 173.87, 33.48, 28.88, 7.89.

200

HRMS m/z calcd for C7H15O2S+ is 163.0793, found 163.0801.

201

EMSP: obtained from ethyl-methyl sulfide and acrylic acid in 44% yield as a colorless oil.

202

1

203

3.38 - 3.29 (m, 1H), 3.01 (t, J = 6.8 Hz, 2H), 2.92 (s, 3H), 1.47 (t, J = 7.4 Hz, 3H); 13C NMR

204

(101 MHz, D2O) δ 173.83, 36.49, 36.29, 28.66, 21.90, 7.79. HRMS m/z calcd for

205

C6H13O2S+ is 149.0636, found 149.0641.

206

cDESP: obtained from tetrahydrothiophene and acrylic acid in 20% yield as a white solid.

207

1

208

3.00 (t, J = 6.8 Hz, 2H), 2.44 – 2.24 (m, 4H); 13C NMR (101 MHz, D2O) δ 174.01, 44.31,

209

37.99, 29.77, 28.22. HRMS m/z calcd for C7H13O2S+ is 161.0636, found 161.0630.

210

2-Methyl DMSP: obtained from DMS and 2-methylacrylic acid in a 15% yield as a white

211

solid.

212

1

213

3.15 – 3.08 (m, 1H), 3.01 (s, 3H), 3.00 (s, 3H), 1.39 (d, J = 7.2 Hz, 3H); 13C NMR (101 MHz,

214

CD3OD) δ 176.37, 48.09, 37.15, 27.28, 26.65, 17.13. HRMS m/z calcd for C6H13O2S+ is

215

149.0636, found 149.0640.

216

3-Methyl DMSP: obtained from DMS and trans-3-methylacrylic acid in a 10% yield as a

217

white solid.

218

H NMR (400 MHz, D2O) δ 3.51 (t, J = 6.9 Hz, 2H), 3.38 (q, J = 7.4 Hz, 4H), 3.00 (t, J = 6.8

H NMR (400 MHz, D2O) δ 3.62 - 3.55 (m, 1H), 3.52 - 3.45 (m, 1H), 3.45 - 3.38 (m, 1H),

H NMR (400 MHz, D2O) δ 3.69 – 3.57 (m, 2H), 3.55 – 3.47 (m, 2H), 3.44 (t, J = 6.8 Hz, 2H),

H NMR (400 MHz, CD3OD) δ 3.57(dd, J = 13.3, 9.1 Hz, 1H), 3.44 (dd, J = 13.3, 5.1 Hz, 1H),

1

H NMR (400 MHz, CD3OD) δ 3.96 – 3.85 (m, 1H), 3.01 – 2.91 (m, 8H), 1.57 (d, J = 6.9 Hz,

219

3H); 13C NMR (101 MHz, CD3OD) δ 172.65, 48.82, 36.87, 23.99, 22.35, 14.63. HRMS m/z

220

calcd for C6H13O2S+ is 149.0636, found 149.0643.

221 222

ACS Paragon Plus Environment

Page 8 of 41

Page 9 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

223

Results

224

Cupin-DLL superfamily members are featured by two histidine motifs and a catalytic

225

metal ion.

226

We have collected a range of putative cupin DMSP lyases sequences, and could thus

227

identify common sequence motifs, including the conserved histidine motifs that are

228

shared by DddK, DddL, DddQ, DddW and DddY13, 33, 34 (Figure 1A). The first shared motif

229

comprises 4 entirely conserved residues – HxHxxxxExY (His268, His270, Glu274 and

230

Tyr277; numbering is for DaDddY). The second motif includes a third entirely conserved

231

histidine, His342. As expected, the two remaining bacterial Ddd+ families, DddD and

232

DddP, and the algal Alma DMSP lyases, do not share these motifs as they belong to

233

completely different superfamilies. As shown below, these motifs include the metal-

234

ligating residues that appear to be absolutely essential, and other residues, including an

235

active site tyrosine (Y131 in DddQ) that seem to play a relatively minor role.

236

Of the 5 cupin Ddd+ families, structures are available for DddQ (PDB 4LA2, Ref.20;

237

5JSO and 5JSP, Ref.21; 4B29 and 5CU1), DddY (PDB 5XKX, 5XKY and 5Y4K, Ref.18), and

238

DddK (PDB 5TG0, Ref.35). The structures indicate that two His residues, one from the

239

first motif and another from the second motif, and the Glu of the first motif, directly

240

chelate the active-site metal ion (His130, Glu134 & His169 in PDB 5CU1). The second His

241

residue of the 1st motif and its Tyr (His270 and Tyr277 in DaDddY, His 265 and Tyr 271 in

242

PDB 5XKX) are also within the active-site and in close proximity to the metal ion (< 4 Å,

243

His125 and Tyr131 in PDB 4LA2; His130 and Tyr136 5CU1; His 129 and Tyr135 in PDB

244

4B29). Other conserved residues are part of both motifs. For example, a Trp/Tyr at the

245

end of the second motif is also an active-site residue, with the Trp ring’s NH placed

246

~4.5Å from the metal ion (Trp363 in DaDddY; Trp178 in DddQ, PDB 4LA2; Trp184 in PDB

247

5CU1; Trp183 in PDB 4B29). The structure alignment showed that, as expected the cupin

248

Ddd+ enzymes also share a β-barrel-fold structure comprised of 8 antiparallel β-strands

249

(Figure 1B). The metal ion and active-site pocket are surrounded by the β-barrel.

250

Together, the two above-described sequence motifs define the Cupin-DLL

251

superfamily (Figure 1A). The location of their residues within the active–site and close to

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

252

the metal indicates that both motifs are critical to function and also explains their

253

absolute conservation in known cupin Ddd+ enzyme families. Thus, based on these

254

motifs, the identification of related sequences, and analysis of what we have defined as

255

the cupin-DLL superfamily, have been performed. However, as described below,

256

mutational analysis indicated that at least one of these residues, the highly conserved

257

tyrosine in Motif 1, may not play a key role in catalysis of DMSP lysis.

258 259

The Cupin-DLL superfamily is a large and widely diverged superfamily.

260

Pfam model PF16867, DMSP lyases, was used to search against the NCBI protein

261

database and the Tara marine metagenomics database.29 The retrieved sequences were

262

confirmed by virtue of having the conserved histidine Motif 1, HxHxxxxExY. Following

263

filtering by similarity, the remaining sequences were aligned (361 sequences in total)

264

and an extended profile that encompassed all potential superfamily members was

265

generated (see Methods and Supporting dataset 1). A phylogenetic tree was also built

266

(Figure S4). This tree represents a wide variability that includes sequences with as low as

267

15% identity and highly variable lengths (90 - 441 amino acids). It thus provided an

268

overview of the content of this superfamily and of where the Ddd+ families map within

269

it, as schematically summarized in Figure 1C.

270

The phylogenetic tree indicates that cupin-DLL superfamily is large and its

271

various clades are widely diverged, suggesting high functional diversity. DddQ is by far

272

the largest family. With the exception of DddW and DddK that are closely related, the

273

other 4 Ddd+ families are sporadically distributed throughout the tree, suggesting

274

multiple independent emergences of DMSP lyases from enzymes with another function.

275

However, the vast majority of Cupin-DLLs are unlikely to be DMSP lyases – as shown

276

later for DddQ, the shared evolutionary origin can give rise to promiscuous DMSP lyase

277

activity.

