The response of Pseudomonas putida to a complex lignolysate

There is strong interest in the valorization of lignin derived from plant biomass to produce valuable products; however, the structural complexity of this biopolymer has been a major bottleneck to conversion. Chemical pretreatment liberates soluble fractions of lignin that may be upgraded by biological conversion. Here, ionic liquid pretreatment was employed to obtain soluble aromatic-rich fractions from sorghum, which were converted by Pseudomonas putida KT2440, a promising host for bioconversion of aromatics derived from lignin. Growth studies and mutational analysis demonstrated that P. putida growth on these soluble lignin-derived fractions, referred to as lignolysate, was dependent on aromatic monomers derived from lignin (p-coumarate and ferulate), but other, unknown factors in the lignolysate contributed to growth. Proteomic and metabolomic analyses provided evidence that these unknown factors were amino acids and residual ionic liquid. Proteomic measurements indicated a coordinated response in which these substrates were catabolized simultaneously. A cholinium catabolic pathway was identified and deletion of five genes in the pathway abrogated the ability of P. putida to grow on cholinium as a sole carbon source. This work demonstrates that lignolysates obtained through biomass pretreatment contain multiple substrates and conversion strategies for lignin-derived should take this complexity into account. Importance Lignin is one of the most abundant biopolymers on Earth and is generated as a co-product in the processing of lignocellulosic biomass. Valorization of these residual lignin streams is a promising method to enhance the economic viability of modern lignocellulosic biorefineries. In this study we developed a process to couple chemical depolymerization of lignin and biological conversion using Pseudomonas putida KT2440. Water-soluble and bioavailable lignolysate was obtained from sorghum and further characterized as a growth substrate for P. putida. Proteomic and metabolomic analyses demonstrated that P. putida metabolized other components of the lignolysate beyond monoaromatic compounds, which illuminates how microbes can process complex lignolysates obtained from plants. Understanding the underlying microbial responses in lignolysates will enable the design of rational strategies for lignin valorization.


Abstract 25
There is strong interest in the valorization of lignin derived from plant biomass to produce 26 valuable products; however, the structural complexity of this biopolymer has been a major 27 bottleneck to conversion. Chemical pretreatment liberates soluble fractions of lignin that may 28 be upgraded by biological conversion. Here, ionic liquid pretreatment was employed to 29 obtain soluble aromatic-rich fractions from sorghum, which were converted by Pseudomonas 30 putida KT2440, a promising host for bioconversion of aromatics derived from lignin. Growth 31 studies and mutational analysis demonstrated that P. putida growth on these soluble lignin-32 derived fractions, referred to as lignolysate, was dependent on aromatic monomers derived 33 from lignin (p-coumarate and ferulate), but other, unknown factors in the lignolysate 34 contributed to growth. Proteomic and metabolomic analyses provided evidence that these 35 unknown factors were amino acids and residual ionic liquid. Proteomic measurements 36 indicated a coordinated response in which these substrates were catabolized simultaneously. 37 A cholinium catabolic pathway was identified and deletion of five genes in the pathway 38 abrogated the ability of P. putida to grow on cholinium as a sole carbon source. This work 39 demonstrates that lignolysates obtained through biomass pretreatment contain multiple 40 substrates and conversion strategies for lignin-derived should take this complexity into 41 account. 42 43 44 Introduction 59 Lignocellulosic biomass, which is primarily composed of cellulose, hemicellulose and 60 lignin, represents a primary renewable feedstock for biofuel and biochemical production (1). 