Increased Butyrate Production in Clostridium saccharoperbutylacetonicum from Lignocellulose-Derived Sugars

ABSTRACT

Butyrate is produced by chemical synthesis based on crude oil, produced by microbial fermentation, or extracted from animal fats (M. Dwidar, J.-Y. Park, R. J. Mitchell, and B.-I. Sang, The Scientific World Journal, 2012:471417, 2012, https://doi.org/10.1100/2012/471417). Butyrate production by anaerobic bacteria is highly favorable since waste or sustainable resources can be used as the substrates. For this purpose, the native hyper-butanol producer Clostridium saccharoperbutylacetonicum N1-4(HMT) was used as a chassis strain due to its broad substrate spectrum. BLASTp analysis of the predicted proteome of C. saccharoperbutylacetonicum N1-4(HMT) resulted in the identification of gene products potentially involved in acetone-butanol-ethanol (ABE) fermentation. Their participation in ABE fermentation was either confirmed or disproven by the parallel production of acids or solvents and the respective transcript levels obtained by transcriptome analysis of this strain. The genes encoding phosphotransacetylase (pta) and butyraldehyde dehydrogenase (bld) were deleted to reduce acetate and alcohol formation. The genes located in the butyryl-CoA synthesis (bcs) operon encoding crotonase, butyryl-CoA dehydrogenase with electron-transferring protein subunits α and β, and 3-hydroxybutyryl-CoA dehydrogenase were overexpressed to channel the flux further towards butyrate formation. Thereby, the native hyper-butanol producer C. saccharoperbutylacetonicum N1-4(HMT) was converted into the hyper-butyrate producer C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]. The transcription pattern following deletion and overexpression was characterized by a second transcriptomic study, revealing partial compensation for the deletion. Furthermore, this strain was characterized in pH-controlled fermentations with either glucose or Excello, a substrate yielded from spruce biomass. Butyrate was the main product, with maximum butyrate concentrations of 11.7 g·L⁻¹ and 14.3 g·L⁻¹, respectively. Minimal amounts of by-products were detected.

IMPORTANCE Platform chemicals such as butyrate are usually produced chemically from crude oil, resulting in the carry-over of harmful compounds. The selective production of butyrate using sustainable resources or waste without harmful by-products can be achieved by bacteria such as clostridia. The hyper-butanol producer Clostridium saccharoperbutylacetonicum N1-4(HMT) was converted into a hyper-butyrate producer. Butyrate production with very small amounts of by-products was established with glucose and the sustainable lignocellulosic sugar substrate Excello extracted from spruce biomass by the biorefinery Borregaard (Sarpsborg, Norway).

INTRODUCTION

Solvent-producing clostridia have been industrially relevant for decades. Nowadays, the production of various chemicals from renewable or sustainable resources is often preferred over petrochemical production, e.g., because of the lack of carcinogen carry-over, the possible use of waste as the substrate, and the high product specificity (1, 2). Therefore, different strains of solvent-producing bacteria were engineered to produce higher yields or to be more resistant to the product, e.g., butanol, which often exhibits cell toxicity (3–7).

Clostridium saccharoperbutylacetonicum N1-4(HMT) was isolated as a hyper-butanol producer (8) and used for the industrial production of butanol and acetone during the 1960s and 1970s (9–12). Later on, it was reintroduced as a new species (13), and its genome was sequenced (14, 15). Sequencing revealed that its genome is the largest known for solventogenic clostridia (16). Similar to Clostridium acetobutylicum, it has a biphasic growth pattern consisting of an acidogenic and a solventogenic growth phase (17). During the acidogenic growth phase, acetate and butyrate are formed, whereas very little lactate is produced (16). During the solventogenic growth phase, acetone, butanol, and ethanol are formed in the so-called acetone-butanol-ethanol (ABE) fermentation (16). This strain has a comprehensive substrate spectrum (8, 13), and therefore, it was investigated for catabolite repression (18). It was previously optimized for butanol (19–24) and hydrogen (25–28) production and used as a chassis for the synthesis of heterologous products such as R-1,3-butanediol, caproate, and hexanol (2, 29). Organic acids in general are important products derivable from clostridial metabolism since they are essential precursors in the production of esters, plastics, or other high-value products (30, 31).

