Refactoring the upper sugar metabolism of Pseudomonas putida for co-utilization of disaccharides, pentoses, and hexoses

Given its capacity to tolerate stress, NAD(P)H/ NAD(P) balance, and increased ATP levels, the platform strain Pseudomonas putida EM42, a genome-edited derivative of the soil bacterium P. putida KT2440, can efficiently host a suite of harsh reactions of biotechnological interest. Because of the lifestyle of the original isolate, however, the nutritional repertoire of P. putida EM42 is centered largely on organic acids, aromatic compounds and some hexoses (glucose and fructose). To enlarge the biochemical network of P. putida EM42 to include disaccharides and pentoses, we implanted heterologous genetic modules for D-cellobiose and D-xylose metabolism into the enzymatic complement of this strain. Cellobiose was actively transported into the cells through the ABC complex formed by native proteins PP1015-PP1018. The knocked-in β-glucosidase BglC from Thermobifida fusca catalyzed intracellular cleavage of the disaccharide to D-glucose, which was then channelled to the default central metabolism. Xylose oxidation to the dead end product D-xylonate was prevented by by deleting the gcd gene that encodes the broad substrate range quinone-dependent glucose dehydrogenase. Intracellular intake was then engineered by expressing the Escherichia coli proton-coupled symporter XylE. The sugar was further metabolized by the products of E. coli xylA (xylose isomerase) and xylB (xylulokinase) towards the pentose phosphate pathway. The resulting P. putida strain co-utilized xylose with glucose or cellobiose to complete depletion of the sugars. These results not only show the broadening of the metabolic capacity of a soil bacterium towards new substrates, but also promote P. putida EM42 as a platform for plug-in of new biochemical pathways for utilization and valorization of carbohydrate mixtures from lignocellulose processing.


Introduction
1 Due to the physicochemical stresses that prevail in the niches in which the soil bacterium 2 Pseudomonas putida thrives (it is typically abundant in sites contaminated by industrial 3 pollutants), this microorganism is endowed with a large number of traits desirable in hosts of 4 harsh biotransformations of industrial interest . The P. putida strain KT2440 5 is a saprophytic, non-pathogenic, GRAS-certified (Generally Recognized as Safe) bacterium; 6 as the most thoroughly characterized laboratory pseudomonad, it has an expanding catalogue Gcd. Xylose metabolism in P. putida was established by implantating of xylose isomerase 8 XylA, xylulokinase XylB, and xylose-proton symporter XylE from E. coli. For the purpose of 9 cellobiose and xylose co-utilization, an expression cassette with the bglC gene was inserted 10 into the chromosome of P. putida EM42 Δgcd, while the synthetic xylABE operon was 11 expressed from the low copy pSEVA2213 plasmid under the constitutive pEM7 promoter. The 12 EM42 Δgcd mutant was used to avoid xylose conversion by glucose dehydrogenase to dead- 13 end product D-xylonate. PQQ, pyrroloquinoline quinone; Glk, glucokinase; TCA cycle, 14 tricarboxylic acid cycle; OM, (outer membrane); IM (inner membrane).  2 All bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli 3 strains used for cloning or triparental mating were routinely grown in lysogeny broth (LB; 10 4 g L -1 tryptone, 5 g L -1 yeast extract, 5 g L -1 NaCl) with agitation (170 rpm) at 37°C. 5 Cloramphenicol (Cm, 30 μg mL -1 ) was supplemented to the medium with E. coli helper strain 6 HB101. Pseudomonas putida recombinants were routinely pre-cultured overnight in 2.5 mL of 7 LB medium with agitation of 300 rpm (Heidolph Unimax 1010 and Heidolph Incubator 1000; 8 Heidolph Instruments, Germany) at 30°C. For initial tests of expression of heterologous genes 9 in P. putida, cells were transferred to 25 