Production of medium‐chain‐length PHA in octanoate‐fed enrichments dominated by Sphaerotilus sp.

Medium‐chain‐length polyhydroxyalkanoate (mcl‐PHA) production by using microbial enrichments is a promising but largely unexplored approach to obtain elastomeric biomaterials from secondary resources. In this study, several enrichment strategies were tested to select a community with a high mcl‐PHA storage capacity when feeding octanoate. On the basis of analysis of the metabolic pathways, the hypothesis was formulated that mcl‐PHA production is more favorable under oxygen‐limited conditions than short‐chain‐length PHA (scl‐PHA). This hypothesis was confirmed by bioreactor experiments showing that oxygen limitation during the PHA accumulation experiments resulted in a higher fraction of mcl‐PHA over scl‐PHA (i.e., a PHA content of 76 wt% with an mcl fraction of 0.79 with oxygen limitation, compared to a PHA content of 72 wt% with an mcl‐fraction of 0.62 without oxygen limitation). Physicochemical analysis revealed that the extracted PHA could be separated efficiently into a hydroxybutyrate‐rich fraction with a higher Mw and a hydroxyhexanoate/hydroxyoctanoate‐rich fraction with a lower Mw. The ratio between the two fractions could be adjusted by changing the environmental conditions, such as oxygen availability and pH. Almost all enrichments were dominated by Sphaerotilus sp. This is the first scientific report that links this genus to mcl‐PHA production, demonstrating that microbial enrichments can be a powerful tool to explore mcl‐PHA biodiversity and to discover novel industrially relevant strains.

PHA can potentially be produced cost-effectively by using microbial enrichments cultures and organic waste streams as feedstock. This approach diminishes the relatively large expenses for sterilization and raw substrates as required for pure culture production (Kleerebezem & van Loosdrecht, 2007), and avoids part of the waste disposal costs (Fernández-Dacosta et al., 2015). To date, at least 19 pilot studies have been conducted, using municipal or industrial organic waste streams as feedstock (Estévez-Alonso et al., 2021). In all these studies, the predominant type of PHA produced was the copolymer poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV).
For thermoplastic applications, the mechanical properties of PHBV are already superior to the homopolymer, polyhydroxybutryate (PHB) (Jain & Tiwari, 2015). However, incorporating mediumchain-length PHA (mcl-PHA) monomers in the polymer is required for elastomeric applications. Mcl-PHA consists of larger monomers ranging from 6 to 14 carbons, while short-chain-length PHA (scl-PHA) consists of monomers with 3-5 carbons. Increasing the mcl-PHA monomer composition in the polymer will result in a lower melting temperature, a lower degree of crystallinity, a lower tensile strength, and a higher elongation to break (Anjum et al., 2016;Rai et al., 2011). The production of these elastomers can expand the product utilization spectrum of PHA in the future, thereby potentially targeting rubber-like materials or adhesives (Elbahloul & Steinbüchel, 2009;Pereira et al., 2019).
The microbial metabolism to produce mcl-PHA is much less widespread than the metabolism for scl-PHA. In general, mcl-PHA production is linked to a special class of PHA synthases (PhaC Class II) possessed by a small number of organisms such as the well-studied Pseudomonas sp. or Comamonas sp. (Rai et al., 2011;Tortajada et al., 2013). However, some studies assert that in rare cases other classes of PHA synthases (Classes I and III) can incorporate hydroxyhexanoate (HHx) or hydroxyoctanoate (HO) monomers in the polymer (Quillaguamán et al., 2010;Sudesh et al., 2000). Because mcl-PHA metabolism appears to be sparse in the microbial world and because the available knowledge is incomplete, applying enrichment techniques offers an adequate approach to explore microbial diversity (Stouten, 2019).
The vast majority of the mcl-PHA research has been conducted with pure cultures. Nevertheless, a small number of laboratory studies focused on using microbial enrichments with feast-famine conditions. This intermittent substrate feeding strategy generates a competitive advantage for bacteria that store PHA as carbon and electron reservoir inside their cell. A proof-of-principle was already established with C8, C9, C12, and C18 fatty acids (Alaux et al., 2022;Chen et al., 2018;Lee et al., 2011;Shen et al., 2015), where it was found that medium-chain fatty acids form a suitable substrate for the enrichment of mcl-PHA producers. The mcl-PHA fractions of the total PHA content ranged from 0.06 to 0.88. However, the maximal obtained PHA contents were still rather low (23-49 wt%) compared to scl-PHA enrichment studies (Johnson, Jiang, et al., 2009;Kourmentza et al., 2017). In addition, an approach to understand and control the ratio between scl-PHA and mcl-PHA is still lacking.
PHA production from organic waste streams typically starts with the anaerobic fermentation of the feedstock where volatile fatty acids (VFAs) are formed, ranging from acetate to hexanoate Silva et al., 2022). Via a chain elongation step, the carbon chain of these VFAs can be extended to synthesize octanoate or more hexanoate Kucek et al., 2016). Although hexanoate is industrially more relevant than octanoate, in this research, octanoate was used as the sole carbon source because octanoate is regarded as a substrate that results in one of the highest mcl-PHA productivities (Li et al., 2021;Sun et al., 2007). Therefore, the aim of this research is to study which enrichment strategies result in a community with a high mcl-PHA storage capacity when feeding octanoate.
To this end, sequencing batch reactors (SBRs) were operated with synthetic octanoate as substrate in a mineral medium. Analysis of the PHA metabolic pathways revealed that the production of mcl-PHA requires less oxygen respiration than scl-PHA production. This resulted in the hypothesis that mcl-PHA production has a competitive advantage over scl-PHA production in an oxygen-limited environment. Therefore, the dissolved oxygen (DO) concentration was varied to understand and modify the enrichment process in terms of PHA content and composition. In addition, the pH of the SBRs was varied to study the effect of the toxicity of undissociated octanoic acid. For every operational steady state, the performance of the enrichments was characterized, including the maximum PHA storage capacity by means of accumulation experiments. An existing metabolic and kinetic model was adapted towards octanoate to derive the stoichiometric and kinetic parameters of all experiments . Finally, physicochemical polymer properties were determined to reveal information about the microbial origin of the produced polymer.

