Biomacromolecules 2001, 2, 211-216
211
Glucose/Lipid Mixed Substrates as a Means of Controlling the Properties of Medium Chain Length Poly(hydroxyalkanoates)† Richard D. Ashby,* Daniel K. Y. Solaiman, Thomas A. Foglia, and Cheng-Kung Liu United States Department of Agriculture, Agricultural Research Service, Eastern Regional Research Center, 600 E. Mermaid Lane, Wyndmoor, Pennsylvania 19038 Received September 27, 2000; Revised Manuscript Received January 30, 2001
Glucose-triacylglycerol (TAG) mixed substrates were used to modulate the physical and mechanical properties of medium-chain-length poly(hydroxyalkanoates) (mcl-PHAs). Pseudomonas resinoVorans NRRL B-2649 grew and produced mcl-PHAs on glucose and TAGs (coconut oil, C; soybean oil, S) after 24 h in a shake flask culture. However, with the exception of coconut oil, maximum cell productivity was not reached in any of the cultures until 72 h post-inoculation. Here, 50:50 mixtures of glucose and coconut oil (glc/C) or glucose and soybean oil (glc/S) resulted in intermediate cell productivities with a maximum of 57% and 48% of the CDW at 72 h, respectively. In addition, mixed substrates resulted in mcl-PHAs with compositions that varied slightly over time. PHA-glc/C and PHA-glc/S were composed of 7 mol % and 8 mol % 3-hydroxydodecenoic acid (C12:1), respectively at 72 h. These concentrations were intermediate to the C12:1 concentration of PHA-glc and respective PHA-TAG. Also, significant amounts of 3-hydroxytetradecanoic acid (C14:0), 3-hydroxytetradecenoic acid (C14:1), and 3-hydroxytetradecadienoic acid (C14:2) were present in PHA-glc/C and PHA-glc/S, which were derived from the respective TAG, as glucose resulted in almost no C14:X monomers. The molar masses of each of the polymers remained relatively constant between 24 and 96 h. At 72 h, the number-average molar masses (Mn) of PHA-glc/C and PHA-glc/S were 178 000 and 163 000 g/mol, respectively, which were also intermediate to the Mn of PHA-glc (225 000 g/mol) and the respective PHA-TAG (PHA-C ) 153 000 g/mol; PHA-S ) 75 000 g/mol). These physical differences caused variations in the mechanical properties of mcl-PHA films, thus providing a new and effective method of modifying their properties. Introduction Poly(hydroxyalkanoates) (PHAs) are naturally occurring polyesters produced by numerous bacteria as carbon and energy reserve materials.1,2 Presently, there are over 100 different structural analogues of PHA whose composition is dependent upon the carbon substrate and the enzymatic specificity of the producing organism.3 Each repeat unit conforms to the [R]-stereochemical configuration resulting in isotactic polyesters.4 This stereoregularity enhances the crystallization of PHAs with short alkyl side chains (sclPHAs), such as poly(3-hydroxybutyrate) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate), while PHAs with longer side chains (mcl-PHA), which are composed of 3-hydroxy acid monomer units ranging in length from C6 to C14, tend to be more amorphous. Presently, there is an industrial push to find replacements for many petrochemical-based polymers because of their negative environmental impact. As such, biodegradable polymers are attractive as substitutes for many synthetic materials provided they can be produced in applicable quantities and exhibit properties that are at least equal to * Corresponding author. Telephone: (215) 233-6483. Fax: (215) 2336795. E-mail address:
[email protected]. † Mention of brand or firm name does not constitute an endorsement by the U.S. Department of Agriculture over others of a similar nature not mentioned. 10.1021/bm000098+
their synthetic counterparts. The fact that PHAs are biodegradable with numerous conformational variations makes them a natural target for a more environmentally benign polymer. Many species of Pseudomonas belonging to the rRNA homology group I accumulate mcl-PHA from free fatty acids (Pseudomonas oleoVorans, Pseudomonas putida),5-8 triacylglycerols (TAGs) (Pseudomonas resinoVorans)9,10 and sugars (P. putida)11 when grown under nutrient limited conditions. Many of these mcl-PHAs contain unsaturated hydrocarbon side chains, which help dictate the properties of the polymers. These olefinic groups have been used to enhance the polymer properties through cross-linking by both chemical (peroxides, sulfur vulcanization, epoxidation and aging) and physical (radiation) means.12-17 While both methods are effective, they each have inherent flaws that make them undesirable. Chemicals introduce unwanted materials into the system while radiation techniques tend to degrade as well as cross-link the polymers. In this study P. resinoVorans was used to produce mclPHA from glucose/TAG mixed substrates in an attempt to modulate the polymer properties without postsynthetic treatments. It was found that the organism could metabolize both glucose and TAGs to produce mcl-PHAs with unique physical properties.