mL of fresh LB medium in Erlenmeyer flask and 10 cultivated as described in section 2.4. For the growth experiments with different carbohydrates, 11 overnight culture was spun by centrifugation (4,000 g, RT, 5 min), washed with M9 minimal 12 medium (per 1 L: 4.25 g Na2HPO4 2H2O, 1.5 g KH2PO4, 0.25 g NaCl, 0.5 g NH4Cl) added 13 with MgSO4 to the final concentration of 2 mM, and with 2.5 mL L -1 trace element solution 14 (Abril et al., 1989). Thiamine HCl (1 mM) was added to the minimal medium for cultures with 15 E. coli recombinants. Cells were resuspended to OD600 of 0.1 in 25 mL of the same medium 16 with kanamycin (Km, 50 µg mL -1 ), in case of recombinants with pSEVA2213 or pSEVA238 17 plasmid, or streptomycin (Sm, 60 µg mL -1 ), in case of P. putida EM42 Δgcd bglC, and with 18 carbon source (glucose, xylose, or cellobiose) of concentration defined in the text or respective

Plasmid and strain constructions
1 DNA was manipulated using standard laboratory protocols (Sambrook and Russell, 2001). 2 Genomic DNA was isolated using GenElute bacterial genomic DNA kit (Sigma-Aldrich, 3 USA). Plasmid DNA was isolated with QIAprep Spin Miniprep kit (Qiagen, USA). The 4 oligonucleotide primers used in this study (Table S1) were purchased from Sigma-Aldrich 5 (USA). The genes of interest were amplified by polymerase chain reaction (PCR) using Q5 6 high fidelity DNA polymerase (New England BioLabs, USA) according to the manufacturer's 7 protocol. The reaction mixture (50 μL) further contained polymerase HF or GC buffer (New 8 England BioLabs, USA), dNTPs mix (0.2 mM each; Roche, Switzerland), respective primers 9 (0.5 mM each), water, template DNA, and DMSO. GC buffer and DMSO were used for 10 amplification of genes from P. putida. PCR products were purified with NucleoSpin Gel and 11 PCR Clean-up (Macherey-Nagel, Germany). DNA concentration was measured with NanoVue 12 spetrophotometer (GE Healthcare, USA). Colony PCR was performed using 2x PCR Master 13 Mix solution of Taq DNA polymerase, dNTPs and reaction buffer (Promega, USA). All used 14 restriction enzymes were from New England BioLabs (USA). Digested DNA fragments were 15 ligated using Quick Ligation kit (New England BioLabs, USA). PCR products and digested 16 plasmids separated by DNA electrophoresis with 0.8 % (w/v) agarose gels were visualised 17 using Molecular Imager VersaDoc (Bio-Rad, USA). Plasmid constructs were confirmed by 18 DNA sequencing (Macrogen, South Korea). Chemocompetent E. coli Dh5α cells were 19 transformed with ligation mixtures or complete plasmids and individual clones selected on LB 20 agar plates with Km (50 µg mL -1 ) were used for preparation of glycerol (20 % w/v) stocks. 21 Constructed plasmids were transferred from E. coli Dh5α donor to P. putida EM42 by tripartite 22 mating, using E. coli HB101 helper strain with pRK600 plasmid (Table1). Alternatively, 23 electroporation (2.5 kV, 4 -5 ms pulse) was used for transformation of P. putida cells with 24 9 selected plasmids using a MicroPulser electroporator and Gene Pulser Cuvettes with 0.2 cm 1 gap (Bio-Rad, USA). Preparation of P. putida electrocompetent cells and electroporation 2 procedure was performed as described elsewhere (Aparicio et al., 2015). P. putida 3 transconjugants or transformants were selected on M9 agar plates with citrate or LB agar plates, 4 respectively, with Km (50 µg mL -1 ) at 30°C overnight. 