The cscRABY genes conferred the ability to metabolize sucrose, but not maltose or lactose

First, we tested whether the new gene cluster was able to give P. putida the ability to grow on sucrose as the sole carbon source. As the putative sucrose permease CscB (PFL_3238) also shows homology to a galactoside permease (Accession number EFK49434.1, E‐score 4e‐71) from E. coli and belongs to the LacY super‐family, we also tested whether the common disaccharides lactose and maltose could serve as a carbon source in the presence of pSEVA221‐cscRABY. All three substrates cannot be metabolized by native P. putida EM178, albeit they might be transported by CscB. The results of cultivations in M9 medium using 3 g l⁻¹ of either of the three disaccharides or citrate as positive control are illustrated in Fig. 2. The only disaccharide able to support the growth of P. putida was sucrose. Moreover, the high growth rate of 0.292 ± 0.016 h⁻¹ indicates sucrose as the preferred substrate. In contrast, the strain bearing only the three genes cscRAB was not able to grow after 3 days of cultivation in M9 with sucrose (data not shown). Therefore, it was tempting to speculate that the porin CscY carried out a vital role during the consumption of sucrose by P. putida, which will be addressed in the following section. Citrate was used at the same (mass) concentration as sucrose, however, it showed lower yields in this experiment, which can be attributed to its higher degree of oxidation.

Open in a separate window

Figure 2

Cell dry weights of P. putida EM178 (pSEVA221‐cscRABY) on sucrose over time. P. putida EM178 (pSEVA221‐cscRABY) was grown in the presence of 3 g l⁻¹ of sucrose (dark blue), lactose (red), maltose (orange) or citrate (light blue) in a microplate reader in M9 minimal medium. The shaded areas represent the 95% confidence bands, which were estimated from three replicates.

The permease CscB also has high similarity to a raffinose permease from E. coli (Accession number EGI19556.1, E‐score 9e‐81). As raffinose and sucrose are structurally related and share the same sucrose (sub)unit, we cannot exclude that the gene cluster might also be able to transport and cleave raffinose. However, a second hydrolase may be needed for efficient cleavage of raffinose for the d‐galactopyranosyl‐(1→6)‐α‐d‐glucopyranoside moiety.

Expression of CscY complemented the loss of growth in P. putida pSEVA221‐cscRAB

The next step was to analyse the role of the porin CscY in more detail, as we had indications from earlier studies that the transport of sucrose across the membrane might be the limiting step (Löwe et al., 2017b). To test this assumption, we compared the growth of P. putida (pSEVA221‐cscRABY) and P. putida (pSEVA221‐cscRAB) on sucrose (Fig. 3). In fact, no growth could be observed on sucrose in the absence of CscY. To confirm that this effect was caused by the absence of the porin and was not a consequence of changes in the promoter region of cscRAB as a cloning artefact, P. putida (pSEVA221‐cscRAB) was complemented with the expression plasmid pSEVA434‐cscY. As can be seen in Fig. 3, the plasmid pSEVA434‐cscY was indeed able to restore the loss of growth associated with a lack of the porin in pSEVA221‐cscRAB, albeit at a slower rate than with the full set of genes. The experiment in Fig. 3 was performed in a microplate reader with LB as the preculture medium. The same experiment with identical precultures was also carried out in shaking flasks yielding similar results (Fig. S1 and Table 1). Following the sugar concentration in the medium over time revealed that sucrose was taken up and split only by strains expressing the porin (Fig. S1).

Open in a separate window

Figure 3

Complementation studies with P. putida EM178. The strain without the porin (pSEVA221‐cscRAB) was complemented with a LacI/P_trc promoter‐controlled expression vector containing the gene encoding the porin CscY (pSEVA434‐cscY). Growth on sucrose of the different strains was monitored in a microplate reader. The plasmid pSEVA221‐cscRABY was used as a positive control and the empty expression vector pSEVA434 as a negative control. The mean values of nine replicates (three biological with three technical replicates each) and 95% confidence bands were estimated from all replicates.

