The starch-based production of lactic acid (LA) from bio-based raw materials has been important in the success of poly-lactic acid (PLA) as a replacement for fossil-based plastics. LA is one of the most successful examples of industrial large-scale microbial fermentation, and it is furthermore the best way to obtain optically pure LA, without relying on downstream processing to separate both enantiomers—L-LA and D-LA (Yi et al., 2016; de Oliveira et al., 2022; Lomwongsopon and Varrone, 2022; Ojo and de Smidt, 2023). The access to pure LA enantiomers makes it possible to tailor ratios of poly-L-LA and poly-D-LA in a PLA polymer, giving advantages such as decreasing brittleness and the slow crystallization of PLA (Abdel-Rahman and Sonomoto, 2016; Rosenboom et al., 2022). There are bioplastic alternatives to most fossil-based plastics, such as polyhydroxyalkanoates (PHA) and bio-polybutylene succinate (bioPBS), but all bio-based and biodegradable options are still more expensive (ranging from 1.5 to 4 times) than fossil-based polymers (Lomwongsopon and Varrone, 2022; Rosenboom et al., 2022; Costa et al., 2023; Yu et al., 2023).

The use of non-food feedstocks (e.g., food, industrial or agricultural waste) instead of starch-based feedstocks (e.g., corn, sugar cane, cassava) may become a strategy to make biopolymer production processes more economically competitive (Abdel-Rahman and Sonomoto, 2016; Abedi and Hashemi, 2020; Abedi and Hashemi, 2020; de Oliveira et al., 2022). Approximately two-thirds of the Swedish land area is forest, with close to 20 million hectares used for production of wood-based products, of which 80% are softwood species (Ekström and Hannerz, 2021). The forest industry is one of the most important sectors in the Scandinavian economy, and therefore it is important to find added-value applications to its wastes. These wood residues can be upcycled as substrates for bio-based industries, to be used to produce diverse platform chemicals, such as LA.

Probably, the biggest challenge in using softwood as raw material for bioprocesses is its high recalcitrance, in comparison to other agricultural wastes (Zhu and Pan, 2010; Galbe and Wallberg, 2019). It is necessary to pretreat wood with chemical and physical methods to disrupt the structure of hemicellulose, cellulose, and lignin, making it more accessible for enzymatic hydrolysis to monomeric sugars (Galbe and Wallberg, 2019; Yankov, 2022). This pretreatment will lead to the formation of degradation products that can be inhibitory for microorganisms in the bioprocess (Liu et al., 2013; van der Pol et al., 2016; Alves de Oliveira et al., 2018). Pediococcus acidilactici ZP26 is a D-LA and exopolysaccharides (EPS) producer, homofermentative, microaerophilic, with proven resistance to lignocellulosic degradation products (Liu et al., 2013; Yi et al., 2016; Campos et al., 2021; 2023). This strain has been reported to produce almost 77 g/L D-LA from pretreated and biodetoxified corn stover (Yi et al., 2016). In that study, Yi and collaborators ran a 25% (w/w) solid content simultaneous saccharification and fermentation (SSF, without separation of the liquid and solid fraction of the pretreated corn stover) with 6 h of pre-hydrolysis, achieving a yield of approximately 60% on cellulose content. Upon inoculation, the biomass started producing LA directly, without any lag phase for adaptation. This was enabled by a combination of three factors: first, the decrease of the concentration of inhibitors during a biodetoxification step of the corn stover hydrolysate with Amorphotheca resinae ZN1; second, the inherent resistance of P. acidilactici ZP26 to any leftover of these lignocellulose degradation products; and finally by the high sugar availability resulting from the pre-hydrolysis of cellulose (Yi et al., 2016).

