Amplification of Ehrlich pathway enhances production of higher alcohols in engineered Yarrowia

lipolytica

Ewelina Celińska1*, Monika Borkowska1, Wojciech Białas1, Monika Kubiak1, Paulina Korpys1, Marta

Archacka1, Rodrigo Ledesma-Amaro2,3, Jean-Marc Nicaud2

¹Department of Biotechnology and Food Microbiology, Poznan University of Life Sciences, ul. Wojska

Polskiego 48, 60-627 Poznań, Poland

²Micalis Institute, INRA-AgroParisTech, UMR1319, Team BIMLip: Integrative Metabolism of Microbial

Lipids, Domaine de Vilvert, 78352 Jouy-en-Josas, France

3 Imperial College London, Centre for Synthetic Biology, South Kensington Campus, London SW7 2AZ,

United Kingdom

*Corresponding author: Ewelina Celińska, phone: +48618466007, e-mail:

ewelina.celinska@up.poznan.pl

Abstract

Microbial cells can produce a vast spectrum of chemical compounds, including those most desired by

the global chemical market, as for example higher alcohols, which are promising alternative fuels and

chemical feedstock. In the current research, we investigated the effects of the Ehrlich pathway

amplification on higher alcohols production in Y. lipolytica, which directly follows our previous

findings concerning elucidation of putative molecular identities involved in this pathway. To this end

we constructed two alternative expression cassettes composed of previously identified genes and

cloned them under the control of constitutive pTEF promoter, and by this - released them from

extensive native regulation. The effects of the pathway amplification were investigated upon

provision of different, Ehrlich pathway-inducing amino acids (L-Phe, L-Leu, L-Ile and L-Val). Depending

on the expression system (cassette-genetic background of the host) we could observe different

effects on the targeted compound production. In general, amplification of the Ehrlich pathway in

many cases led to increased formation of a respective higher alcohol from its precursor by on

average 1.5 to >> 2-fold. We observed interesting effects of bat2Δ on 2-PE and 2-PEAc formation,

significant relationship between L-Val and L-Phe catabolic pathways, and extensive “cross-induction”

of the derivative compounds synthesis by a non-direct precursors.

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

12

Keywords:

higher alcohols, amino acid metabolism, Ehrlich pathway, bioconversion, Yarrowia lipolytica,

metabolic engineering

Abbreviations

AA – Amino Acid

BCAA – Branched Chain Amino Acid (L-Leu, L-Val, L-Ile)

aroma AA – Amino Acid with aromatic residue (L-Phe, L-Trp, L-Tyr)

GGFs– Golden Gate Fragments

GGVs – Golden Gate Vectors

GGVA – Golden Gate Vector bearing Assembly

GGA - Golden Gate Assembly

2-MEB – 2-Methyl-1-Butanol

2-MEBAc - 2-Methyl-1-Butyl Acetate

2-MEP – 2-Methyl-1-Propanol

2-MEPAc - 2-Methyl-1-Propyl Acetate

2-MEPacid - 2-Methyl Propanoic acid/Isobutyric acid

3-MEB – 3-Methyl-1-Butanol

3-MEBAc - 3-Methyl-1-Butyl Acetate

3-MEBacid - 3-Methyl Butanoic acid/ Isovaleric acid

2-PE – 2-Phenylethanol

2-PEAc – 2-Phenethylacetate

ARO8 [YALI0E20977p] - aminotransferase

BAT2 [YALI0F19910p] – aminotransferase

ARO10 [YALI0D06930p] - decarboxylase

Adh [YALI0F24937p]- alcohol dehydrogenase

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

34

Introduction

Growing awareness of the environmental burden imposed by chemical and fuel industry, as

well as uncertain supply of the fossil fuels are the major driving forces for development in the green

chemistry area. Apart from lower environmental impact, biotechnological processes employing

indigenous or engineered potential of microbial metabolism, may turn out to be superior in terms of

efficacy and economic feasibility compared to the chemical synthesis, as already demonstrated by

several examples (Nakamura and Whited 2003; Schrader et al. 2004). Amongst top most desired

chemicals on the global market, the authorities and stakeholders name several diols (2,3-butanediol,

1,3-propanediol), organic acids (succinic acid, lactic acid), lipid / alkene-derived metabolites

(ethylene, fatty acids, fatty alcohols) and higher alcohols (butanol, isobutanol) (Sauer et al. 2008;

Alper and Stephanopoulos 2009; Borodina and Nielsen 2014; Jansen and van Gulik 2014; Dai and

Nielsen 2015; Generoso et al. 2015; Nielsen and Keasling 2016). Higher alcohols are aliphatic,

medium-chain compounds of more than three carbon atoms with branched or aromatic side-

residues. Due to their useful chemical properties, higher alcohols have found applications in

numerous industrial processes, starting from production of alternative fuels with superior

physicochemical properties, through serving as chemical feedstock, to aromatization of various

goods (Nozzi et al. 2014). Microbial cells can produce a vast spectrum of chemical compounds,

including those most desired by the global chemical market. Higher alcohols can be produced in

microbial cells de novo by sequestering carbon skeletons from the central carbon metabolism, or in a

course of bioconversion from structurally-related precursors. While de novo synthesis seems more

reasonable in terms of the culture media costs (no need for the precursors), it is by far less efficient

than bioconversion (Schrader et al. 2004). Bioconversion pathway per se is decoupled from the core

carbon metabolism and thus the precursor is not dissipated in the competing pathways. As such,

carefully optimized bioconversion processes may approach perfect atom economy, expressed in high

yield of the product per consumed precursor. Higher alcohols, either synthesized de novo or in a

course of bioconversion, are inherently coupled with AA metabolism. Catabolism of L-leucine (L-Leu),

L-isoleucine (L-Ile), L-valine (L-Val) or L-phenylalanine (L-Phe) leads to the formation of higher

alcohols, namely 3-MEB (Isoamyl alcohol; CAS 123-51-3; 3-MBE), 2-MEB (Active Amyl Alcohol; CAS

137-32-6; 2-MBE), 2-MEP (Isobutanol; CAS 78-83-1; 2-MPE) and 2-Phenylethyl Alcohol (2-PE; CAS 60-

12-8; 2-PE), respectively. The metabolic pathway of branched chain AAs (L-Val, L-Leu, L-Ileu),

aromatic AAs (L-Phe, L-Tyr, L-Trp) and L-Met has been initially described by Ehrlich (Ehrlich 1907),

and scrutinized in the following studies employing Saccharomyces cerevisiae (Dickinson et al. 2003;

Hazelwood et al. 2008) Figure 1. In terms of the yeast cell physiology, the Ehrlich pathway is used to

seize nitrogen from AA. In terms of the chemical industry – it leads to formation of valuable chemical

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

56

compounds of significant interest. These chemical compounds derive from the carbon skeletons of

the AA which basically remain intact. The Ehrlich pathway is initiated by the action of

aminotransferase, which uses 2-oxoglutarate or other 2-oxo acids as amino acceptors to the

corresponding keto acids (phenylpyruvic acid from L-Phe, α-ketoisocaproic acid from L-Leu, α-

ketoisovaleric acid from L-Val, and α-ketomethylvaleric acid from L-Ile). For some AAs (e.g., L-Ala),

the resulting 2-oxo acid (pyruvate), can be readily metabolized in the central carbon metabolism by

the producer cell. However, in the other cases, such as aroma AA and BCAA, the 2-oxo acids resulting

from transamination are not intermediates of the central metabolism, implying that they cannot be

used as auxiliary carbon sources. The 2-oxo acids are secreted out of the cells as they are, or may be

also transformed into higher alcohols or derivatives by the yeast cells before they are excreted into

the growth medium (Vuralhan et al. 2005). Synthesis of each of the higher alcohols may be

accompanied by concomitant formation of its ester (3-MEBAc (CAS 123-92-2), 2-MEBAc (CAS 624-41-

9), 2-MEPAc (CAS 110-19-0)) or fusel acid (3-MEBacid (CAS 503-74-2) or 2-MEPacid (CAS 79-31-2)),

which also find numerous applications in industrial processes. Noteworthy, higher alcohols and their

esters constitute an important fraction of aroma compounds used in aromatization of various

commodities, including foodstuff and cosmetics. In contrast to fuel, chemical or polymer industry,

aroma production is a specific branch which is strongly influenced by the consumer acceptance. Thus,

highly-valued natural origin of the compound is an additional parameter driving its final market price.

According to European regulation on flavors (EEC No 1334/2008), aroma chemical compound can be

labelled as “natural” when “the flavoring substance is obtained by appropriate physical, enzymatic or

microbiological processes from material of vegetable, animal or microbiological origin either in the

raw state or after processing”. This in turn, makes biotechnology a powerful alternative compared to

chemical synthesis, contributing to substantial burden imposed on the natural environment and

providing “synthetic” compounds, or to exploitation of natural resources, which are highly

susceptible to seasonal variations, and/or hard-to-foresee climate-geographical factors.

Amongst microbial species potentially assisting production of higher alcohols, yeasts occupy

an important place, with the leading role assigned to a model species S. cerevisiae. A number of

studies investigating molecular background of the compounds synthesis (Dickinson et al. 2003;

Hazelwood et al. 2008; Styger et al. 2013), inventive strategies of the pathway engineering (Brat et al.

2012; Avalos et al. 2013; Nozzi et al. 2014) or intensification of the production processes employing

this species, have been published to date. Apart from S. cerevisiae, the other yeast species are

becoming attractive alternatives in microbial production of higher alcohols (Sibirny and Scheffers

2002; Forti et al. 2015; Grondin et al. 2015; Celińska et al. 2018). Although their superior potential

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

78

towards high-level production of higher alcohols has been demonstrated, the number of studies

within this area stay far behind the model species.

