Abstract
Bioconversion process with a single target product often lacks economic competitiveness owing to incomplete use of raw material and high costs of downstream processing (DSP). Here, we show with the microbial conversion of crude glycerol that an integrated strain engineering and catalytic conversion of the so-called byproducts can greatly improve DSP and the process economy. Specifically, Clostridium pasteurianumwas first adapted to increased concentration of crude glycerol in a novel automatic laboratory evolution system. At m3 scale bioreactor the strain achieved a simultaneous production of 1,3-propanediol (PDO), acetic and butyric acids at 81.21, 18.72 and 11.09 g/L within only 19 h, respectively, representing the most efficient fermentation of crude glycerol to targeted products. A heterogeneous catalytic step was developed and integrated into the DSP process to obtain high-value methyl esters from acetic and butyric acids at high yields. The co-production of the esters also greatly simplified the recovery of PDO. For example, a cosmetic grade PDO (96% PDO) was easily obtained by a simple single-stage distillation process (with an overall yield more than 77%). This integrated approach provides an industrially attractive route for a complete use of the raw material with the simultaneous production of three appealing products which greatly improve the process economy and ecology.
KEYWORDS: Clostridium pasteurianum , automatic adaptive evolution, 1,3-Propanediol , organic acid esters, co-production
INTRODUCTION
1,3-Propanediol (PDO) is a vital value-added monomer which is wildly used in plastics industry for synthesizing polytrimethylene terephthalate (PTT), polytrimethylene naphthalate (PTN), polytrimethylene isophthalate (PTI), and thermoplastic polyester elastomer (Sun et al., 2018 ;Fokum et al., 2021 ). It also has extensive applications for manufacturing cosmetics, foods, medicines, detergents, solvents, glues and resins (Sabra et al., 2015 ). The global PDO market is expected to reach $776.3 million in 2022, with a compound annual growth rate (CAGR) of 10.4% from 2014 to 2022 (Zabed et al., 2019 ). Currently, PDO can be produced by both chemical and biological synthesis. Due to ecological concerns, limited petrochemical resources and the safety risks in production, the microbial production of PDO with many advantages such as mild reaction conditions, and using renewable bio-substrates, has gained growing interest in both academic and industrial fields.
In particular, one promising route for the efficient and safe bio-production of PDO is the anaerobic fermentation of glycerol using the non-pathogenic strain Clostridium pasteurianum (Groeger et al., 2016 ; Kaeding et al., 2015 ;Jensen et al., 2012 ). Previously, our group has successfully demonstrated the production of PDO by a newly isolated low-butanol-producing C. pasteurianum K1 in a fed-batch fermentation of crude glycerol at up to 1 m3 pilot scale, achieving a final PDO titer of 55 g/L, a yield of 0.52 g/g and an overall productivity of 2.3 g/L/h (Kaeding et al., 2015 ). The whole process was done under non-sterile conditions without N2 aeration and without using yeast extract in the medium, which demonstrated the robustness and cost-effectiveness for the PDO production by C. pasteurianum . However, in order to avoid possible growth inhibition caused by the toxicity of crude glycerol, low initial glycerol concentration (<30 g/L) in the medium and low glycerol feeding rate (<20g/L/h) were used in the fermentation, which highly limited the final PDO titer and productivity. As a by-product of biodiesel production, crude glycerol is cheap and abundant for the economical bioconversion of PDO (Laura et al., 2020 ), but it also contains a lot of impurities such as high degree of unsaturated fatty acids and soluble proteins which are reported to have strong inhibitory effect on the cell growth and glycerol utilization of C. pasteurianum (Venkataramanan et al., 2012 ;Samul et al., 2014 ). The low tolerance of the strain to crude glycerol is a major bottleneck hindering a high-level PDO bio-production. One solution to this bottleneck could be the metabolic engineering for direct strain improvement. However, the lack of adequate genetic engineering tools still limits its application on C. pasteurianum (Jensen et al., 2012; Schmitz et al., 2019 ). As an alternative, adaptive laboratory evolution (ALE) that mimics the natural adaptation of microorganisms to the artificial stress in the lab, could be helpful (Liang et al., 2020 ). Briefly, ALE subjects targeted strain to repeated or continuous cultivation under stress conditions for many generations to improve tolerance towards inhibiting environmental conditions (Sun et al., 2018 ). In recent years, many ALE experiments have been reported to improve the crude glycerol tolerance of the microbes for enhanced PDO production (Szymanowska-Powałowska, 2015 ;Zhang et al., 2019 ). For example, a highly crude glycerol tolerant C. butyricum was selected after more than 50 repeated cultivations in medium with serially increasing crude glycerol concentration from 30-110 g/L (Zhang et al., 2019 ). Fermentation with the evolved strain led to overall productivity of PDO increasing from 0.97 to 2.14 g/L/h by 114%. However, an obvious disadvantage of the traditional ALE experiment is that it requires the passaging of cells manually for many generations to enrich favorable genetic changes at the expense of a lot of manpower. Moreover, the timing of cell passage in ALE is usually not well-determined due to the lack of real-time monitoring of cell growth (Zhang et al., 2019 ;Alves et al., 2021 ). Therefore, it is necessary to develop a smart ALE device which can regulate the cell growth and conduct long-term ALE in an automatic manner for achieving better and more stable crude glycerol tolerance of C. pasteurianum.
