Scientific Papers

Two-step conversion of polyethylene into recombinant proteins using a microbial platform | Microbial Cell Factories


Media design and growth of Pseudomonas strains using hexadecane as sole carbon source

Two strains of Pseudomonas, P. aeruginosa RR1 and P. oleovorans, were chosen for this work based on their previously demonstrated ability to utilize pyrolyzed polyethylene as a sole carbon source during growth (all strains, primers, and plasmids used in this work can be found in Additional file 1: Table S1) [22]. The growth of these strains was first characterized by using a model alkane, hexadecane (C16), as the sole carbon source. Hexadecane served as a proxy for hydrocarbons derived from polyethylene, as it can be a major alkane constituent resulting from the depolymerization of polyethylene via pyrolysis [22]. A minimal salt media that contained either ammonium chloride (NH4Cl) or ammonium nitrate (NH4NO3) as the nitrogen source was used, and hexadecane added was the sole source of elemental carbon in the media. The amount of hexadecane and elemental nitrogen supplied was based on theoretical calculations for the minimum amount of elemental carbon and nitrogen required to reach high cell densities of at least 10 g/L of wet pellet mass [37, 38]. Calculations were performed by assuming that dry cell mass accounted for 25% of wet pellet mass, and within that dry mass 50% was elemental carbon and 14% was elemental nitrogen [37, 38]. Thus, the theoretical minimum amount of hexadecane needed to supply enough elemental carbon to reach a wet pellet mass of 10 g/L is 1.47 g/L (0.147% w/v). Likewise, the theoretical minimum amount of ammonium chloride or ammonium nitrate required for a desired wet pellet mass of at least 10 g/L, is 1.34 and 1.00 g/L respectively.

A total of twelve different combinations of culture conditions were tested with three different amounts of hexadecane (0.46%, 4.6%, or 9.2% w/v), two different amounts of elemental nitrogen (0.65 g/L or 1.3 g/L elemental N), and two different nitrogen sources (NH4Cl or NH4NO3). Table 1 depicts the combinations of culture conditions that were tested for growth. To promote cell growth and recombinant protein production, the levels of supplied carbon and nitrogen were several times above the calculated minimum requirements for achieving a wet pellet mass of 10 g/L. Flask cultures were inoculated with a single plated colony of either P. aeruginosa RR1 or P. oleovorans and kept at 37 °C with 225 rpm shaking. As others have noted, the growth of Pseudomonas on hexadecane resulted in the formation of millimeter-sized white, waxy particulates that were numerous enough to interfere with OD600 measurements [39]. Thus, cell growth was quantified by performing serial dilutions of cultures with subsequent plating on LB agar to determine the density of colony-forming-units (cfu/ml).

Table 1 Design of Media for Growth on Hexadecane as Sole Carbon Source*

Figure 1 depicts the average cfu/ml of P. aeruginosa RR1 and P. oleovorans cultures 96 h after inoculation when using hexadecane as the sole carbon source under different culture conditions. For comparison, data for overnight cultures (18 h) grown in rich LB media can be found on the far left, averaging 2.38 ± 0.61 × 109 cfu/ml for P. aeruginosa RR1 and 1.65 ± 0.45 × 109 cfu/ml for P. oleovorans. This corresponds to OD600 values in LB media of 2.62 ± 0.19 for P. aeruginosa RR1 and 2.2 ± 0.16 for P. oleovorans. Seven out of the twelve different culture conditions facilitated the growth of P. aeruginosa RR1 to levels above 1 × 109 cfu/ml, which is comparable to the dense growth observed for overnight cultures in LB media (Fig. 1). The five remaining conditions facilitate growth of to at least 5.9 × 108 cfu/ml (Fig. 1). P. oleovorans was more limited in the number of conditions that facilitated robust growth, as only two conditions supported growth above 1 × 109 cfu/ml and eight conditions showed cfu/ml below 1 × 108 (Fig. 1). The two best conditions both contained the lowest concentration of hexadecane tested (0.46% w/v) and NH4Cl as the nitrogen source (Fig. 1). P. oleovorans showed a preference for NH4Cl versus NH4NO3, as three conditions using NH4Cl supported growth above 1 × 108 cfu/ml but only one condition yielded such results when NH4NO3 was supplied (Fig. 2). With substantially higher cfu/ml across all but two culture conditions, P. aeruginosa RR1 was found to grow robustly on a wider range of culture conditions versus P. oleovorans (Fig. 1).