The unbalanced expression of the porin might provide an explanation for the slower growth rate of the strain complemented with CscY (see Fig. 3): The backbone plasmid pSEVA434 with its pBBR1 ori (moderate copy number) and the control of expression by the leaky LaqI/P_trc promoter (Balzer et al., 2013) might produce CscY at an unfavourably high rate. Furthermore, the artificial expression of membrane proteins is known to have a negative influence on cell vitality (Wagner et al., 2007). In fact, formation of flocks – probably biofilm formation, a common stress response in P. putida – was observed in the strains expressing CscY from pSEVA434‐cscY in shaking flasks. This was accompanied by a reduction in the cell dry weight of the planktonic cells (compare Fig. S1).

From these experiments, we concluded that, at least under these experimental conditions, the native outer membrane of P. putida was not permeable to sucrose. Compared to E. coli where transport of small molecules across the outer membrane is mainly facilitated by rather unspecific porins, P. putida has an outer membrane that more closely resembles the membrane of P. aeruginosa: homologues of the unspecific porins from E. coli (OpmF and OmpC) are not present. Instead, the general porin is OprF, which has a diffusion rate of two orders of magnitude lower than OpmF and OmpC of E. coli (Saravolac et al., 1991; Nikaido, 2003; Eren et al., 2012). This lower permeability is believed to be a major contributor to the remarkable resistance to toxic agents in fluorescent Pseudomonads (Nikaido, 2003) and could also be important for P. putida's resistance to chemical stresses.

Sucrose transport across the outer membrane was dependent on preculture medium

While carrying out the complementation experiments, we realized that, when using M9 medium‐grown precultures instead of LB‐grown precultures for otherwise identical experiments, the growth of P. putida (pSEVA221‐cscRAB) was not impaired on M9 sucrose, but the cells grew at almost the same rate as P. putida (pSEVA221‐cscRABY) (Fig S2 and Table 1). This unexpected finding might be explained by the different metabolic regimes of the cultures: Cells grown in minimal medium with only glucose as a carbon source have a glycolytic regime whereas those grown in LB need to carry out gluconeogenesis. For these different lifestyles, a different set of proteins and also a different composition of the outer membrane is required, which depends on the culture environment (Thompson et al., 2010; Choi et al., 2014). We speculate that the outer membrane of cells grown in M9 medium has a different set of outer membrane proteins and might already possess sucrose transport activity, whereas the one of LB‐grown cells does not. To test this hypothesis, we conducted a periplasm swelling experiment that is based on a method described by Nakae et al. (1989). Cells were grown in different media (LB, M9 with glucose, or M9 with citrate) and subsequently washed with a hypertonic solution containing glycerol (which should pass the outer membrane), sucrose and xylose. The substances that passed the outer membrane should be present in the periplasm after this first washing step.

Subsequently, these compounds now present in the periplasm were washed out again in a second washing step with deionized water. Then, the sugar concentrations in the supernatant were determined, which allowed us to draw conclusions on the permeability of the outer membrane for the detected sugars. As expected, the glycerol concentration was independent of preculture conditions, corroborating the assumption that it can diffuse freely across the membrane. However, sucrose was twice as concentrated in the supernatants of cells grown in M9 medium (Fig. 4). This can be explained by the faster diffusion rate of sucrose through the outer membrane of M9‐grown cells compared with LB‐grown cultures and was a further sign that using LB as preculture medium led to a lower permeability of the outer membrane for sucrose. The negative values for Xylose in Fig. 4 are an artefact from the calculation (Eq. (2)) and indicate that xylose is metabolized by the cells. In fact, the loss of xylose in the chromatogram is accompanied by the appearance of another peak close by, which is very likely an oxidized intermediate (data not shown). This behaviour of xylose was also reported by (Dvorak and de Lorenzo, 2018) and makes it an unsuitable internal standard for swelling experiments like this. Therefore, GFP and glycerol were used as internal standards.