This article not subject to U.S. Copyright. Published 2001 by the American Chemical Society Published on Web 02/21/2001
212
Biomacromolecules, Vol. 2, No. 1, 2001
Materials and Methods Materials. All simple salts, glucose, coconut oil, and soybean oil were obtained from Sigma Chemical Co. (St. Louis, MO). All organic solvents were HPLC grade and purchased from Burdick and Jackson (Muskegon, MI). The silylation reagent was N,O-bis(trimethylsilyl)trifluoroacetimide (BSTFA) and was purchased from Aldrich Chemical Co. (Milwaukee, WI). Strain Information and Polymer Synthesis. P. resinoVorans NRRL B-2649 was obtained from the NCAUR, ARS, United States Department of Agriculture, Peoria IL, and used as the producer strain in all experiments. Stock cultures were prepared by aseptically scraping an individual colony of P. resinoVorans NRRL B-2649 off of a nutrient agar plate and inoculating it into 50 mL of sterilized nutrient broth. This stock culture was then allowed to incubate at 30 °C overnight with shaking at 250 rpm. Fifty milliliters of 20% (v/v) glycerol was aseptically added to the bacterial cells and the cells placed into cryovials for storage at -70 °C until needed as the inocula for the polymer production (shake flask) experiments. Shake flask experiments were conducted in 500 mL volumes in Medium E* (pH 7.0) (for medium composition see ref 5). Glucose was added into sterile Medium E* from a 20% (w/v) filter-sterilized stock solution to a concentration of 1% (w/v). Filter-sterilized coconut oil was melted at 55 °C for 15 min and added to the Medium E*, while filter-sterilized soybean oil was brought to room temperature and added directly. Each TAG was added at a concentration of 1% (w/v). The mixed substrate flasks contained 0.5% (w/v) glucose and 0.5% (w/v) TAG for a final substrate concentration of 1% (w/v). The inoculum (0.5 mL) for each shake flask came directly from the frozen stock cultures described above. Each shake flask was incubated at 30 °C with shaking at 250 rpm for 24, 36, 48, 72, or 96 h. At the appropriate time the cells were pelleted by centrifugation (8000g, 20 min, 4 °C), washed twice in deionized water and lyophilized (∼24 h) to a constant weight. Carbon Source Utilization. Both glucose and the TAGs were monitored for preferential utilization by the bacterium in the mixed substrate cultures. Glucose levels were monitored by using the Glucose (HK) 20 Detection Kit (Sigma Diagnostics, St. Louis, MO) according to the manufacturers specifications. In short, 1 mL of glucose detection reagent was pipetted into a clean test tube followed by 0.01 mL of deionized water (blank) or culture supernatant. The tubes were gently agitated. After 5 min the absorbance of each sample was read at 340 nm and the residual glucose was expressed in percent. TAG utilization was monitored by a modified Environmental Protection Agency (EPA) method, Method 1664 for the Determination of Oil and Grease and Nonpolar Material in EPA’s Wastewater and Hazardous Waste Programs. This method is a gravimetric procedure where the culture supernatants (approximately 500 mL) containing the residual TAGs were washed 3 times with 75 mL of hexane (total 225 mL hexane). The hexane layer (containing the residual TAGs) was separated from the aqueous phase through a 1 L separatory funnel, placed into a 500 mL tared round-bottom flask and rotoevaporated to remove the hexane. The flask
Ashby et al.
