Microbes exhibit varying growth condition requirements when metabolizing different carbon sources^[11], necessitating an understanding of the requisite O₂ limitation for converting individual lignin compounds. As monomeric compounds in alkaline lignin liquor comprise linear carboxylic acids and aromatic compounds, three monomers identified in alkaline lignin liquor^[21] - acetate, vanillic acid, and ferulic acid - were selected as the representative substrates to investigate the effect of O₂ limitation on lignin bioconversion. Different volumes of media (ranging from 50 to 200 mL) were used in the shaking flasks to establish varying levels of O₂ limitation, as the rate of oxygen diffusion into the medium decreases with increasing volume (surface aeration)^[27]. The 50 mL condition was used as the positive control, representing non-oxygen-limited conditions. Figure 2 presents the effect of limiting O₂ availability on cell growth (OD₆₀₀). To characterize the impacts of O₂ limitation on redirecting carbon fluxes (ratio of PHA to cell growth), the resulting residual bacterial biomass (dcw excluding PHA concentration) and PHA production are shown in Figure 3. The results presented in Figures 2A and 3 indicate that O₂ limitation did not restrict the growth (maximum specific growth rate μ_max of 0.03OD₆₀₀/h) of P. putida on acetate, resulting in similar carbon flux redirection with a PHA concentration of 19% PHA/dcw across all the flasks. Although acetate was slightly volatile, limiting O₂ supplementation reduced substrate loss, marginally improving terminal bacterial growth from 0.3 g/L in 50 mL medium to 0.4 g/L in 150 mL medium, concomitant with PHA production of about 0.1 g/L. This unrestricted cell growth on acetate by varying O₂ availability was attributed to its simple degradation pathway, which could directly enter central carbon metabolism via acetyl-CoA [Supplementary Figure 2]. Acetate was measured at the end of fermentation, and no residual carbon source was detected (data not shown).

Aromatic compounds undergo upper degradation to protocatechuic acid, followed by aromatic ring opening via the β-ketoadipate pathway to central carbon metabolism^[19]. Molecular oxygen is not only necessary for energy metabolism; the catabolism of vanillic acid and ferulic acid involves multiple oxygen-dependent enzymatic cleavages and oxidation reactions prior to entering central carbon metabolism [Supplementary Figure 2]. When utilizing vanillic acid as the sole carbon source, increasing O₂ limitation reduced the cell growth rate with μ_max of 0.06, 0.03, and 0.01OD₆₀₀/h in 50, 100, and 200 mL medium [Figure 2B], respectively, resulting in the final bacterial biomass decreasing from 0.82 g/L in 50 mL medium to 0.26 g/L in 200 mL medium [Figure 3], albeit vanillic acid being completely consumed in all the flasks (data not shown). This restricted cell growth led to higher carbon flux towards PHA production, reaching a maximum of 0.21 g/L in the 200 mL medium flask, which was 2.5 times the PHA production in the 50 mL medium flask.

The catabolism of ferulic acid, a more recalcitrant aromatic monomer than vanillic acid, involves initial enzymatic cleavages to acetyl-CoA and vanillin. Acetyl-CoA is directly utilized in central carbon metabolism, while vanillin is further degraded following the same catabolic pathway as vanillic acid [Supplementary Figure 2]. Ferulic acid degradation can also occur extracellularly by enzymes secreted via outer membrane vesicles, leading to the appearance of acetate and vanillic acid in the extracellular environment^[28,29]. Considering that acetyl-CoA derived from ferulic acid degradation theoretically supports a maximum cell growth of about 0.6 OD₆₀₀ (acetate utilization in Figure 2A), the varied cell growth on ferulic acid within the first 24 h - from the highest at 0.74 OD₆₀₀ in 50 mL flasks to the lowest at 0.31 OD₆₀₀ in 200 mL flasks [Figure 2C] - demonstrates that the cleavage of ferulic acid to acetyl-CoA and vanillic acid was affected by O₂ limitation. As the appearance of the easily utilizable carbon source acetyl-CoA, which could accelerate the utilization of aromatic compounds, increasing O₂ limitation by increasing medium volume from 50 to 100 mL did not significantly restrict cell growth (P > 0.05), resulting in similar carbon flux redirection with PHA content of about 20% PHA / dcw [Figure 2C and Figure 3]. The marginal increase in cell growth and PHA production in the 100 mL flask compared to the 50 mL flask was attributed to reduced acetate volatilization. Further increases of O₂ limitation in the 150 and 200 mL flasks reduced both cell growth and PHA production due to substrate leftovers. At the end of the 72-h fermentation, intermediate vanillic acid from ferulic acid catabolism was detected in the 150 mL and 200 mL flasks (data not shown). This observation is consistent with the ferulic acid catabolic pathway and the hierarchical utilization of heterogeneous carbon sources, with the more recalcitrant vanillic acid remaining unconsumed under O₂ limitation. To gain further insights into the reason behind the remaining substrate, a repeat experiment on ferulic acid fermentation was conducted to observe cell growth and DO levels [Supplementary Figure 3]. The results demonstrated that higher medium volumes in flasks limited O₂ diffusion from the gas phase into the liquid, thereby restricting cell growth. The 50 mL culture showed the highest growth (OD₆₀₀ = 0.36 at 24 h; DO = 46%), whereas the 200 mL culture exhibited the lowest growth (OD₆₀₀ = 0.21; DO = 20%). Notably, DO in the 150 and 200 mL flasks later recovered to >80% without a corresponding increase in growth.