A list of valuable biomolecules synthesized by either native or engineered metabolism of

Yarrowia lipolytica is continuously increasing (Alper and Stephanopoulos 2009; Groenewald et al.

2014; Madzak 2015; Ledesma-Amaro and Nicaud 2016). Apart from a number of bioprocesses

employing Y. lipolytica’s metabolism shown to be highly promising in the research context, several

technologies are operating in the industry, like the production of citric acid (Pfizer Inc. and ADM,

USA), erythritol (Baolingbao Biology Co., China), eicosapentaenoic acid-rich oils (DuPont, USA),

nutrient-rich fodder yeast (Skotan SA, Poland; (Białas et al. 2015), or enzymes (Oxyrane, Belgium

and UK; Artechno, Belgium; Mayoly Spindler, France) (Darvishi Harzevili 2014; Groenewald et al.

2014), demonstrating high propensity of this yeast to serve as an industrial workhorse. For a long

time, Y. lipolytica has been employed in industrial production of creamy-peach aroma compound γ-

decalactone from ricinoleic acid, commercialized by BASF. Recently, we have shown that Y. lipolytica

is an efficient producer of another highly-valued higher alcohol – 2-PE in a course of bioconversion

from L-Phe (Celińska et al. 2013). In the following study we performed whole-proteome analysis in

order to fish-out molecular identities involved in the bioconversion of the AA to the higher alcohol

(Celinska et al. 2015). As mentioned above, the Ehrlich pathway is executed via three subsequent

reactions: deamination, decarboxylation and oxidation, catalyzed by transaminase (like ARO8/9 or

BAT1/2 [EC:2.6.-.-.]), 2-oxo acid decarboxylase (like ARO10 or PDC1/3/5 [EC:4.1.1.-]), and

dehydrogenase (like ADH1–5 in S. cerevisiae; [EC:1.1.1.90]), respectively. It is known, that in S.

cerevisiae each step can be conducted by several possible enzymes implying that the Ehrlich pathway

is highly redundant (Lilly et al. 2000; Dickinson et al. 2003; Vuralhan et al. 2003; Vuralhan et al. 2005;

Lilly et al. 2006; Hazelwood et al. 2008). Although the general overview of the Ehrlich pathway was

known, involvement of particular molecular identities in this pathway had not been elucidated

before in Y. lipolytica. Based on the proteomic data, we observed presence or over-representation of

specific proteins which, based on their sequence, function and similarity to S. cerevisiae’s “template”

pathway, could be directly linked to the general pathway of higher alcohols formation. The first step

of the Ehrlich pathway in Y. lipolytica could be catalyzed by ARO8 [YALI0E20977p; 2.6.1.5] and BAT2

[YALI0F19910p; 2.6.1.57] homologues whose presence was identified in the analyzed proteome.

YALI0D06930p [EC:4.1.1.43] was a single homolog of ARO10/PDCs which has been identified with

good matching results only in samples where the Ehrlich pathway was induced. This protein has been

also identified with good matching results, when blasting all the homologous sequences

(ARO10/PDC1/PDC5/PDC6) from S. cerevisiae against Y. lipolytica protein database. The last activity

that is directly involved in AAs catabolism via Ehrlich pathway is dehydrogenase. Determining the

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

910

molecular identities of the dehydrogenase involved in the Ehrlich pathway presents a challenge

(Hazelwood et al. 2008), Dickinson, Salgado and Hewlins (2003) have shown that any of the alcohol

dehydrogenases (Adh1–5 and Sfa1) encoded within S. cerevisiae genome can catalyze this final

reaction, demonstrating high redundancy of this step. In the previous study we have observed high

over-representation of a single dehydrogenase having broad-range substrate specificity -

YALI0F24937p. Significant induction of its expression, implied its involvement in the catabolism of

AAs via the Ehrlich pathway. Consequently, based on comparative whole proteome analysis we

identified candidate genes for genetic manipulations oriented towards enhanced formation of higher

alcohols in Y. lipolytica.

In the current research, we investigated the effects of the Ehrlich pathway amplification on

higher alcohols production in Y. lipolytica. To this end we constructed two alternative expression

cassettes composed of previously identified aminotransferases encoded by YALI0E20977g or

YALI0F19910g (either ARO8 or BAT2, depending on the cassette variant), decarboxylase encoded by

YALI0D06930g, and YALI0F24937g coding for broad substrate specificity dehydrogenase. The genes

were cloned under the control of constitutive promoter pTEF and Lip2 terminator, thus released

from native regulation mechanisms. The three-gene assemblies were transformed in two Y. lipolytica

strains (JMY2101 and JMY3150) differing in the genetic background related to the Ehrlich pathway

(JMY3150). The effects of the pathway amplification were investigated upon provision of different,

Ehrlich pathway-inducing AAs in comparison to the reference strain. It is known that at greater

concentrations, higher alcohols exert toxic effect toward microbial cells. Thus, to avoid toxic effect of

higher alcohols overproduction, we applied suboptimal conditions of the precursor feeding, to

enable observation of the modifications effects.

Materials and Methods

Strains and routine culturing conditions

E. coli and Y. lipolytica strains used in this study are listed in Table S1. All cultivations required

for molecular biology protocols complied with the standards described in (Barth and Gaillardin 1996;

Sambrook and Russell 2001). Briefly, E. coli strains were routinely maintained on LB medium (g/L: 10,

bacteriological peptone, 10, NaCl, 5, yeast extract; liquid or solidified with agar, 15 g/L)

supplemented with appropriate antibiotic when necessary (ampicillin at 100 µg/L; kanamycin 40

µg/L), at 37°C, 250 rpm in rotary shaking incubator (Biosan). Y. lipolytica strains were cultured in YNB

(g/L: 5, ammonium sulfate, 1.7, YNB without AA and ammonium sulfate, 20, glucose) or YPD (g/L: 10,

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

1112

yeast extract, 20, bacteriological peptone, 20, D-glucose) medium (liquid or solidified with agar, 15

g/L), at 30°C, 250 rpm.

Molecular biology protocols

If not stated otherwise, all the molecular biology protocols followed the methodologies

described in (Sambrook and Russell 2001). All oligonucleotides and plasmids used in this study are

listed in Table S2.

E. coli and Y. lipolytica transformations were conducted according to the heat-shock or LiAc

methodologies described in (Sambrook and Russell 2001) and (Barth and Gaillardin 1996),

respectively. Genomic DNA isolation from Y. lipolytica, plasmid isolation from E. coli, DNA fragments

extraction from agarose gel or purification of DNA fragments were all conducted using appropriate

kit from A&A Biotechnology (Poland) – Genomic Mini AX Yeast, Plasmid Mini, Gel-Out, Clean-Up.

Restriction digestion of DNA fragments was done using either NotI enzyme (Thermo Scientific) or

BsaI (New England Biolabs). Routine colony PCR with E. coli biomass was run using Taq DNA

polymerase (A&A Biotechnology; Poland) while colony PCR with Y. lipolytica biomass was conducted

using Phire Hot Start II DNA Polymerase (Thermo Scientific). Phire Hot Start II DNA Polymerase was

also used for amplification of any GGFs from Yarrowia’s genomic DNA, and for assembly PCR

procedure enabling elimination of internal BsaI recognition sites found within ARO8 (YALI0E20977p),

BAT2 (YALI0F19910p), ARO10 (YALI0D6930g) nucleotide sequences. In silico restriction analysis was

done using BioEdit’s Restriction Map tool (Hall 1999) and confirmed with control restriction

digestion. Elimination of undesired recognition sites from the targeted sequences was confirmed

through sequencing (GATC Biotech, Germany; Genomed, Poland), and BsaI restriction digestion. The

reactions were conducted according to protocols provided by the manufacturers.

Modular cloning – Golden Gate Assembly and positive clones selection

The modular cloning protocol followed a previously set standard (Celinska et al. 2017). All the

plasmids and primers used in this study are listed in Table S2. Briefly, a set of thirteen 4 nt overhangs

was developed together with the corresponding destination vector pSB1A3-RFP,

(http://parts.igem.org/Collections). Genomic DNA sequence of Y. lipolytica CLIB122 used in this study

as GGFs amplification template can be acquired from GRYC database (http://gryc.inra.fr/). All the

GGVs (GGF-bearing donor vectors) were constructed on a pCR Blunt II TOPO vector’s backbone

(Thermo Scientific) and a corresponding GGF amplicon, according to the instruction provided by the

manufacturer. Ligation mixture was transformed into one of E. coli host strains (JM109, TOP10 or

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

1314

DH5α), and correct GGVA organization was confirmed through restriction digestion and

electrophoresis. DNA concentration was assessed using NanoDrop device (Thermo Fisher Scientific).

The Golden Gate reaction mixture contained precalculated equimolar amounts of each GGF and the

destination vector (50 pmoles of ends), 2 µL of T4 DNA ligase buffer (New England Biolabs), 5 U of

BsaI, 200 U of T4 and ddH2O up to 20 µL. The following thermalcycling profile was applied: [37°C - 5

min, 16°C - 2 min] x 60, 80°C - 5 min, 15°C ∞. Subsequently, the reaction mixtures were used for E.

coli JM109 transformation. White colonies were screened for identification of complete GGA through

colony PCR covering adjacent elements (e.g. P1_F and G1_R), followed by plasmid isolation,

restriction digestion and multiplex PCR. Complete GGAs were linearized with NotI and transformed

into JMY2101 and JMY3150 strains. Clones appearing after 48 h incubation at 30°C on YNB-selection

plates were replica-plated on fresh YNB plates. All the clones were screened for the GGA presence

through colony PCR covering adjacent GGFs. All the strains confirmed to bear a complete GGA were

deposited as glycerol stocks at -80°C.