In the close relative C. acetobutylicum, ABE fermentation has been explored in detail. Acetyl-CoA is converted to acetate via acetyl phosphate by the enzymes phosphotransacetylase and acetate kinase (32–34). Thiolase can convert acetyl-CoA to acetoacetyl-CoA (35). This can be converted to acetone via acetoacetate by CoA transferase and acetoacetate decarboxylase or to butyryl-CoA via 3-hydroxybutyryl-CoA and crotonyl-CoA by 3-hydroxybutyryl-CoA dehydrogenase, crotonase, and butyryl-CoA dehydrogenase (36). The latter three enzymes are encoded in the so-called butyryl-CoA synthesis (bcs) operon (36). Butyryl-CoA is converted to butyrate via butyryl phosphate by phosphotransbutyrylase (ptb) and butyrate kinase (buk) (37, 38). Acetyl-CoA and butyryl-CoA are converted to ethanol and butanol via acetaldehyde and butyraldehyde by mono- and bifunctional aldehyde and alcohol dehydrogenases, respectively (39). The gene encoding the bifunctional aldehyde/alcohol dehydrogenase (adhE) together with the genes encoding the CoA transferase (ctfA and ctfB) are located in the so-called solvent (sol) operon (39). The gene encoding the acetoacetate decarboxylase (adc) is located adjacent to ctfB, with converging directions of transcription (39). This cluster is located on the megaplasmid pSOL1 of C. acetobutylicum (40, 41).

The genes encoding the above-described enzymes were also identified in the genome sequence of C. saccharoperbutylacetonicum, and some of them are present in several copies (16, 42). In contrast to C. acetobutylicum, the bcs operon of C. saccharoperbutylacetonicum is adjacent to a gene coding for an alcohol dehydrogenase, and the sol operon consists of the genes encoding a butyraldehyde dehydrogenase (bld), a CoA transferase, and an acetoacetate decarboxylase (16, 43).

There are several methods available for genetic modification in C. saccharoperbutylacetonicum (44–47). In this study, the genes encoding enzymes involved in ABE fermentation of C. saccharoperbutylacetonicum N1-4(HMT) were identified using BLASTp and transcriptome analyses. Furthermore, overproduction and subsequent optimization of native butyrate formation were performed using the obtained data for the determination of deletion and overexpression targets.

Since this strain was isolated for commercial butanol production, the formation of butyrate by the best-performing constructed strain was also tested by pH-controlled fermentations using glucose and Excello as carbon sources. Excello is a lignocellulosic hydrolysate extracted from spruce biomass containing a monomeric sugar mixture, and it is a by-product of the lignin extraction process performed by the largest sustainable biorefinery, Borregaard (Sarpsborg, Norway) (48).

RESULTS

Identification of deletion and overexpression targets.

Starting from annotated and previously characterized proteins (7, 16, 43, 49), the translated genome of C. saccharoperbutylacetonicum N1-4(HMT) was searched for homologous proteins using the BLASTp algorithm provided by the Integrated Microbial Genomes and Microgenomes (IMG/M) system of the U.S. Department of Energy’s Joint Genome Institute (50) (see Fig. S1 in the supplemental material). Proteins can be regarded as paralogues above ∼30% identity and a query coverage of ≥50% (51).

BLASTp analyses revealed one phosphotransacetylase (pta [Cspa_c13010]) and two phosphotransbutyrylases (ptb [Cspa_c02520] and Cspa_c21460), whereas two acetate kinases (Cspa_c13020 and Cspa_c30040) and three butyrate kinases (buk [Cspa_c02530], Cspa_c21470, and Cspa_c53320) are present in the genome of C. saccharoperbutylacetonicum (Fig. S1). Five butyryl-CoA dehydrogenases were identified, with the enzymes encoded by Cspa_c04340 and Cspa_c20900 sharing 98% identity, which is not the case for the enzymes encoded by Cspa_c03050, Cspa_c12400, and Cspa_c41420 (Fig. S1). Three 3-hydroxybutyryl dehydrogenases were identified, two of which share 73% identity, namely, the enzymes encoded by Cspa_c04370 and Cspa_c20880, whereas the enzyme encoded by Cspa_c22690 shares <40% identity with these enzymes (Fig. S1). All identified CoA transferases are encoded in operons of genes encoding the α and β subunits (Fig. S1). Seventeen mono- and bifunctional aldehyde and alcohol dehydrogenases were found as paralogues for either the alcohol dehydrogenase located adjacent to the bcs operon encoded by Cspa_c04380 or the butyraldehyde dehydrogenase (bld) located in the sol operon encoded by Cspa_c56880 (Fig. S1). A third cluster with identities of around 35% was found, with some genes encoding aldehyde dehydrogenases (Fig. S1).