278

The shared cupin motifs, and the role and location of their conserved residues,

279

indicate that DddY, DddQ, DddL, DddK, and DddW belong to the same superfamily. To

280

complete the picture with respect to activity levels, we first tested the DMSP lyase

ACS Paragon Plus Environment

Page 10 of 41

Page 11 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

281

activity of members from all five Ddd+ families, as described in the next Sections.

282

Further on, we describe the exploration of representatives from the other cupin-DLL

283

families identified in the tree.

284 285

Recombinant DddY’s exhibit DMSP lyase activity with kcat/KM in the range of 106 M-1s-1

286

We began with exploring the DddY family. Although DddY enzymes seem to exhibit the

287

highest reported DMSP lyases activity,15 at the time we began this work, no DddY has

288

been characterized in detail and/or has been expressed as a recombinant protein (very

289

recently, DddY from Acinetobacter bereziniae has been recombinantly expressed, and

290

the structure solved18). DddY sequences were collected from the NCBI database. After

291

filtering (≥ 90% coverage and ≥ 30% amino acids identity) and Muscle alignment,30 a

292

phylogenetic tree was built indicating two major clades that are related although with

293

considerable divergence (Figure 2A). The 1st clade included the previously described

294

Alcaligenes and Desulfovibrio DddY’s,15, 16 while the 2nd clade that on average shows ≤38%

295

identity to the first one included primarily Ferrimonas and Shewanella genes. Two

296

additional sequences were identified that appear as outgroup of the 2nd clade and

297

notably belong to Synechococcus, widely spread marine photosynthetic cyanobacteria

298

(31% and 33% sequence identity compared to AfDddY).

299

Representatives of these two clades were cloned and over-expressed in E. coli,

300

including Shewanella putrefaciens CN-32, Ferrimonas balearica DSM 9799, Desulfovibrio

301

acrylicus (DaDddY), Alcaligenes faecalis J481(AfDddY) and Synechococcus sp. KORDI-100

302

DddY genes. Alcaligenes DddY was reported to have a periplasmic signal peptide,17 and

303

a homologous N-terminal region seems to appear in all DddY’s. Initially, the wild-type

304

genes including their putative signal peptides were expressed (see Methods). Of all

305

tested candidates, DaDddY exhibited the highest activity. The putative mature enzyme

306

sequence was then recloned fused to PelB – a widely used E. coli periplasmic signal

307

peptide. This recombinant DaDddY variant expressed at higher yield and was purified

308

yielding the 50 kDa mature protein at ≥ 95% purity (Figure S5A).

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

309

Recombinant DaDddY exhibited maximal activity at pH 8.5, although the pH-rate

310

profile indicated multiple titratable groups (Figure S5B) The origins of this double-bell-

311

shaped pH-rate profile are unclear, but a very similar profile was reported for

312

endogenous AfDddY.36 The recombinant AfDddY variant was expressed with its native

313

signal peptide.

314

The activity of metallo-enzymes varies with their metal composition. The highest

315

activity was observed when manganese (Mn2+) was added to the growth medium.

316

However, to ensure the metal identity, the metal was removed from the purified

317

enzymes by chelation (described below) and the enzymes were reconstituted with a

318

specific metal ion. The highest activity was indeed observed with manganese (Figure S6)

319

and all data below relate to manganese DddY’s. Both enzymes followed Michaelis-

320

Menten kinetics (Figure 2B). DaDddY exhibited a kcat/KM value of ~1.3 × 106 M-1s-1, and

321

its orthologue, AfDddY, showed ~4-fold lower kcat/KM of 0.35 × 106 M-1s-1. These values

322

are highly similar to the values originally reported for the enzymes isolated from their

323

endogenous organisms (Table 1). Thus, as indicated by previously reported specific

324

activities, DddY shows the highest catalytic efficiency among all known bacterial DMSP

325

lyases. DddY’s catalytic efficiency is also considerably higher than that of Alma algal

326

DMSP lyases (0.8 × 105 M-1s-1 for Emiliana huxleyi Alma1 and 2.7 × 104 M-1s-1 for

327

Symbiodinum-A1 Alma1).6, 8

328

Whilst the 1st clade that includes AfDddY and DaDddY’s clearly encompasses

329

highly active DMSP lyases, the identity of the 2nd clade of putative DddY’s, including the

330

Shewanella, Ferrimonas and Synechococcus genes, is unclear. Upon expression in E. coli,

331

few of the representative genes but these showed low soluble expression and no

332

activity. We have also attempted to express the inferred ancestor of the Shewanella

333

clade, because inferred ancestors seem to consistently yield proteins with high

334

foldability and stability37. However, the ancestor protein exhibited no DMSP lyase

335

activity, neither in crude cell lysates nor in purified preparations. Hence, at this stage it

336

remains unclear whether the 2nd clade is an integral part of the DddY family although its

337

members are unable to correctly fold in E. coli. Alternatively, this clade might represent

ACS Paragon Plus Environment

Page 12 of 41

Page 13 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

338

a related enzyme family with some residual, promiscuous DMSP lyase activity. As

339

described below, this ambiguity applies for other members of the Cupin-DLL superfamily

340

that exhibit weak or no DMSP lyase activity upon expression in E. coli.

341 342

DddL is likely a highly active DMSP lyase

343

To our knowledge, reports of DddL’s specific activity, let alone of its kinetic parameters,

344

are not available. Here, two DddL orthologues were examined; A previously described

345

Rhodobacter sphaeroides DddL (RsDddL)11, and a DddL gene we identified in the

346

Thioclava pacifica genome (TpDddL; 78% amino acid identity to R. sphaeroides DddL),

347

and Sulfitobacter EE-36 DddL (50% amino acid to RsDddL).11 RsDddL was cloned and

348

expressed in E. coli with His-tags at either the N- or C-terminus, or an N-terminal Strep-

349

tag. High DMSP lyase activity was repetitively observed in crude cell lysates, indicating a

350

soluble, correctly folded and active enzyme. However, all attempts to purify RsDddL

351

failed. Subsequently, TpDddL and Sulfitobacter DddL were cloned and expressed with an

352

N-terminal Strep-tag, but showed similar behavior. Nonetheless, the specific activity of

353

both DddL’s in freshly prepared E. coli crude cell lysates, as estimated from their protein

354

concentrations and DMSP lyase activities, is high, in the range of 70 Units (Figure S1).

355

The kinetics indicated a linear increase of initial rates with DMSP concentrations up to

356

10 mM suggesting a high KM and accordingly high kcat. This value of specific activity (70

357

Units) is probably underestimated, as the soluble fraction is likely to contain misfolded

358

enzyme. Nonetheless, this specific activity is up to 10-fold higher than DddK and DddW’s,

359

and is well above 1000-fold higher than DddQ’s (Table 1).