Batch bioconversion cultures

Batch bioconversion cultures to be analyzed through GC were conducted with at least two

subclones of each type selected from deposited library: JMY2101_GGA077, JMY3150_GGA077,

JMY2101_GGA097 and JMY3150_GGA097; each cultured in triplicate. JMY2900 control strain was

cultivated in parallel with each culture run as a non-transformed reference. The cultures were

conducted in shake flasks of total volume 50 mL and working volume 5 mL, at 30°C, 250 rpm. The

basal production medium contained in g/L: 50, glycerol, 1 mL, mineral salt solution II (Barth and

Gaillardin 1996), 15, KH2PO4, 0.5, MgSO4x7H2O, 0.2, yeast nitrogen base. The medium was

supplemented with appropriate AA, either: L-Phe, BCAA (Sigma Aldrich) up to 2 g/L. The supplements

were prepared as 10-fold concentrated stocks, sterilized through filtering through 0.22 µm (Millex,

Millipore). The cultures were continued for 72 h and the samples were collected at 24 h intervals and

centrifuged (15 min, 15 min, 22,639 x g; Biofuge R, Hareus). Cellular pellets and supernatants were

frozen separately at -20°C until analysis. The samples were used for analysis of biomass production

and targeted aroma compounds concentration. Biomass accumulation was assessed through OD600

measurements using UV/Vis Spectrophotometer SPECORD (Analytik Jena, Germany). Concentration

of targeted metabolites retained in the culture medium was determined through gas

chromatography analysis (GC; procedure described below).

Batch bioconversion cultures to be analyzed through GC-MS were conducted in 300 mL

Erlemeyer flasks in a working volume of 100 mL, in the same medium and culturing conditions as

described above. Three independent cultivations were conducted for each recombinant strain type

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

1516

(JMY2101_GGA097, JMY3150_GGA097, JMY2900). Samples were collected at 48 h of culturing and

subjected to GC-MS analysis.

Extraction of volatile compounds in the culture media

After thawing, the supernatant samples were vortexed, centrifuged (15 min, 22,639 x g;

Biofuge R, Hareus) and passed through 0.45 µm filter (Millex, Millipore). Extraction of the aroma

compounds was conducted through mixing the filtrate with equal quantities (GC analysis) or 9:1 ratio

(GC-MS analysis) of dichloromethane, vigorous shaking, and phases separation via centrifugation (15

min, 22,639 x g; Biofuge R, Hareus). The organic phase was collected and subjected to either GC or

GC-MS analysis. Comparison of the extraction efficiency was assessed by conducting parallel

cultivations in different culture volumes (5 mL vs 100 mL) followed by extraction with either 0.5 or

0.1 volume of dichloromethane.

GC analysis

The GC analysis of the organic phase (1 µL) was conducted using Agilent Technologies 7890A

System gas chromatograph equipped with a Zebron™ ZB-WAXplus™ GC capillary column and FID

detector. Mobile phase: hydrogen (30 mL/min), air (300 mL/min), helium (20 mL/min). The oven

programme was as follows: 80°C for 1 min, 80°C to 120°C, at 20°C/min, followed by 120°C to 220°C,

at 120°C/min, and final hold at 220°C for 3 min. Temperature of detector - 250°C. Data were analyzed

using Agilent Chemstation software. Compounds were identified and quantified through comparison

with analytical standards of known concentrations: 1) L-Leu derivatives: 3-MEB (CAS 123-51-3), 3-

MEBAc (CAS 123-92-2), 3-MEBacid (CAS 503-74-2); 2) L-Val derivatives: 2-MEP (CAS 78-83-1), 2-

MEPAc (CAS 110-19-0), 2-MEPacid (CAS 79-31-2); 3) L-Ile derivatives: 2-MEB (CAS 137-32-6), 2-

MEBAc (CAS 624-41-9); 4) L-Phe derivatives: 2-Phenylethyl Alcohol (CAS 60-12-8), 2-Phenylethyl

Acetate (CAS 103-45-7). All analytical standards were purchased from Sigma-Aldrich. Presence and

quantity of all the compounds was determined in all analyzed samples, however, only the “non-zero

concentration” compounds are shown in the respective figures in the Results section.

GC-MS analysis

GC-MS analysis was conducted at Wielkopolska Center of Advanced Technologies (Poznan,

Poland). GC-MS SCION TQ (Bruker) equipped with BP5MS capillary column (5% Phenyl

Polysilphenylene-siloxane) 30 m x 0,25 mm (length x inner diameter), 0,25 µm mobile phase

thickness was used. Mobile phase: helium at 1 mL/min. Oven temperature profile: initial column

temperature: 40°C - 3.5 min, followed by 10°C/min up to 250°C, 250°C - 5 min hold. Total analysis

time: 29.5 min. Temperature of injector: 185°C, ion source: 250°C, transfer line: 200°C. Ioninzation 70

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

1718

eV, scanning range: 25-300 m/z. Each sample was analyzed in technical triplicate. The obtained raw

data were analyzed using Bruker MS Workstation 8.0 Software. Compounds contained in the sample

were identified through comparison of spectra against NIST database.

Results

Design and construction of recombinant strains

DNA constructions were designed and assembled according to previously described

methodology (Figure 2AB). DNA sequence analysis of the ARO-genes revealed that ARO8, BAT2 and

ARO10 all bear two internal BsaI sites, which were eliminated through assembly PCR with

corresponding “BsaI del” primers (Table S2) and verified through sequencing and in silico translation.

Assemblies were first constructed and amplified in E. coli cells. Subsequently, correct constructions

were transformed into Y. lipolytica JMY2101 and JMY3150 strains. Clones bearing complete

assemblies were selected by colony PCR with primers covering adjacent biobricks (gene and

regulatory element) (Figure 2C). Positive clones of each of the four types (JMY2101 bearing either

GGA077 and GGA097 assemblies, or JMY3150 bearing GGA077 and GGA097 assemblies) were

deposited as glycerol stocks and subjected to further studies.

Phenotype evolution and analyses

Representative clones of the recombinant strains: JMY2101_GGA0077, JMY2101_GGA0097,

JMY3150_GGA0077 and JMY3150_GGA0097, were selected and subjected to bioconversion cultures

with L-Phe, L-Leu, L-Ile and L-Val as the sole nitrogen sources. The results are presented in Figures 3

and 4, as well as in supplementary material S1 and S2.

In general, presuming that overall biomass accumulation reflects feasibility of a given AA

assimilation, L-Val was the most easily utilized AA amongst all strains (13.82±0.5 OD600), followed by

L-Leu (13.64±0.35 OD600 average in JMY2900 and JMY2101), which assimilation was impaired in

bat2Δ JMY3150 strains (9.64±0.63 OD600; Figure S1.C.G, S2.C.G). In GGY0077 strains, the weakest

growth was observed upon feeding with L-Ile and L-Phe (9.18±0.06 and 8.61±0.39 OD600), while in

GGY0097 strains – the lowest OD600 readouts were observed in cultures with L-Phe as the sole

nitrogen source (7.1±0.15 OD600; Figure S1.A.E, S2.A).

With respect to the strains transformed with GGA0077 cassette cultured in the bioconversion

medium supplemented with L-Phe (Figure 3.A), more than 50% increase in 2-PE formation was

observed when compared to the reference strain (average for the recombinants 0.65±0.064 vs

0.424±0.078 g/L for the reference strain), while the growth was not significantly compromised in the

engineered strains (8.6±0.52 vs 10.04±1.04 OD600; Figure S1.A). The final specific titer of 2-PE was

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

1920

exactly 2-fold higher for the strains bearing recombinant ARO genes under the control of constitutive

pTEF promoter (average 0.084±0.001 vs 0.042 g/OD600, for the recombinants and the reference

strain, respectively; Figure S1.B). Considering the accompanying metabolites, all the recombinants

synthesized ~50% higher quantities of 2-MEPacid (co-product of L-Val bioconversion) while cultured

in L-Phe-supplemented medium, which was demonstrated by both the final concentration and

specific titer (average 0.2±0.0014 vs 0.13±0.02 g/L±SD (1.54-fold increase) and average 0.023 vs

0.012 g/OD600 (1.92-fold increase), for the recombinants and the reference strain, respectively;

Figure 3.B and S1.B). The levels of the other analyzed compounds (in g/L; 3-MEBacid, 2-PEAc) were

more deviating than actually different between the recombinants and the reference strain, however,

the final concentration and specific titer of 2-MEB (L-Ile derivative) were uniformly more than 2-fold

higher for all the recombinants than for the reference strain (0.011±0.001 vs 0.0038±0.0003 g/L±SD

and 0.0012 vs 0.00038 g/OD600 for the recombinants and the reference strain; Figure 3.B, S1.B).