61 For decades, conversion strategies have focused on the polysaccharides cellulose and 62 hemicellulose, whereas lignin, which makes up around 15-30 wt % of biomass, is usually 63 combusted to provide heat and electricity to support pulping operations or recovered as kraft 64 lignin, vanillin or lignosulfonates (2, 3). Despite the fact that lignin is the only large-volume 65 renewable aromatic feedstock on the Earth (4, 5), the bioprocessing of lignin into bioproducts 66 is a major bottleneck because of its intrinsic heterogeneity and recalcitrance to 67 depolymerization (5). However, with the emergence of lignocellulosic biorefineries, lignin 68 conversion is a crucial component of integrated biorefineries with respect to economics and 69 sustainability (5, 6). 70 A promising lignin valorizing strategy couples chemical lignin depolymerization with 71 microbial catabolism of aromatic monomers by hosts that have been engineered to upgrade 72 the depolymerized lignin (7)(8)(9), since only these low molecular weight products of lignin can 73 be assimilated as carbon sources by microbes (9-12). To achieve this objective, chemical 74 lignin depolymerization methods have been developed (12-15). Among these methods, a 75 base-catalyzed depolymerization (BCD) process has been demonstrated to produce high 76 yields of aromatic monomers, primarily p-coumarate and ferulate, that acylate lignin (12, 16, 77 17). Previously, the BCD process was employed on solid lignin-rich residue derived from 78 corn stover via deacetylation, mechanical refining, and enzymatic hydrolysis to release 79 lignin-derived aromatic monomers that can be further upgraded into value-added molecules 80 (12). Ionic liquids (ILs) have been proposed as pretreatment chemicals for lignocellulosic 81 biomass fractionation due to their highly tunable physicochemical and compatibility with 82 biology (18,19). Significant advances have also been made in improvement of ILs recovery 83 and recycling to overcome high cost of ILs (18)(19)(20)(21). Combining IL pretreatment and the BCD 84 process may increase the bioavailable depolymerized lignin that can be converted by 85

microbes. 86
Pseudomonas putida KT2440 is as a promising host for engineering lignin 87 bioconversion, due to its ability to catabolize numerous aromatic compounds and its 88 amenability to genetic manipulation. Bioconversion studies with aromatics have mostly 89 focused on purified model compounds (7,22), whereas the molecular mechanism of 90 bioconversion using aromatics directly derived from plant lignin is less well-understood. 91 Therefore, a fundamental understanding of the biological conversion of complex mixtures 92 derived from lignin, referred to as lignolysates, is critical for rational strain engineering and 93 upgrading of lignin-derived substrates to bioproducts. While genomic and transcriptomic 94 analyses have been carried out to characterize the lignin-degrading mechanisms (23-25), 95 proteomic analysis is able to offer insights into the protein abundance which is important for 96 understanding microbial functions. Therefore, proteomic analysis is crucial for understanding 97 the microbial conversion of lignin-derived substrates in greater depth. 98 In this work, we obtained size-defined soluble aromatic-rich streams from sorghum 99 using biologically-derived ILs and assessed their biocompatibility with P. putida. Growth 100 studies, mutational analyses and proteomic measurements demonstrated that the lignolysate 101 derived from sorghum was a complex mixture composed of monoaromatics, amino acids and 102 residual ILs that supported growth in P. putida. 103 104 105

Generation of size-defined aromatic fractions from sorghum lignin 107
Multiple solubilized fractions of sorghum-derived lignin were generated to identify a 108 fraction that would be bioavailable for conversion by P. putida KT2440 (Figure 1). In the 109 first approach, the soluble fraction from pretreatment with a cholinium IL, cholinium 110 aspartate, was treated with acid and then equilibrated for 2 days at 5 °C. This acid treatment 111 produced a precipitated solid fraction enriched in lignin (54.2%) and hemicellulose (19.2%) 112 (Table 1). The hemicellulose was removed by enzymatic hydrolysis, providing a fraction, 113 referred to as acid-precipitated lignin (AP lignin). Compositional analysis revealed that AP 114 lignin, which was soluble in aqueous solution at neutral pH, was ~70% lignin, and <5% 115 residual carbohydrate (Table 1). Gel-Permeation Chromatography (GPC) analysis 116 demonstrated that AP lignin contained aromatic molecules with a molecular weight 117 distribution