360 361

The conserved Tyr of Motif 1 may not be essential for DMSP lyase activity

362

To confirm the role of the most conserved amino acids in the active site of DddY, a

363

series of site-directed mutants to Ala was generated at DaDddY’s background. The

364

mutations H268A, H270A, E274A and H342A all showed complete loss of the activity, as

365

reported for other cupin Ddd+ enzymes such as DddW.34 The histidine residues of the

366

two cupin motifs are therefore essential for enzymatic activity, foremost, for the ligation

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

367

of the active-site metal ion.38 However, the conserved tyrosine that belongs to the first

368

motif presents a puzzle. The Y277A mutant of DaDddY retained around 20% of the wild-

369

type’s activity indicating that its contribution to catalysis is minor. However, when

370

Tyr277 was replaced by phenylalanine whose side-chain is similar to tyrosine, enzymatic

371

activity was completely lost (Figure 2C). The CD spectra of all these mutants were

372

essentially identical to wild-type, suggesting that the mutations did not perturb the

373

enzyme’s overall fold (Figure S7A). Zhang et al18 suggested this tyrosine acts as the

374

deprotonation base thus initiating DMSP cleavage. However, given the above results,

375

and the high pKa of Tyr (10.5), it appears more likely that this active site tyrosine plays a

376

critical role in the other function(s) that prevail in Cupin-DLLs. Thus, although this

377

tyrosine has been retained in DMSP lyases such as DddY or DddL, its role in catalysis is

378

relatively minor.

379

To get a clearer understanding of the role of this active-site tyrosine, its mutants

380

in DddK, DddW and DddL were examined. Similarly, to DddY, the Tyr-to-Ala mutants of

381

these three Ddd+ enzymes showed a relatively mild decrease in activity, around 10%

382

residual activity. In DddL and DddW, the Phe mutants showed near-complete loss of

383

activity as in DaDddY, whereas in DddK replacements to either Ala or Phe had a similar

384

effect (around 10% residual activity, see Figure S7B). Overall, it appears that the

385

tyrosine of the 1st motif plays a role in DaDddY’s DMSP lyase activity, as well is in the

386

other Cupin-DLL enzymes, but its role is secondary, certainly in comparison to the metal-

387

ligating residues. The complete inactivation in the more conservative substitution to Phe

388

may be due to the absence of the interacting hydroxyl of Tyr, or because a hydrophobic

389

moiety in the metal’s close vicinity is deleterious. Specifically, a water molecule may

390

bind the Ala mutant and substitute the missing Tyr’s hydroxyl group, whereas Phe might

391

block water binding.

392 393

Cupin-DLL enzymes are selectively inhibited by the metal chelator, TPEN

394

The activity of any Cupin-DLL enzyme is expected to be metal-dependent, and thus

395

inactivation by a metal chelator comprises a facile way of identifying these enzymes,

ACS Paragon Plus Environment

Page 14 of 41

Page 15 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

396

including in marine organisms and environments. However, metal chelators vary in their

397

ability to inactivate metallo-enzymes. Additionally, many enzymes, including metallo-

398

enzymes may possess additional structural metal sites, typically calcium or magnesium,

399

whose removal causes the loss of structure and thereby of enzymatic activity. We found,

400

for example, that some algal Alma DMSP lyases,6, 39 and DaDddY, require calcium to

401

maintain activity. Hence, to identify a selective inhibitor of cupin-DLLs, we examined

402

chelators that can chelate transition metal ions such as manganese, zinc, or iron, in the

403

presentence of alkali metal ions such as magnesium or calcium. To this end, DddY, DddL,

404

DddQ, DddK and DddW were incubated with various transition metal chelators in a

405

reaction buffer that also contained CaCl2 to avoid Alma or DddY losing activity (due to a

406

structural calcium site; see ref.

407

huxleyi8 was examined as a control for a non-cupin-DLL DMSP lyase. We found that all

408

cupin Ddd+ enzymes were fully inhibited by TPEN while Alma DMSP lyases were

409

unaffected (Figure 3).

6, 39

Figure S6). The Alma DMSP lyase from Emiliania

410

Selective inhibition by TPEN may therefore serve as a probe for the identification

411

of enzymes belonging to the cupin-DLL superfamily. Additionally, chelation with TPEN

412

allowed us to reconstitute cupin Ddd+ enzymes with a specific metal ion (Figure S6),

413

thus ensuring that the reported properties relate to a known metal composition.

414 415

Genome context suggests that DddQ’s native activity is not DMSP lysis

416

Although DddY, DddL, DddQ, DddK and DddW share an evolutionary origin, and

417

structural and functional properties, as shown above, some of these enzyme families

418

show abnormally low DMSP lyases activity. This suggests that their primary,

419

physiological activity might not be as DMSP lyases. In bacterial genomes, enzymes of the

420

same pathway are typically found in gene clusters or even in the same operon. Thus,

421

genome context can provide important hints regarding the enzymatic function of ddd+

422

genes. Specifically, acrylate, the verified product of all cupin Ddd+ enzymes (Table 1), is

423

toxic,40,

424

catabolism. These include: acuI, a zinc/iron dependent alcohol reductase that converts

41

and previous studies indicated proximal genes that relate to acrylate

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

425

acrylyl-CoA into propanoyl-CoA;17, 40 acuK (a dehydratase) and acuN (a CoA transferase)

426

that can jointly convert acrylate into 3-hydroxypropionate;42 or cytochrome dependence

427

oxi-reductases that could also catabolize acrylate.43

428

We applied few common tools of genome context analysis, including EFI-GNT,44

429

STRINGE,45 EASYFIG,46 and RODEO.47 Of these, RODEO provided the most systematic

430

results for the examined Ddd+ genes, including the previously studied DMSP CoA-

431

transferase/lyase dddD that serves as benchmark for a bona fide DMSP lyase.9

432

Accordingly, dddD is characterized by proximal dddB and dddC genes, encoding an iron-

433

containing dehydrogenase and a methyl-malonate semi-aldehyde dehydrogenase-like

434

protein, respectively (Figure S8A; Refs.2, 7). The product of DddD, 3-hydroxypropionate-

435

CoA, is converted into malonate semi-aldehyde (DddB) and then to acetyl-CoA (DddC).2

436

Accordingly, dddY’s genome neighborhoods repetitively contain putative acrylate

437

utilizing genes (Ref

438

resides (Desulfovibrio acrylicus) converts acrylate into propionate;48 and Alcaligenes

439

faecalis M3A strain that contains AfDddY converts acrylate into 3-hydroxypropionate.49

17

; Figure S8B). Indeed, the bacterium in which DaDddY originally

440

In DddL’s case, however, the genome context is not conserved, making it hard to

441

draw a clear conclusion (in 4 out of 15 genomes an acuI-like zinc containing reductase is

442

present just next to DddL; Figure S9A). Similarly, because dddK genes are only found in

443

several strains of Pelagibacter ubique (SAR11),12 it is difficult to get a systematic

444

prediction based on genome context. In the genome of SAR11, the proximity of enoyl-

445

ACP-reductate, β-ketoacyl-ACP-synthase and β-hydroxydecanoyl-ACP dehydratase

446

indicate a relation to fatty acid or polyketide biosynthesis.50 (Figure S9B)

447

The genome neighborhoods of both dddQ and dddW could be consistently

448

derived. For dddW gene, the proximal D-alanyl-D-alanine carboxypeptidase gene (Figure

449

S9C) suggests a gene cluster that is involved in bacterial peptidoglycan biosynthesis.

450

Finally, the highly conserved neighborhood of dddQ includes a putative

451

mandelate racemase-like protein and a putative dimethylglycine dehydrogenase (Figure

452

4). The mandelate racemase-like gene exhibits present distinct homology (40% and 35%

453

amino acid identities, respectively) to two recently identified enzymes: cis-3-hydroxy-L-

ACS Paragon Plus Environment

Page 16 of 41

Page 17 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

454

proline dehydratase and 4-hydroxyproline betaine 2-epimerase.51,

52

455

neighbors of dddQ, and the dimethylglycine dehydrogenase-like neighbor, suggest that

456

DddQ takes part in the catabolism of proline-betaine or/and hydroxyproline-betaine. Its

457

DMSP lyase activity is therefore likely to be promiscuous, as also indicated by its

458

markedly low specific activity (Table 1) and its lack of selectivity as elaborated in the

459

substrate profiling section below.