5 Construction of cellobiose metabolism module. The ccel_2454 gene encoding β- PstI resulting in pSEVA238_ccel2454. The β-glucosidase encoding bglX gene was PCR 10 amplified from the genomic DNA of P. putida KT2440 using bglX fw and bglX rv primers, 11 digested with NdeI and HindIII and cloned into corresponding restriction sites of modified 12 pSEVA238 resulting in pSEVA238_bglX. The NdeI site and a consensus RBS were previously 13 introduced into the standard SEVA polylinker of pSEVA238 (unpublished plasmid). The bglC 14 gene encoding β-glucosidase from Thermobifida fusca with N-terminal 6xHis tag was 15 subcloned from pET21a_bglC construct into NdeI and HindIII restriction sites of modified 16 pSEVA238. The bglC gene with the RBS and His tag was subsequently PCR amplified using 17 primers bglC fw and bglC rv, the PCR product was cut with SacI and PstI and subcloned into 18 pSEVA2213 giving rise to pSEVA2213_bglC. 19 Insertion of bglC gene into P. putida chromosome. The bglC gene with consensus RBS was 20 subcloned into SacI and PstI sites of mini-Tn5-vector pBAMD1-4 (Martínez-García et al., 24 g L -1 , tryptone 20 g L -1 , KH2PO4 0.017 M, K2HPO4 0.072 M) at 30°C with shaking (170 1 rpm). Cells were collected by centrifugation (4000 rpm, 10 min) and resuspended in 100 mL 2 of selection M9 medium with 5 g L -1 cellobiose and streptomycin (50 μg mL -1 ). After four days 3 of incubation at 30°C with shaking (170 rpm), cells were spun (4000 rpm, 15 min) and plated 4 on selection M9 agar plates with 5 g L -1 cellobiose and streptomycin (50 μg mL -1 ). Three fastest 5 growing clones were re-streaked on fresh M9 agar plates with streptomycin or with 6 streptomycin (50 μg mL -1 ) and ampicillin (500 µg mL -1 ) to rule out insertion of the whole 7 pBAMD1-4 plasmid. The growth of three candidates in liquid minimal medium with cellobiose 8 was verified. The insertion site of expression cassette (pEM7 promoter, bglC gene, T500 9 transcriptional terminator, and aadA gene) in chromosome of the fastest growing clone was 10 determined by two-round arbitrary primed PCR with Arb6, Arb2, ME-O-Sm-Ext-F, and ME- 11 O-Sm-Int-F primers ( pSEVA2213_xylAB. The gene of xylose-proton symporter (xylE) was amplified from the 20 genomic DNA of E.coli BL21 (DE3) using two-step PCR protocol. In the first step, the gene 21 was amplified using xylE fw 1 and xylE rv 1 primers. The sample of the reaction mixture with 22 the PCR product (1 μL) was transferred into the second reaction with xylE fw 2 and xylE rv 2 23 primers. Final PCR product was digested with BamHI and HindIII and cloned downstream 24 xylAB operon in pSEVA2213_xylAB resulting in pSEVA2213_xylABE. For the purpose of 1 construction of the plasmid allowing translational fusion of XylE to monomeric superfolded 2 GFP (msfGFP), gfp gene was initially amplified without its own RBS but with STOP codon 3 from pSEVA238_gfp plasmid (SEVA collection) using gfpC fw and gfpC rv primers. The PCR 4 product was digested with HindIII and SpeI and ligated into pSEVA238, cut with the same pair 5 of enzymes, giving rise to pSEVA238_gfpC. The xylE gene was amplified from 6 pSEVA2213_xylABE with its synthetic RBS but without STOP codon using xylE-gfp fw and 7 xylE-gfp rv primers. The PCR product was digested with BamHI and HindIII and cloned 8 upstream the gfp gene in pSEVA238_gfpC, resulting in pSEVA238_xylE-gfpC.

9
Preparation of deletion mutants of P. putida EM42. Deletion mutants were prepared using  gene gcd (PP_1444) was deleted accordingly in P. putida EM42, resulting in P. putida EM42 5 Δgcd, and in P. putida EM42 Δgts, resulting in P. putida EM42 Δgts Δgcd, using a set of TS 6 primers listed in Table S1. Expression of I-SceI in selected co-integrates was induced with 1 7 mM 3MB for 6 hrs. Check(-)gcd fw and check(-)gcd rv primers were used to confirm deletion 8 of gcd gene. PCR product size in case of deletion was 1500 bp. P. putida recombinants were 9 cured of pSW-I plasmid after several passes in LB medium lacking Amp. 10 11 Biomass was determined as dry cell weight. Samples of cultures grown in M9 minimal 12 medium with 5 g L -1 glucose were transferred into 2 mL pre-dried and pre-weighed Eppendorf 13 tubes and pelleted at 13,000 g for 10 min. The pellets were washed twice with distilled water 14 and dried at 80°C for 48 h. Based on the prepared standard curve, one