Upon cultivation with L-Phe as the sole nitrogen source, the bat2Δ genotype in GGY0077s slightly

improved 2-PE formation, suggesting a kind of competition for the NH4+ residues between “aromatic

AA” and “BCAA” bioconversions, initiated by ARO8 and BAT2 aminotransferases, respectively, which

was relieved upon deletion of BAT2 gene in JMY3150. This statement is corroborated by the results

obtained with the JMY2101 and JMY3150 strains transformed with GGA0097 cassette (Figure 4.A and

S2.B). Additional copy of BAT2 aminotransferase in JMY2101 (altogether 2 copies – one under native

regulation, and the other – under pTEF), ultimately led to slightly lower 2-PE formation in a course of

L-Phe bioconversion (g/L±SD: JMY2101_GGA097: 0.37±0.05 / 1.37-fold decrease, JMY3150_GGA097:

0.48±0.042 / 1.06-fold decrease, reference strain: 0.508±0.018), which was also illustrated by specific

titers of 2-PE per OD600 (0.053 2-PE g/OD600 giving 1.36–fold decrease for JMY2101-derivatives,

0.066 2-PE g/OD600 giving 1.1–fold decrease for JMY3150-derivatives and 0.072 2-PE g/OD600 for

JMY2900; Figure S2.B). Consequently, it can be stated that pTEF-driven expression of ARO10 and Adh

did not result in higher bioconversion of L-Phe to 2-PE when BAT2 was concomitantly expressed,

either in JMY2101 or JMY3150. On the other hand, amplification of ARO8 and the following genes of

the Ehrlich pathway, enhanced 2-PE synthesis by more than 50% in the recombinants in general.

Further enhancement by more than 60% was observed upon deletion bat2Δ.

It is commonly now known that AA metabolism is more an extensive net of biochemical

reactions than a linear pathway. Hence, we also studied performance of here developed Y. lipolytica

recombinants upon provision of the other AA, known to be metabolized through Ehrlich pathway,

namely BCAA (L-Leu, L-Val, L-Ile). Considering recombinants bearing constitutive expression of the

Ehrlich pathway in GGA077 variant cultured in BCAA-containing media, expectedly production of 2-

PE was rather low, reaching on average 0.01 – 0.02 g/L (Figure 3.D,F,H). Nevertheless, the highest 2-

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

2122

PE concentrations were observed in the case of JMY3150_GGA077 strains (doubled ARO8, ARO10

and Adh, in bat2Δ) cultured in L-Val containing media (0.02±0.004 g/L / 4.35-fold increase for the

JMY3150-derivatives, 0.007±0.001 / 1.52-fold increase for the JMY2101-derivatives, 0.0046±0.001

g/L for JMY2900; g/L±SD), again suggesting that bat2Δ genotype is advantageous for 2-PE formation

(Figure 3.H). Cultivation of the GGA077-bearing strains in L-Ile-containing media showed that the

GGA077 modification was the enabling factor for detectable 2-PE formation, as the reference strain

did not formed this metabolite at all, as shown by 2-PE and its corresponding ester levels (2-PEAc)

(Figure 3.F). Also, an increase in 2-PE and its ester formation was observed for the GGA077-bearing

strains cultured in L-Leu-supplemented medium (average for the recombinants 0.0083±0.0014 vs

0.0043±0.0011 g of 2-PE/L±SD for the reference; 1.93-fold increase; Figure 3.D).

Interestingly, amplification of the Ehrlich pathway (GGA077) uniformly led to increased

formation of a respective higher alcohol, being derivative of the provided BCAA: 3-MEB from L-Leu

(0.76±0.038 / 2.38-fold increase for JMY2101-derivatives, 0.56±0.042 / 1.75-fold increase for

JMY3150-derivatives, and 0.32±0.04 for JMY2900; g/L±SD) (Figure 3.C), 2-MEP from L-Val

(0.077±0.012 / 2-fold increase for the JMY2101-derivatives, 0.086±0.049 / 2.26-fold increase for the

JMY3150-derivatives, and 0.038±0.004 for JMY2900; g/L±SD) (Figure 3.G), and 2-MEB from L-Ile

(0.025±0.002 / 1.6-fold increase for the JMY2101-derivatives, 0.023±0.002 / 1.5-fold increase for the

JMY3150-derivatives, and 0.0157±0.002 for JMY2900; g/L±SD) (Figure 3.E), which was also reflected

by specific titers of the derivative compounds (Figure S1.D, F, H). It seems that amplification and

constitutive overexpression of the broad range decarboxylase and oxidoreductase and relatively

broad range aminotransferase ARO8, improved formation of the higher alcohols. Even more

interestingly, the accompanying fusel acid production was concomitantly decreased for the

recombinants when compared to the reference strain: 3-MEBacid from L-Leu (0.021±0.0005 / 1.43-

fold decrease for the JMY2101-derivatives, 0.01±0.005 / 3-fold decrease for the JMY3150-derivatives,

and 0.03±0.003 for JMY2900; g/L±SD) and 2-MEPacid from L-Val (0.026±0.0044 / 3.88-fold decrease

for the JMY2101-derivatives, 0.043±0.0041 / 2.35-fold decrease for the JMY3150-derivatives,

0.101±0.0042 for JMY2900; g/L±SD). Which was also demonstrated by corresponding specific titer

values (Figure S1.D, H). Furthermore, we noted that for GGY0077 strains, the titers of 2-MEB (L-Ile

derivative) were increased upon supplementation with its “not direct” precursor BCAA - L-Leu

(0.85±0.043 / 2.36-fold increase for the JMY2101-derivatives, 0.54±0.07 / 1.5-fold increase for the

JMY3150-derivatives, and 0.36±0.076 for JMY2900; g/L±SD) (Figure 3.D). Both concentrations and

specific titers of the two isoforms: 2-MEB and 3-MEB in a course of L-Leu bioconversion were highly

corresponding (0.058 / 2.5-fold increase and 0.0637 / 2.5-fold increase in g of 3-MEB and

2-MEB/OD600 for the JMY2101-derivatives, 0.0657 / 2.6-fold increase and 0.0585 / 2.3-fold increase

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

2324

in g of 3-MEB and 2-MEB/OD600 for the JMY3150-derivatives, and 0.023 and 0.0259 g of 3-MEB and

2-MEB/OD600 for the reference; Figure S1.D). Moreover, it seems that ARO8 aminotransferase

presence / constitutive expression (GGY0077) does not impose such an unfavorable effect on the

BCAA-derived higher alcohols production, as it was observed for the BAT2 and 2-PE/2-PEAc synthesis.

It was rather observed, that the BCAA-derived higher alcohol formation could benefit from higher

abundance of the broad substrate specificity decarboxylase and the dehydrogenase. Upon

overexpression of the Ehrlich pathway with ARO8 aminotransferase in the G1 position (GGY0077), we

also observed an interplay between L-Phe and L-Val catabolic pathways. In general, provision of L-Val

promoted production of L-Phe derivatives - 2-PE and its derivative ester (which was more

pronounced in bat2Δ genotype background), when compared to L-Leu and L-Ile, while L-Phe

enhanced production of isobutyrate and 2-MEP (derivatives of L-Val) in these recombinant strains.

The second expression cassette GGA0097 differed from the GGA0077 by bearing BAT2

aminotransferase gene in the G1 position. Upon cultivation of GGA097-bearing strains in L-Leu-

supplemented medium, it was observed that BAT2 gene under native regulation is required for

maintaining growth phenotype corresponding to the reference strain, while lack of this gene resulted

in impaired growth (1.36-fold decrease in OD600 value; Figure S2.C). Surprisingly, such phenomenon

was not observed for the cultures supplemented with L-Ile and L-Val, where the recombinants’

growth remained at the native level by pTEF-BAT2 (Figure S2.E, G). The lower growth rate in the

JMY3150-derivatives was accompanied by slightly higher level of 3-MEB synthesis when expressed

per volume unit (peak concentration at 48 h 0.341±0.0343 vs 0.26±0.034 g/L±SD for JMY3150-

derivatives and the reference strain; 1.31-fold increase), which altogether resulted in importantly

increased specific titer (JMY3150-derivatives vs JMY2900: 0.103±0.026 and 0.028 of 3-MEB

g/OD600±SD at 48h; 3.67-fold increase; Figure 4.B, S2.D). On the other hand, for the JMY2101-

derivatives cultured in L-Leu-based medium the growth characteristic was maintained, still the peak

concentration of the derivative higher alcohol at 48 h was only slightly improved, when expressed in

g/L (0.32±0.025 vs 0.26±0.034 g/L±SD for JMY2101-derivatives and the reference strain; 1.23-fold

increase), but anyway improved specific titer (0.0377±0.0016 3-MEB g/OD600±SD at 48h; 1.34-fold

increase; Figure 4.B, S2.D). With respect to the accompanying metabolites, only 2-PE and 2-PEAc

were synthesized at detectable levels by GGY0097 strains in L-Leu-based cultures. While for the

JMY2101-derivatives, the increase in 2-PE and 2-PEAc formation was relatively low (2-PE:

0.079±0.014 vs 0.049 g/L±SD / 1.6-fold increase, 2-PEAc: 0.33±0.048 vs 0.154 g/L±SD / 2.14-fold

increase, for the recombinants vs the reference), in bat2Δ strains 2-PE and 2-PEAc formation was

strongly enhanced (more than 4-fold in both cases) up to 0.21±0.023 and 0.745±0.047 g/L±SD at 72 h

of culturing, even upon supplementation with L-Leu and not L-Phe (Figure 4.C); which was also

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

2526

reflected by significantly enhanced specific titers: 0.074±0.011 / 6.7-fold increase in 2-PEAc

g/OD600±SD at 72h, and 0.021±0.004 / 6-fold increase in 2-PE g/OD600±SD at 72h for

JMY3150_GGA0097 (Figure S2.D).