| Enrichment in SBRs
Two double-jacket glass bioreactors with a working volume of 1.4 L (Applikon Biotechnology) were operated for the enrichment of a PHA-storing microbial culture on octanoate. The setup and operation of these bioreactors were based on the conditions described by Johnson, Jiang et al. (2009). The bioreactors were operated as nonsterile SBRs, subjected to a feast-famine regime with a cycle length of 12 h and solids retention time (SRT) and hydraulic retention time of 24 h, which implies that every cycle 50% of the SBR volume is replaced with fresh medium. The inoculum of the SBRs was aerobicactivated sludge of a wastewater treatment plant (WWTP) (Harnaschpolder Delfluent). Furthermore, the airflow rate to the bioreactors was set to 0.2 L N /min by means of a mass flow controller (MX4/4; DASGIP ® ), and the stirring speed was set to 800 rpm (TC4SC4; DASGIP ® ). The temperature in the bioreactor was controlled at 30 ± 0.5°C with the water jacket around the bioreactor and an external thermostat bath (ECO RE 630S). The pumps for feeding, effluent removal, and pH control, the stirrer, and the airflow were controlled by a hardware abstraction layer (TU Delft), which, in turn, was controlled by a PC using a custom scheduling software (D2I; TU Delft). The D2I was also used for data acquisition of the online measurements: DO, pH, temperature, acid and base dosage, in-and off-gas composition, and feed/water balances. The addition of nutrients and carbon sources at the beginning of each cycle was followed by a phase of a high oxygen uptake rate, of which the duration and the magnitude were extracted from the off-gas data.
Moreover, the bioreactors were cleaned about twice per week to remove biofilms from the glass walls and the sensors of the bioreactor. The medium consisted of a separate carbon and nutrient source. The carbon source concentration in the SBR was 4.75 mM octanoate. The nutrients concentrations in the SBR were composed of 6.74 mM NH 4 Cl, 2.49 mM KH 2 PO 4 , 0.55 mM MgSO 4 ·ּH 2 O, 0.72 mM KCl, 1.5 ml/L trace elements solution according to Vishniac and Santer (1957), and 5 mg/L allylthiourea (to prevent nitrification).
To characterize the operational performances, the SBRs were subjected to multiple cycle analysis experiments.