Alkaline lignin comprises a collection of linear and aromatic compounds. Considering that P. putida’s response to O₂ limitation might differ between utilizing sole and mixed substrates, three selected monomeric compounds - acetate, vanillic acid, and ferulic acid - were mixed at equivalent molar concentrations (6.67 mM each) to examine the effect of O₂ limitation on the bioconversion of mixed lignin monomers. As shown in Figure 2D and Figure 3, increasing O₂ limitation resulted in a monotonic decrease in cell growth while increasing PHA production, with PHA content (PHA/dcw) rising from 10% in 50 mL flasks to 18% in 200 mL flasks. This demonstrated the redirection of P. putida carbon fluxes towards higher PHA production on mixed substrates through O₂ limitation. Additionally, it is important to note that P. putida’s cell growth on mixed substrates was enhanced compared to that on individual sole carbon sources. Under the highest level of O₂ limitation (200 mL medium), fermentation using mixed substrates doubled the maximum specific growth rate (μ_max of 0.027 OD₆₀₀/h) compared to that on sole vanillic acid (μ_max of 0.013) or ferulic acid (μ_max of 0.015) [Figure 2]. This demonstrates that the simple carbon source acetate facilitated the utilization of the aromatic compounds. This result suggested that fermentation on mixed substrates such as alkaline lignin liquor might require further increasing O₂ limitation to restrict cell growth and reconfigure more fluxes towards PHA production.

Appropriate O₂ limitation has been confirmed to restrict cell growth on aromatic monomers and enhance PHA production. In this section, the monomers were replaced by 10 g/L alkaline lignin liquor to examine the effects of O₂ limitation on lignin bioconversion. Similar to the examination of lignin monomer bioconversion, varying O₂ limitation was achieved by using different volumes of medium (50-200 mL) in 250 mL shaking flasks. With the same level of O₂ limitation as in the fermentation of lignin monomers, the results of residual cell growth, PHA production, and lignin degradation in Figure 4A-C show that O₂ limitation during lignin fermentation did not exhibit a clear trend in either cell growth or PHA production, resulting in a similar final PHA production across all the flasks, at approximately 0.2 g/L. This insignificant carbon flux redirection was attributed to an unrestricted cell growth rate. As shown in Figure 4A, cell growth in all the flasks increased rapidly to above OD₆₀₀ of 1.0 within 24 h, reaching over 80% of the total cell biomass in 120 h. This trend was consistent with the findings from fermentation on mixed monomers [Figure 2D], where the cell growth on mixed substrates was enhanced compared to that on the sole carbon sources. Alkaline lignin liquor represents a more complex mixture of substrates, consisting of many linear compounds like acetate and malonic acid, as well as aromatics^[21]. During fermentation of alkaline lignin liquor, the utilization of linear compounds could facilitate the degradation of aromatics, explaining the insignificant restriction of cell growth at the same level of O₂ limitation used for restricting cell growth on aromatic monomers.

To further restrict cell growth on alkaline lignin, O₂ limitation was intensified by performing lignin fermentation in flasks with only medium stirring rather than flask shaking. This approach of only stirring the medium effectively distinguished the maximum specific cell growth rate and resulting bacterial biomass (within the first 72 h) [Figure 4D]. Maximum PHA production within 72 h reached 0.24 g/L [Figure 4E], surpassing the total PHA produced by fermentation in shaking flasks. After 72 h of fermentation in stirring flasks, substrate utilization tended to stagnate [Figure 4F]; therefore, the fermentation was transferred into a shaking incubator for shaking flasks, thereby increasing oxygen transfer into the medium and facilitating the degradation of the most recalcitrant compounds. At the end of the shaking stage, flasks containing 50 and 100 mL of medium showed a similarly high level of cell growth at OD₆₀₀ of 1.2. Flasks with 150 and 200 mL of medium restricted cell growth to about OD₆₀₀ of 1.0 and enhanced the terminal PHA production to 0.31 and 0.34 g/L, respectively, approximately 45% higher than in 50 and 100 mL of medium [Figure 4E].