between 1000 and 10000 Da (Figure 2A). In a second procedure, the solid 118 remaining after cholinium ionic liquid pretreatment was enzymatically hydrolyzed and the 119 residual material was treated with NaOH for base catalyzed depolymerization (BCD). The 120 aqueous fraction, referred to as base-catalyzed depolymerized lignin (BCD liquor) had 20.5% 121 lignin and 1.5% carbohydrate (Table 1). The low recovery of lignin in the BCD liquor 122 indicated that the majority of the lignin was depolymerized. GPC analysis confirmed that the 123 BCD fraction had peaks corresponding to monoaromatics (0.1-0.3 kDa) as well as a broad 124 distribution of higher molecular weight species (Figure 2 The aromatic/unsaturated (δ H /δ C 6.0 -8.0/90 -160) and aliphatic (δ H /δ C 2.5 -6.0/50 -135 90) regions of HSQC NMR spectra of the AP lignin and BCD liquor were analyzed to 136 provide chemical information related to their composition characterized by interunit linkages 137 ( Figure 3). Signals from the aromatic ring correlations from syringyl (S) lignin (derived from 138 sinapyl alcohol), guaiacyl (G) units (derived from coniferyl alcohol), and p-hydroxyphenyl 139 (H) lignin (derived from p-coumaryl alcohol) were observed in the spectra of both the BCD 140 and AP fractions. The aromatic region of the HSQC spectrum indicated that BCD fraction 141 consisted of S (8.2%), G (88.3%), and H (3.5%) units, which represents a S/G ratio of 0.1 142 ( Figure 3A). The AP fraction consisted of S (19%), G (73.7%), and H (7.3%) units, 143 representing a S/G ratio of 0.26 ( Figure 3B). Prominent signals corresponding to pCA were 144 also observed in both the BCD and AP lignin (26, 27); in addition, signals for FA were 145 observed in the spectrum of the AP lignin. Since the LC-MS measurements only detected free 146 pCA and FA in the BCD liquor, the HSQC spectra corresponding to p-coumarate and ferulate 147 found in AP lignin are likely to be pCA-and FA-end groups attached to oligomers. The intermediates in pCA and FA catabolic pathways, respectively, followed by complete 169 consumption ( Figure S1). Furthermore, the MW distributions after the bacterial treatment 170 were also performed to examine catabolism of LMW lignin derived compounds as well as 171 depolymerization of HMW lignin. The MW profile in the uninoculated control exhibited a 172 major peak in the LMW regions (Figure 4(B)). After microbial cultivation of the BCD liquor, 173 the LMW species were absent in the GPC trace, consistent with the complete utilization of 174 the aromatic monomers; however, the peaks in the HMW regions were unchanged after 175 microbial culturing. These findings indicated that P. putida was able to degrade lignin-176 derived monoaromatic monomers but not higher molecular weight aromatics derived from 177 lignin depolymerization. In comparison to growth on BCD liquor, P. putida growth on pCA 178 under the same conditions and at the same concentration as present in the BCD liquor was 179 lower (Figure 4(A)). This difference was unexpected, as it was assumed that pCA was 180 responsible for almost all the P. putida growth observed in the BCD liquor. To determine if 181 additional substrates were present in the BCD liquor, the ability of P. putida to grow on pCA 182 and FA was abrogated by disrupting the hydroxycinnamoyl-CoA hydratase-lyase (ech, 183 PP_3358) gene, whose gene product dehydrates and liberates acetyl-CoA from 184 hydroxycinnamic acids. As expected, the ΔPP3358 mutant was not able to grow with pCA-as 185 the sole carbon source (Figure 4(C)), and a major peak corresponding to LMW aromatics 186 were still observed in the BCD liquor after microbial conversion using the P. putida mutant 187 (Figure 4(D)). In parallel with the GPC profile, HPLC demonstrated that pCA and FA in the 188 BCD liquor were not consumed during mutant cultivation ( Figure S2). Nonetheless, the 189 ΔPP3358 mutant strain was still capable of growing in the BCD liquor to an optical density 190 approximately half of what was observed with the wild type P. putida, confirming that the 191 BCD liquor contained additional substrates for P. putida. 192 193 A previous study demonstrated that the plant-derived amino acids in biomass hydrolysates 194 enhanced P. putida growth and production of fatty acid-derived molecules (28). Therefore, 195 amino acids liberated by the BCD process may serve as additional substrates for P. putida. 