Fig. 1
figure 1

Average cfu/ml of P. aeruginosa RR1 and P. oleovorans 96 h after inoculation when using hexadecane as the sole carbon source (*18 h for LB media). Hexadecane (C16) was supplemented in the cultures at either 0.46%, 4.6%, or 9.2% (w/v). a Cfu/ml when 2.5 or 5 g/L of NH4Cl was supplied as the nitrogen source (2.5 Cl and 5 Cl). b Cfu/ml when 1.9 or 3.8 g/L of NH4NO3 was supplied as the nitrogen source (1.9 NO3 and 3.8 NO3). Error bars represent standard deviations from the mean values of three replicates

Fig. 2
figure 2

a Growth of P. aeruginosa RR1 RK2-GFPuv at 24 and 48 h post inoculation with and without kanamycin supplementation. b P. aeruginosa RR1 RK2-GFPuv plasmid maintenance at 24 and 48 h post inoculation in the presence of 50 µg/ml kanamycin. c Growth of P. oleovorans RK2-GFPuv at 24 and 48 h post inoculation with and without kanamycin supplementation. d P. oleovorans RK2-GFPuv plasmid maintenance at 24 and 48 h post inoculation in the presence of 50 µg/ml kanamycin. For both strains, culture conditions that yielded the best growth on hexadecane were used. Namely 0.46% w/v hexadecane with 1.9 g/L NH4NO3 for P. aeruginosa RR1 and 0.46% w/v hexadecane with 2.5 g/L NH4Cl for P. oleovorans. Error bars represent standard deviations from the mean values of three replicates

Growth kinetics for the two strains can be found in Additional file 1: Figures S2 and S3. For both Pseudomonas strains, the cfu/ml in cultures directly after inoculation with a plated colony was in the 105 range. At 24 h post inoculation, cfu/ml of P. aeruginosa RR1 was found to be in the 107 range, with an increase to the lower end of the 108 cfu/ml range by 48 h (Additional file 1: Fig. S2). Measurements at 72 h showed growth increases to the high end of the 108 cfu/ml range (Additional file 1: Fig. S2). At 96 h the cfu/ml in P. aeruginosa RR1 cultures was either maintained in the 108 cfu/ml range or had increased to above 1 × 109 cfu/ml (Figs. 1 and Additional file 1: Figure S2). The shape of the growth curves indicates that cells had reached either a late log phase or early stationary phase at 96 h (Additional file 1: Fig. S2). The growth kinetics of P. oleovorans showed more variability across culture conditions and required more time to reach the 107–108 cfu/ml range, versus P. aeruginosa RR1 (Figs. 1 and Additional file 1: Figure S3). At 24 h, P. oleovorans cfu/ml was observed to be in the low 106 range, with either no substantial increase or an increase to the high 106 range at 48 h (Additional file 1: Fig. S3). At 72 h, growth ranged from 6 × 106 to 7.4 × 108 cfu/ml depending on culture conditions (Additional file 1: Fig. S3). Likewise, at 96 h cultures ranged from 2.63 × 107 to 3.1 × 109 cfu/ml (Figs. 1 and Additional file 1: Figure S3). The shapes of the P. oleovorans growth curves indicate that the cultures may have been in a late log phase and shown additional growth over a longer time frame (Additional file 1: Fig. S3). Notwithstanding, this data shows that P. aeruginosa RR1 has more favorable growth kinetics (higher cfu/ml at 24 and 48 h) across the tested culture conditions versus P. oleovorans. These findings, combined with 1–2 order of magnitude differences in the final cfu/ml for the two species, lead to the conclusion that culture conditions must be specifically designed for a given species when an alkane is employed as the sole carbon source and growth trends observed in one species are likely not replicable in another.