| PHA accumulation in fed-batch bioreactor
The PHA accumulation experiments were performed in the same bioreactors as the enrichment but operated in fed-batch mode. The pH, temperature, and aeration rate were copied from the corresponding SBR. Half of the content of the SBR (700 ml) of the final cycle was used as seeding material for the accumulation experiment. In addition, 700 ml medium was supplied as described above, but without ammonium and carbon source. After 30 min, to ensure a temperature of 30°C, a pulse of 6.7 mmol octanoates was supplied to each bioreactor. To prevent carbon source depletion throughout the PHA accumulation, octanoic acid (undiluted) and NaOH (1 M) were used to control the pH. In addition, pulses of 0.67 mmol octanoates were supplied every 2 h from t = 4 h to the end of the experiment. One accumulation was performed by replacing octanoate and octanoic acid with hexanoate and hexanoic acid. Nitrogen was limited during most of the accumulation since no nitrogen source was supplied to the bioreactors and only a small amount (<1.7 mM of NH 4 + ) remained from the previous SBR cycle.
In this way, growth in the fed-batch bioreactor was limited. If necessary, a few drops of (10X diluted) Antifoam C (Sigma-Aldrich) were added to inhibit the formation of foam. The experiments were terminated after 24 h.

| PHA extraction and fractionation
During the accumulation experiments, 50 ml cell suspension was collected from the bioreactor after 12 h. After centrifuging and freezedrying, ±350 mg of the dried biomass was mixed with 12.5 ml chloroform.
The suspension was incubated for 3 h at 60°C while manually shaken every 30 min. Then, the suspension was filtered (0.45 µm filter). Lastly, the chloroform was evaporated in a fume hood to obtain purified PHA.
Part of the purified PHA from the accumulation at pH 8 (A8) was fractionated by selective precipitation. To this end, the PHA was divided into five parts of ±20 mg and redissolved in 0.8 ml chloroform. Different volumes of antisolvent (1-heptane) were added (0.9, 1.1, 1.4, 1.6, and 15).
After 15 min incubation, the tubes were centrifuged and the supernatant was decanted into another tube. Only the tube with 15 volumes of 1-heptane was incubated for 48 h. Then, the solvent-antisolvent mixture in all tubes was evaporated in a fume hood. The obtained PHA fractions were analyzed by gas chromatography (GC).