Another bottleneck affecting the process economy of PDO bio-production from crude glycerol by C. pasteurianum is that the oxidation and reduction reactions of glycerol will take place simultaneously in the strain to maintain the redox and energy balance (Sabra et al., 2016 ). As a result, the conversion of glycerol to PDO will be inevitably accompanied by the formation of different types of organic acids, which leads to a theoretical maximum product yield lower than 0.72 mol PDO/mol glycerol (Dietz and Zeng, 2014 ). Specifically, acetic acid and butyric acid are the major by-products of glycerol metabolism in C. pasteurianum(Sabra et al., 2014 ;Zhang et al., 2021 ). In this case, one strategy that can further improve the process economy could be the integrated process design towards the co-production of PDO with multiple value-added products from the organic acids. In fact, many co-product processes have been developed recently for reducing the production cost of PDO from glycerol fermentation (Groeger et al., 2016 ; Suppuram et al., 2019 ;Huang et al., 2017 ). For instance, Huang et al . constructed a recombinant strain ofCorynebacterium glutamicum and creatively developed a co-production process for PDO and glutamate (Huang et al., 2017 ). The reducing equivalents generated during glutamate fermentation can be recycled for PDO production, which resulted in the maximum yield of 1.0 mol PDO/mol glycerol and 1.0 mol glutamate/mol glucose. Glutamate can be easily separated from PDO by acidification and crystallization in the fermentation broth. Another study done by Groeger et al . reported the co-production of PDO and butanol by the high-butanol producingC. pasteurianum DSM 525 with a co-substrate fermentation strategy using glucose and glycerol (Groeger et al., 2016 ). With gas stripping for in-situ butanol removal, 53.7 g/L PDO and 39.2 g/L butanol can be simultaneously produced. However, to date, very few studies have focused on the integrated process for the co-production of PDO and organic acids due to the difficulty of complete acid removal from PDO (Zhang et al., 2021 ). The fiber grade PDO for high quality polymer synthesis requires a purity > 99.5% without any organic acid residue (Kurian, 2005 ). To overcome this challenge, several industrially applicable methods have been proposed to separate the organic acids from fermentation broth, such as ion exchange (Adkesson et al., 2011 ), electro-dialysis (Gong et al., 2006 ;Wu et al., 2011 ) and two phase salting-out extraction (Song et al., 2013 ; Li et al., 2019 ). However, these methods produce a large amount of waste water containing the medium salts and organic acid salts (Zhang et al., 2021 ), which reduces the economic interest to recover the organic acids from the waste water.
Recent research from our group has demonstrated that acetate and butyrate in ammonium form can be completely separated from PDO and partially recovered from fermentation broth without producing any high-salt waste water (Zhang et al., 2021 ). The process involves ultrafiltration to remove cell and protein, evaporation to remove water, vacuum distillation under glycerol suspension to separate PDO and organic acids from the medium salts, where the free acids are liberated, followed by a second vacuum distillation to recover the free acids from PDO. Finally, alkaline hydrolysis to eliminate the PDO esters impurities, and the rectification of highly pure PDO were performed. Although the obtained PDO purity was over 99.6% with a satisfactory overall yield more than 76%, the total recovery yield of acetic and butyric acid were less than 50% due to the loss in water evaporation step and the esterification with PDO during the vacuum distillation step. Therefore, further smart design of DSP to improve the recovery yield of both PDO and organic acids from crude glycerol fermentation broth is urgently needed.
In this study, a novel fully automatic ALE system capable of online monitoring of cell growth was firstly developed to perform a long-term crude glycerol adaption of C. pasteurianum . This system can significantly increase the number of adaptation cycles during the ALE and thus shorten the time for obtaining a stable improved phenotype compared to the traditional manual adaption process. The improved PDO production by the adapted strain was further confirmed in non-sterilized fed-batch fermentations of crude glycerol using a simple medium without yeast extract. Moreover, a new DSP for the co-production of PDO, methyl acetate (MA) and methyl butyrate (MB) from the crude glycerol fermentation broth was proposed for the first time. By dissolving the concentrated broth in acidified methanol, followed by adding a heterogeneous catalyst Amberlyst 15, the medium salt can be separated due to salt crystallization, and the free acetic and butyric acid can be esterified with methanol simultaneously. The novelty of this strategy lies on the significant reduction in the boiling points of the acids and the reduction in the reactivity of the acids with PDO by converting the free acids to their corresponding value-added methyl esters, which is proved to be crucial for achieving high recovery yield for both PDO and organic acids from the crude glycerol fermentation broth.
MATERIALS AND METHODS
Materials
Crude glycerol was obtained from PT Musim Mas (Indonesia). It contained by weight 80% glycerol, 10.41% water, 3.9% ash, 2.21% non-glycerol organic matter, and 3.54% salts. The pH value of the crude glycerol was 6.8. The acidic ion exchange resin (Amberlyst-15) was purchased from Rohm and Haas Company (USA), which has a structure of macroporous polystyrene crosslinked with divinylbenzene. The resin has a dry weight capacity (min.) of 5.2 eq/kg (H+ form), moisture retention of 51-59 % (H+ form), and surface area of 20-40 m²/g. Pure glycerol (≥98%), standards of 1,3-PDO (≥98%), methyl acetate (≥99%), and methyl butyrate (≥99%), and all other reagents of analytical grade were commercially purchased from Carl Roth (Germany).