Open in a separate window

Figure 4

Permeability of the outer membrane of P. putida EM178 estimated from the swelling experiment using a hypertonic solution of 0.15 M sucrose, xylose and glycerol in M9 medium. The bars represent the estimated substrate amounts (as quantified by their peak areas in HPLC) after washing the cells with deionized water, normalized with the fluorescence signal of GFP in supernatants according to Eqs (1) and (2). They should therefore represent the amount of substrate that is able to permeate through the outer membrane. Mean values and standard deviations are derived from three replicates. From left to right: blue bars: precultures grown in LB; red bars: precultures grown in M9 glucose; orange bars: precultures grown in M9 citrate.

The reason for the differences in the uptake of substrates through the outer membrane can probably be attributed to a different set of porins as a result of the preculture medium. In fact, previous proteome studies showed a strong dependency of most outer membrane porins on medium composition and carbon source in P. putida F1 and KT2440 (Thompson et al., 2010; Choi et al., 2014). From our experiments, we cannot conclude which particular porin might be responsible for sucrose uptake, but outer membrane fractions and purified porins of other Pseudomonads have been shown to be able to transport sucrose to a certain extent (Trias et al., 1988; Shrivastava et al., 2011; van den Berg, 2012). The major outer membrane porin of P. aeruginosa is OprF. As P. putida and P. aeruginosa share similar features in the organization of their outer membrane (Saravolac et al., 1991) and OprF is able to transport molecules as large as raffinose (Nikaido, 2003), it might well be responsible for sucrose transport in this work.

In any case, the whole cscRABY operon was used to construct a sucrose‐consuming P. putida strain to avoid phenotypic variations and to ensure sucrose uptake independent of the preculture medium.

Genomic integration of cscRABY genes resulted in stable and rapid sucrose‐dependent growth

As a next step, we aimed to integrate the gene cluster into the chromosome of P. putida in order to avoid the need for a plasmid and the associated antibiotic selection pressure. Using antibiotics may not be desirable or suitable for some applications. Therefore, we integrated the cscRABY genes into the attTn7‐site of P. putida EM178 via the mini‐Tn7 transposon vector pTn7‐M and evaluated the growth of the resulting strain on sucrose (Fig. 5). In shaking flasks growth rates of 0.45 ± 0.02 h⁻¹ and sucrose uptake rates of 5.9 ± 0.7 mmol g_CDW ⁻¹  h⁻¹ (calculated from three replicates) could be reached. For a more detailed overview of the growth rates, yields and substrate uptake rates, the key growth parameters are listed in Table 1. This compares well to growth rates measured with the monomers glucose and fructose and is around 60% higher than previously reported growth rates (Löwe et al., 2017b). Time courses of cell dry weight and sugar concentration in the supernatant are depicted in Fig. 5. Interestingly, fructose accumulated in the growth medium during fermentation, which might be explained by leakage of fructose after intracellular splitting of sucrose or by extracellular hydrolase activity. The latter option is not very likely because no splitting activity could be measured when the porin was lacking (Fig. S1). Leakage of fructose is also supported by the fact that P. putida does not show any fructokinase activity (Sawyer et al., 1977). Instead, fructose is phosphorylated during uptake through the native phosphotransferase system (PTS^Fru) (Velázquez et al., 2007). In fact, the absence of FruB, the cytoplasmatic component of the PTS^Fru, completely abolishes growth on fructose (Velázquez et al., 2007). Thus, after the intracellular cleavage of sucrose into glucose and fructose, glucose could be metabolized directly by glucokinase Glk, whereas fructose has to be secreted and taken up again via the fructose‐PTS system in order to be accessible to P. putida's metabolism. Alternatively, glucose could also be secreted and then oxidized in to gluconate before uptake. We did not measure any relevant amount of intermediates of glucose metabolism via HPLC, which indicates that glucose either did not accumulate because of rapid oxidation and re‐uptake or never left the cell.