Deleterious effects of antibiotic selection and plasmid loss in hexadecane media

Upon identifying conditions that yielded growth of P. aeruginosa RR1 and P. oleovorans above 1 × 109 cfu/ml with hexadecane as the carbon source, focus was turned to recombinant protein production. Green fluorescent protein (GFPuv) was chosen as a reporter protein to screen for proof-of-concept results before transiting to a recombinant silk construct. A plasmid-based approach was first tested, using vector pBb(RK2)1 k-GFPuv (RK2-GFPuv) [29]. The RK2-GFPuv vector was developed for use in Pseudomonas and contains an IPTG-inducible genetic circuit to produce a GFP mutant (GFPuv or eGFP) that has increased fluorescence over wild-type GFP for easier detection and quantification [29, 40]. The vector also contains a kanamycin resistance gene for selection [29]. The RK2-GFPuv plasmid was transformed into the Pseudomonas strains to create strains P. aeruginosa RR1 RK2-GFPuv and P oleovorans RK2-GFPuv. Production of GFPuv was first screened in rich LB media, with P. aeruginosa RR1 RK2-GFPuv and P. oleovorans RK2-GFPuv producing the recombinant protein at titers of 10.9 ± 2.7 and 12.4 ± 2.3 mg/L, respectively (Additional file 1: Fig. S4). From here, culture conditions that yielded the highest average cfu/ml at 96 h for strains grown on hexadecane were chosen for the plasmid-based conversion of hexadecane to a recombinant protein, namely, 0.46% w/v hexadecane with 1.9 g/L NH4NO3 for P. aeruginosa RR1 and 0.46% w/v hexadecane with 2.5 g/L NH4Cl for P. oleovorans.

However, upon addition of the antibiotic selection agent (kanamycin) to the hexadecane-based media, a severe inhibition of the growth of strains P. aeruginosa RR1 RK2-GFPuv and P oleovorans RK2-GFPuv was observed (Fig. 2a and c). At 24 h culture time, both P. aeruginosa RR1 RK2-GFPuv and P. oleovorans RK2-GFPuv showed approximately an order of magnitude decrease in cfu/ml when kanamycin was included in the media. By 48 h, there was a decrease in cfu/ml of approximately two orders of magnitude. Further examination showed that despite the presence of selective kanamycin, plasmid maintenance levels were found to be low in the system (Fig. 2b and d). Plasmid maintenance 24 h post inoculation was 9.3 ± 5.7% for P. aeruginosa RR1 RK2-GFPuv and decreased to 1.3 ± 0.9% by 48 h (Fig. 2b). Plasmid maintenance at 24 h post inoculation for P. oleovorans RK2-GFPuv was 12.0 ± 1.6% and decreased to 0.0 ± 0% by 48 h (Fig. 2d). This contrasts with cultures grown overnight (18 h) in LB media with kanamycin, which show plasmid maintenance levels of 79.3 ± 16.4% and 97.3 ± 0.9% for P. aeruginosa RR1 RK2-GFPuv and P. oleovorans RK2-GFPuv, respectively. The culture conditions for both strains were changed to 0.46% w/v hexadecane with 5 g/L NH4Cl (a condition that supported robust growth for both strains) to ascertain if the inhibited growth and plasmid maintenance could be alleviated by altering media composition. However, this made no substantial difference for growth or plasmid maintenance at 24 or 48 h post inoculation (Additional file 1: Figure S5).

It was hypothesized that the low copy number (3) of the RK2-GFPuv vector may cause difficulties with passage of the plasmid to daughter cells during growth in the hexadecane media [41]. Two new vectors, pBb(B5)1 k-GFPuv (B5-GFPuv) and pBb(RSF1010)1 k-GFPuv (RSF-GFPuv), were tested in P. aeruginosa RR1 (transformation of these vectors into P. oleovorans was unsuccessful) [29]. Plasmid B5-GFPuv has a copy number of 70, and RSF-GFPuv has a copy number of 30–60 in Pseudomonas, with both vectors containing an IPTG-inducible GPFuv gene and a kanamycin resistance gene [29, 42]. These increased copy numbers did not yield improved outcomes, as both vectors still showed only single digit percentages of plasmid maintenance 48 h post inoculation and there was no substantial improvement in culture cfu/ml at when selective kanamycin was present (Additional file 1: Fig. S6). Thus, it appears that the use of plasmids and antibiotic selection in Pseudomonas is not a feasible option for producing recombinant proteins in cultures with an alkane as a sole carbon source, independent of the culture conditions, specific species, or plasmid backbone.