As mentioned above, constitutive expression of the Ehrlich pathway in GGA0097 variant,

reconstituted growth impairment observed for bat2Δ strains cultured in L-Ile and L-Val-containing

media. Upon cultivation with L-Ile, the GGA0097 modification increased production of 2-MEB, its

major bioconversion product, only in the JMY2101 strains (at 24 h: 0.35±0.02 vs 0.27±0.015 g/L±SD /

1.3-fold increase for JMY2101 derivatives vs the reference), while in the JMY3150-derivatives, its

production remained the reference level (Figure 4.D). In all strains, 2-MEB was converted to its

corresponding acetate ester, starting from approx. 48 h. In GGY0097 strains cultures with L-Ile, we

observed slightly increased titers of 3-MEBacid (derivative of L-Leu; 0.0132±0.002 / 1.4-fold increase

for the JMY2101 derivatives, 0.011±0.0008 / 1.16-fold increase for the JMY3150 derivatives, and

0.0095±0.0017 g/L±SD for JMY2900; Figure 4.E); while production of 2-PE was observed only in bat2Δ

strains, which was an unique quality. Surprisingly, none of GGY0097s differed significantly from

JMY2900 in production of major L-Val products (2-MEP and its ester) (Figure 4.F). However,

JMY3150_GGA0097 strains cultured with L-Val produced more 2-PE and 2-PEAc, when compared to

the JMY2900 reference strain (2-PEAc at 24 h: 1.163±0.003 / 1.43-fold increase for the JMY2101

derivatives, 0.96±0.08 / 1.18-fold increase for the JMY3150 derivatives, and 0.813±0.04 g/L±SD for

JMY2900; 2-PE at 48 h: 0.42±0.05 / 2-fold increase for the JMY2101 derivatives, 0.22±0.03 / 1.05-fold

for the JMY3150 derivatives, 0.21±0.05 g/L±SD for JMY2900; Figure 4.G). At the following time

points, the 2-PEAc concentration dropped from over 1 g/L to zero in the JMY3150 cultures, which

was not the case for the remaining strains. Moreover, the production of 2-PE was continued only in

the JMY3150-derivatives which again indicates important effect of bat2Δ on 2-PE formation and also

important relationship between L-Val and L-Phe pathways.

Discussion

The current study directly follows our previous findings concerning elucidation of putative

molecular identities involved in bioconversion of precursory AA to higher alcohols via Ehrlich

bioconversion pathway in Y. lipolytica (Celinska et al. 2015). Consequently, here we modified

previously identified elements of the Ehrlich pathway in Y. lipolytica cells by either its overexpression

under strong, constitutive promoter or knock-out, and investigated effects of such manipulations on

the yeast cell metabolism. To the best of our knowledge, this is the first study on engineering the

Ehrlich pathway in Y. lipolytica, as previous reports concerning characterization and manipulation of

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

2728

the Ehrlich pathway were mostly conducted in S. cerevisiae, with only few examples conducted with

other yeast (Hua and Xu 2011; Bolat et al. 2013; Kim et al. 2014b; Romagnoli et al. 2015; Stribny et al.

2016). Previous attempts of the Ehrlich pathway manipulation in S. cerevisiae mainly relied on single

gene deletion or overexpression which led to corresponding reduction or increase in the derivative

bioconversion products (2-MEB, 3-MEB, 3-MEBacid, 2-MEPacid) (Vuralhan et al. 2005; Lilly et al.

2006; Styger et al. 2013). More complex strategy relied on deletion of Ald3 gene, responsible for

phenylacetaldehyde oxidation, competing with 2-PE production, in a strain overexpressing ARO9 and

ARO10 and ARO80 transcription factor, resulted in increased formation of the derivative higher

alcohol (Kim et al. 2014a). Another interesting approach aimed at improvement of de novo

isobutanol production (2-MEP) by re-location of part of Ehrlich pathway to mitochondria

(decarboxylation and reduction) (Brat et al. 2012). Natively, the upstream de novo 2-MEP pathway is

located in mitochondria, while the downstream steps, executed by α-ketoacid decarboxylase and

alcohol dehydrogenase of Ehrlich pathway, are localized in cytoplasm. In order to secure availability

of intermediates without the need for transportation between the compartments, the researchers

attempted mitochondrial engineering. Targeting the entire pathway to mitochondria significantly

improved titer, yield and productivity of the higher alcohol, compared with a partly cytoplasmic

equivalent (Brat et al. 2012).

Due to possible differences in the experimental setups, direct comparison of own and

literature data concerning any bioprocess is not always fully justified. Nevertheless, such

comparisons allow to locate own data in relation to state-of-art. The aim of the present study was

manipulation with putative genes of Ehrlich pathway in Y. lipolytica, as identified in our previous

study, and examination of its effect on higher alcohols production. Thus, the bioconversion cultures

were conducted under suboptimal conditions in terms of amount of the precursory AA. The reason

behind such approach was to avoid a toxicity effect exerted by higher alcohols towards the producer

cells prior to observation of differences caused by the introduced modifications. Considering overall

titers of higher alcohols that were be obtained from the cultures, the highest levels of 2-PE were

obtained in L-Phe supplemented cultures of JMY3150_GGAA0077 and JMY2101_GGA0077 (> 0.7

g/L), followed by approx. 0.5 g/L produced by JMY3150_GGA0097. The highest levels of 2-PE’s

acetate ester (2-PEAc), which is highly valued aroma compound, were reached surprisingly in L-Val or

L-Leu supplemented cultures (and not L-Phe) of JMY3150_GGA0097 strain (> 1.0 and > 0.74 g/L).

Isobutanol (2-MEP) was accumulated at the highest concentration (0.347 g/L) in L-Val supplemented

cultures of JMY3150_GGA0097, while 2-MEB and its isoform (3-MEB) – in L-Leu supplemented

cultures of JMY2101_GGA0077 (0.85 and 0.75 g/L, respectively). As reported previously 2-MEP could

be accumulated up to 0.16 g/L by wild type S. cerevisiae, and up to 0.45 g/L by mutant strain, out of 2

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

2930

g/L of L-Leu supplemented into the culture medium (Abe and Horikoshi 2005). Its isomer – n-butanol,

could be accumulated from provided glycine via artificial pathway up to 0.092 g/L in engineered S.

cerevisiae (Branduardi et al. 2013). Industrial S. cerevisiae JP1 strain cultured in L-Leu supplemented

(at 9.8 g/L) bioreactor cultivations, produced up to 1.1 g/L of 3-MEB, together with 0.08 g/L of 2-MEP

and 0.05 g/L of 2-PE (Espinosa Vidal et al. 2015). Up to date, the highest levels of 2-MEP that have

been synthesized de novo (up to 0.63 g/L) were reported for recombinant S. cerevisiae strain having

part of Ehrlich pathway re-located to mitochondria (Brat et al. 2012).

In here designed expression cassettes, aminotransferase-encoding genes (ARO8 or BAT2)

were located in the first transcription unit. Depending on the assembly variant, either YALI0E20977g

or YALI0F19910g gene, encoding ARO8 or BAT2 aminotransferase homologs from S. cerevisiae,

respectively, was cloned in this position. ARO8 aminotransferase is mainly associated with

deamination of L-Phe, however, it has been experimentally shown that it possess relatively broad

substrate specificity, as it can act on L-Phe as the amino group donor with α-ketoisocaproate and α-

keto-γ(methylthio)butyrate, providing the connection node between L-Phe and L-Leu and L-Met

metabolic pathways (Urrestarazu et al. 1998; Iraqui et al. 1999). On the other hand, BAT2 is known to

be involved in both the first step of BCAA decomposition, as well as the last step of their synthesis (

BAT1 isozyme in mitochondria), while no earlier reports on its activity with aroma AA have been

published to date (according to www.uniprot.org; UniProtKB - P47176). In the current study we have

observed that while overexpression of the Ehrlich pathway in GGA0077 variant had significant

modulatory effect on several higher alcohols formation upon cultivation in L-Phe but also BCAA-

supplemented media (higher production of a respective higher alcohol and increase / decrease in the

fusel acid formation), the opposite setup (GGA0097 and L-Phe supplemented medium) had the

adverse effect on targeted higher alcohol production. These results well correspond with the

abovementioned findings on substrate specificity of the two aminotransferases – ARO8 and BAT2 in

S. cerevisiae. It seems unlikely that YALI0E20977p (ARO8) could directly act on BCAA, however, our

data suggest it is highly probable that it can interact with intermediates of BCAA catabolism, as it was

shown for S. cerevisiae ARO8. Interestingly, our current results indicate that overrepresentation of

BAT2 activity decreases production of aroma AA-derived metabolites (2-PE and 2-PEAc), while its

reduction (by generation of bat2Δ genotype) – leads to their higher formation. This observation

uniformly applied to all the analyzed experimental variants (different GGAs, BAT2+ or Δbat2

genotype, different AA fed to the culture media). Altogether it can be concluded that aroma AA and

BCAA pathways initiated by ARO8 and BAT2 aminotransferases, respectively, compete / act

synergistically in terms of sequestering amino groups from AA. This phenomenon can be seen as

competition, since it decreases titers of targeted higher alcohol and its ester, but considering the cell

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

3132

physiology, it dissipates the carbon skeletons and amino groups into different pathways, making the

process more efficient. Such dissipation of amino groups into different products has been earlier

observed in many studies on higher alcohols production (e.g. Espinosa Vidal et al. 2015; Celińska et

al. 2018). Moreover, it has been shown in S. cerevisiae that generation of Aro8Δ genotype underlies

the transcriptional upregulation of ARO10, which leads to increased de novo production of 2-PE

during growth on glucose and causes substantial changes in the absolute and relative intracellular

concentrations of AAs (Romagnoli et al. 2015). Our data suggest existence of similar phenomenon

operating between BAT2 aminotransferase and ARO10 decarboxylase (as seen in bat2Δ), however,

further research is need to support such conclusion.