Examination of DO levels facilitated understanding of the relationship between fermentation outcomes and O₂ availability. Figure 4G illustrates the DO levels in the 150 mL flasks. During the stirring stage, DO decreased to below 1% of the saturation level in water, during which cell growth and PHA synthesis could be maintained, as shown in Figure 4D and E. This was consistent with the observation that P. putida could sustain PHA synthesis at 0% DO when utilizing simple carbon sources^[18]. When the stirring stage proceeded to 72 h, lignin degradation stagnated, and DO rose, guiding the decision to change the oxygen supply approach to using shaking flasks, where DO gradually rose above 20% and finally reached above the level of 80%. Molecular weight distribution analysis identified two dominant peaks in the lignin liquor [Figure 4H], corresponding to low-molecular-weight compounds (< 1,000 Da) and a higher-molecular-weight fraction centered at approximately 5,000 Da. P. putida was able to utilize compounds from both fractions. The molecular weight profile of the abiotic aeration control remained unchanged relative to the raw lignin feedstock, indicating that lignin degradation was primarily attributable to microbial activity rather than aeration-induced oxidation [Supplementary Figure 4].

Industrial fermentation typically employs impellers and air spargers for agitation and oxygen supplementation. To validate this approach for industrial applications, lignin fermentation was scaled up in a 1 L tank reactor equipped with a three-blade impeller and bottom air sparger. Different levels of air supplementation ranging from 0.3 to 0 VVM were examined [Figure 4I and Supplementary Figure 5]. Results indicated that reducing air supplementation from 0.3 to 0.01 VVM monotonically increased both cell growth and PHA production, achieving PHA of 0.13, 0.20, and 0.24 g/L with air supply at 0.3, 0.1, and 0.01 VVM, respectively. This was probably caused by substrate volatilization and reduced cell viability at high oxygen supply^[30], as reflected by the lack of corresponding increases in cell growth and PHA despite reduced lignin levels [Supplementary Figure 4]. The aeration-induced substrate loss was further confirmed through a separate abiotic experiment evaluating substrate loss under different aeration conditions [Supplementary Figure 6]. Further reduction of air supply from 0.01 to 0 VVM resulted in a decrease in cell growth by 21% and an increase in PHA production to 0.26 g/L. Implementing a two-stage air supplementation strategy (0 VVM air sparging for 72 h followed by 0.01 VVM air sparging), akin to the two-stage oxygen supply in flask fermentation (initial stirring the medium in fixed flasks followed by shaking flasks), further enhanced PHA production to 0.33 g/L. These results validated the effectiveness of tuning O₂ limitation to enhance lignin bioconversion to PHA in a tank reactor.

As traditional approaches to enhance PHA production often involved redirecting carbon flux by limiting nutrient supplementation, particularly nitrogen^[8,9], it was appropriate to compare the approaches of tuning O₂ limitation with nitrogen limitation on lignin bioconversion. For this comparison, different concentrations (2-25 g/L) of lignin were examined to compare the results from nitrogen-limited fermentation in Figure 1. Figure 5 and Supplementary Figure 7 present the results of lignin bioconversion, cell growth, and PHA production using a two-stage tuning O₂ limitation approach during the stirring stage (0-72 h) and shaking stage (0-144 h), respectively. At the end of the stirring stage, fermentation with low lignin concentrations (2 and 5 g/L) had achieved their maximum degradation of 37% and 30%, respectively, which were not significantly enhanced during the shaking stage (P > 0.05). Lignin degradation decreased to 18% with 10 g/L lignin, while 15 g/L lignin showed slightly higher degradation (attributed to increased availability of simple compounds for bacterial growth). Bacterial growth exhibited a monotonic relationship with increasing lignin concentration from 2 to 20 g/L, as greater availability of utilizable carbon sources promoted growth. However, the use of 25 g/L lignin, with its highest concentration of simple compounds, led to reduced bacterial growth compared to 20 g/L lignin, underscoring inhibition by concentrated lignin. PHA production similarly showed a monotonic increase with increasing lignin concentration from 2 to 20 g/L, although production decreased with 25 g/L lignin. After 72 h of stirring, the flasks were transferred to a shaking incubator to enhance O₂ availability and further lignin degradation. Results in Figure 5B demonstrated that degradation of 10 and 15 g/L lignin improved to above 30%, approaching the upper limit of reported lignin degradation (35%)^[13,26]. Degradation of 20 and 25 g/L lignin improved to 27% and 24%, respectively. Both cell growth and PHA production increased monotonically with increasing lignin concentrations, though the enhancement was less pronounced at higher lignin concentrations, highlighting the inhibitory effects of concentrated lignin on bacterial metabolism. PHA yields (g-PHA/g-lignin) calculated across different lignin concentrations ranged between 22 and 45 mg/g-lignin.