Genomic integration of recombinant GFPuv gene

A strategy that excluded antibiotic selection and enabled higher maintenance of recombinant genes over a 96-h period was required to achieve measurable levels of recombinant protein production when Pseudomonas was grown on hexadecane. Genomic integration was identified as a method that would meet these criteria, and an IPTG-inducible recombinant GFPuv gene with a tac promoter was genomically integrated into P. aeruginosa RR1 to create the strain P. aeruginosa RR1 g-GFPuv. Integration of the GFPuv construct was confirmed with colony PCR (Additional file 1: Fig. S7). The integration method utilizes two suicide vectors for the integration of constructs at a single attTn7 site within the Pseudomonas genome [30]. Mutants are isolated from wild type cells via the co-integration of a gentamycin resistance gene and subsequent plating on LB-gentamicin plates [30]. However, gentamicin is not required to maintain the genomic integration, which is reported to be stable for at least 100 generations [30]. Despite the successful transformation of the RK2-GFPuv vector into P. oleovorans, a genomic mutant of this species was unable to be created as neither transformation nor conjugation techniques could introduce the genomic integration suicide vectors into P. oleovorans. Thus, all subsequent work was performed with genomic mutants of P. aeruginosa RR1, however, future work should investigate new genomic integration techniques to analyze the performance of a genomically altered P. oleovorans strain. Experiments were first performed with rich LB media to ensure that recombinant expression of the integrated GFPuv gene could be achieved without gentamicin selection. Single colonies of P. aeruginosa RR1 g-GFPuv taken from LB-gentamicin plates were used to inoculate flask cultures of LB media (without gentamicin) that contained either 0, 0.3, or 1 mM IPTG. After overnight incubation (18 h) at 37 °C cells were harvested and the presence of fluorescent GFPuv can be seen when the lysates are exposed to UV light (Additional file 1: Figure S8). Titers of GFPuv expressed from the genome were found to be 9.0 ± 1.6 mg/L with 0.3 mM IPTG and 9.9 ± 1.1 mg/L with 1 mM IPTG, which is similar to the GFPuv titer of 10.9 ± 2.73 mg/L when using the low-copy number RK2-GFPuv plasmid.

Conversion of hexadecane to recombinant GFPuv and spider silk protein

Although certain hexadecane culture conditions yielded more growth versus others for strain P. aeruginosa RR1, all tested conditions facilitated growth beyond 5 × 108 cfu/ml by 96 h and it was unknown if the best growth conditions would also result in the highest yields of recombinant protein. Likewise, the optimal time at which to induce cells with IPTG was unknown. Therefore, a fluorescence assay was developed to measure GFPuv titer directly from unpurified cell lysates to enable the testing of many conditions. Using 96 h total culture times, all twelve culture conditions that were tested for growth were also tested for recombinant GFPuv production. The twelve different culture conditions were each tested using three different expression conditions. They were induced with 0.3 mM IPTG at either 24-, 48-, or 72-h culture time (corresponding to 72-, 48-, and 24- hour expression times), thus, resulting in data for 36 different conditions for the conversion of hexadecane to recombinant GFPuv.