| Analytical methods
The performance of the cycle and accumulation experiments were characterized by online measurements (DO, pH, acid/base dosage, and in-/off-gas composition) with the equipment and software described above, and with offline samples (VFAs, ammonium, PHA, and total and volatile suspended solids). The composition of the active biomass was assumed to be CH 1.8 O 0.5 N 0.2 (Beun et al., 2002).
A detailed description of the analytical methods is given by Johnson, Jiang et al. (2009). A modification has been made to the ammonium measurement. These samples were measured with a Gallery™ Plus Discrete Analyzer (Thermo-Fisher Scientific).
The method to analyze the PHA composition of the biomass by GC was also modified to include mcl-PHA. In brief, the PHA in the biomass was hydrolyzed and esterified in the presence of concentrated acid, propanol, and dichloroethane with a ratio of 1/4/5 (vol/ vol/vol) for 3 h at 100°C. In this research, H 2 SO 4 was used as acid instead of HCl. The formed propylesters, which accumulated in the organic phase, were analyzed by a GC (model 6890N; Agilent). The PHA analysis method was expanded to include the quantification of poly(3-hydroxyhexanoate) (PHHx) and poly(3-hydroxyoctanoate) (PHO) by using methyl 3-hydroxyhexanoate (Sigma-Aldrich) and methyl 3-hydroxyoctanoate (Santa Cruz Biotechnology) as standard.
GC-mass spectrometry (MS) analysis was carried out on a 7890A GC coupled to a 5975C quadrupole mass selective detector (both from Agilent) to identify PHB, PHHx, and PHO. Samples were pretreated in the same way as described for GC analysis. However, methanol was added instead of propanol to form methylesters. The mass spectrum of these methylesters was processed by the Mass Hunter Quantitative Analysis Software for compound identification.
A detailed description of the analytical protocol is described by Velasco Alvarez et al. (2017).
Furthermore, the method for measuring volatile suspended solids was substituted by a thermogravimetric analysis (TGA) using a Perkin Elmer TGA 8000. Around 2 mg of freeze-dried biomass, the sample was heated from 35°C to 105°C (10°C/min), followed by an isothermal step (100 min), followed by second heating run from 105°C to 550°C (10°C/min), followed by a second isothermal step (60 min). All steps were under a nitrogen atmosphere, while in the last isothermal step, the nitrogen gas was switched to air. The TGA data was also applied as an alternative method to determine the PHA weight percentage of the biomass (Chan et al., 2017).  Pabst et al. (2021). Various genomic fragments were aligned with published sequences from GenBank using the NCBI BLAST tool to find related genes and strains (Altschul et al., 1990). All sequence data of this study have been deposited in GenBank with BioProject ID PRJNA831682.

| Metabolic model and parameter identification
A metabolic and kinetic model proposed by  was used as starting point for this study. The previous model was transformed from acetate uptake and PHB production to octanoate (and hexanoate) uptake and PHB, PHHx, and PHO production (Supporting Information: Figure S1 and

| Performance of enrichment at varying pH
The pH of the bioreactor was increased from pH 7 to pH 8 and pH 9 in two consecutive enrichment experiments. At each pH, detailed VERMEER ET AL. More detailed information can be found in Table 1, and Supporting Information Online Materials: Figure S2 and Table S4.
The PHA content at the end of the feast phase appeared to be slightly higher for the enrichment at pH 8 (S8), than for the enrichment at pH 7 (S7) and pH 9 (S9 Nevertheless, it must be noted that the community at pH 8 was enriched for more cycles than the community in S7.

| The impact of oxygen limitation
The oxygen supply rate was decreased in the bioreactor to test the hypothesis that oxygen limitation favors the production of mcl-PHA

| Accumulation with hexanoate
One accumulation experiment was performed with hexanoate (A8-Hx) as the sole carbon source inoculated with biomass enriched on octanoate (S8) to see if high mcl-PHA contents could be produced from a substrate that is more relevant from an industrial perspective.  (Table 1). Moreover, the PHA consisted predominantly of PHB while PHO was absent. The PHHx content was again very similar to A8.