Open in a separate window

Figure 5

Cell dry weight and sugar concentrations of a culture of P. putida EM178 attTn7::cscRABY over time. Cells were grown in M9 medium supplemented with 3 g l⁻¹ sucrose. Mean values and standard deviations of three independent biological replicates are shown. Experiments were performed in shaking flasks filled with 50 mL liquid and at an agitation rate of 220 rpm at 30°C.

For future applications, it might be worth considering the addition of a functional fructokinase to the metabolism of P. putida. This would result in a hybrid system consisting of the csc genes from P. protegens Pf‐5 and the ones of E. coli W that also comprise a fructokinase, and could be a way to make sucrose metabolism more efficient. Engineering the central carbon metabolism of P. putida was recently demonstrated (Sánchez‐Pascuala et al., 2017) and offers the potential to replace biomodules that are inefficient in an industrial context with streamlined versions for the desired application.

Genetic manipulations

All genetic constructs were created by restriction/ligation cloning and are listed in Table S1. The cscRABY gene cluster was amplified from genomic DNA of P. protegens Pf‐5 using the primers fwP_I_scr_P_pro and rvP_scr_P_pro (compare Table S3). The PCR product was cut with restriction enzymes SacI and XbaI (New England Biolabs, Ipswich, MA, USA) for the full gene cluster and with SacI and PstI to obtain only the region spanning the genes cscRAB. These fragments were subsequently ligated into the multiple cloning site of the likewise digested pSEVA221. Plasmid pTn7‐M‐cscRABY was created by the digestion of pSEVA221‐cscRABY with restriction enzymes SacI and XbaI and insertion into the identically cut multiple cloning site of pTn7‐M. To create pSEVA434‐cscY, pSEVA221‐cscRABY was digested with PstI and XbaI and ligated into the multiple cloning site of pSEVA434 that was previously cut with PstI and SpeI.

Constructs cloned in E. coli were transferred to P. putida EM178 via conjugation as described by others (de Lorenzo and Timmis, 1994). The method was simplified as follows: first, 200–300 μl of overnight grown cultures (LB medium) of plasmid donor, helper strain [E. coli HB101 (pRK600)] and recipient (P. putida EM178) was mixed and subsequently centrifuged (10 min at 8000 g), the supernatant was discarded, and the pellet was resuspended in the remaining droplet. It was transferred to LB agar plates without antibiotics and incubated as a sitting drop for 6–10 h. The cells were then simply streaked onto an M9 citrate agar plate with suitable antibiotics.

Sugar and alcohol determination by HPLC

High‐performance liquid chromatography was used to quantify the substrates and intermediates of sucrose metabolism: sucrose, glucose and fructose. Standards of 2 g l⁻¹ were used of each substrate. Glycerol and xylose were only determined semiquantitatively by considering their peak areas. Samples were prepared as follows: Cells were separated from samples by centrifuging at least 400 μl of cultivation broth at 17 000 g for 5 min. Supernatants were filtered through 0.22‐μm regenerated cellulose filter plates and injected (20 μl) into the HPLC (Agilent 1100 series, Waldbronn, Germany). Analytes were separated in a Shodex SH 1011 column at a flow rate of 0.45 ml min⁻¹ with 0.5 mM sulfuric acid as the mobile phase at 30°C. Concentrations were calculated by integration of the peak area of each peak and correlation to the corresponding standards.

We gratefully thank Professor Kirstin Jung from the Microbiology group, LMU Munich, for gratefully providing us a strain of Pseudomonas protegens Pf‐5 and Victor de Lorenzo, CNB, Spain, and his laboratory for access to the pSEVA plasmid collection and the genetically streamlined versions of P. putida KT2440.