These two

460 461

DMSP lyases activity coincides in most cupin-DLL enzymes

462

As indicated by the phylogenetic analysis (Figure 1) and the genome context (above

463

section), the Cupin-DLL superfamily is a large and highly diverse in sequence, and in

464

most likelihood, also in function. All the so far explored members of the Cupin-DLL

465

superfamily were initially identified owing to their DMSP lyase activity (the Ddd+

466

families). However, given how distant these families are, yet unknown DMSP lyase

467

families may exist within this superfamily. Alternatively, since within enzyme

468

superfamilies, the native activity of one family may appear as promiscuous activity in

469

related families whose enzymatic function is different, DMSP lyase activity may be a

470

hallmark of all Cupin-DLLs, either as their native function, or as a weak, promiscuous

471

activity. To obtain a more comprehensive view of the Cupin-DLL superfamily, we

472

examined several genes that represent the different non-Ddd+ clades identified in our

473

phylogenetic tree (Figure 1). Overall, we examined two sequences from each of the yet

474

unexplored clades, and one sequence from the tree’s outgroup as a control. Synthetic

475

genes were overexpressed in E. coli, the proteins were purified and their DMSP lyase

476

activity was tested. The yields of purified enzymes varied, but all 9 candidates could be

477

isolated as soluble proteins. Three of the tested genes failed to exhibit detectable DMSP

478

lyase activity (hollow circles in Figure 1: Figure S10): the outgroup sequence, and the 2

479

representatives of Cupin-DLL-3 family. The six other genes showed weak DMSP lyases

480

activity, in the range of ≤20 nmol*min-1*mg-1 protein, roughly in the range of activity

481

exhibited by DddQ.

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

482

Random sampling, and especially of sequences obtained via metagenomics, is

483

problematic. Further, low activity can also be the outcome of poor folding capacity

484

or/and reaction conditions. Nonetheless, the above results indicate that families #1, #2,

485

and #4 exhibit weak, promiscuous DMSP lyase activity. Indeed, these families are

486

phylogenetically close to families with the highest known DMSP lyase activity (DddY and

487

DddL, respectively). On the other hand, Family #3, which is closer to DddQ that exhibits

488

very low DMSP activity, exhibits no activity. Overall, it appears that the DMSP lyase

489

activity is shared by most members of the Cupin-DLL superfamily, either as their native

490

function, or as weak promiscuous activity.

491 492 493

Substrate profiling indicates that DMSP lyase is DddL and DddY’s native activity, but not DddQ’s

494

As indicated above, the DMSP lyase activity levels of cupin-DLLs are widely distributed.

495

The specific activities range from less than 1 Units up to 675 Units, and accordingly, the

496

kcat/KM values range from 0.7 M-1s-1 (ref. 21) up to 106 M-1s-1 (ref. 16) (Table 1). The kinetic

497

parameters of enzymes are widely distributed with the average kcat/KM value being ~ 105

498

M-1s-1.53 Accordingly, kcat/KM values of 1 M-1s-1 with the native substrate are exceedingly

499

rare, yet are characteristic to promiscuous enzymatic activities. The genome context

500

analysis presented above supports this hypothesis at least with respect to DddQ that

501

also exhibits the lowest specific activity in comparison to all other Ddd+ enzymes (Table

502

1). However, low specific activity can also be the outcome of suboptimal folding and/or

503

reaction conditions (e.g. missing metal ions). Thus, if a low kcat/KM value is observed two

504

possible scenarios apply that cannot be distinguished; namely: (i) this activity is the

505

native one yet the enzyme is poorly active due to misfolding or suboptimal reaction

506

conditions; (ii) the enzyme preparation is fully active, and there exists another activity,

507

the native one, that exhibits high kcat/KM. Additionally, some promiscuous activities

508

occur with kcat/KM values that are similar to the native activities.25 Thus, to determine

509

whether a given enzyme’s activity is native or promiscuous can be challenging.

ACS Paragon Plus Environment

Page 18 of 41

Page 19 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

510

Substrate profiling can be applied to distinguish in an enzyme exhibiting multiple

511

activities which activity is native and which one is promiscuous. Specifically, the native

512

function seems to coincide with substrate structure dictating enzyme reactivity (rather

513

than the substrates’ inherent reactivity or coincidental changes in substrate structure)54-

514

57

515

enzyme’s native substrate, the enzyme’s active-site would largely be shaped to fit this

516

particular substrate. Thus, relatively small changes in the substrate’s structure are

517

expected to induce relatively large drops in activity.22, 58-61 True, the essence of enzyme

518

promiscuity is acceptance of alternative substrates, sometimes with little resemblance

519

to the original substrate and also in a selective manner.62 Nonetheless, selectivity with

520

respect to the native substrate is observed in most enzymes including promiscuous ones,

521

and especially in enzymes that evolved for one particular substrate (in contrast to

522

broad-specificity enzymes). As a rule of thumb, substrates smaller than the original one

523

will be accepted more readily than bigger ones (due to steric clashes) and increasingly

524

larger substrates will show increasingly lower activity, and, perturbations next to where

525

the key chemical step occurs would have larger impact.

. The rationale behind substrate profiling is that, if a given substrate were the

526

Based on this rationale, we developed an approach of substrate profiling for

527

DMSP lyases. A series of DMSP analogues was synthesized in which the potential for β-

528

elimination and release of a dialkylsulfide product was retained (Figure 5A). The

529

modifications included increasingly large dialkylsulfonyl leaving groups (EMSP, DESP,

530

cDESP) and addition of methyl groups to the propionate moiety, either on the α-carbon

531

from which a proton is abstracted (2-methyl-DMSP) or on the adjacent carbon (3-

532

methyl-DMSP). None of these analogues matches a known natural metabolite.

533

We tested representatives of 6 bacterial Ddd+ families (all bacterial families

534

except DddD which is a CoA-ligase/DMSP lyase confirmed as a bona fide DMSP lyase 9)

535

and, as control the algal E. huxeleyi Alma18. Foremost, our substrate profiling approach

536

is based on relative activities of DMSP analogues compared to DMSP, and hence the

537

results should be relevant regardless of differences in specific activities. Indeed, the

538

specific activities of the tested enzymes with DMSP vary by >3,000-fold (Table 1; Figure

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

539

S9) and were thus tested at different concentrations in accordance with their specific

540

activities. The initial rates observed with each substrate analogue, for release of DMS, or

541

of the corresponding dialkyl sulfide, were compared to the initial rate of the same

542

enzyme with DMSP.

543

As can be seen in Figure 5B, the enzymes tested clearly divide to two groups:

544

highly selective enzymes (Alma1, DddY and DddL) and less, or evidently non-selective

545

enzymes (DddK, DddW, and DddQ respectively). Specifically, Alma and DddL failed to

546

accept any of the applied DMSP analogues under the tested conditions (enzyme

547

concentrations and initial rates for DMS release with DMSP are noted in the legend).