As demonstrated previously, overexpression or deletion of BAT2 gene in S. cerevisiae

triggered respectively enhanced or decreased formation of derivative higher alcohols and acids from

their corresponding precursory BCAA (Lilly et al. 2006). In the present study, overexpression of BAT2

under control of strong pTEF promoter enhanced formation of 3-MEB from L-Leu in terms of specific

titer, as well as 3-MEBacid from L-Ile by both JMY2101- and JMY3150-derivatives. Interestingly, for

unbiased induction of 2-MEB formation from L-Ile, an additional BAT2 copy under native regulation

was necessary, since in JMY3150-derivatives bearing GGA0097 cassette the metabolite production

remained at the native level, while in JMY2101-derivatives a significant increase in 2-MEB formation

from L-Ile was observed. Importantly, the concentrations of the keto-acid (isovalerate) and the

downstream higher alcohol (2-MEB) were highly correlated in the GGA0097-bearing strains, cultured

under L-Ile provision. L-Ile promoted production of 2-PE solely in JMY3150-derived strains, which was

not observed for the reference and JMY2101-derived strains. Such “cross-induction” of the derivative

compounds synthesis by a non-direct precursors has been shown earlier for S. cerevisiae (Lilly et al.

2006; Espinosa Vidal et al. 2015). In a study by (Lilly et al. 2006), it was observed L-Val promotes

formation of 2-MEP and 2-MEPacid (straightforward bioconversion products), L-Leu and L-Ile are

both converted to 3-MEB and 3-MEBacid, and also, that L-Leu indirectly contributes to 2-PE

formation. It was indicated that supplementation of the growth media with individual BCAAs results

in increased levels of the corresponding higher alcohols and acids but also – the other, non-direct

bioconversion products (Lilly et al. 2006). Our current results are in agreement with those presented

by (Lilly et al. 2006), as we have observed that for example L-Val improved production of 2-PE and its

derivative ester in JMY3150_GGA077/GGA097 strains; L-Leu enhances production of 2-MEB (L-Ile

metabolism), especially in the case of JMY2101_GGA077 strains; L-Phe stimulates production of

isobutyrate (derivative of L-Val) in JMY2101-/JMY3150_GGA077 strains; ARO8 overexpression in

presence of native BAT2 (JMY2101) – strongly increases titers of 2-MEB (L-Ile derivative) in a course

of L-Leu bioconversion. As observed earlier in S. cerevisiae (Vuralhan et al. 2005), induction of an

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

3334

Ehrlich pathway for catabolism of a given AA leads to the formation of significant amounts of fusel

alcohols and acids from the other precursor, which implies that conversion of BCAA, aroma AA and

sulfur-containing AAs to the corresponding fusel alcohols and acids, via an Ehrlich pathway, involves

common broad-substrate-specificity enzyme activities (Vuralhan et al. 2003; Vuralhan et al. 2005;

Hazelwood et al. 2008). Such construction of the pathway enables flexible interchanging between the

individual sub-pathways. In our previous study (Celinska et al. 2015), we have observed high over-

representation of 2-oxoisovalerate dehydrogenase E1 (YALI0D08690g and YALI0F05038g; primarily

associated with L-Val), isovaleryl-CoA dehydrogenase (YALI0E12573g; primarily associated with L-Leu)

and 3-oxoacid CoA transferase (YALI0F26587g; primarily associated with L-Leu) as well as 3-

hydroxyisobutyryl-CoA hydrolase (YALI0D06215g; primarily associated with L-Val) upon provision of

L-Phe as the sole nitrogen source. This suggests that apart from tight connection between BCAAs

catabolic sub-pathways, such tight linkage functions also for the other, more structurally distinct

substrates metabolized via the Ehrlich pathway in Y. lipolytica, as well.

Although the crucial role in the Ehrlich bioconversion pathway is attributed to the initializing

aminotransferase (Romagnoli et al. 2015), the important contribution of the following keto acid

decarboxylase and dehydrogenase cannot be neglected (Romagnoli et al. 2012). Indeed, previous

studies with Kluyveromyces marxianus have shown that upon overexpression of ARO10 and the

following alcohol dehydrogenase (ADH2) genes from S. cerevisiae significantly improved de novo

synthesis of 2-PE (Kim et al. 2014b) In the present study we focused on a single gene, encoding

decarboxylase activity - YALI0D06930p (a homolog of ARO10 from S. cerevisiae), which was the only

protein, putatively catalyzing the keto acid decarboxylation reaction, identified in the previous

proteomic study (Celinska et al. 2015). As earlier proved in S. cerevisiae (Vuralhan et al. 2005), ARO10

is responsible for a broad- substrate-specificity decarboxylase activity involved in production of fusel

alcohols and acids, showing superior kinetic properties when compared with the other isozymes –

PDCs (Romagnoli et al. 2012). In S. cerevisiae, several different decarboxylases can operate in the

Ehrlich pathway. For L-Leu catabolism, KID1 decarboxylase acts with keto isocaproate , for L-Val – any

of the three isozymes of pyruvate decarboxylase encoded by PDC1, PDC5, and PDC6 can

decarboxylate α–ketoisovalerate, while the broadest range of isozymes can conduct this reaction

during L-Ile degradation, including PDC1, PDC5, PDC6, YDL080c, or YDR380w (Dickinson et al. 2003;

Hazelwood et al. 2008). Nevertheless, only ARO10-encoding gene’s expression was strongly up-

regulated in the presence of L-Phe, L-Leu and L-Met (the transcript level was at least 15-fold higher

than in cultures grown with ammonium sulfate), which coincided with induction of the Ehrlich

pathway. Such strong up-regulation was not observed for PDC5, PDC6, PDC1 and THI3. Since only

ARO10 was differentially and highly transcribed upon cultivation in different nitrogen sources, it was

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

3536

concluded that it operates as the most potent and flexible decarboxylase in the Ehrlich pathway

(Vuralhan et al. 2005). Moreover, overexpression of ARO10 in S. cerevisiae strain, in which the five

chromosomal decarboxylase genes had been deleted, was sufficient to restore a broad-substrate-

specificity decarboxylase activity. More research is needed to elucidate molecular bases of this step

of Ehrlich pathway in Y. lipolytica. It might be useful to identify specific decarboxylases in order to

more precisely manipulate with the Ehrlich pathway, to ultimately obtain narrowed spectrum of

metabolites. On the other hand, it might be challenging to reach more “focused” Ehrlich pathway, as

it is known that the individual steps are highly redundant, the enzymatic activities have broad-

substrate specificity, and the intermediates are shared between different branches, as for example α-

oxoisovalerate (L-Val metabolism) which is known, shared intermediate of BCAAs turnover.

It was previously shown that 2-MEP, 3-MEB, 2-MEB can all be re-metabolized by S. cerevisiae

cells in a specific culturing conditions, namely stationary phase, after exhaustion of glucose, ethanol

or acetate carbon sources in long-term batch cultures, using minimal medium with glucose as the

initial carbon source and L-Leu as the sole nitrogen source (Vuralhan et al. 2005). As neither growth

nor ester formation were observed concomitantly with depletion of the higher alcohols, the authors

concluded that this re-metabolisation might play a role in the maintenance of the redox balance

through reduction of NAD+, while oxidizing the higher alcohols into their corresponding aldehydes in

the opposite direction to the final reductive step of the Ehrlich pathway, and could potentially be

used as carbon sources after further oxidation to the corresponding acid and assimilation by β-

oxidation. Our cultures were conducted until the major carbon source was exhausted from the media

(72 h), and we have not observed such a phenomenon. Probably, prolonged cultivations could

answer whether such a phenomenon occurs also in Y. lipolytica cells. However, we observed that 2-

MEB (GGA097 L-Ile), 2-MEP (GGA097 L-Val) are converted to its acetate ester which production is

correlated with decrease in the alcohol concentration. Such correlation was not observed for 2-MEB

(GGA077, cultured with L-Ile), 2-MEP (GGA077, cultures with L-Val), and also 2-MEP (both GGAs, L-

Leu) where we determined just presence of the corresponding acetate ester, but not concerted with

decrease in the alcohol concentration.

Our current data also demonstrate that, similarly to S. cerevisiae, in Y. lipolytica the AA

precursors can be partly transformed into the corresponding fusel acid. It was earlier suggested that

the ratio between the final concentrations of synthesized acid or alcohol depends on the redox state

of the cell (Dickinson and Dawes 1992). Interestingly, we observed that overexpression of the Ehrlich

pathway in GGA077 variant and cultivation of the resulting strains in L-Leu, L-Ile and L-Val-containing

media led to significantly decreased synthesis of the fusel acid, when compared to the reference

strain. In the view of former findings in S. cerevisiae it might indicate favorable change in the redox

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

3738

state of the cell followed by the modification or, overexpression of dehydrogenase YALI0F24937p

with high preference for reduction reaction and low reversibility.

Conclusions and Perspectives

In summary, the present study reports effects of amplification of putative Ehrlich pathway in

Y. lipolytica on production of valuable higher alcohols. Here adopted modification strategy

contributed to improved production of targeted higher alcohol in many cases, but also enabled

observation of several interesting phenomena. First of all, we observed that bat2Δ genotype

promotes formation of aromatic higher alcohol and its derivative acetate ester, suggesting either

competition between the aroma AA- and BCAA-initiated pathways in sequestering ammonium

residues, or existence of regulation mechanism between aminotransferase and decarboxylase genes,

as it was observed for S. cerevisiae. Correspondingly, doubling BAT2 gene (JMY2101_GGA0097)

decreases formation of 2-PE from L-Phe. Moreover, we observed an interplay between L-Val and L-

Phe pathways, especially in GGA0077-bearing strains (cross-induction of derivative higher alcohols).

Interestingly, upon overexpression of GGA0077 cassette genes, we observed enhanced formation

higher alcohols deriving from L-Leu and L-Val, and concomitantly decreases fusel acid production.