| Bioreactor data validation
The carbon and electron balances of the SBR cycles closed on average for 99.9 ± 5.3%. For the accumulation experiments, a gap in the mass balance developed over time, resulting in an average closure of 66.5 ± 13%. Although these gaps have been observed in accumulation experiments of previous research (Marang et al., 2016;Vermeer et al., 2022), the gaps in this study are larger. A possible explanation for the large size of these gaps forms the feeding method during the accumulation which entails the addition of undiluted octanoic or hexanoic acid. It was hypothesized that the undiluted acid was not able to dissolve completely in the medium. To confirm this hypothesis, undiluted octanoic acid was fed to a bioreactor without biomass to aim for a concentration of 500 mg/L. Nevertheless, measurements showed that the medium contained 420 ± 6 mg/L after 2 h of incubation with a stirrer speed F I G U R E 3 Main results of the enrichment experiments at varying pH. "PHA SBR" is the average PHA wt% at the end of the feast phase over the last three samples of the SBR enrichment. "PHA Accu." is the PHA wt% of the maximal value in the accumulation experiment. "Sphaerotilus" is the relative abundance of this genus in the 16S amplicon data over the last three samples of the enrichment. q PHA, max is calculated from the SBR cycle analysis in the enrichment. In Supporting Information: Figure S2, all experimental results fitted with the metabolic model are shown, including off-gas, ammonium, and biomass. HB, hydroxybutyrate; HHx, hydroxyhexanoate; HO, hydroxyoctanoate; PHA, polyhydroxyalkanoate; SBR, sequencing batch reactors.  of 800 rpm at pH 8, a difference of 17%. This result makes it plausible that part of the octanoic acid did not dissolve in the above-described accumulation experiments as well.
TGA was used as a secondary PHA quantification method to validate the PHA analysis results of the GC (Chan et al., 2017). On average, the TGA measured 3.0 ± 2.6 wt% higher than the GC which confirms that the applied GC method is reliable for mcl-PHA quantification (Table 1).

| Metagenomic analysis
One biomass sample, taken at the time of the cycle analysis of pH 7 (S7), was subjected to metagenomic analysis. The outcome revealed a relative abundance of 54% Sphaerotilus natans, which is in accordance with the 16S amplicon results. The remaining microbial community is very fractionated with the second most abundant microorganism only having a relative abundance of 0.6%.

F I G U R E 5
Overview of PHA and substrate analysis of S8 (a), S8-OL (b), A8 (c), A8-OL+ (d), A8-OL (e), and A8-Hx (f). The symbols represent the measured data, while the lines represent the modeled data. The yellow data points represent the total amount of PHA in the bioreactor, which is a sum of the individual monomers (HO, HHx, HB). For the accumulations, the substrate consumption is presented as a cumulative curve starting and ending at an arbitrary value. In Supporting Information: Figure S2, all experimental results fitted with the metabolic model are shown, including off-gas, ammonium, and biomass. A8-OL, oxygen-limited accumulation; HB, hydroxybutyrate; HHx, hydroxyhexanoate; HO, hydroxyoctanoate; PHA, polyhydroxyalkanoate; S8-OL, SBR under oxygen limitation; SBR, sequencing batch reactors.
Two copies of the PHA synthase gene were found both in the genomic data of this study and in the online available genome of S.
natans subsp. sulfidivorans D-507. The first synthase gene has 100% similarity with a PhaC class I synthase found in S. natans (McCool & Cannon, 2001). The second PHA synthase is classified as alpha/beta fold hydrolase in S. natans (Ollis et al., 1992), but it reveals a high similarity with PHA synthases in other microorganisms which are not assigned to a specific class of PHA synthases (e.g., 96% identities with PHA synthase in Leptothrix sp. C29) (Whitman et al., 2015).
Although both PHA synthases have roughly the same size (∼1200 nucleotides), the similarity between the two genes is low. Only two fragments of the genes revealed reasonable similarity. The first alignment had 154/271 (71%) identities, while the second alignment had 166/251 (66%) identities. The DSC curves revealed a high degree of similarity between all samples. Two melting temperature (T m ) peaks were measured in each sample with an average value of 143 ± 6°C and 157 ± 4°C. A higher HB content seems to result in a slightly higher T m , although the trend is not very distinct (Supporting Information: Figure S3). A8-Hx forms an exception: It does not follow this trend and has only one peak. A reference sample consisting mainly of HB monomers (94%) with a small fraction of hydroxyvalerate (HV) (2%) and HHx (4%) monomers also gave two melting peaks at slightly lower temperatures (137°C and 152°C).