548

DddY showed some cross-reactivity. However, a significant and systematic decrease in

549

activity was observed with increased size of substituents on the dialkylsulfide leaving

550

group. Foremost, almost no activity was observed with analogues contacting an extra

551

methyl group close to where proton abstraction occurs (2-methyl-DMSP and 3-methyl-

552

DMSP). In contrast, in DddQ, modifications of the leaving group, including the bulkiest

553

one (cDESP) hardly affected the rate. Strikingly, DddQ catalyzes the β-elimination of 3-

554

methyl-DMSP nearly 20 times faster compared to DMSP. Considering that 3-methyl-

555

DMSP is not known to be a natural metabolite, DddQ is highly likely to be an enzyme

556

that catalyzes a reaction other than DMSP lysis.

557

Metal compositions may have a different effect on the native compared to the

558

promiscuous activities of metallo-enzymes.63 We thus examined whether and how

559

metals alter the activity profiles of DddY, a bona fide DMSP lyase) and of DddQ,

560

presumably a promiscuous DMSP lyase. Both enzymes as purified from E. coli, most

561

likely containing Mn2+ (the metal yielding the highest specific activity in reconstitution

562

experiments). In parallel, we tested the reconstituted enzymes with either Zn2+ or Co2+,

563

(Figure 5C and D; Figure S6). DddY’s DMSP lysis activity was the highest with Mn2+,

564

and >10-fold higher with Co2+ compared to Zn2+. The activity with the closest analogue,

565

EMSP, largely followed the same trend, while the very low activity with 3-methyl-DMSP

566

remained almost the same regardless of the metal. Overall, the selectivity pattern of

567

DddY is essentially the same with all three metals. For DddQ, the highest activity with 3-

ACS Paragon Plus Environment

Page 20 of 41

Page 21 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

568

methyl-DMSP compared to all other substrates was also retained regardless of the

569

metal composition. Overall, the substrate profiling supports the notion that DddQ is not

570

bona fide DMSP lyases irrespective of the active site metal.

571 572

The most active DMSP lyases also show the highest selectivity

573

Having obtained the substrate selectivity profiles of 6 different enzymes, we could

574

examine whether their substrate selectivity and DMSP lyase activities correlate. To

575

examine such a correlation, a quantitative measure of enzyme selectivity had to be

576

developed. Previous work provided quantitative measures of promiscuity, but these

577

measures were designed to assess how broad is the substrate acceptance of a given

578

enzyme.64 In contrast, we aimed at examining how likely is DMSP to be the native

579

substrate, with the expectation being that the more selective an enzyme is toward

580

DMSP, the higher would its DMSP lyase activity be. To this end, we defined a simple

581

measure of substrate selectivity, dubbed cross-reactivity index, ICR, which was defined as:

582

[Eq. 1]

583

; Whereby ei is the activity relative to DMSP per each DMSP analogue (as in Figure 5B)

584

and N is the number of tested DMSP analogues (N = 5, in this case). If an enzyme only

585

cleaves DMSP, ICR =0, as is the case with Alma1. In principle, ICR is expected to be in the

586

range of 0 up to 1. However, for DddQ the ICR value is >>1 due to the activity with one

587

analogue being ~19-fold higher than with DMSP.

ICR =

∑   

588

For the cupin-DLL family members compared here, cross-reactivity seems to be

589

anti-correlated with specific activity: the most active DMSP lyases are also the ones

590

showing the clearest signature of a DMSP tailored active-site, i.e, a low ICR, and vice

591

versa (Figure 6). The correlation is, as expected, not perfect. Some of the deviations may

592

relate to suboptimal expression or/and reaction conditions; foremost, given the failure

593

to purify DddL, this enzyme’s specific activity is most probably underestimated.

594

Foremost, DddQ clearly stand out by both criteria – i.e., they exhibit the lowest DMSP

595

lyase activity as well as the lowest selectivity toward DMSP. Both criteria therefore

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

596

indicate that DddQ’s native function is likely to be other than DMSP lyase. This

597

conclusion is also in agreement with the genome context analysis (Figure 4, S8, S9).

598 599

Discussion

600

With the exception of DddD and DddP, all bacterial ddd+ genes implicated as DMSP

601

lyases belong to this newly defined superfamily dubbed Cupin-DLL (Figure 1C). All

602

members of this superfamily share two conserved active site motifs that jointly mediate

603

the binding of a transition metal ion that is essential to their enzymatic activity. We

604

have accordingly identified TPEN as a selective inhibitor of all known Cupin-DLL enzymes

605

(Figure 3).

606

The phylogenetics suggest that DMSP lyases evolved within this superfamily

607

independently along multiple lineages. This phenomenon of parallel evolution is

608

commonly observed in enzyme superfamilies, and its origins are in shared promiscuous

609

activities - relatively distant superfamily members often share the same promiscuous

610

activity.65, 66 Such promiscuous activities comprise the starting point for the divergence

611

of new enzymes by turning a latent, promiscuous activity into a physiologically relevant

612

function, initially alongside the enzyme’s original function. Indeed, bi-functional,

613

‘generalist’ enzymes commonly comprise evolutionary intermediates, although they

614

may persist for long periods.22 As the divergence process proceeds, highly active,

615

selective ‘specialist’ enzymes evolve, typically by duplication and sub-functionalization

616

of the bi-functional generalist ancestor.67

617

Our view as evolutionary biochemists is therefore that the Cupin-DLL families

618

likely represent an entire range of evolutionary states – they may be fully diverged

619

DMSP lyase specialists (DddY and DddL), bi-functional intermediates (possibly DddK/W),

620

or enzymes with a different function that only exhibit latent, promiscuous DMSP lyase

621

activity. The latter seems to apply to DddQ, in most likelihood, and possibly to members

622

of other Cupin-DLL families, e.g., Cupin-DLL #1, #2 and #4, which we found exhibit weak

623

DMSP lyase activity (Figure 1C). The level of activity is an important parameter with

624

respect to physiological relevance. However, low specific activity may also be due to

ACS Paragon Plus Environment

Page 22 of 41

Page 23 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

625

suboptimal folding or/and reaction conditions. We thus pursued an independent test of

626

substrate profiling. It appears, however, that activity levels and substrate selectivity are

627

correlated – the least active enzymes, foremost DddQ, also show the weakest signature

628

of a DMSP tailored active-site. DddK and DddW reside in the mid-range by both criteria,

629

activity and selectivity, and could therefore be bi-functional or ‘generalist’ intermediates

630

that have evolved toward DMSP lyase activity while retaining an additional native

631

function that remains unknown at this stage. Enzymes in secondary metabolism tend to

632

exhibit lower catalytic efficiency compared to central metabolism enzymes53 as well as

633

multi-functionality. Indeed, while the DMSP lyase activity of DddK/W is modest (kcat/KM

634

≈ 103 M-1s-1), these rates are notably comparable to the rates of DMSP demethylation by

635

the corresponding enzyme (DmdA).35 This comparison is relevant because DMSP is

636

catabolized via two routes, lysis to give DMS as described here, or demethylation to give

637

methylmercaptopropionate,2,

638

dominating one.2, 3, 69

4, 68

and the demethylation route is actually the

639

Convergent evolution is also dominant in DMSP lyases, as exemplified by the

640

independent evolutionary origins of the algal Alma1 DMSP lyase. Convergence also

641

relates to DMSP lysis being a facile reaction,2 and foremost to the fact that lysis is

642

initiated by abstraction of a proton from a carbon next to a carboxylate (α-carbon). This

643

is one of the most common steps in enzyme catalysis, and several large enzyme

644

superfamilies share it as the key catalytic step, including the Asp/Glu racemase

645

superfamily to which the Alma DMSP lyases belong. Accordingly, DMSP lyases that

646

belong to the enolase superfamily – where α-proton abstraction is the hallmark,

647

highly likely to exist.