Comparing the results of lignin bioconversion under these two strategies, the approach of tuning O₂ limitation demonstrated several advantages over the nitrogen-limited strategy in redirecting carbon fluxes towards PHA production, achieving higher lignin degradation, and enhancing treatment capacity for high lignin concentrations. Firstly, O₂ limitation during the stirring stage imposed a stronger restriction on cell growth, resulting in increased PHA production compared to the nitrogen-limited strategy. Evidence from fermentations with 2 and 5 g/L lignin under O₂-limited conditions already achieved the maximum degradation during the stirring stage, with PHA contents (PHA/dcw) of 41% and 38%, respectively, higher than that using the nitrogen-limited strategy (26% and 29%, respectively). Secondly, the O₂-tuning approach improved the extent of lignin degradation, particularly at high lignin concentrations. A limitation of the nitrogen-limited strategy was that nitrogen consumption during cell growth conflicted with increased nutrient requirements for catabolizing recalcitrant compounds^[9], resulting in low lignin degradation. In contrast, the approach of two-stage tuning O₂-limitation - initially with higher O₂ limitation facilitating the bioconversion of simple compounds, followed by improved O₂ availability for catabolizing recalcitrant compounds - aligned well with the hierarchical utilization of heterogeneous compounds in alkaline lignin liquor, thereby enhancing lignin bioconversion to PHA.

Understanding PHA composition across different strategies is crucial for determining bioplastic properties and potential applications. For instance, short-chain-length PHA (C4-C6) exhibits brittleness, high melting temperature, crystallinity, and tensile strength, making it suitable for packaging, whereas medium-chain-length PHA (C8-C10) is flexible and finds applications in adhesives and biomedical applications^[31,32]. In this study, when O₂ was limited, 3-hydroxydecanoic acid (C10) accounted for over 80% of the PHA composition with lignin at 2 and 5 g/L, followed by 3-hydroxyoctanoic acid (C8) at around 20% [Figure 6 and Supplementary Figure 8]. The proportion of C10 decreased to about 70% with increased lignin concentrations of 10 to 25 g/L. Interestingly, polyhydroxybutyrate (C4) appeared in fermentations with lignin at 20 and 25 g/L, comprising approximately 6% of total PHA production. Traditionally, P. putida was regarded as rare in producing short-chain-length PHA^[33,34]. The emergence of polyhydroxybutyrate was likely induced by inhibition caused by concentrated lignin, activating the pathway of synthesizing short-chain-length PHA, warranting further investigation. Under nitrogen-limited conditions in lignin fermentation, 3-hydroxydecanoic acid (C10) predominated, constituting approximately 80% of the PHA, followed by 3-hydroxyoctanoic acid (C8) at around 20%. Varying concentrations of lignin feeding did not significantly alter PHA compositions (P > 0.05), as P. putida metabolism was primarily influenced by simple compounds under nitrogen-limited conditions. The medium-chain-length PHA has often been criticized for its low melting temperature and tensile strength, which limited its applications. To address that, medium-chain-length PHA was commonly blended with short-chain-length PHA or other materials to enhance its strength^[31,35]. This study’s findings demonstrate the tunability of PHA compositions by varying lignin concentrations, offering potential for producing plastic products tailored to diverse applications.

Limiting nitrogen supply enables rapid cell growth to a moderate density, at which cells retain the capacity to catabolize the same carbon source for PHA production. However, due to the heterogeneous nature of lignin-derived substrates and their hierarchical utilization, nitrogen depletion restricts the catabolism of more complex compounds^[12]. To explore an alternative approach, this study examined the effect of oxygen availability on the fermentation of compounds in alkaline lignin liquor. During the low-oxygen stage (initial 72 h; Figures 4 and 5), carbon source utilization was reduced, while PHA accumulation increased. This phenomenon has also been observed for other substrates, such as glucose and butanol, although the underlying genetic mechanisms remain unclear^[16,17]. In the subsequent stage (72-144 h; Figures 4 and 5), where oxygen was no longer limiting, fermentation proceeded under non-limited conditions, enabling the utilization of more recalcitrant compounds and further enhancement of PHA production. Notably, 1 L tank reactor experiments demonstrated that lignin fermentation can be sustained under surface aeration without additional air sparging, and that reducing aeration rates helps minimize substrate loss [Figure 4I and Supplementary Figure 5], which is critical for future industrial applications. Further work is recommended to elucidate the mechanisms of metabolic flux redistribution, particularly the decoupling between reduced carbon utilization rates and enhanced PHA accumulation. A techno-economic analysis is warranted to evaluate industrial-scale implementation. Nutrient-limited fermentation benefits from reduced nutrient consumption but may experience carbon losses during aeration, whereas the oxygen-tuning strategy enables higher lignin conversion. The trade-off between nutrient savings and improved carbon utilization should be assessed to determine the most economically viable approach.