A600 unit is equivalent 15 to 0.38 g L -1 of dry cell weight. Specific growth rate (μ) was determined during exponential 16 growth as a slope of the data points obtained by plotting the natural logarithm of A600 values 17 against time. Substrate consumption rate (r) was determined for initial 12 and 24 h of culture 18 as r = (c substrate at t0 -c substrate at t1) / (t1 -t0). Biomass yield (YX/S) was determined 24 h 19 after each culture started to grow exponentially as YX/S = c biomass at t1 / (c substrate at t0 -c 20 substrate at t1). Specific carbon (C) consumption rate (qs) was determined during exponential 21 growth on glucose or cellobiose as qs = (mmol C at t0 -mmol C at t1) / ((t1 -t0) * (g biomass at 22 t1 -g biomass at t0)). For initial screening of β-glucosidase activities, 25 mL of LB medium was inoculated from 2 overnight cultures to A600 = 0.05 and cells were grown for 3 h at 30°C with shaking (170 rpm). 3 Expression of β-glucosidase genes from pSEVA238 plasmid was then induced with 1 mM 4 3MB. After induction, cells were grown in the same conditions for additional 5 h and then  Bioprocessing Reagent and lysed for 15 min at RT with slow agitation. Cell lysates for xylose 14 isomerase and xylulokinase activity determination were prepared by sonicating concentrated 15 cell solutions prepared by spinning (4,000 g, 4°C, 10 min) 25 mL of cells grown in LB medium 16 to A600 = 1.0. Cell pellets were washed by 5 mL of ice-cold 50 mM Tris-Cl buffer (pH 7.5) 17 resuspended in 1 mL of the same buffer, placed in ice bath and disrupted by sonication. In all 18 cases, cell lysates were centrifuged at 21,000 g for 30 min at 4°C and supernatants, termed here  nm with UV/Vis spectrophotometer Ultrospec 2100 (Biochrom, UK) and activity was 10 calculated using calibration curve prepared with p-nitrophenol standard (Sigma-Aldrich, 11 USA). β-glucosidase activity in culture supernatants was measured correspondingly with 166 12 μL of culture supernatant in 600 μL of reaction mixture. 13 Activity of xylose isomerase (XylA) was measured as described by Le Meur and co-   23 2 mM ATP, 2 mM MgCl2, 0.2 mM phosphoenolpyruvate, 10 U of pyruvate kinase and 10 U, 24 lactate dehydrogenase, and 10 mM D-xylulose. Reaction at 30°C was started by addition of 5 1 μL of 20-fold diluted CFE. 2 Both in xylose isomerase and xylulokinase assay, the depletion of NADH was measured 3 spectrophotometrically at 340 nm with Victor 2 1420 Multilabel Counter (Perkin Elmer, USA). 4 Molar extinction coefficient of 6.22 mM -1 cm -1 for NADH was used for activity calculations. 5 1 unit (U) of activity corresponds to 1 μmol of substrate (pNPG or NADH) converted by 6 enzyme per 1 min. 7 Activity of glucose dehydrogenase (Gcd) in P. putida EM42 and P. putida EM42 Δgcd was 8 determined by measuring conversion of 5 g L -1 xylose to xylonate by cell suspension of A600 = 9 0.55 in 25 mL of M9 medium at 30°C. The time course of the reaction was 6 h. 1 U of enzyme 10 activity corresponds to 1 μmol of xylonate produced per 1 minute. 11

SDS-PAGE and Western blot analyses
12 CFE for determination of expression levels of selected enzymes were prepared using cell 13 pellets from cultures induced with 1 mM 3MB and lysed with BugBuster Protein Exctraction 14 Reagent as described above. Samples of CFE containing 5 μg of total protein were added with 15 5x Laemmli buffer, boiled at 95°C for 5 min and separated by SDS-PAGE using 12 % gels. 16 CFE prepared from P. putida cells with empty pSEVA238 plasmid was used as control. Gels 17 were stained with Coomassie Brilliant Blue R-250 (Fluka/Sigma-Aldrich, Switzerland). 18 The staining step was omitted for Western blotting. Instead, proteins were 19 electrotransferred from gel onto Immobilon-P membrane (Merck Millipore, Germany) of pore       1 . Column temperature was 30°C. Chemicals were identified using pure compound standards.