This study deals with primary engineering of aroma AA and BCAA three-step bioconversion

pathway, ultimately leading to many valuable higher alcohols. Obviously, production of higher

alcohols from pure AA is not economically feasible. Still, as demonstrated by several studies

(discussed above) application of metabolic engineering strategies may successfully redirect necessary

carbon skeletons from central metabolism to be further converted by the Ehrlich pathway activities.

On the other hand, to add to sustainability of higher alcohols production from AA, a protein-rich raw

materials like e.g. whey, which is not completely absorbed by the market, could serve as the cost-

effective feedstock. Such solution would definitely constitute an added value of the higher alcohols

production process, but also would complement the biorefinery concept.

Acknowledgements

This study was co-funded by the European Union via the European Regional Development Fund

within the framework of the Innovative Economy Operational Program 2007–2013 (No. 01.01.02-00-

074/09), and was also financially supported by the Ministry of Science and Higher Education Project

entitled “Exploitation of biotechnological processes in production of safe food commodities and

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

3940

chemical compounds” (No. 508.864.00). We thank to Eng. Martyna Przybylak for her technical

assistance.

References

Abe F, Horikoshi K (2005) Enhanced production of isoamyl alcohol and isoamyl acetate by

ubiquitination-deficient Saccharomyces cerevisiae mutants. Cell Mol Biol Lett 10:383–388.

Alper H, Stephanopoulos G (2009) Engineering for biofuels: exploiting innate microbial capacity or

importing biosynthetic potential? Nat Rev Microbiol 7:715–723. doi: 10.1038/nrmicro2186

Avalos JL, Fink GR, Stephanopoulos G (2013) Compartmentalization of metabolic pathways in yeast

mitochondria improves the production of branched-chain alcohols. Nat Biotechnol 31:335–341.

doi: 10.1038/nbt.2509

Barth G, Gaillardin C (1996) Yarrowia lipolytica. In: Wolf K (ed) Nonconventional Yeasts in

Biotechnology A Handbook. Springer Berlin Heidelberg, pp 313–388

Białas W, Czerniak A, Dobrowolska A, Wojciechowska J, Grajek W (2015) Improvement of microbial

stability of innovative dietary supplement based on the biomass of Yarrowia lipolytica. Prog Res

Technol Agro-Food Ind 70:5–20.

Bolat I, Romagnoli G, Zhu F, Pronk JT, Daran JM (2013) Functional analysis and transcriptional

regulation of two orthologs of ARO10, encoding broad-substrate-specificity 2-oxo-acid

decarboxylases, in the brewing yeast Saccharomyces pastorianus CBS1483. FEMS Yeast Res

13:505–517. doi: 10.1111/1567-1364.12051

Borodina I, Nielsen J (2014) Advances in metabolic engineering of yeast Saccharomyces cerevisiae for

production of chemicals. Biotechnol J 9:609–620. doi: 10.1002/biot.201300445

Branduardi P, Longo V, Berterame NM, Rossi G, Porro D (2013) A novel pathway to produce butanol

and isobutanol in S. cerevisiae. Biotechnol Biofuels 6:68.

Brat D, Weber C, Lorenzen W, Bode HB, Boles E (2012) Cytosolic re-localization and optimization of

valine synthesis and catabolism enables inseased isobutanol production with the yeast

Saccharomyces cerevisiae. Biotechnol Biofuels 5:65. doi: 10.1186/1754-6834-5-65

Celińska E, Bonikowski R, Białas W, Dobrowolska A, Słoma B, Borkowska M, Kubiak M, Korpys P,

Grajek W (2018) Pichia cactophila and Kluyveromyces lactis are Highly Efficient Microbial Cell

Factories of Natural Amino Acid-Derived Aroma Compounds. Molecules. doi:

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

4142

10.3390/molecules23010097

Celińska E, Kubiak P, Białas W, Dziadas M, Grajek W (2013) Yarrowia lipolytica: The novel and

promising 2-phenylethanol producer. J Ind Microbiol Biotechnol 40:389–392. doi:

10.1007/s10295-013-1240-3

Celinska E, Ledesma-Amaro R, Larroude M, Rossignol T, Pauthenier C, Nicaud JM (2017) Golden Gate

Assembly system dedicated to complex pathway manipulation in Yarrowia lipolytica. Microb

Biotechnol 10:450–455. doi: 10.1111/1751-7915.12605

Celinska E, Olkowicz M, Grajek W (2015) L-Phenylalanine catabolism and 2-phenylethanol synthesis

in Yarrowia lipolytica-mapping molecular identities through whole-proteome quantitative mass

spectrometry analysis. FEMS Yeast Res. doi: 10.1093/femsyr/fov041

Dai Z, Nielsen J (2015) Advancing metabolic engineering through systems biology of industrial

microorganisms. Curr Opin Biotechnol 36:8–15. doi: 10.1016/j.copbio.2015.08.006

Darvishi Harzevili F (2014) Biotechnological Applications of the Yeast Yarrowia lipolytica. Springer,

Cham

Dickinson JR, Dawes IW (1992) The catabolism of branched-chain amino acids occurs via 2-oxoacid

dehydrogenase in Saccharomyces cerevisiae. J Gen Microbiol 138:2029–2033. doi:

10.1099/00221287-138-10-2029

Dickinson JR, Salgado LEJ, Hewlins MJE (2003) The catabolism of amino acids to long chain and

complex alcohols in Saccharomyces cerevisiae. J Biol Chem 278:8028–8034. doi:

10.1074/jbc.M211914200

Ehrlich F (1907) Über die Bedingungen der Fuselölbildung und über ihren Zusammenhang mit dem

Eiweißaufbau der Hefe. Berichte der Dtsch Chem Gesellschaft 40:1027–1047. doi:

10.1002/cber.190704001156

Espinosa Vidal E, de Morais MA, François JM, de Billerbeck GM (2015) Biosynthesis of higher alcohol

flavour compounds by the yeast Saccharomyces cerevisiae: Impact of oxygen availability and

responses to glucose pulse in minimal growth medium with leucine as sole nitrogen source.

Yeast 32:47–56. doi: 10.1002/yea.3045

Forti L, Di Mauro S, Cramarossa MR, Filippucci S, Turchetti B, Buzzini P (2015) Non-Conventional

Yeasts Whole Cells as Efficient Biocatalysts for the Production of Flavors and Fragrances.

Molecules 20:10377–10398. doi: 10.3390/molecules200610377

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

4344

Generoso WC, Schadeweg V, Oreb M, Boles E (2015) Metabolic engineering of Saccharomyces

cerevisiae for production of butanol isomers. Curr. Opin. Biotechnol. 33:1–7.

Groenewald M, Boekhout T, Neuvéglise C, Gaillardin C, Van Dijck PWM, Wyss M (2014) Yarrowia

lipolytica: safety assessment of an oleaginous yeast with a great industrial potential. Crit Rev

Microbiol 40:187–206. doi: 10.3109/1040841X.2013.770386

Grondin E, Shum A, Sing C, Caro Y, Raherimandimby M, Lalaniaina A, James S, Nueno-palop C, Marie

J, Petit T (2015) A comparative study on the potential of epiphytic yeasts isolated from tropical

fruits to produce flavoring compounds. Int J Food Microbiol 203:101–108. doi:

10.1016/j.ijfoodmicro.2015.02.032

Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for

Windows 95/98/NT. Nucleic Acids Symp. Ser. 41:95–98.

Hazelwood LA, Daran J-M, van Maris AJA, Pronk JT, Dickinson JR (2008) The Ehrlich pathway for fusel

alcohol production: a century of research on Saccharomyces cerevisiae metabolism. Appl

Environ Microbiol 74:2259–2266. doi: 10.1128/AEM.02625-07

Hua D, Xu P (2011) Recent advances in biotechnological production of 2-phenylethanol. Biotechnol

Adv 29:654–660. doi: 10.1016/j.biotechadv.2011.05.001

Iraqui I, Vissers S, André B, Urrestarazu A (1999) Transcriptional induction by aromatic amino acids in

Saccharomyces cerevisiae. Mol Cell Biol 19:3360–3371.

Jansen MLA, van Gulik WM (2014) Towards large scale fermentative production of succinic acid. Curr

Opin Biotechnol 30:190–197. doi: 10.1016/j.copbio.2014.07.003

Kim B, Cho BR, Hahn JS (2014a) Metabolic engineering of Saccharomyces cerevisiae for the

production of 2-phenylethanol via Ehrlich pathway. Biotechnol Bioeng 111:115–124. doi:

10.1002/bit.24993

Kim TY, Lee SW, Oh MK (2014b) Biosynthesis of 2-phenylethanol from glucose with genetically

engineered Kluyveromyces marxianus. Enzyme Microb Technol 61–62:44–47. doi:

10.1016/j.enzmictec.2014.04.011

Ledesma-Amaro R, Nicaud JM (2016) Yarrowia lipolytica as a biotechnological chassis to produce

usual and unusual fatty acids. Prog Lipid Res 61:40–50. doi: 10.1016/j.plipres.2015.12.001

Lilly M, Bauer FF, Styger G, Lambrechts MG, Pretorius IS (2006) The effect of increased branched-

chain amino acid transaminase activity in yeast on the production of higher alcohols and on the

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

752

753

754

755

756

4546

flavour profiles of wine and distillates. FEMS Yeast Res 6:726–743. doi: 10.1111/j.1567-

1364.2006.00057.x

Lilly M, Lambrechts MG, Pretorius IS (2000) Effect of increased yeast alcohol acetyltransferase

activity on flavor profiles of wine and distillates. Appl Environ Microbiol 66:744–753. doi:

10.1128/AEM.66.2.744-753.2000

Madzak C (2015) Yarrowia lipolytica: recent achievements in heterologous protein expression and

pathway engineering. Appl Microbiol Biotechnol 99:4559–4577. doi: 10.1007/s00253-015-6624-

z

Nakamura CE, Whited GM (2003) Metabolic engineering for the microbial production of 1,3-

propanediol. Curr. Opin. Biotechnol. 14:454–459.