196 The concentration of amino acids in the BCD liquor was measured at 0.37 g/L by  Alanine (0.19 g/L) and serine (0.07 g/L) were the most abundant amino acids, and other low-198 abundance amino acids were present at ~ 0.1 g/L in total ( respectively. All of these increased proteins were clustered into 18 functional groups 213 indicating particular metabolic processes responsible for the utilization of the different 214 substrate sources ( Figure S3). According to the COG analysis, significantly increased 215 proteins in P. putida in response to BCD liquor were grouped into the categories "energy 216 production and conversion", "lipid transport and metabolism", "amino acid transport and 217 metabolism" and "secondary metabolites biosynthesis, transport and catabolism" in cells 218 grown in the BCD liquor. 219 220

Differentially increased proteins in P. putida KT2440 grown in BCD liquor 221
Growth of P. putida in a M9 minimal medium containing BCD liquor led to the 222 significant induction of proteins associated with aromatic catabolic and β-ketoadipate 223 pathways ( Figure 5 and Table S1). More specifically, the proteins involved in the conversion 224 of p-coumarate and ferulate to hydroxybenzoate: 4-coumarate: CoA ligase (Fcs), 225 hydroxycinnamoyl-CoA hydratase-lyase (Ech) and vanillin dehydrogenase (Vdh) were 226 significantly increased (5.12-to 7.51-log 2 FC) when cells were grown in BCD liquor 227 compared to the control culture grown from sugar only. 4-hydroxybenzoate hydroxylase 228 (PobA), which transforms 4-hydroxybenzoate into protocatechuate, was also significantly 229 increased (6.75-log 2 FC). The subsequent enzymes encoded by the pca genes further catalyze 230 the protocatechuate ortho-cleavage pathway (29). The pca genes are arranged in four 231 different clusters, pcaHG, pcaBDC, pcaIJ, and pcaF. Herein, enzymes (PcaHG, PcaB and 232 PcaD) belonging to the protocatechuate branch of the β-ketoadipate pathway were 233 significantly increased (1.41-to 5.96-log 2 FC). 4-carboxymuconolactone decarboxylase 234 (PcaC) required for the transformation of 4-carboxymucono-lactone to beta-ketoadipate-enol-235 lactone was not detected in this study. Lastly, proteins (PcaIJ and PcaF) involved in two 236 further steps of converting β-ketoadipate into tricarboxylic acids (TCA) cycle intermediates 237 were also significantly increased (3.04 and 2.88-log 2 FC, respectively). Similar changes in the 238 level of these enzymes involved in the aromatic catabolic pathway was also observed in 239 pCA-only (3.3-to 7.71-log 2 FC) and pCA/AA (1.89-to 7.84-log 2 FC) controls ( Figure  240 S4(B),(C) and Table S2). Degradation of p-coumarate through the protocatechuate-branch of 241 the β-ketoadipate pathway yields acetyl-CoA and succinyl-CoA, which enter the TCA cycle 242 (29). TCA cycle enzymes citrate synthase (GltA, 0.76-log 2 FC), aconitate hydratase (AcnA-2, 243 2.07-log 2 FC), isocitrate dehydrogenase (Icd, 0.96-log 2 FC), succinate dehydrogenase 244 (SdhAB, 0.93-to 1.16-log 2 FC) and the first enzyme of glyoxylate shunt (isocitrate lyase 245 (AceA), 4.51-log 2 FC) were at significantly higher abundance when cells were grown in BCD 246 liquor. Similar results were observed in succinate dehydrogenase and glyoxylate shunt 247 enzymes in pCA-only (1.15-to 4.4-log 2 FC) and pCA/AA (0.6-to 3.99-log 2 FC) controls 248 ( Figure S4(B),(C) and Table S2).  were significantly increased probably due to the utilization of plant-derived amino acids. The 264 enzymes were also increased in the control groups with amino acids present (2.06-to 2.69-265 log 2 FC in AA-only; 0.56-to 2.0-log 2 FC in pCA/AA). 266 267 Another possible substrate in the BCD liquor could be residual IL from the initial 268 pretreatment. We considered this an unlikely possibility, since the solids remaining after both 269 IL pretreatment and enzymatic hydrolysis was extensively washed. However, other 270 Pseudomonas species have been shown to catabolize cholinium (30-32), and growth on P. 271 putida on 0.2% cholinium aspartate as the sole carbon source was demonstrated ( Figure S5). 