22
Glucose and xylose concentrations in culture supernatants were also determined by Glucose 23 (GO) Assay Kit (Sigma-Aldrich, USA) and Xylose Assay Kit (Megazyme, Ireland), 24 respectively. Xylonic acid was measured using the hydroxamate method (Lien, 1959). Samples    Initial tests with P. putida EM42 in minimal medium with cellobiose showed no growth 3 of this platform strain on the disaccharide (Fig. 2A). The absence of cellobiose assimilation 4 implied either lack of transport through the cell membranes or missing enzymatic machinery   %) with the well-characterized E. coli -glucosidase BglX (Yang et al., 1996). Each of these 10 genes was cloned into the pSEVA238 plasmid downstream of the inducible XylS/Pm promoter. 11 We tested the effect of gene expression on EM42 viability, as well as soluble protein production 12 and enzyme activity in cell-free extracts (CFE). 13 All three enzymes were produced in the soluble fraction of the P. putida chassis grown 14 in LB medium (Fig. S1A), but only Ccel_2454 and BglC showed measurable -glucosidase 15 activity. No activity was detected in CFE containing endogenous BglX, whose overexpression 16 also had a clear toxic effect on the host (Fig. S2). Absence of BglX -glucosidase activity cellobiose as a sole carbon source. In this preliminary shake flask experiment, the OD600 at 48 5 h was 3.6 (data not shown). 6 The bglC gene was subsequently subcloned into the low copy plasmid pSEVA2213 7 with the constitutive promoter pEM7 that functions well in both P. putida and E. coli (Zobel et 8 al., 2015). The P. putida EM42 bearing the pSEVA2213_bglC construct grew rapidly in 9 minimal medium with 5 g L -1 cellobiose, with only a ~2 h adaptation period and a specific 10 growth rate of 0.35 ± 0.02 h -1 (Fig. 2B, Table 2). The substrate consumption rate and biomass 11 yield parameters were about 40 % and 20 % lower, respectively, than those of P. putida EM42 12 pSEVA2213 cultured in minimal medium with glucose ( Table 2). As all disaccharide was 13 consumed within 24 h of culture, the P. putida strain outperformed the best engineered E. coli 14 strain CP12CHBASC30, which assimilated 3.          The second is a periplasmic pathway formed by membrane-bound PQQ-dependent glucose 20 dehydrogenase Gcd (PP_1444). This enzyme oxidizes D-glucose to D-glucono-1,5-lactone, in EM42 strain, we deleted the gtsABCD operon that encodes the ABC glucose transporter.

9
The mutant transformed with the pSEVA2213_bglC plasmid showed a substantially prolonged 10 (~7 h) adaptation phase on cellobiose (Fig. 2C). Three determined growth parameters were 11 reduced when compared with P. putida EM42 pSEVA2213_bglC ( Table 2). The results

12
suggested that the glucose ABC transporter plays an important role in cellobiose uptake, but is 13 not the only access route for the disaccharide in P. putida. 14 Closure of the second glucose uptake route by deleting the glucose dehydrogenase gene 15 gcd in P. putida EM42 had no notable effect on growth on cellobiose, but slightly prolonged 16 the lag phase (~3 h; Fig. 2D). Substrate consumption during the initial 12 h of growth was 17 nonetheless significantly reduced ( Table 2), which might be attributed to the uptake of 18 cellobiose only through the direct phosphorylation route and slower initial expression of ABC 19 transporter components that is normally induced by monomeric glucose (del Castillo et al., 20 2007). It is worth noting here that neither the genes which encode two out of three 21 carbohydrate-selective porins (oprB-1 and oprB-2) adjacent to gtsABCD and gcd, respectively, 22 nor their regulatory sequences were affected by the scarless deletions. 23 Growth on disaccharide was completely abolished when the deletions in the direct 1 phosphorylation and oxidative routes were combined (Fig. 2E, Table 2). It can be thus argued 2 that the peripheral glucose pathway also takes part in cellobiose assimilation by P. putida. One 3 could speculate that Gcd also has cellobiose dehydrogenase activity (EC 1.1.99.18) and 4 converts cellobiose to cellobiono-1,5-lactone (Henriksson et al., 2000). Much like glucono-1,5-5 lactone, in the presence of water cellobiono-1,5-lactone might be hydrolyzed spontaneously to 6 cellobionic acid, which would be transported to the cytoplasm. Intracellular -glucosidase 7 could then cleave cellobionate to glucose and gluconic acid (Li et al., 2015), two molecules 8 easily metabolized by P. putida. Nonetheless, neither cellobionic acid formation nor its further 9 metabolism in P. putida can be confirmed based on currently available experimental data. 10 Hence, the detailed functioning of the peripheral oxidative route in upper cellobiose 11 metabolism in P. putida remains to be elucidated by our future experiments.