In the present study, a SO₂-catalysed steam-explosion pretreatment was applied to spruce chips. The application of this physicochemical pretreatment in softwood has been extensively investigated since it favors fiber separation and the hydrolysis of hemicelluloses. The gaseous SO₂ is preferred to a liquid acid since it is more evenly distributed through the biomass (Alvira et al., 2010; Galbe and Wallberg, 2019). The high temperature and low pH lead to the formation of sugar and lignin degradation products. The influence of inhibitors present in the filter-pressed liquid fraction of pretreated softwood in the production of optically pure LA was presented in previous work (Campos et al., 2023). Enzyme and cell culture inhibition are usually associated not only with toxic compounds produced in the lignocellulosic biomass pretreatment but also with high solids loadings (>15% solids (w/w) in the reaction medium) (Kadić et al., 2014; da Silva et al., 2020). However, the first SSF tests showed an unexpectedly long lag phase between inoculation and exponential LA production, even though the usual inhibitors (organic acids and furans) were not detected, and solid loadings were kept low (7% on weight [w/w] of water-insoluble solids [WIS]). The focus in the current work was therefore to find methods to decrease the lag phase experienced after the addition of pretreated softwood solids at the beginning of the fermentation time (0 h) and propose a hypothesis for this extended lag phase.

An unexpected delay in the start of D-LA production from pretreated softwood solids by P. acidilactici ZP26 was found in the present work (Figure 1A). The presence of pretreated softwood solids led to a longer delay, whether or not the solids were used as carbon source. In this study, a lag phase was defined as the time from inoculation until the beginning of the exponential phase. In the search for the reason for such a long lag phase, the addition of Tween 80 to the pretreated softwood solids was found to decrease the lag phase by half (see Supplementary Figure S2). No, or only a short, lag phase (0–4 h after inoculation) was observed in glucose controls. The addition of other solid materials, such as a mixture of Avicel cellulose and Kraft lignin, or sawdust, also gave no delay in the start of the exponential production of D-LA (see Supplementary Figure S2). Very importantly, there was no lag phase if pretreated solids were added to exponentially growing cells (3 h after inoculation on glucose). Extensive (pre-)hydrolysis of the pretreated softwood solids (48 h) also removed the lag phase, but a substantial analogous phase of 40 h or more was observed with a pre-hydrolysis time of 12 h. This suggests that there was an interaction between pretreated but not enzymatically hydrolyzed fibers and the stationary cells causing an extended lag phase.

The addition of a higher enzyme concentration to the SSF led to faster substrate availability but did not decrease the 40-hour-long lag phase (Figure 1A). This was followed by tests with pre-hydrolysis of the solids (Figure 1B), which would make cellulosic glucose readily available for consumption after inoculation. 12 h of pre-hydrolysis led to a slow production period soon after inoculation, while the volumetric production rate was still very low in comparison with other conditions. This low production period was not caused by a lack of enzyme activity, as shown by measurements of glucose concentrations in the broth, as assessed by measurement of the glucose concentration in the broth (see Supplementary Figure S1). In these reactors, the glucose concentration in the medium started to decrease rapidly at 48 h after inoculation, while the D-LA production rate increased, which hinted at an unidentified change in the medium. To allow for maximum solids hydrolysis before inoculation, a separate hydrolysis and fermentation (SHF) mode was simulated in bioreactors with 48 h pre-hydrolysis (Figure 1B). To confirm that the longer lag phase was caused by any problem with the inoculum, control bioreactors using a glucose solution as the sole carbon source, without solids, were performed. The 48 h pre-hydrolysis tests (Figure 1B), together with the glucose control (Figure 1C), made it possible to rule out the possibility that the longer lag phase was caused by a deficient inoculum since the production of D-LA started right after inoculation for both cases.

Optimizing the hydrolysis yield of the pretreated softwood was not a target for the study. Moreover, the presence of inhibitors such as aldehydes, furans, and organic acids in aqueous solution has been demonstrated to decrease the efficiency of enzymatic hydrolysis (Palmqvist and Hahn-Hägerdal, 2000; Arora et al., 2013; Galbe and Wallberg, 2019). The tests with pre-hydrolysis left the question if the lower amount of glucose was the cause of a longer lag phase. Therefore, cultivations were run with synthetic sugars and solids, without enzymatic hydrolysis (Figure 1C). The results confirmed that the source of the monomeric glucose (synthetic medium or softwood solids) was not the reason for the longer lag phase, since bacteria were not able to quickly start consuming the monomeric sugar available.