Nielsen J, Keasling JD (2016) Engineering Cellular Metabolism. Cell 164:1185–1197. doi:

10.1016/j.cell.2016.02.004

Nozzi NE, Desai SH, Case AE, Atsumi S (2014) Metabolic engineering for higher alcohol production.

Metab Eng 25:174–182. doi: https://doi.org/10.1016/j.ymben.2014.07.007

Romagnoli G, Knijnenburg TA, Liti G, Louis EJ, Pronk JT, Daran JM (2015) Deletion of the

Saccharomyces cerevisiae ARO8 gene, encoding an aromatic amino acid transaminase,

enhances phenylethanol production from glucose. Yeast 32:29–45. doi: 10.1002/yea.3015

Romagnoli G, Luttik MAH, Kötter P, Pronk JT, Daran JM (2012) Substrate specificity of thiamine

pyrophosphate-dependent 2-oxo-acid decarboxylases in Saccharomyces cerevisiae. Appl

Environ Microbiol 78:7538–7548. doi: 10.1128/AEM.01675-12

Sambrook J, Russell D (2001) Molecular Cloning: A Laboratory Manual, 3rd Editio. Cold Spring Harbor

Laboratory Press

Sauer M, Marx H, Mattanovich D (2008) Microbial Production of 1,3-Propanediol. Recent Pat

Biotechnol 2:191–197.

Schrader J, Etschmann MM, Sell D, Hilmer JM, Rabenhorst J (2004) Applied biocatalysis for the

synthesis of natural flavour compounds--current industrial processes and future prospects.

Biotechnol Lett 26:463–472. doi: 10.1023/B:BILE.0000019576.80594.0e

Sibirny A, Scheffers L (2002) Thematic section: Biochemistry, Genetics, Biotechnology and Ecology of

Non-conventional Yeasts. FEMS Yeast Res. doi:

http://onlinelibrary.wiley.com/doi/10.1111/j.1567-1364.2002.tb00097.x/pdf

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

4748

Stribny J, Romagnoli G, Pérez-Torrado R, Daran J-M, Querol A (2016) Characterisation of the broad

substrate specificity 2-keto acid decarboxylase Aro10p of Saccharomyces kudriavzevii and its

implication in aroma development. Microb Cell Fact 15:51. doi: 10.1186/s12934-016-0449-z

Styger G, Jacobson D, Prior BA, Bauer FF (2013) Genetic analysis of the metabolic pathways

responsible for aroma metabolite production by Saccharomyces cerevisiae. Appl Microbiol

Biotechnol 97:4429–4442. doi: 10.1007/s00253-012-4522-1

Urrestarazu A, Vissers S, Iraqui I, Grenson M (1998) Phenylalanine- and tyrosine-auxotrophic mutants

of Saccharomyces cerevisiae impaired in transamination. Mol Gen Genet 257:230–237.

Vuralhan Z, Luttik MA, Tai SL, Boer VM, Morais MA, Schipper D, Almering MJ, Kotter P, Dickinson JR,

Daran JM, Pronk JT (2005) Physiological characterization of the ARO10-dependent, broad-

substrate-specificity 2-oxo acid decarboxylase activity of Saccharomyces cerevisiae. Appl

Environ Microbiol 71:3276–3284. doi: 10.1128/AEM.71.6.3276-3284.2005

Vuralhan Z, Morais MA, Tai SL, Piper MDW, Pronk JT (2003) Identification and characterization of

phenylpyruvate decarboxylase genes in Saccharomyces cerevisiae. Appl Environ Microbiol

69:4534–4541. doi: 10.1128/AEM.69.8.4534-4541.2003

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

4950

Figure Captions:

Figure 1. The Ehrlich pathway. Metabolic pathway of branched chain AAs (L-Val, L-Leu, L-Ileu),

aromatic AAs (L-Phe, L-Tyr, L-Trp) and L-Met catabolism initialized by transamination by

aminotransferase to alpha-keto acid, followed by decarboxylation by decarboxylase to fusel

aldehyde, and either oxidation or reduction by dehydrogenase to fusel alcohol or fusel acid,

respectively. CAS numbers are provided for those compounds that were analyzed in this study.

Figure 2. Y. lipolytica recombinant strains construction system. A.B. Golden Gate Assembly platform

dedicated to Y. lipolytica with three transcription units, bearing homologous genes encoding

activities involved in higher alcohols synthesis pathway: ARO8 [YALI0E20977p] /BAT2 [YALI0F19910p]

– aminotransferases; ARO10 [YALI0D06930p] – decarboxylase; Adh [YALI0F24937p]- dehydrogenase.

A. Complete set of donor vectors constructed on a backbone of pCR Blunt II-TOPO vector (Thermo‐

Scientific) bearing kanamycin resistance gene (KanR). Each building block is flanked with respective

BsaI recognition sites and predesigned 4 nt overhangs with signatures A–M. The overhangs allow to

introduce all the elements of the expression cassette in the correct position and orientation.

Abbreviations: InsUP/InsDOWN – sequences targeting insertion in the 5'/3' region (zeta); M –

selection marker (URA3); P1/2/3– constitutive promoters (pTEF) governing expression of the gene of

interest in transcription unit 1, 2 and 3, respectively; G1/2/3 – gene of interest in transcription unit 1,

2 and 3 respectively; T1/2/3– constitutive terminators (tLip2) located in the transcription unit 1, 2

and 3 respectively. B. Complete GGVAs (Golden Gate Assembles in destination vector pSB1A3-RFP

from iGEM collection) in 3-transcription unit (TU)-bearing format. Each TU is composed of a

promoter – pTEF, a gene of interest: BAT2/ARO8/ARO10/Adh, and a terminator – tLip2); Complete

GGVA bears a selection marker (URA3) and is flanked with integration targeting sequences (InsUP

and InsDOWN); signatures A to M and the following nucleotide sequences mark position and

sequence of the 4 nt overhangs, mentioned in part A; NotI: NotI recognition sites for release of the

expression cassettes prior to Y. lipolytica transformation. C. Electrophoretic separation of amplicones

obtained in colony PCR with biomass Y. lipolytica recombinant or parental strains, corresponding to

regions: ARO8-tLip2T1 (lanes: 1, 5, 9, 13, 17), ARO10-tLip2T2 (lanes: 2, 6, 10, 14, 18), Adh-tLip2T3

(lanes: 3, 7, 11, 15, 19) and ura3 (lanes: 4, 8, 12, 16, 20); lanes 1-4 (clone 1), 5-8 (clone 2), 9-12 (clone

3), 13-16 (clone 4), 17-20 (negative control – not transformed strain). Positive clones: 1 and 3. M:

Gene Ruler Express.

Figure 3. Metabolite profiles generated by GGY0077-bering Y. lipolytica strains in bioconversion

cultures supplemented with L-Phe (A.B), L-Leu (C.D.), L-Ile (E.F), and L-Val (G.H). The graphs represent

either kinetic (A.C.E.G) or end-point (B.D.F.H.) data. Legends indicate the type of detected metabolite

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

847

5152

and type of strain. X axis: time of culturing in h (A.C.E.G) or type of strain (B.D.F.H.). Y axis:

concentration of a given metabolite in g/L. Mind main and auxiliary Y axes. Error bars indicate mean

values from biological duplicate, each analyzed in triplicate.

Figure 4. Metabolite profiles generated by GGY0097-bering Y. lipolytica strains in bioconversion

cultures supplemented with L-Phe (A.), L-Leu (B.C.), L-Ile (D.E.), and L-Val (F.G.). The graphs represent

either kinetic (A.B.D.F.G) or end-point (C.E.) data. Legends indicate the type of detected metabolite

and type of strain. X axis: time of culturing in h (A.B.D.F.G) or type of strain (C.E.). Y axis:

concentration of a given metabolite in g/L. Mind main and auxiliary Y axes. Error bars indicate mean

values from biological duplicate, each analyzed in triplicate.

Supplementary Material

Figure S1. Biomass growth and specific titers of metabolites generated by GGY0077-bering Y.

lipolytica strains in bioconversion cultures supplemented with L-Phe (A.B), L-Leu (C.D.), L-Ile (E.F), and

L-Val (G.H). The graphs represent either kinetic (A.C.E.G) or single-point (B.D.F.H.) data. Legends

indicate the culturing time or type of detected metabolite. X axis: type of strain. Y axis: absorbance at

600 nm in OD units (A.C.E.G) or specific titer in g/OD600 (B.D.F.H.). Mind main and auxiliary Y axes.

Error bars indicate mean values from biological duplicate, each analyzed in triplicate.

Figure S2. Biomass growth and specific titers of metabolites generated by GGY0097-bering Y.

lipolytica strains in bioconversion cultures supplemented with L-Phe (A.B), L-Leu (C.D.), L-Ile (E.F), and

L-Val (G.H). The graphs represent either kinetic (A.C.E.G) or single-point (B.D.F.H.) data. Legends

indicate the culturing time or type of detected metabolite. X axis: type of strain. Y axis: absorbance at

600 nm in OD units (A.C.E.G) or specific titer in g/OD600 (B.D.F.H.). Mind main and auxiliary Y axes.

Error bars indicate mean values from biological duplicate, each analyzed in triplicate.

Table S1. List of E. coli and Y. lipolytica strains used in this study

Table S2. List of plasmids and primers used in this study

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

5354