12
From the acquired growth parameters of these P. putida mutants, it can be deduced that 13 the direct phosphorylation route is of major importance for cellobiose assimilation. On the 14 other hand, experimental evidence shows that glucose enters P. putida cells predominantly 15 through the peripheral oxidative pathway . Despite these opposing access 16 route preferences, when P. putida EM42 pSEVA2213_bglC was exposed to a mixture of the 17 two sugars (2 g L -1 each), we observed diauxic growth (Fig. 2F). Glucose was utilized first 18 during the initial 8 h of the experiment. When all hexose was removed from the medium, 19 cellobiose was consumed rapidly during the next 6 h of the culture. To conclude, these 20 experiments suggest that glucose and cellobiose share the same acess routes in P. putida and 21 that monomeric hexose is a preferred substrate in the mixture of the two carbon sources.

3.3.Probing energetic benefit of cellobiose metabolism in engineered P. putida
1 ATP is a universal energy source and a major driving force for biochemical processes 2 in microbial cell factories (Hara and Kondo, 2015). Due to its variant of glycolysisthe Entner- 3 Doudoroff pathway -P. putida yields only one net ATP per one mole of assimilated glucose 4 . In the case, for instance, of E. coli with its characteristic Embden-  two means at P < 0.05 (*) or P < 0.01 (**). We determined ATP levels in EM42 pSEVA2213 and EM42 pSEVA2213_bglC strains 13 grown on glucose and cellobiose, respectively, to evaluate the effect of altered substrate on P. 14 putida energy status (Fig. 3A). Indeed, the ATP level in P. putida grown on cellobiose was 15 almost double that of the cells cultured on glucose. ATP savings could partially stem from 16 cellobiose transport. One ATP per molecule of glucose is theoretically saved when cellobiose To test the capacity of P. putida to co-utilize cellobiose with pentoses, we established  (Fig. 1). We aimed at following the same strategy in our study. 16 We first verified the absence of xylose catabolism in P. putida EM42. Our 17 bioinformatic analysis confirmed that the KT2440 genome has no genes that encode 18 homologues of E. coli XylA or XylB proteins. The EM42 strain was then incubated in minimal 19 medium with xylose, and the cell density and substrate concentration were measured for five 20 consecutive days (Fig. 4A). No growth was detected, despite the fact that only 10% of the 21 starting xylose concentration was detected in the culture medium after five days. Meijnen and 22 co-workers (2008) described a similar phenomenon for engineered P. putida S12. In that case, 23 the majority of D-xylose was oxidized to the dead-end product D-xylonate by the periplasmic 24 glucose dehydrogenase Gcd. In fact, xylonate concentrations determined in samples from the 1 five-day experiment with the strain EM42 suggested that all xylose was converted to the acid, 2 which was not assimilated by the cells (Fig. 4A). Xylonate formation was accompanied by a 3 decrease in the culture pH from 7.0 to 6.2. This initial experiment provided additional evidence 4 of Gcd broad substrate specificity in P. putida KT2440 and its derivatives. The phenomenon 5 of xylonate formation was not discussed in the study by Le Meur and colleagues (2012), who 6 implanted the isomerase pathway in the KT2440 strain. To avoid accumulation of an undesirable metabolite, we transplanted the xylAB 8 fragment of the xyl operon from E. coli BL21 (DE3), which encodes xylose isomerase XylA 9 and xylulokinase XylB, directly to P. putida EM42 Δgcd. The xylAB fragment was amplified 10 as a whole. The xylA gene (SI sequences) was provided with a consensus RBS, with the native 11 RBS maintained upstream of the xylB gene. The fragment was cloned into pSEVA2213 and 12 xylAB expression was verified in the EM42 Δgcd strain (Fig. S3). XylA and XylB activities 13 determined in CFE were higher than those reported for engineered P. putida S12 growing on 14 xylose ( limited growth and substrate uptake in minimal medium with 5 g L -1 xylose (Table 3, Fig. 4B). 16 When xylonate accumulation no longer hindered efficient cell use of xylose, we found substrate 17 transport to be another bottleneck to xylose metabolism in this host. This was not anticipated 18 based on a previous study with the xylose-utilizing KT2440 strain, which reported only 19 implantation of the XylAB metabolic module with no transport system (Le Meur et al., 2012). 20 In another report, an upregulated glucose ABC transporter was nonetheless defined as one of 21 the major changes that shaped laboratory-evolved P. putida S12 towards rapid growth on      The xylose-proton symporter XylE from E. coli is a relatively small (491 amino acids) 11 single-gene transporter with a known tertiary structure and a well-described transport  Once we confirmed correct XylE expression and localization in P. putida, the EM42

11