In other studies that use softwood as a carbon source for bacteria, degradation products formed during the harsh physicochemical pretreatment are usually considered the most significant reasons source of microbial inhibition (Asada et al., 2012; Dietrich et al., 2019; Campos et al., 2023). These are usually found in the liquid fraction of the pretreated biomass, together with hemicellulose sugars. The longer lag phase observed in the first tests (Figure 1) exceeded the 2–4 h registered in previous studies with the liquid fraction of the pretreated softwood (Campos et al., 2023), despite the considerable amounts of lignocellulose degradation products present in the previous study. Some of the water-soluble inhibitors were also analyzed in the present study (succinic acid, lactic acid, formic acid, acetic acid, levulinic acid, HMF, and furfural) but the amounts were lower than the detection limits. The longer lag phase is therefore unlikely to be caused by inhibition by any of these compounds. Since most of the SSF studies are focused on the use of the whole pretreated lignocellulosic biomass, possible effects of the presence of solids in microorganisms’ metabolism are possibly masked.

From the results, it was possible to hypothesize that inoculum cells interact with non-hydrolyzed solids detrimentally in the initial stage of D-LA production. Enzymatic hydrolysis before inoculation (e.g., 48 h pre-hydrolysis) decreased the lag phase before the start of D-LA formation. Interestingly, there was a change in the interaction between the bacterial cells and the softwood solids after 40-plus hours of fermentation even without enzymatic hydrolysis (Figure 1C), which allows P. acidilactici to start fermenting the glucose into D-LA.

Previous studies have reported a decrease in particle size during agitation of pretreated softwood (Kadić et al., 2014). The referenced study showed that higher agitation rates (up to 600 rpm) were the cause of particle size reduction in high WIS loading (13%) spruce hydrolysis. Lower solid loadings (7% (w/w) WIS) did not show such prominent effects during the conversion from glucan to glucose and were therefore chosen for the present study to limit the high solid effect influence on bacterial performance. However, after the observation of a longer lag phase, parallels can be drawn between the challenges of working with such recalcitrant and fibrous materials. Kadić et al. (2014) observed that softwood fibrous structure is damaged by agitation in an aqueous solution increasing its surface area and the exposure of the cellulosic fibers. This aperture of the softwood structure, even without enzymatic hydrolysis, might be one of the reasons for the delayed start of fermentation of glucose into D-LA shown in Figure 1C. However, the results presented by Kadić et al. (2014) refer to higher solid loadings than the present study, and therefore conclusions cannot be directly applied to the present study. Other studies have been conducted on the rheological characterization of pretreated softwood (Wiman et al., 2011), but these were also coupled with enzymatic hydrolysis for cellulosic sugar utilization. A more thorough study on the structural placement of the pretreated softwood fibers without enzymatic hydrolysis would have to be conducted to understand its influence on cell behavior.

The potential of cell adsorption to wood solid particles has previously been shown to be a cause for microbial inhibition. Adherence of cells to surfaces can happen due to van der Waals, electrostatic, and acid-base interactions, and it is an important field of research since it is the first step for biofilm formation (Soumya et al., 2013; Munir et al., 2019). In the case of non-pretreated wood, extractives have been proven to have anti-microbial properties (Tomičić et al., 2020). Reactor experiments were conducted with sawdust (untreated spruce fine particles), using a synthetic glucose solution as a carbon source, and no inhibition was observed. P. acidilactici ZP26 was able to start consuming glucose for D-LA production right after inoculation (see Supplementary Figure S2). Therefore, neither wood extractives nor untreated softwood solids showed a negative influence on the culture. This indicated that the inhibition was caused by one or more characteristics of the pretreated solids. Cultivations with Avicel cellulose and Kraft lignin powder were conducted to assess the influence of model solid compounds on the process. However, these showed very similar results to tests with sawdust. It is known that lignin has a complex structure that cannot be modeled after these types of reagents. The results indicate that inhibition was not caused by chemical interactions between the cells and the lignin and cellulose molecules, or that the model Kraft lignin used is not representative of that present in the dilute-acid pretreated softwood solids.