272 To determine if the catabolism of cholinium occurred during growth on the BCD liquor, a 273 cholinium catabolic pathway was identified in the P. putida genome by reference to a 274 characterized pathway in Pseudomonas aeruginosa (30,31). Cholinium is oxidized to glycine 275 betaine by genes encoding choline dehydrogenase (BetA) and betaine aldehyde 276 dehydrogenase (BetB), followed by demethylation of glycine betaine to dimethylglycine by 277 serine hydroxymethyltransferase (GlyA-1). The demethylation of dimethylglycine is carried 278 out by DgcA and DgcB. Sarcosine demethylation is conducted by a heterotetrameric enzyme, 279 SoxBDAG. Proteomic analysis revealed that the some of the proteins of the cholinium 280 catabolic pathway was significantly increased in BCD liquor (Table S1) The biocompatible lignin depolymerization method described here enables a route to 291 biologically upgrading lignin into value-added bioproducts, a process that could potentially 292 be run in parallel with sugar conversion or as a separate stream. We employed choline-based 293 IL pretreatment to obtain solubilized and insoluble lignin fractions that were then further 294 processed into size-defined fractions of lignin, which were examined for microbial utilization. 295 In one approach, a low molecular weight fraction was obtained by BCD reaction of the solid 296 fraction after saccharification, while a relatively high molecular weight fraction was 297 produced by another approach using acid precipitation of the IL-solubilized lignin. HSQC 298 NMR of the AP lignin was consistent with condensation of lignin, which may arise during IL 299 pretreatment or subsequent acid precipitation (33), whereas BCD provided depolymerized 300 lignin streams without condensation. 301 The substrates p-coumarate and ferulate were the main lignin-derived aromatic 302 monomers present in BCD liquor, but p-coumarate was present at much higher levels than 303 ferulate. The higher abundances of p-coumarate compared to ferulate support previous 304 observations that p-coumarate is predominantly attached to the lignin while ferulate is mostly 305 attached to the polysaccharides. Previous studies have indicated that the bulk of p-coumarate 306 is esterified to the lignin side chains and acylates the γ-OH of the lignin side chain in grasses 307 (34-37). On the other hand, ferulate has been shown to be involved in lignin-polysaccharides 308 linkages (38). While previous studies focused mainly only on aromatic compounds in BCD 309 liquor (12,13,17), this study further revealed that BCD liquor contains fatty acid and 310 residual ILs (choline) as well as aromatics and amino acids, which revealed not only specific, 311 differentially increased proteins of KT2440 using the mass spectrometry-based proteomic 312 approach, but also aromatic-independent growth in by a P. putida mutant strain that was 313 unable to metabolize p-coumarate and ferulate. The amino acids in BCD liquor were 314 probably liberated during the BCD reaction by hydrolysis of plant proteins in the solid 315 fraction after one-pot pretreatment. Some covalent linkages have also been demonstrated 316 between lignin, polysaccharides and structural proteins of grass cell walls (39). The plant-317 derived amino acids (serine, valine, aspartate, phenylalanine and tryptophan) were also 318 shown in hydrolysates obtained from Arabidopsis, switchgrass and sorghum (28). Likewise, 319 in this study, free fatty acids in BCD liquor were probably produced by BCD reaction since a 320 common feature of the lipid membrane components in plant cells is esters of moderate to 321 long chain fatty acids, and acid or base-catalyzed hydrolysis yields the component fatty acid 322 from the lipid components (40, 41). 323 When P. putida KT2440 were grown in the BCD liquor, complete utilization of 324 aromatic monomers was observed during aromatic catabolism corresponding with the 325 disappearance of LMW lignin peaks as shown by GPC, which is in agreement with a 326 previous study (12). We observed the complete list of proteins responsible for aromatic 327 catabolism via the protocatechuate ortho-cleavage pathway and β-ketoadipate pathway. It 328 should be noted that the proteins participating in the catabolic pathways for aromatic 329 compounds to central metabolism increased significantly in the proteome, indicating that 330 lignin-derived aromatics in the BCD liquor are the main carbon and energy sources, which is 331 consistent with the observations in the presence of p-coumarate (pCA-only and pCA/AA 332 controls). However, depolymerization of HMW lignin by P. putida KT2440 was not 333 observed in our system. These results are in contrast to those of a previous study (8) For the BCD reaction, the lyophilized substrate was added as 10% (w/v) solids to a 5% 389 NaOH solution, loaded into a 350 mL stainless steel Miniclave drive 3 pressure reactor 390 (Buchiglas, Switzerland), which was equipped with an impeller and temperature controller. 391 The reaction proceeded through a 35 min ramp from 25 to 120 °C, a 30 min reaction at 392 120 °C, and a 25 min ramp from 120 to 40 °C, while keeping the stirrer speed constant at 393 1500 rpm as described previously (12)