was then reweighed and the residual TAG expressed as a percent. The 100% recovery of both glucose and TAGs was determined in duplicate by harvesting two uninoculated flasks at time 0 and carrying out both the glucose and TAG recovery assays to completion. PHA Purification and Isolation. Supercritical fluid extraction (SFE) was used to remove any residual TAG from the bacterial cells prior to polymer isolation. The dried bacterial cells (5 g per extraction) were packed tightly into a 24 mL stainless steel extraction vessel (rated at 680 atm). The extraction vessel was placed in the SFE apparatus oven (Applied Separations, Allentown, PA). The oven was adjusted to 60 °C and the restrictor valve temperature was set at 100 °C. When the vessel temperature reached 60 °C, carbon dioxide was allowed to flow through the extraction vessel at 1.5 L/min at 340 atm (1 h), 476 atm (1 h), and 612 atm (1 h), consecutively. The bacterial cells were removed from the extraction vessel and the PHA isolated by chloroform extraction.18 Instrumental Procedures. Each mcl-PHA was characterized with respect to repeat unit composition, molar mass and thermal properties. PHA repeat unit compositions were determined by gas chromatography/mass spectrometry (GC/ MS) of the silylated 3-hydroxymethyl esters. Samples were prepared according to Brandl et al.5 and were silylated by reacting 10 µL of each sample with 250 µL of BSTFA and 200 µL of pyridine. The mixtures were heated at 70 °C for 30 min and allowed to cool to room temperature. Finally, 150 µL of hexane was added to each sample, and the samples were analyzed by GC/MS as described elsewhere.17 Percent composition was obtained by selecting the 175 ion, indicative of silylated 3-hydroxymethyl esters, and identifying the molecular ion - 15 (CH3 group) (M - 15).19 Molar mass averages were determined by gel permeation chromatography (GPC) as described elsewhere.10 Styragel HR1, HR3, HR4, and HR6 columns (Waters Corp. Milford, MA) were connected in series and polystyrene standards (Polyscience, Warrington, PA) with narrow polydispersities were used to generate a calibration curve. Chloroform was used as the eluent at a flow rate of 1 mL/min. The sample concentration and injection volumes were 0.3% (w/v) and 200 µL, respectively. Thermal properties were measured for each PHA sample using a Pyris 1 differential scanning calorimeter (PerkinElmer, Norwalk, CT) at a heating rate of 10 °C/min with a dry nitrogen purge. The instrument was calibrated using both indium (Tm ) 156 °C) and cyclohexane (transition temperatures at -87 and +6 °C). The Tg was taken as the midpoint temperature and the Tm as the peak of the melting endotherm. Tensile properties of PHA film specimens (7 cm × 5 mm × 0.1 mm) were measured at 23 °C and 50% relative humidity with a gauge length of 25 mm on an Instron tensile testing machine, model 1122 (Canton, MA) as described elsewhere.17 Measurements included tensile strength, percent elongation, and Young’s modulus. All the data were collected and analyzed by TestWorks version 3.10 software, developed by Material Testing Systems Corp. (Eden Prairie, MN). Film Preparation. Films were cast from solutions of PHA-glucose (glc), PHA-TAG, and PHA-glc/TAG. Solu-
Mixed-Substrate PHA Property Control
Biomacromolecules, Vol. 2, No. 1, 2001 213
tions were prepared by dissolving 1.5 g of purified PHA in 15 mL of chloroform. Films were cast in glass Petri dishes (100 mm × 15 mm), the solvent evaporated in a nitrogen atmosphere, and the films dried under vacuum at 20 °C for 7 h. The mechanical properties of the films were measured immediately. Results and Discussion An organism’s ability to synthesize mcl-PHAs is dependent upon the enzymatic synthesis of (R)-3-hydroxyacyl CoA intermediates from the starting substrate. Some strains of Pseudomonas can use simple sugars to form these compounds through fatty acid biosynthesis, while others β-oxidize free fatty acids to form them. This complexity led us to believe that the simultaneous metabolism of glucose and a TAG could be realized by some strains of Pseudomonas to produce mcl-PHAs with modified properties. We chose P. resinoVorans NRRL B-2649 for our study because it is one of only a few known bacteria that can metabolize intact TAGs to make mcl-PHA.9,10 The organism utilizes a cellassociated esterase (lipase) to cleave the fatty acids from the glycerol backbone and then β-oxidizes the fatty acids for growth and mcl-PHA production.10 Our results showed that P. resinoVorans was able to grow and produce mcl-PHA from glucose (PHA-glc) and TAG (coconut oil, PHA-C, and soybean oil, PHA-S) individually and from a 50:50 mixture of glucose and TAG (PHA-glc/C or PHA-glc/S). These findings contradict previous information, which reported that P. resinoVorans ATCC 14235 grew but did not produce mcl-PHA from glucose.8 In our experience using Medium E*, both strains of P. resinoVorans (ATCC 14235 and NRRL B-2649) grew and produced mcl-PHA from glucose and TAGs equally well. The cell growth and polymer yield from glucose increased during the first 36 h in shake flask culture, reached a maximum of 2.5 ( 0.1 g/L and 1.1 ( 0.1 g/L, respectively, by 48 h and remained relatively constant through 96 h (Figure 1). In contrast, the cell growth in the coconut oil and soybean oil experiments continued to increase up to 6.6 ( 0.3 and 5.5 ( 0.2 g/L (polymer yields ) 4.6 ( 0.3 and 3.1 ( 0.2 g/L), respectively at 96 h. The mixed substrate cultures resulted in cell growth and polymer yields that were comparable to those for the glucose experiments. However, the cell productivity of the glc/C and glc/S cultures reached a maximum of 57% and 48% of the cell dry weight (CDW) at 72 h, intermediate to the glucose (cell productivity ) 45% of the CDW) and TAG (coconut oil, C, cell productivity ) 65% of the CDW; soybean oil, S, cell productivity ) 60% of the CDW) experiments. The generation of metabolic intermediates from alkanoic acids depends on the organism’s ability to utilize the glyoxylate shunt to bypass the irreversible pyruvate dehydrogenase complex.20 If this were not so, both carbon atoms that constitute acetyl CoA (the β-oxidation product) would be lost as CO2, thus inhibiting the tricarboxylic acid cycle (TCA cycle). The high cell growth and polymer yield in the coconut oil and soybean oil cultures indicated that P. resinoVorans efficiently utilizes β-oxidation (including the glyoxylate shunt) when grown on TAGs alone.