Production of GFPuv from hexadecane can be seen in Additional file 1: Figure S9, in which lysates of induced P. aeruginosa RR1 g-GFPuv show a green fluorescence under UV light as compared to pale blue fluorescence of uninduced cultures (the pale blue color is due to production of a natural pigment, pyocyanin) [43]. Figure 3 shows the GFPuv titer across all culture and expression conditions tested. The highest titer of 8.3 ± 0.5 mg/L was identified in culture conditions of 0.46% w/v hexadecane with 5 g/L NH4Cl that had been exposed to expression conditions of 0.3 mM IPTG added 24 h post inoculation (72-h expression time). This is approximately 92% of the GFPuv production level observed for P. aeruginosa RR1 g-GFPuv when grown in LB media with 0.3 mM IPTG. The second highest titers achieved for the conversion of hexadecane to recombinant GFPuv were approximately 25% less than this (Fig. 3). Therefore, 0.46% w/v hexadecane with 5 g/L NH4Cl was selected as the optimal culture condition for the conversion of hexadecane to a recombinant silk protein. Figures 3 and Additional file 1: Fig. S9 also show that the time of induction affected GFPuv titer, with substantial decreases in visible fluorescence for titers below 4 mg/L. IPTG added at 24 or 48 h post inoculation (longer expression times) yielded increased titers versus IPTG added at 72 h for ten of the twelve different media compositions (Fig. 3). Longer expression times may be favorable for titers due to the slow growth kinetics of this system in comparison to cultures grown in rich media (Fig. 1). Results also show that when comparing cultures with identical hexadecane concentration, elemental nitrogen concentration, and IPTG induction times, those containing NH4Cl achieved higher GFPuv titers versus those containing NH4NO3 in 15 out of 18 cases (Fig. 3). This finding is significant when considering that there is no significant difference in growth of P. aeruginosa RR1 on hexadecane when either NH4Cl or NH4NO3 is used (Figs. 1 and Additional file 1: Fig. S2). Counterintuitively, the single highest GFPuv levels achieved for both nitrogen sources used the lowest concentration of C16 (0.46% w/v) (Fig. 3). It is known that PHA synthesis in Pseudomonas is promoted by increasing the ratio of carbon to nitrogen in media, and this may shuttle cellular resources away from non-essential recombinant protein synthesis thus lowering titers [44, 45]. Future work using 13C-labeled hexadecane to do a metabolic-flux analysis may resolve differences in metabolism at various hexadecane concentrations and lead to a more developed understanding of this observation [46].

Fig. 3
figure 3

Titers of recombinant GFPuv when strain P. aeruginosa RR1 g-GFPuv is grown on hexadecane under a variety of culture conditions. a Cultures grown with either 2.5 or 5 g/L of NH4Cl (2.5 Cl and 5 Cl). b Cultures grown with either 1.9 or 3.8 g/L of NH4NO3 (1.9 NO3 and 3.8 NO3). Error bars represent standard deviations from the mean value of three replicates. 0.3 mM IPTG was added at 24, 48, or 72 h post inoculation. Total culturing time was 96 h, resulting in expression times of either 72, 48, or 24 h. Hexadecane concentrations (C16) are provided as % w/v

We have previously produced the de novo designed recombinant silk protein, A5 4mer, in E. coli and this construct was genomically integrated into P. aeruginosa RR1 to form P. aeruginosa RR1 g-A5 [47]. The A5 4mer primary sequence (Additional file 1: Fig. S1) was designed to represent a truncated version of dragline silk protein produced by spiders [47]. Expressions carried out in LB media demonstrated that P. aeruginosa RR1 g-A5 was able to produce the A5 4mer protein at a titer of 14.7 ± 1.57 mg/L when overnight cultures (18 h) inoculated from a single colony were grown at 37 ℃ in the presence of 1 mM IPTG. This also represented the first production of recombinant silk in the Pseudomonas genus. Unlike GFPuv, for which small amounts could be measured with fluorescence, a larger amount of A5 4mer protein was required for accurate measurements of titer. Therefore, larger culture volumes were employed for the conversion of hexadecane to recombinant silk. A single colony of P. aeruginosa RR1 g-A5 was used to inoculate a 25 ml starter culture of 0.46% w/v hexadecane with 5 g/L NH4Cl. After 72 h of culture time, the entire 25 ml starter culture was used to inoculate a 200 ml expression culture containing 0.46% w/v hexadecane with 5 g/L NH4Cl and supplemented with either 0.3 or 1 mM IPTG. The expression culture was then incubated for 48, 60, or 72 h.