LC-MS analysis of phenolic compounds 467
All metabolites were quantified using HPLC-electrospray ionization (ESI)-time-of-468 flight (TOF) mass spectrometry (MS). An aliquot of the culture medium was cleared by 469 centrifugation (21,000 x g, 5 min, 4°C) and filtered using Amicon Ultra centrifugal filters 470 (3,000 Da MW cut off regenerated cellulose membrane; Millipore, Billerica, MA, USA) prior 471 to analysis. The separation of metabolites was conducted on the fermentation-monitoring 472 HPX-87H column with 8% cross-linkage (150-mm length, 7.8-mm inside diameter, and 9-473 μm particle size; Bio-Rad, Richmond, CA, USA) using an Agilent Technologies 1100 Series 474 HPLC system. A sample injection volume of 10 μl was used throughout. The sample tray and 475 column compartment were set to 4 and 50°C, respectively. Metabolites were eluted 476 isocratically with a mobile-phase composition of 0.1% formic acid in water at a flow rate of 477 0.5 ml/min. The HPLC system was coupled to an Agilent Technologies 6210 series time-of-478 flight mass spectrometer (LC-TOF MS) via a MassHunter workstation (Agilent Technologies, 479 CA, USA). Drying and nebulizing gases were set to 13 liters/min and 30 lb/ in 2 , respectively, 480 and a drying-gas temperature of 330°C was used throughout. ESI was conducted in the 481 negative ion mode and a capillary voltage of -3,500 V was utilized. All other MS conditions 482 were described previously (53). 483

LC-MS analysis of amino acids 485
For the measurement of plant-derived amino acids in the BCD fraction, liquid 486 chromatographic separation was conducted using a Kinetex HILIC column (100-mm length, 487 4.6-mm internal diameter, 2.6-μm particle size; Phenomenex, Torrance, CA) using a 1200 488 Series HPLC system (Agilent Technologies, Santa Clara, CA, USA) as described previously 489 (28). The injection volume for each measurement was 2 μL. The sample tray and column 490 compartment were set to 6°C and 40°C, respectively. The mobile phase was composed of 20 491 mM ammonium acetate in water (solvent A) and 10 mM ammonium acetate in 90% 492 acetonitrile and 10% water (solvent B) (HPLC grade, Honeywell Burdick & Jackson, CA, 493 USA). Ammonium acetate was prepared from a stock solution of 100 mM ammonium acetate 494 and 0.7 % formic acid (98-100% chemical purity, from Sigma-Aldrich, St. Louis, MO, USA) 495 in water. Amino acids were separated with the following gradient: 90% to 70%B in 4 min, 496 held at 70%B for 1.5 min, 70% to 40%B in 0.5 min, held at 40%B for 2.5 min, 40% to 90%B 497 in 0.5 min, held at 90%B for 2 min. The flow rate was varied as follows: held at 0.6 mL/min 498 for 6.5 min, linearly increased from 0.6 mL/min to 1 mL/min in 0.5 min, and held at 1 499 mL/min for 4 min. The total run time was 11 min. The mass spectrometry parameters have 500 been previously described (54). 501

GC-MS analysis for fatty acid 503
Fatty acid was quantified using a method as described previously (55). Specifically, 504 0.5 mL of supernatant was acidified with 50 μL of concentrated HCl (12N). The fatty acids 505 were extracted twice with 0.5 mL ethyl acetate. The extracted fatty acids were derivatized to 506 fatty acid methyl esters (FAME) by adding 10 μL concentrated HCl, 90 μL methanol and 100 507 μL of TMS-diazomethane, and incubated at room temperature for 15 min. Gas 508 chromatography-mass spectrometry (GC-MS) analysis of FAME was performed on Agilent 509 5975 system (Agilent, USA) equipped with a capillary column (DB5-MS, 30 m X 0.25 mm). 510 Sample solutions were analyzed directly by GC-MS at a flow rate of 0.8ml min -1 , column 511 was equilibrated at 75°C for 1 min, with a 30°C min -1 increase to 170°C, 10°C min -1 increase 512 to 280°C for holding 2 min. Final FAME concentration was analyzed on the basis of the 513 FAME standard curve obtained from standard FAME mix (GLC-20 and GLC-30, Sigma 514 Aldrich). 515 516 Culture media, cultivation conditions and sample preparation 517 P. putida KT2440 (ATCC 47054) was obtained from ATCC and grown in a 518 chemically defined mineral medium containing the following (per liter): (NH 4 ) 2 SO 4 1.0 g/L, 519 KH 2 PO 4 1.5 g/L, Na 2 HPO 4 3.54 g/L, MgSO 4 ·7H 2 O 0.2 g/L, CaCl 2 ·2H 2 O 0.01 g/L, 520 ammonium ferric citrate 0.06 g/L and trace elements (H 3 BO 3