| Physicochemical properties
F I G U R E 6 Determination of polymer structure. (a) The degree of polymerization versus monomer composition for all accumulation experiments. A trendline was drawn through both the HB and HO data points. (b) PHA fractionation test. Different volumes of 1-heptane were added to chloroform with PHA dissolved from the accumulation at pH 8 at 12 h (S8). PHA was measured in both the pellet (precipitated) and supernatant (dissolved were dominant (i.e., an HHx fraction of 0.14 and an HO fraction of 0.65). This is considerably higher than the 34 wt% reported in previous research investigating octanoate-fed enrichments, although the mcl-PHA fraction was similar (i.e., HHx fraction of 0.08 and HO fraction of 0.62) (Chen et al., 2018). Li et al. (2021) also obtained a lower PHA content (53 wt%) when enriching with octanoate as a cosubstrate, which also resulted in a significantly lower mcl-PHA fraction (i.e., HHx fraction of 0.08 and HO fraction of 0.26).
In scl-PHA studies, it was established that minimizing the number of SBR cycles per SRT is a crucial factor for the enrichment of communities with high PHA productivity. The reason is that the substrate-to-biomass ratio is increased leading to a higher PHA content at the end of the feast phase provided that substrate is predominantly used for PHA production (Jiang et al., 2011). In this study, the number of SBR cycles per SRT was deliberately set at a low value of 2 which is lower than the value extracted from other mcl-PHA studies whenever data were available (Table 2). Therefore, we argue that the low number of SBR cycles per SRT in this study was key for the high final PHA content compared to other studies where mcl-PHA is produced with enrichments.

| Mcl-PHA production as a selective strategy
Analysis of the pathways for octanoate metabolism revealed that an active beta-oxidation pathway is accompanied by the production of reduced electron carriers such as NADH (and FADH 2 ) (R3-5 in Supporting Information: Figure S1). For PHB production, more cycles of the beta-oxidation are required and, therefore, more NADH is produced than for PHHx and PHO production. The production of biomass from octanoate also results in a net NADH production.
NADH can be regenerated via oxidative phosphorylation resulting in the production of ATP (R12 in Supporting Information: Figure S1).
When the degree of reduction of the substrate is high compared to the storage polymer, oxidative regeneration of NADH is coupled to more ATP production than is actually needed in storage polymer production. Alternatively, ATP can be used in the maintenance reactions of the bacteria. However, under feast-famine conditions, the biomass-specific PHA production rates are relatively high compared to the biomass-specific maintenance reaction rates when the substrate is present (Table 1 and Supporting Information: - Figure S4). Therefore, maintenance reactions offer a relatively small sink for the surplus of ATP.
In Figure 7, the modeled surplus of ATP is shown for a hypothetical community that produces 100% PHB but with the kinetic parameters found in the enrichment at pH 8 (S8 by the blue line in Figure 7. This shortage of ATP can be replenished by the complete oxidation of substrate via the tricarboxylic acid cycle. A third scenario is the production of a mixture of PHO, PHHx, and PHB as has been found in the experimental results of this study (S8). In this scenario indicated by the yellow line in Figure 7, the ATP surplus is minimal compared to the 100% PHB scenario. Overall, it appears that the scenario with mcl-PHA production does not result in a large energy surplus. Therefore, this theory suggests that mixed PHO, PHHx, and PHB production may be the direct result of the electron distribution between energy-consuming product formation pathways and respiration-driven energy production.