Figure 4 shows A5 4mer silk protein purified from a culture of P. aeruginosa RR1 g-A5 grown using hexadecane as the sole carbon source and an expression time of 60 h. The purified protein can be seen at an identical molecular weight as standards of the protein produced in E. coli (Fig. 4). As previously documented, the 16 kDa A5 4mer silk protein appears at ~ 38 kDa due to its high level of structural disorder and subsequent aberrant mobility through SDS PAGE [47]. This represents the first conversion of an alkane to a recombinant silk protein or protein-based material. The highest titers of approximately 10 mg/L were achieved at 60- or 72-h expression time and 1 mM IPTG, with only a negligible difference between two expression times (Fig. 5). Interestingly, the similarity in titer between the 60- and 72-h mark coincided with a plateau of the wet pellet mass at harvest. The change in wet pellet mass from 48 to 60 h was + 1.9 g/L, but only + 0.3 g/L from 60 to 72 h (Fig. 5). Increasing the IPTG concentration from 0.3 mM to 1 mM increased the silk protein titer by approximately 41% and 36% for the 48- and 60/72-h expressions, respectively. The expression cultures were observed to be relatively clear directly after inoculation with the starter culture. However, progressive increases in turbidity were observed over time in accordance with increases in cfu/ml, with cultures becoming completely opaque at the time of harvest (Additional file 1: Fig. S10). Expression cultures developed pale orange color that progressed to a darker, burnt orange tone over time (Fig. S10).

Fig. 4
figure 4

SDS PAGE after nickel chromatography showing the A5 4mer silk protein purified from lysates of P. aeruginosa RR1 g-A5 that were grown using hexadecane as the sole carbon source (culture conditions of 0.46% w/v hexadecane with 5 g/L NH4Cl and induced with 1 mM IPTG). Expression time was 60 h. Lanes (1) Protein ladder, with mass listed to left in kDa (2) flow through (3) wash (4) first elution fraction (5) second elution fraction (6–9) standards of A5 4mer protein, 2, 1, 0.5, and 0.25 mg/L. The band in lane 4 (first elution fraction) is shown at an identical molecular weight (~ 38 kDa) as the standards of A5 4mer protein produced in E. coli (lanes 6–9). As previously documented, the 16 kDa A5 4mer silk protein appears at ~ 38 kDa due to its high level of structural disorder and subsequent aberrant mobility through SDS PAGE [47]. A slight smiling effect was observed on the SDS PAGE, resulting in a slightly slower migration of outer lanes (Lanes 1,2 and 5–8) compared to inner lanes (Lane 4) [48, 49]. The faint protein band to the right of Lane 4 is the remainder of histidine-tagged silk protein that was still bound to the purification resin after the Lane 4 elution

Fig. 5
figure 5

a Titer of the A5 4mer produced in P. aeruginosa RR1 g-A5 grown on hexadecane (culture conditions of 0.46% w/v hexadecane with 5 g/L NH4Cl). Expression times were either 48, 60, or 72 h and cultures were induced with either 0.3 or 1 mM IPTG. b Wet pellet mass at the end of A5 4mer expressions. Error bars represent standard deviations from the mean value of three replicates

Conversion of polyethylene-derived hydrocarbons to recombinant silk protein

Samples of depolymerized polyethylene were obtained to determine if our results could be extended to plastic-derived substrates. Polyethylene (Mw = 4,000 g/mol, Mn = 1,700 g/mol, and Đ = 2.35) was depolymerized via catalytic hydrogenolysis using a Pt/SrTiO3 catalyst in a batch reactor at 300 °C, in presence of 170 psi of H2 for 72 h under solvent-free conditions [50, 51]. This method confers control over the size and dispersity of the products and was leveraged to produce hydrocarbons within the range that P. aeruginosa can metabolize (approximately C8-C27) [22, 50, 51]. This is the first report to use such a method for the generation of a microbial feedstock from a plastic substrate. Previous reports have used methods such as pyrolysis or alkaline hydrolysis [14, 22]. It should be noted that unlike the monodisperse and linear hexadecane used previously herein, the chemically depolymerized polyethylene sample created in this study contains a distribution of alkanes (Mw = 305 g/mol, Mn = 240 g/mol, and Đ = 1.26) and a non-negligible degree of branching (Nbranch/1000C = 111). The sample is white and has the consistency of a soft wax (Additional file 1: Fig. S11). Several experiments were performed to establish robust growth of P. aeruginosa RR1 g-A5 when using this sample as the sole carbon source. For all growth experiments, the supplemental nitrogen was held constant at 5 g/L NH4Cl, as this provided the best GFPuv titers and facilitated the conversion of hexadecane to recombinant silk.