Δgcd strain was transformed with the pSEVA2213_xylABE construct to test functioning of the 12 whole synthetic operon. Co-expression of the exogenous transporter with xylose isomerase and 13 xylulokinase genes improved the specific growth rate on xylose by 8-fold when compared with 14 the recombinant without xylE (Fig. 4C and Table 2); improvement was also observed for other 15 growth parameters ( Table 2). The faster growth of the recombinant bearing the xylABE 16 synthetic operon could not be attributed to changes in XylA or XylB activitity ( Table 2). We 17 observed a >2-fold decrease in XylB activity when xylE was subcloned downstream of xylB 18 and the position of the gene in operon was changed. Finally, we verified the importance of the 19 gcd deletion for complete xylose utilization by P. putida recombinants in an experiment with 1 the EM42 pSEVA2213_xylABE strain (Fig. 4D). It is clear from the time course of the culture 2 that ~50 % of all uptaken xylose was still converted non-productively to xylonate in the 3 bacterium with functional Gcd, despite the presence of the heterologous machinery that funnels 4 the substrate to the pentose phosphate pathway. 5 We thus demonstrate that efficient xylose metabolism can be established in the platform 6 strain P. putida EM42 when two major bottlenecksthe peripheral oxidative pathway and the 7 missing transport systemare removed. Neither of these bottlenecks was rationally engineered other substrates such as organic acids or amino acids are preferred to sugars (Rojo, 2010). 6 Moreover, the isomerase pathway and the xylose transporter introduced into P. putida EM42 7 are of exogenous origin and their expression is not governed by the host. We thus anticipated 8 that our recombinant strains would have co-utilized xylose with glucose or cellobiose with no 9 restrictions. 10 We first sought to verify simultaneous utilization of glucose and xylose in P. putida 11 EM42 Δgcd pSEVA2213_xylABE. This experiment with two monomeric sugars was an 12 essential prerequisite for co-utilization of xylose and cellobiose in engineered P. putida. The 13 strain with the gcd deletion was used to avoid xylose oxidation to the dead-end by-product 14 xylonate. As explained in section 3.2., this deletion is not detrimental either for glucose or for 15 cellobiose uptake in P. putida, and both molecules can enter the cell with the help of ABC 16 transporter. Equal concentrations of monosacharides (2 g L -1 ) were used to better visualize the 17 differences in glucose and xylose consumption. Pseudomonas putida EM42 Δgcd 18 pSEVA2213_xylABE assimilated glucose and xylose simultaneously, and no sugar was 19 detected in culture supernatants after 24 h (Fig. 6B). In cultures of the negative control P. 20 putida EM42 Δgcd pSEVA2213 lacking the xylABE operon, the xylose concentration dropped 21 by only 13% in the same time period (Fig. 6A). It is possible that some xylose entered the cells 22 by non-specific transport routes. The presence of the additional carbon source significantly 23 accelerated xylose assimilation by recombinant P. putida (Fig. 6B). On average, 2 g L -1 of 24 37 xylose were consumed during the initial 24 h of co-utilization experiments, while <1 g L -1 was 1 mineralized in cultures with pentose alone at a starting concentration of 5 g L -1 (Fig. 4C). The 2 substrate consumption rate of xylose was nonetheless still lower than that of glucose, which 3 caused two-phase growth of the EM42 Δgcd pSEVA2213_xylABE strain on two sugars at the 4 same starting concentration (Fig. 6B). This experiment demonstrated the ability of engineered 5 P. putida to co-utilize hexose and pentose without CCR.  For the cellobiose and xylose co-utilization experiments, new metabolic modules had 1 to be combined in a single P. putida cell. As the synthetic bglC-xylABE operon borne on a 2 single pSEVA2213 plasmid appeared to be unstable, we integrated the bglC gene directly into 3 the P. putida EM42 Δgcd chromosome. The gene with the pEM7 promoter and consensus RBS 4 was subcloned into the mini-Tn5-vector pBAMD1-4, which allows for random chromosomal 5 insertions and subsequent selection of the optimal phenotype from a broad expression bearing the bglC gene on the plasmid; substrate consumption rate was even faster ( Table 2, 16 Fig. S4). The insertion had no effect on cell viability either in rich LB medium or in minimal 17 medium with citrate as the gluconeogenic carbon source (Fig. S5). 18 Hence, P. putida EM42 Δgcd bglC was transformed with the pSEVA2213_xylABE 19 construct, and the resulting recombinant used for co-utilization experiments. Cells were 20 incubated in minimal medium with cellobiose and xylose at equal concentrations (2 g L -1 ). 21 Once again, both substrates were co-utilized rapidly; no residual carbohydrate was detected in 22 supernatants after 28 h culture (Fig. 6C). Xylose was assimilated slower than cellobiose. We 23 also tested performance of EM42 Δgcd bglC pSEVA2213_xylABE strain in mixture of 24 cellobiose, xylose, and glucose at 2 g L -1 concentration each (Fig. 6D). While xylose 1 consumption remained unafected by presence of glucose in medium, cellobiose uptake was 2 notably accelerated. In contrast, glucose concentration in the culture supernatants increased 3 untill cellobiose was consumed and then all the hexose was co-utilized with remaining xylose. 4 We argue that glucose, present in the medium already at the beginning of the culture, allowed Such induced expression of the ABC transporter in P. putida Δgcd mutant could thus lead to 10 the inverted substrate preference observed in the last co-utilization experiment.