| Oxygen limitation as a tool to enhance mcl production
In aerobic feast-famine SBRs, the oxygen uptake rate profile during the cycle is a key indicator of the functional performance of the microbial community. When the community predominantly converted the substrate directly into biomass, the majority of the oxygen (∼80%) was consumed in the feast phase ( Figure 2b). A smaller fraction (∼60%) of the oxygen was consumed in the feast phase when an scl-PHA production strategy was adopted by the community. The reason is that the produced PHA was oxidized in the famine phase to support growth. As explained in the previous section, mcl-PHA production in the feast phase results in a smaller surplus of ATP than scl-PHA production. Therefore, oxidative phosphorylation can be operated at a lower rate, thereby consuming less oxygen. This was confirmed in the first enrichment where an even smaller fraction (∼30%) of the oxygen was consumed in the feast phase when mcl-PHA was produced. First, this makes oxygen consumption an important process performance indicator. Second, limiting oxygen supply becomes a potential selective tool to increase the mcl-PHA fraction. Although this tool has been successfully employed for enhancing mcl-PHA production in pure cultures (Blunt et al., 2017;Fernández et al., 2005) and for enhancing PHA yields in scl-PHA producing enrichments (Pratt et al., 2012;Third et al., 2003), it has never been demonstrated as a selective strategy to enhance mcl-PHA has been linked to PHB production (Takeda et al., 1995), this is, to our knowledge, the first report that links this genus to mcl-PHA production. Furthermore, Sphaerotilus natans is known to grow relatively well under conditions of low oxygen concentration (Pellegrin et al., 1999), which could explain why this species was also dominant in the oxygen-limited SBR (S8-OL). In a recent study by Grabovich et al. (2021), a comparative genome analysis was conducted on S. natans subp. sulfidivorans and S. natans subsp.
natans. It was proposed that the two bacteria should be reclassified as separate species, S. sulfidivorans sp. nov. and S. natans, based on a significant difference in genome characteristics and metabolic versatility.

| Polymer structure and microbial origin
The PHA fractionation experiment revealed that the enrichments in this study produced at least an HB-rich polymer, which could be effectively separated from the mcl-PHA fraction. For this reason, the polymer properties of the produced PHA are likely to be highly customizable. A study by Furrer et al. (2007) showed that poly(3hydroxyhexanoate-co-octanoate (PHHxO) can solubilize reasonably well in 1-hexane, which presumably also explains why the HHx/HOrich fraction did not precipitate when 15 volumes of 1-heptane were added in this study.
The DSC data also suggests the presence of an HB-rich polymer, because a high melting temperature was found similar to the PHBVHx reference. From the literature, it is known that mcl-PHA exists predominantly in an amorphous state (Cai & Qiu, 2009;Fernández et al., 2005;Ruiz et al., 2019). Therefore, no clear melt transition exists when heat is applied, which could explain why no independent melting peaks appear on the DSC spectrum corresponding to the mcl-PHA.
The low PDI of 1.7 of the HHx/HO-rich fraction suggests that the HHx and HO monomers are embedded in one copolymer.
In the genome of S. sulfidivorans sp. nov., two PhaC genes were found with significant differences. Alvarez-Santullano et al. (2021) state that members of the order Burkholderia, which is the order where Sphaerotilus sp. belongs, generally possess two or more different copies of the PhaC gene. It was also proposed that PhaC classification is more diverse than was previously known with the existence of additional classes. For S. sulfidivorans sp. nov., it is possible that one PhaC gene encodes for a PHA synthase that produces mainly PHB, while the other encodes for a PHA synthase that produces mainly PHHxO resulting in two different polymers instead of one copolymer in one microorganism. Both synthases will presumably have different characteristics, which may contribute to explaining the relation between the monomer content and the degree of polymerization (Figure 6a). Different physiological conditions result in differences in the activity of the two PHA synthases and subsequently different values of the overall M w . Nevertheless, it is also possible that a second microbial species is responsible for part of the produced PHA.

| Outlook
To date, octanoate is present in most chain elongation studies and processes as a minor component (Holtzapple et al., 2022). However, in recent years, the octanoate yield in these studies increased (Kucek et al., 2016), and it is expected that this trend will continue.
Therefore, it is believed that octanoate valorization routes such as those described in this study will become more relevant in the future.
Alternatively, the findings from this study (e.g., a low number of cycles per SRT or oxygen limitation) can possibly be extrapolated toward platform chemicals that are currently omnipresent in waste valorization routes such as hexanoate (Chen et al., 2017).
The bacterial species that were enriched at pH 7 and 8 (S. sulfidivorans sp. nov.) and at pH 9 (Thauera sp. or Phreatobacter sp.) were not yet linked to mcl-PHA production. This illustrates that large parts of the mcl-PHA biodiversity are still unexplored, and that microbial enrichments can be a powerful tool to explore these parts and to seek novel industrially relevant strains.

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.