70

are

648

To conclude, our preliminary mapping of the cupin-DLL superfamiliy has

649

somewhat clarified the picture with respect to the DMSP lyases. However, a key

650

remaining question is the primary function of the cupin-DLL superfamily, and specifically

651

of DddQ that also comprises its largest clade (Figure 1C). Future research may shed light

652

on the function of Cupin-DLLs that are not DMSP-lyases, and may thus also shed light on

653

the evolutionary history and function of the Ddd+ enzyme families.

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

654 655 656 657

ACS Paragon Plus Environment

Page 24 of 41

Page 25 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

658

Biochemistry

References

659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704

1. 2.

3. 4.

5.

6. 7.

8.

9.

10.

11.

12.

13.

14.

Sievert, S. M., Kiene, R. P., and Schulz-Vogt, H. N. (2007) The Sulfur Cycle, Oceanography 20. Curson, A. R., Todd, J. D., Sullivan, M. J., and Johnston, A. W. (2011) Catabolism of dimethylsulphoniopropionate: microorganisms, enzymes and genes, Nat Rev Microbiol 9, 849-859. Reisch, C. R., Moran, M. A., and Whitman, W. B. (2011) Bacterial Catabolism of Dimethylsulfoniopropionate (DMSP), Front Microbiol 2, 172. Johnston, A. W., Green, R. T., and Todd, J. D. (2016) Enzymatic breakage of dimethylsulfoniopropionate-a signature molecule for life at sea, Curr Opin Chem Biol 31, 58-65. Curson, A. R., Liu, J., Bermejo Martinez, A., Green, R. T., Chan, Y., Carrion, O., Williams, B. T., Zhang, S. H., Yang, G. P., Bulman Page, P. C., Zhang, X. H., and Todd, J. D. (2017) Dimethylsulfoniopropionate biosynthesis in marine bacteria and identification of the key gene in this process, Nat Microbiol 2, 17009. Alcolombri, U., Lei, L., Meltzer, D., Vardi, A., and Tawfik, D. S. (2016) Assigning the algal source of dimethylsulfide using a selective lyase inhibitor, ACS Chemical Biology. Todd, J. D., Rogers, R., Li, Y. G., Wexler, M., Bond, P. L., Sun, L., Curson, A. R., Malin, G., Steinke, M., and Johnston, A. W. (2007) Structural and regulatory genes required to make the gas dimethyl sulfide in bacteria, Science 315, 666-669. Alcolombri, U., Ben-Dor, S., Feldmesser, E., Levin, Y., Tawfik, D. S., and Vardi, A. (2015) MARINE SULFUR CYCLE. Identification of the algal dimethyl sulfide-releasing enzyme: A missing link in the marine sulfur cycle, Science 348, 1466-1469. Alcolombri, U., Laurino, P., Lara-Astiaso, P., Vardi, A., and Tawfik, D. S. (2014) DddD is a CoA-transferase/lyase producing dimethyl sulfide in the marine environment, Biochemistry 53, 5473-5475. Todd, J. D., Curson, A. R., Dupont, C. L., Nicholson, P., and Johnston, A. W. (2009) The dddP gene, encoding a novel enzyme that converts dimethylsulfoniopropionate into dimethyl sulfide, is widespread in ocean metagenomes and marine bacteria and also occurs in some Ascomycete fungi, Environ Microbiol 11, 1376-1385. Curson, A. R., Rogers, R., Todd, J. D., Brearley, C. A., and Johnston, A. W. (2008) Molecular genetic analysis of a dimethylsulfoniopropionate lyase that liberates the climate-changing gas dimethylsulfide in several marine alpha-proteobacteria and Rhodobacter sphaeroides, Environ Microbiol 10, 757-767. Sun, J., Todd, J. D., Thrash, J. C., Qian, Y., Qian, M. C., Temperton, B., Guo, J., Fowler, E. K., Aldrich, J. T., Nicora, C. D., Lipton, M. S., Smith, R. D., De Leenheer, P., Payne, S. H., Johnston, A. W., Davie-Martin, C. L., Halsey, K. H., and Giovannoni, S. J. (2016) The abundant marine bacterium Pelagibacter simultaneously catabolizes dimethylsulfoniopropionate to the gases dimethyl sulfide and methanethiol, Nat Microbiol 1, 16065. Todd, J. D., Curson, A. R., Kirkwood, M., Sullivan, M. J., Green, R. T., and Johnston, A. W. (2011) DddQ, a novel, cupin-containing, dimethylsulfoniopropionate lyase in marine roseobacters and in uncultured marine bacteria, Environ Microbiol 13, 427-438. Todd, J. D., Kirkwood, M., Newton-Payne, S., and Johnston, A. W. (2012) DddW, a third DMSP lyase in a model Roseobacter marine bacterium, Ruegeria pomeroyi DSS-3, ISME J 6, 223-226.

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751

15.

16.

17.

18.

19.

20.

21.

22. 23. 24.

25.

26.

27.

28. 29.

de Souza, M. P., and Yoch, D. C. (1995) Purification and characterization of dimethylsulfoniopropionate lyase from an alcaligenes-like dimethyl sulfide-producing marine isolate, Appl Environ Microbiol 61, 21-26. van der Maarel, M. J. E. C., Aukema, W., and Hansen, T. A. (1996) Purification and characterization of a dimethylsulfoniopropionate cleaving enzyme from Desulfovibrio acrylicus, FEMS Microbiology Letters 143, 241-245. Curson, A. R., Sullivan, M. J., Todd, J. D., and Johnston, A. W. (2011) DddY, a periplasmic dimethylsulfoniopropionate lyase found in taxonomically diverse species of Proteobacteria, ISME J 5, 1191-1200. Li, C. Y., Zhang, D., Chen, X. L., Wang, P., Shi, W. L., Li, P. Y., Zhang, X. Y., Qin, Q. L., Todd, J. D., and Zhang, Y. Z. (2017) Mechanistic Insights into Dimethylsulfoniopropionate Lyase DddY, a New Member of the Cupin Superfamily, J Mol Biol 429, 3850-3862. Alcolombri, U., Elias, M., Vardi, A., and Tawfik, D. S. (2014) Ambiguous evidence for assigning DddQ as a dimethylsulfoniopropionate lyase and oceanic dimethylsulfide producer, Proc Natl Acad Sci U S A 111, E2078-2079. Li, C. Y., Wei, T. D., Zhang, S. H., Chen, X. L., Gao, X., Wang, P., Xie, B. B., Su, H. N., Qin, Q. L., Zhang, X. Y., Yu, J., Zhang, H. H., Zhou, B. C., Yang, G. P., and Zhang, Y. Z. (2014) Molecular insight into bacterial cleavage of oceanic dimethylsulfoniopropionate into dimethyl sulfide, Proc Natl Acad Sci U S A 111, 1026-1031. Brummett, A. E., and Dey, M. (2016) New Mechanistic Insight from Substrate- and Product-Bound Structures of the Metal-Dependent Dimethylsulfoniopropionate Lyase DddQ, Biochemistry 55, 6162-6174. Khersonsky, O., and Tawfik, D. S. (2010) Enzyme promiscuity: a mechanistic and evolutionary perspective, Annu Rev Biochem 79, 471-505. Pandya, C., Farelli, J. D., Dunaway-Mariano, D., and Allen, K. N. (2014) Enzyme promiscuity: engine of evolutionary innovation, J Biol Chem 289, 30229-30236. Miao, Y., Rahimi, M., Geertsema, E. M., and Poelarends, G. J. (2015) Recent developments in enzyme promiscuity for carbon-carbon bond-forming reactions, Curr Opin Chem Biol 25, 115-123. Van Loo, B., and Hollfelder, F. (2010) Enzyme Promiscuity and Evolution of New Protein Functions, In Manual of Industrial Microbiology and Biotechnology, Third Edition, American Society of Microbiology. Roodveldt, C., and Tawfik, D. S. (2005) Shared promiscuous activities and evolutionary features in various members of the amidohydrolase superfamily, Biochemistry 44, 12728-12736. Schnoes, A. M., Brown, S. D., Dodevski, I., and Babbitt, P. C. (2009) Annotation error in public databases: misannotation of molecular function in enzyme superfamilies, PLoS Comput Biol 5, e1000605. Eddy, S. R. (2011) Accelerated Profile HMM Searches, PLoS Comput Biol 7, e1002195. Sunagawa, S., Coelho, L. P., Chaffron, S., Kultima, J. R., Labadie, K., Salazar, G., Djahanschiri, B., Zeller, G., Mende, D. R., Alberti, A., Cornejo-Castillo, F. M., Costea, P. I., Cruaud, C., d'Ovidio, F., Engelen, S., Ferrera, I., Gasol, J. M., Guidi, L., Hildebrand, F., Kokoszka, F., Lepoivre, C., Lima-Mendez, G., Poulain, J., Poulos, B. T., Royo-Llonch, M., Sarmento, H., Vieira-Silva, S., Dimier, C., Picheral, M., Searson, S., Kandels-Lewis, S., Tara Oceans, c., Bowler, C., de Vargas, C., Gorsky, G., Grimsley, N., Hingamp, P., Iudicone, D., Jaillon, O., Not, F., Ogata, H., Pesant, S., Speich, S., Stemmann, L., Sullivan, M. B., Weissenbach, J., Wincker, P., Karsenti, E., Raes, J., Acinas, S. G., and Bork, P. (2015)