Figure 6a depicts the average cfu/ml 96 h after inoculation for P. aeruginosa RR1 g-A5 grown using the depolymerized polyethylene sample as the sole carbon source supplied at either 12.5 or 25 g/L. Preliminary experiments were performed by simply depositing centimeter-scale chunks of the depolymerized polyethylene sample into culture flasks, with cohesion of the sample resulting in a floating bolus (Fig. 6a “bolus”, Fig S12a). Growth of P. aeruginosa RR1 g-A5 using this method was suboptimal compared to results obtained with hexadecane, reaching cfu/ml levels several times below that of optimized conditions using hexadecane (Figs. 1 and 6a). It was hypothesized that increasing the surface area of the depolymerized polyethylene sample may allow the cells improved access to the substrate and facilitate increased growth. As such, multiple alternative morphologies of the depolymerized polyethylene sample within the culture flask were tested (Additional file 1: Fig. S12). A hot plate was used to melt the sample and coat the bottom of the culture flasks prior to the addition of media and cells. This resulted in coatings that contained no gaps in the coverage of the flask within their area (Additional file 1: Fig. S12b). Although the sample surface area was increased relative to a bolus, no improvements in growth over 96 h were observed (Fig. 6a “coated”). However, growth was substantially improved if the coating was manually disrupted with a sterile metal laboratory scoop to form an imperfect coating with gaps in the coverage of the flask (Fig. 6a “imperfect coating”, Fig. 6b, Additional file 1: Fig. S12d). This method yielded an average cfu/ml of 3.9 ± 0.6 and 4.2 ± 0.9 × 109 at 96 h post inoculation for 12.5 and 25 g/L of substrate, respectively. While functional, it should be acknowledged that the preparation of the “imperfect coating” via manual disruption with a laboratory scoop was not standardized and was thus subject to variability in terms of final morphology. Future work should seek to prepare hydrocarbon substrates into well-defined geometries with high surface area, such as micro or nano-scale particles. Notwithstanding, this imperfect coating method surpasses the highest levels of growth achieved with optimized conditions using hexadecane as the sole carbon source (Figs. 1 and 6a). To our knowledge, this is the highest reported level of cell growth achieved for a microbial strain grown on a polyethylene-derived, or alkane-based, substrate as the sole carbon source. A final sample preparation method was also tested, in which a bolus of the sample was melted on top of heated culture media placed in a flask (prior to inoculation). Upon cooling, the sample formed a solid, thin floating disk that could be broken into smaller pieces with a pipette tip or prolonged agitation in an incubator (Fig. 6a “floating disk”, Additional file 1: Fig. S12c). This method also yielded robust cfu/ml at 96 h, although slightly lower than an imperfect coating (Fig. 6a). Moreover, it was observed that pieces of the “floating disk” substrate tended to adhere to upper regions of the flask upon shaking incubation, effectively removing them from the culture media unless manually pushed down. For these reasons, the method of generating an imperfect coating of the substrate was chosen for silk protein expressions. Likewise, 12.5 g/L was chosen instead of 25 g/L due to similar growth between the two conditions and the ability to conserve the depolymerized polyethylene sample (Fig. 6a “imperfect coating”).

Fig. 6
figure 6

a Average cfu/ml 96 h post inoculation for P. aeruginosa RR1 g-A5 grown using depolymerized polyethylene as the sole carbon source. Multiple concentrations (12.5 and 25 g/L) and morphologies of the depolymerized polyethylene were tested, with 5 g/L NH4Cl supplied as the nitrogen source. Error bars represent standard deviations from the mean values of three replicates. b Depiction of the “Imperfect Coating” preparation method. Samples of the depolymerization polyethylene were added to flasks and gently melted to form a coating on the bottom. Upon solidification, the coatings were agitated to create an imperfect coating coverage. Sections of coating that were removed were added back into the flask as small boluses to maintain consistency in the total amount of sample per liter of culture