11
Small amounts of extracellular glucose were detected in all the cultures with P. putida 12 EM42 Δgcd mutant grown on cellobiose (Figs. 2D, 6C, 6D, and S4). Cellobiose streamed into 13 the cells only through the ABC transporter might cause temporal accumulation of intracellular 14 glucose, which is released into the medium and later transported back to the cytoplasm. In 15 contrast to many other -glucosidases, BglC is not inhibited by glucose (Spiridonov and 16 Wilson, 2001). We thus hypothesize that glucose accumulation in the EM42 Δgcd mutant stems 17 from an imbalance between -glucosidase activity and P. putida glycolysis, namely its the 18 upper part encompassing glucokinase (PP_1011) and glucose-6-phosphate 1-dehydrogenase 19 (PP_1022, PP_4042, PP_5351). The system must accomodate all the glucose in its native form 20 rather than its oxidized intermediates gluconate and 2-ketogluconate, which prevail in the cells 21 with functional Gcd . The imbalance could be reduced for instance by   are cultured on glucose. The cellobiose-utilizing P. putida would thus be an even more robust 15 host than the template strain for accommodating heterologous or engineering endogenous 16 anabolic pathways for biosynthesis of value-added chemicals directly from the disaccharide or 17 a co-substrate (Hara and Kondo, 2015). Finally, we identify the ability of P. putida, following 18 introduction of the xylABE synthetic operon, to co-utilize cellobiose or glucose with pentose, 19 with no need for further interventions in the regulatory mechanisms of central carbon assimilation and co-utilization with glucose in our P. putida KT2440 derivative is demonstrated 1 in this study. 2 Although there is indeed room for improvement and further testing of the strains 3 constructed here, we argue that this study increases the value of P. putida for the 4 biotechnological recycling of lignocellulosic feedstocks, specifically for processes that include 5 partial hydrolysis of the input material. Using synthetic and systems biology approaches, 6 carbon from new (hemic)cellulosic substrates -cellobiose and xylose-can be streamlined 7 towards valuable chemicals whose production has been reported in P. putida, such as mcl-PHA Acknowledgements 16 We would like to thank prof. Edward A. Bayer for providing us with pET21a_bglC plasmid, 17 to Dr. Esteban Martínez-García for P. putida EM42 strain, and to Dr. Alberto Sánchez-Pascuala 18 for pEMG_gtsABCD and pEMG_gcd plasmids and corresponding oligonucleotide primers. 19 The project has received funding from the EU's Horizon 2020 research and innovation    Microbiol. 42, 295-301. 30 Stephanopoulos, G., 2007. Challenges in engineering microbes for biofuels production. 31 Science 315, 801-804. 32 Taha, M., Foda, M., Shahsavari, E., Aburto-Medina, A., Adetutu, E., Ball, A., 2016. 33 Commercial feasibility of lignocellulose biodegradation: possibilities and challenges. 34 Curr. Opin. Biotechnol. 38, 190-197. 35 Teugjas, H., Väljamäe, P., 2013. Product inhibition of cellulases studied with 14 C-labeled 36 cellulose substrates. Biotechnol. Biofuels 6, 104.