ACS Paragon Plus Environment

Page 26 of 41

Page 27 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799

Biochemistry

30. 31.

32. 33.

34.

35.

36.

37. 38. 39.

40.

41.

42.

43.

44.

45.

Ocean plankton. Structure and function of the global ocean microbiome, Science 348, 1261359. Edgar, R. C. (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput, Nucleic Acids Res 32, 1792-1797. Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D. L., Darling, A., Hohna, S., Larget, B., Liu, L., Suchard, M. A., and Huelsenbeck, J. P. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space, Syst Biol 61, 539542. Walker, J. M. (1994) The bicinchoninic acid (BCA) assay for protein quantitation, Methods Mol Biol 32, 5-8. Sullivan, M. J., Curson, A. R., Shearer, N., Todd, J. D., Green, R. T., and Johnston, A. W. (2011) Unusual regulation of a leaderless operon involved in the catabolism of dimethylsulfoniopropionate in Rhodobacter sphaeroides, PLoS One 6, e15972. Brummett, A. E., Schnicker, N. J., Crider, A., Todd, J. D., and Dey, M. (2015) Biochemical, Kinetic, and Spectroscopic Characterization of Ruegeria pomeroyi DddW--A Mononuclear Iron-Dependent DMSP Lyase, PLoS One 10, e0127288. Schnicker, N. J., De Silva, S. M., Todd, J. D., and Dey, M. (2017) Structural and Biochemical Insights into Dimethylsulfoniopropionate Cleavage by Cofactor-Bound DddK from the Prolific Marine Bacterium Pelagibacter, Biochemistry 56, 2873-2885. de Souza, M. P., and Yoch, D. C. (1995) Comparative Physiology of Dimethyl Sulfide Production by Dimethylsulfoniopropionate Lyase in Pseudomonas doudoroffii and Alcaligenes sp. Strain M3A, Appl Environ Microbiol 61, 3986-3991. Trudeau, D. L., Kaltenbach, M., and Tawfik, D. S. (2016) On the Potential Origins of the High Stability of Reconstructed Ancestral Proteins, Mol Biol Evol 33, 2633-2641. Dunwell, J. M., Purvis, A., and Khuri, S. (2004) Cupins: the most functionally diverse protein superfamily?, Phytochemistry 65, 7-17. Nishiguchi, M. K., and Goff, L. J. (1995) Isolation, Purification, and Characterization of Dmsp Lyase (Dimethylpropiothetin-Dethiomethylase-(4.4.1.3)) from the Red Alga Polysiphonia-Paniculata, Journal of Phycology 31, 567-574. Todd, J. D., Curson, A. R., Sullivan, M. J., Kirkwood, M., and Johnston, A. W. (2012) The Ruegeria pomeroyi acuI gene has a role in DMSP catabolism and resembles yhdH of E. coli and other bacteria in conferring resistance to acrylate, PLoS One 7, e35947. Curson, A. R., Burns, O. J., Voget, S., Daniel, R., Todd, J. D., McInnis, K., Wexler, M., and Johnston, A. W. (2014) Screening of metagenomic and genomic libraries reveals three classes of bacterial enzymes that overcome the toxicity of acrylate, PLoS One 9, e97660. Todd, J. D., Curson, A. R., Nikolaidou-Katsaraidou, N., Brearley, C. A., Watmough, N. J., Chan, Y., Page, P. C., Sun, L., and Johnston, A. W. (2010) Molecular dissection of bacterial acrylate catabolism--unexpected links with dimethylsulfoniopropionate catabolism and dimethyl sulfide production, Environ Microbiol 12, 327-343. Mikoulinskaia, O., Akimenko, V., Galouchko, A., Thauer, R. K., and Hedderich, R. (1999) Cytochrome c-dependent methacrylate reductase from Geobacter sulfurreducens AM-1, Eur J Biochem 263, 346-352. Gerlt, J. A., Bouvier, J. T., Davidson, D. B., Imker, H. J., Sadkhin, B., Slater, D. R., and Whalen, K. L. (2015) Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST): A web tool for generating protein sequence similarity networks, Biochim Biophys Acta 1854, 1019-1037. Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J., Simonovic, M., Roth, A., Santos, A., Tsafou, K. P., Kuhn, M., Bork, P., Jensen, L. J., and

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846

46. 47.

48.

49.

50. 51.

52.

53.

54.

55.

56.

57.

58.