Generation of ΔPP3358 deletion strains in KT2440 543
Hydroxycinnamoyl-CoA hydratase-lyase (ech; PP_3358) gene deletion mutants in P. 544 putida were constructed by homologous recombination and sacB counterselection using the 545 allelic exchange vector pMQ30 (58). Briefly, homology fragments 1kb up-and downstream 546 of the target gene, including the start and stop codons respectively, were cloned into pMQ30 547 via Gibson assembly. Plasmids were then transformed via electroporation in E. coli S17 and 548 then mated into P. putida via conjugation. Transconjugants were selected for on LB Agar 549 plates supplemented with gentamicin 30 mg/mL, and chloramphenicol 30 mg/mL. 550 Transconjugants were then grown overnight on LB media also supplemented with gentamicin 551 30 mg/mL, and chloramphenicol 30 mg/mL, and then plated on LB Agar with no NaCl 552 supplemented with 10% (w/v) sucrose. Putative deletions were restreaked on LB Agar with 553 no NaCl supplemented with 10% (w/v) sucrose, and then were screened via PCR with 554 primers flanking the target gene to confirm gene deletion (Table S3). Plasmids and primers 555 were designed using Device Editor (59) and j5 software (60), and plasmids were assembled 556 with Gibson assembly (61). The strain (JPUB_013613) is available from the JBEI Public 557 Registry (https://public-registry.jbei.org/; Table 2). 558 559 Generation of ΔPP_0308-0313 and ΔPP_5063-5064 deletion strains in KT2440 560 First, deletion mutant P. putida ΔPP5063_5064 was assembled via the same method 561 as previously described in creation of ΔPP3358 deletion strains in KT2440 above. 562 Homologous recombination and sacB counterselection was used to construct the double 563 knockout deletion mutants targeting the genes NAD-dependent betaine aldehyde 564 dehydrogenase (betB; PP5063) and choline dehydrogenase (betA-II; PP5064). 565 Transconjugants were screened via colony PCR with primers flanking the target gene 566 deletion regions (Table S4). The deletion of dimethylglycine dehydrogenase (dgc operon; 567 PP0308-0313) in P. putida was further constructed by homologous recombination and sacB 568 counterselection using the vector pBF_ΔPP_0308-0313. Homologous fragments of 500 bps 569 flanking the up-and downstream regions of the target genes were amplified through PCR and 570 assembled into the pKD019MobSacB vector with restriction enzyme digestion and T4 571 ligation. The assembled plasmid was transformed via heat-shock into E. coli S17 cells and 572 transformants were selected on LB Agar plates supplemented with 50 mg/mL kanamycin. 573 Positive transformants were grown overnight in LB broth supplemented with 50 mg/mL 574 kanamycin and the plasmid pBF_ΔPP0308-0313 was extracted using the Qiagen Miniprep kit. 575 The extracted pBF_ΔPP0308-0313 plasmid was further transformed into the P. putida 576 ΔPP5063_5064 mutant via electroporation. Selection on single and double crossover events 577 were screened by plating on LB Agar supplemented with 50 mg/mL kanamycin and LB Agar 578 with no NaCl supplemented with 25% (w/v) sucrose. Deletion mutants that did not grow 579 when restreaked onto kanamycin supplemented plates were confirmed via colony PCR using 580 primers that flank the deletion region in the genome (Table S4). Plasmids and primers were 581 designed using SnapGene (GSL Biotech; available at snapgene.com). In order to compare the 582 genotype of the P. putida ΔPP5063_5064_0308-0313 mutant and the wild type P. putida 583 KT2440 in the presence of choline, each strain was grown in M9 medium supplemented with 584 0.2% (w/v) choline aspartate and monitored over 24 hours using a TECAN F200 microplate 585 reader (TECAN, Switzerland) at 30 °C, agitated at 200 rpm. The Strain (JPUB_013615) is 586 available from the JBEI Public Registry (https://public-registry.jbei.org/; Table 2). 587 588