The complex 3D structure of the pretreated softwood solids seemed to play an important role in the observed microbial inhibition. Steam explosion pretreatment opens the wood fiber structure, making cellulose more accessible to enzymes and microorganisms (Galbe and Wallberg, 2019), but also exposing lignin. Cellulolytic enzymes bind to lignin, making them useless for the enzymatic hydrolysis of lignocellulose (Pan et al., 2005). The addition of surfactants to the enzymatic hydrolysis process has been shown as a possible solution to decrease the concentration of enzyme added for increased hydrolysis yields (Kandhola et al., 2017; Hou et al., 2022). Eriksson et al. (2002) reported a reduction of enzyme adsorption to pretreated lignocellulosic biomass when Tween 20 was added during enzymatic hydrolysis. The study proposes that the positive effects come from the interaction of the exposed lignin with the hydrophilic part of the surfactant, making it inaccessible to cellulolytic enzymes (Eriksson et al., 2002; Alkasrawi et al., 2003). The original MRS medium has 1% Tween 80, which was withdrawn for medium simplification (Yi et al., 2016). However, this surfactant has been shown to increase nutrient intake by improving cell wall permeability (Nielsen et al., 2016). Tween 80 was then added to the cultivation to assess its effect on D-LA production. The addition of this surfactant had a positive effect on the SSF process by reducing the lag phase from 48 to 24 h. This might either suggest a possible interaction between the cells and the surface of the pretreated softwood solids be the source for this inhibitory effect or an improvement of the nutrient intake by the cells. Further study is required to fully understand the mechanism of action of Tween 80 in the cultivation. Non-etheless, a lag phase of 24 h is still too long for possible industrial production of D-LA. The present study focused on a way to overcome this inhibition, to shorten the lag phase in the presence of pretreated softwood solids. The specific nature of the solid-cell interaction remains to be elucidated.

P. acidilactici ZP26 has been reported to produce EPS in association with growth (Liu et al., 2015; Campos et al., 2021). When in the late exponential or stationary phase, the flocculation of the culture caused by EPS production is evident. It was hypothesized a culture that has started to produce EPS, which is both secreted to the medium and retained in the cellular membrane for cell protection (Korcz and Varga, 2021), might be adsorbed to the pretreated wood (Santore, 2022), and thereby showing negative effects on D-LA metabolism. The interaction between biofilms and wood might change not only because of the increase in the number of cells that do not attach to the biofilms after a certain density but also might be related to changes in the physiological state of the cells, which might lead to cell disaggregation from the biofilm and subsequent improvement of mass transfer and metabolism (Santore, 2022). Since EPS production is not observed in an early growth stage, a study was conducted where the culture was initiated in a synthetic medium supplemented with glucose until it reached an early exponential phase, when pretreated solids were added to the bioreactor, with and without enzyme addition (Figure 2). The lag phase decreased from more than 40 h to less than 2 h when the solids were added after 3 h of fermentation. These results showed that the use of pretreated softwood solids for the production of D-LA was possible without a long start-up phase, as long as these solids are only added to the reactor after cells are grown to an early exponential phase.

A longer lag phase was observed in SSF experiments with pretreated softwood solids compared with previous experiments using the liquid fraction of the steam-pretreated softwood. It was found that the lag phase could be reduced by adding the pretreated solids and starting the SSF when cells were in an early stage of exponential growth. The significant decrease demonstrated in this work is substantial when designing a process for an industrial scale bioprocess.

It is inferred that the decrease in inhibitory effect was likely due to a change in the interaction between the cells and the solid particles, possibly due to a changed EPS content as a function of the physiological status of the culture. It is necessary to study further the SSF of softwood for the production of D-LA by optimizing the enzymatic hydrolysis conditions to achieve maximum hydrolysis yield, using the liquid fraction instead of defined glucose to start the culture before the addition of the pretreated solids, and following downstream processing for product purification, for example. Nevertheless, this study sheds light on a type of inhibition that is not commonly addressed (cell-solid interaction) and opens the door to the utilization of softwood for biochemical production through SSF without the need for a detoxification step to decrease inhibition effects.