59. 60.

von Mering, C. (2015) STRING v10: protein-protein interaction networks, integrated over the tree of life, Nucleic Acids Res 43, D447-452. Sullivan, M. J., Petty, N. K., and Beatson, S. A. (2011) Easyfig: a genome comparison visualizer, Bioinformatics 27, 1009-1010. Tietz, J. I., Schwalen, C. J., Patel, P. S., Maxson, T., Blair, P. M., Tai, H. C., Zakai, U. I., and Mitchell, D. A. (2017) A new genome-mining tool redefines the lasso peptide biosynthetic landscape, Nat Chem Biol 13, 470-478. van der Maarel, M. J. E. C., van Bergeijk, S., van Werkhoven, A. F., Laverman, A. M., Meijer, W. G., Stam, W. T., and Hansen, T. A. (1996) Cleavage of dimethylsulfoniopropionate and reduction of acrylate by Desulfovibrio acrylicus sp. nov, Archives of Microbiology 166, 109-115. Ansede, J. H., Pellechia, P. J., and Yoch, D. C. (1999) Metabolism of acrylate to betahydroxypropionate and its role in dimethylsulfoniopropionate lyase induction by a salt marsh sediment bacterium, Alcaligenes faecalis M3A, Appl Environ Microbiol 65, 50755081. Massengo-Tiasse, R. P., and Cronan, J. E. (2009) Diversity in enoyl-acyl carrier protein reductases, Cell Mol Life Sci 66, 1507-1517. Zhao, S., Kumar, R., Sakai, A., Vetting, M. W., Wood, B. M., Brown, S., Bonanno, J. B., Hillerich, B. S., Seidel, R. D., Babbitt, P. C., Almo, S. C., Sweedler, J. V., Gerlt, J. A., Cronan, J. E., and Jacobson, M. P. (2013) Discovery of new enzymes and metabolic pathways by using structure and genome context, Nature 502, 698-702. Zhang, X., Kumar, R., Vetting, M. W., Zhao, S., Jacobson, M. P., Almo, S. C., and Gerlt, J. A. (2015) A unique cis-3-hydroxy-l-proline dehydratase in the enolase superfamily, J Am Chem Soc 137, 1388-1391. Bar-Even, A., Noor, E., Savir, Y., Liebermeister, W., Davidi, D., Tawfik, D. S., and Milo, R. (2011) The moderately efficient enzyme: evolutionary and physicochemical trends shaping enzyme parameters, Biochemistry 50, 4402-4410. Khersonsky, O., and Tawfik, D. S. (2005) Structure-reactivity studies of serum paraoxonase PON1 suggest that its native activity is lactonase, Biochemistry 44, 63716382. Cummings, J. A., Nguyen, T. T., Fedorov, A. A., Kolb, P., Xu, C., Fedorov, E. V., Shoichet, B. K., Barondeau, D. P., Almo, S. C., and Raushel, F. M. (2010) Structure, mechanism, and substrate profile for Sco3058: the closest bacterial homologue to human renal dipeptidase, Biochemistry 49, 611-622. Illing, A. C., Shawki, A., Cunningham, C. L., and Mackenzie, B. (2012) Substrate profile and metal-ion selectivity of human divalent metal-ion transporter-1, J Biol Chem 287, 30485-30496. Kronen, M., Sasikaran, J., and Berg, I. A. (2015) Mesaconase Activity of Class I Fumarase Contributes to Mesaconate Utilization by Burkholderia xenovorans, Appl Environ Microbiol 81, 5632-5638. Khan, I. H., Kim, H., Ashida, H., Ishikawa, T., Shibata, H., and Sawa, Y. (2005) Altering the substrate specificity of glutamate dehydrogenase from Bacillus subtilis by site-directed mutagenesis, Biosci Biotechnol Biochem 69, 1802-1805. Tardif, K. D., and Horowitz, J. (2004) Functional group recognition at the aminoacylation and editing sites of E. coli valyl-tRNA synthetase, RNA 10, 493-503. Patel, S. S., and Walt, D. R. (1987) Substrate specificity of acetyl coenzyme A synthetase, J Biol Chem 262, 7132-7134.

ACS Paragon Plus Environment

Page 28 of 41

Page 29 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872

Biochemistry

61.

62.

63.

64. 65.

66. 67. 68.

69.

70.

Ganta, S. R., Perumal, S., Pagadala, S. R., Samuelsen, O., Spencer, J., Pratt, R. F., and Buynak, J. D. (2009) Approaches to the simultaneous inactivation of metallo- and serinebeta-lactamases, Bioorg Med Chem Lett 19, 1618-1622. James, L. C., and Tawfik, D. S. (2003) The specificity of cross-reactivity: promiscuous antibody binding involves specific hydrogen bonds rather than nonspecific hydrophobic stickiness, Protein Sci 12, 2183-2193. Baier, F., Chen, J., Solomonson, M., Strynadka, N. C., and Tokuriki, N. (2015) Distinct Metal Isoforms Underlie Promiscuous Activity Profiles of Metalloenzymes, ACS Chem Biol 10, 1684-1693. Nath, A., and Atkins, W. M. (2008) A quantitative index of substrate promiscuity, Biochemistry 47, 157-166. Elias, M., and Tawfik, D. S. (2012) Divergence and convergence in enzyme evolution: parallel evolution of paraoxonases from quorum-quenching lactonases, J Biol Chem 287, 11-20. Glasner, M. E., Gerlt, J. A., and Babbitt, P. C. (2006) Evolution of enzyme superfamilies, Curr Opin Chem Biol 10, 492-497. Soskine, M., and Tawfik, D. S. (2010) Mutational effects and the evolution of new protein functions, Nat Rev Genet 11, 572-582. Reisch, C. R., Moran, M. A., and Whitman, W. B. (2008) Dimethylsulfoniopropionatedependent demethylase (DmdA) from Pelagibacter ubique and Silicibacter pomeroyi, J Bacteriol 190, 8018-8024. Reisch, C. R., Stoudemayer, M. J., Varaljay, V. A., Amster, I. J., Moran, M. A., and Whitman, W. B. (2011) Novel pathway for assimilation of dimethylsulphoniopropionate widespread in marine bacteria, Nature 473, 208-211. Gerlt, J. A., Babbitt, P. C., Jacobson, M. P., and Almo, S. C. (2012) Divergent evolution in enolase superfamily: strategies for assigning functions, J Biol Chem 287, 29-34.

873 874

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 41

875 876 877

Table 1: Features of the explored DddY and DddL enzymes compared to other CupinDLL family members

878

Family

Species

Alcaligenes faecalis J481

DddY

Desulfovibrio acrylicus

NCBI Accession

E7DDH2.1

SHJ73420.1

Thioclava pacific WP_051692700.1 Rhodobacter sphaeroides Q3J6L0.1 2.4.1 Ruegeria AAV94883.1 pomeroyi DSS-3 DddQ Ruegeria lacuscaerulensis D0CY60 ITI-1157 Candidatus Pelagibacter WP_011281678.1 DddK ubique HTCC1062 Ruegeria DddW WP_011046214.1 pomeroyi DSS-3

879 880 881 882 883 884 885 886 887

401

403

418

5

KM (mM)

kcat -1 (s )

2.56

0.9*10

Metal cofactor 3

902±35

3.5*10

390 15 (Ref. )

2.3*10 1.41 0.3*10 15 15 15 (Ref . ) (Ref. ) Ref.

5

1391±76

Ferrimonas WP_028114584.1 kyonanensis Shewanella putrefaciens CN- WP_011920089.1 32 Acinetobacter WP_004831354.1 bereziniae

DddL

Specific Length activity kcat/KM (amino -1 -1 -1 (μmol min mg (M s ) acids) -1 enzyme )

2100 16 (Ref. ) Unclear (1)

1.32*10

6

0.85

2+

3

1.13*10

6

Mn (2)

3 2+

3

4.60*10 0.45 2.07*10 16 16 16 (Ref. ) (Ref. ) (Ref. )

Mn (2)

N.D.

N.D.

N.D.

N.D.

N.D.

N.D.

N.D.

N.D.

5.06 18 (Ref )

8.3*10 18 (Ref )

414

Unclear (1)

401

1.1*10 18 (Ref )

1.66*10 18 (Ref )

232

≥83 (3)

N.D.

N.D.

N.D.

Mn

2+

232

≥70 (3)

N.D.

N.D.

N.D.

Mn

2+

201

2~5*10 13, 19 (Refs. )