An expression strategy derived from results using hexadecane was applied for the conversion of polyethylene to recombinant silk. A 25 ml starter culture was grown for 72 h, and the entire culture was used to inoculate 200 ml of fresh media (with 1 mM IPTG) for a 72-h expression period. Both the starter cultures and expression cultures contained 12.5 g/L of depolymerized polyethylene prepared using the imperfect coating method. Akin to expressions using hexadecane, cultures utilizing depolymerized polyethylene were clear upon inoculation and increased in turbidity over time in accordance with cell growth, resulting in extremely turbid cultures at harvest (Additional file 1: Fig. S13). At harvest, however, expression cultures utilizing depolymerized polyethylene were observed to be a pale brown color versus the burnt orange observed when hexadecane was used (Additional file 1: Fig. S13d). Nickel chromatography on lysates of expression cultures of P. aeruginosa RR1 g-A5 grown using depolymerized polyethylene as the sole carbon source showed the A5 4mer protein in the elution fraction, (Additional file 1: Fig. S14) thus demonstrating the conversion of polyethylene to a recombinant protein-based polymer material. Titers of the silk protein were 11.3 ± 1.1 mg/L, which is approximately 77% of the production level observed in rich LB media for strain P. aeruginosa RR1 g-A5. Samples of the A5 4mer spider silk protein produced via the microbial conversion of depolymerized polyethylene were dialyzed into water and lyophilized, resulting in white, cohesive biopolymer material (Fig. 7).

Fig. 7
figure 7

Purified A5 4mer spider silk protein produced through the microbial conversion of depolymerized polyethylene. The silk protein was dialyzed into water and lyophilized, resulting in several discrete and cohesive chunks of a white, polymeric material that ranged in length from approximately 1.5–4 cm. This image was obtained directly after the freeze-drying process, and the protein sample had not been spun into fibers or put through any additional material processing

In conclusion, this work demonstrates the first reported microbial platform for the conversion of polyethylene-derived substrates to a bespoke recombinant protein. Our results also demonstrate the first recombinant silk production in the Pseudomonas genus. Additionally, culture conditions and specific polyethylene depolymerization parameters that yield levels of microbial growth using a plastic-derived substrate as the sole carbon source are identified. Although a plasmid-based expression system was inefficient, genomic integration of recombinant genes was found to be a viable strategy for facilitating the microbial conversion of depolymerized polyethylene into recombinant proteins. Taken together, these findings demonstrate the modularity this novel system, which can potentially be used to upcycled waste polyolefins into any recombinant protein that can be integrated into the genome of P. aeruginosa RR1.

Future work will seek to increase production levels of recombinant constructs through optimization of the recombinant gene. In this context, multi-copy integrations of heterologous genes may increase production levels compared to the single copy integrations reported in this work. Likewise, the use of a stronger promoter on recombinant integrations (such as T7) may yield increased titers compared to the tac promoter used in this work. Additionally, future work may seek to decrease the time required for P. aeruginosa RR1 to reach > 1 × 109 cfu/ml when depolymerized polyethylene is the sole carbon source, thus increasing the efficiency and decreasing the cost of the process. Within this scope, fundamental work that can identify metabolic bottlenecks when Pseudomonas is grown on alkane-based substrates by way of metabolic flux analysis (using 13C-labeled feedstocks) and quantify the amount of substrate consumed by the cells and its conversion rate to recombinant protein are also important aspects of future work [46]. Potential strategies to increase the rate of growth include the testing of additional morphologies of the depolymerized polyethylene sample or upregulation of endogenous alkane hydroxylases within Pseuodmonas [28]. Furthermore, the use of a depolymerization process that yields oxidized hydrocarbon substrates of increased solubility may be advantageous by promoting faster uptake of the carbon source into the cells. Oxidized substrates can also bypass the first, and possibly rate-limiting, step of terminal oxidation that occurs in the alkane degradation pathway used by Pseudomonas [52, 53]. Since P. aeruginosa is an opportunistic pathogen, future work should also seek to implement non-pathogenic strains of P. aeruginosa to promote the optimization and scale-up of this process by ensuring a high degree of safety [54]. Recent work shas hown that the virulence of P. aeruginosa can be attenuated by deleting five chromosomal genes (toxA, plcH, phzM, wapR, and aroA) through a plasmid-based methods [55]. This resultant strain exhibits a mortality rate of 0% in mice upon injection, as compared to a 95% mortality rate of wild-type P. aeruginosa [55]. However, it is unknown if these genetic changes will impact the efficiency of cell growth and recombinant protein production when hydrocarbon-rich substrates are used as the sole carbon source.



Source link