Figure 1. Bacterial cell growth (A), polymer yield (B), and cell productivity (C) of P. resinovorans NRRL B-2649 grown on glucose, triacylglycerols (coconut oil, soybean oil), and mixed substrates (50: 50 mixtures of glucose/coconut oil or glucose/soybean oil).
However, in the presence of the mixed substrate glucose can be more easily utilized for cell growth because it results in more efficient energy yields by alleviating the energy intensive reactions necessary for the formation of metabolic intermediates from fatty acid oxidation. To more precisely determine the preferential use of glucose or TAG in the mixed substrate cultures, the utilization of both glucose and TAG was monitored during cell growth (Figure 2) by measuring each residual carbon source in the supernatant after harvesting the biomass. In both cases (glc/C and glc/ S) glucose was more readily utilized initially (within the first
214
Biomacromolecules, Vol. 2, No. 1, 2001
Figure 2. Recovery of glucose and coconut oil (A) and glucose and soybean oil (B) in the mixed substrate shake flask cultures over time.
12 h). However, no polymer could be isolated at this time, probably due to low cell numbers and the lack of a stressful environment to induce PHA production. This indicated that glucose was a more favorable growth substrate, however the loss of TAG in the culture supernatants between 24 and 96 h, and the variation in PHA composition of the polymers showed that coconut oil and soybean oil were also being utilized by the bacterium. In addition, as the duration of the cultures increased, the slopes of the lines representing recoverable carbon substrates equalized. This was further proof that the rate of carbon source utilization became more equal after the initial 12 h and hinted at the formation of copolymers rather than polymer blends, which theoretically would have caused a greater discrepancy in the slopes of the lines. In addition, these results pointed to glucose as the primary carbon source for cell growth, while both glucose and TAG were utilized for mcl-PHA biosynthesis in the mixed substrate cultures beginning between 12 and 24 h postinoculation. Further proof of the use of both substrates in mcl-PHA biosynthesis was seen in the polymer compositions (Table 1). As expected, all of the mcl-PHAs contained appreciable concentrations of 3-hydroxyoctanoic acid (C8:0) and 3-hydroxydecanoic acid (C10:0). Interestingly, the composition of the PHA-glc remained constant throughout the duration of the shake flask experiments. In contrast, the PHA-TAG polymers showed some compositional variation over time. Specifically, the C8:0 concentration decreased while the C10:0
Ashby et al.
concentration increased as the duration of the shake flask experiments increased. The most telling compositional differences between the mixed substrate polymers (MSPs) were in the concentrations of 3-hydroxydodeca(e)noic acid (C12: X), 3-hydroxytetradecanoic acid (C14:0), 3-hydroxytetradecenoic acid (C14:1) and 3-hydroxytetradecadienoic acid (C14: 2). PHA-glc/C and PHA-glc/S were found to contain appreciable concentrations of C14:X monomers. Coconut oil is composed of greater than 50% lauric acid and is primarily a saturated oil, containing only 5% oleic acid (C18:1∆9) and 1% linoleic acid (C18:2∆9,12), while soybean oil contains 21% oleic acid and 57% linoleic acid. The presence of C14:X monomers were unique to PHA-C and PHA-S, indicating that at least a portion of the 3-hydroxyalkanoic acids making up the MSPs from 24 to 96 h originated from TAG as glucose metabolism resulted in