Subsequently, transcript levels were determined in the early exponential growth phase (after 12 h, at an optical density at 600 nm [OD₆₀₀] of 0.7), after the butyrate peak (after 24 h, at an OD₆₀₀ of 3.6), after the acetate peak (after 29.5 h, at an OD₆₀₀ of 6.1), and in the stationary growth phase (after 52 h, at an OD₆₀₀ of 6.7) by transcriptome analysis (Fig. S2) to identify possible enzymes to compensate for the loss of genes involved in central metabolism. Glucose was used as a carbon source, and samples sent for sequencing were chosen based on growth curves and analysis of the culture samples by gas chromatography (GC) for the butyrate peak and the acetate peak (Fig. S2). Figure 1 and Fig. S3 show the transcript levels of genes whose gene products are involved in ABE fermentation and central metabolism of C. saccharoperbutylacetonicum, respectively. All determined log₂ fold change values compared to transcription in the early exponential growth phase and the normalized read counts (NRCs) calculated by DeSeq2 (52) of the proteins identified via BLASTp are shown in Fig. 1.

FIG 1.

Transcriptomic data from wild type C. saccharoperbutylacetonicum. Data are shown as log₂ fold changes compared to the early exponential growth phase and as normalized read counts determined by DeSeq2 (52). Filled mauve bars, after the butyrate peak; filled orange bars, after the acetate peak; filled yellow bars, stationary growth phase compared to the early exponential growth phase; open purple bars, early exponential growth phase; open mauve bars, after the butyrate peak; open orange bars, after the acetate peak; open yellow bars, stationary growth phase. (Data modified from reference 81.) Pta, phosphotransacetylase; Ptb, phosphotransbutyrylase (of interest is Cspa_c02520); Ack, acetate kinase; Buk, butyrate kinase (of interest is Cspa_c02530); Crt, crotonase; Hbd, 3-hydroxybutyryl-CoA dehydrogenase (of interest is Cspa_c04370); CtfA, CoA transferase subunit α; CtfB, CoA transferase subunit β; Adc, acetoacetate decarboxylase; Ald/Adh/AdhE, mono- and bifunctional aldehyde and alcohol dehydrogenases (of interest is butyraldehyde dehydrogenase encoded by Cspa_c56880); Bcd, butyryl-CoA dehydrogenase; Ldh, lactate dehydrogenase; ThlA, thiolase.

Acetate production is catalyzed by phosphotransacetylase (Pta) and acetate kinase (Ack) encoded by Cspa_c13010 and Cspa_c13020, which show the highest expression levels at the sampling point after the butyrate peak, which is easily seen by the log₂ fold changes and the NRC data (Fig. 1). Butyryl-CoA is produced by thiolase (Thl), 3-hydroxybutyryl-CoA dehydrogenase (Hbd), crotonase (Crt), and butyryl-CoA dehydrogenase (Bcd) encoded by Cspa_06180, Cspa_c04370, Cspa_c04330, Cspa_c04340, and Cspa_c20900, which are transcribed at the highest levels during the early exponential growth phase, decreasing slightly after the butyrate peak (Fig. 1). Furthermore, the transcription of the involved genes is decreased strongly after the acetate peak as well as during the stationary growth phase, as indicated by the negative log₂ fold change values (Fig. 1). Paralogues identified by BLASTp analysis (Fig. S1) show very little transcription and regulation in other growth phases, suggesting no involvement in butyrate production (Fig. 1). Starting from butyryl-CoA, butyrate is formed by phosphotransbutyrylase (Ptb) and butyrate kinase (Buk) (encoded by Cspa_c02520 and Cspa_c02530), as they are the most highly transcribed during the early exponential growth phase and shortly after the butyrate peak, decreasing sharply after the acetate peak, whereas the transcription of the butyrate kinase encoded by Cspa_c21460 starts after the butyrate peak (Fig. 1). Solvents, i.e., acetone, butanol, and ethanol, were formed after the acetate peak and during the stationary growth phase (Fig. S2). The genes encoding CoA transferase and acetoacetate decarboxylase organized in the sol operon (Cspa_c56890, Cspa_c56900, and Cspa_c56910) are transcribed after the acetate peak and during the stationary growth phase (Fig. 1). Although the log₂ fold change of Cspa_c56880 (bld) does show comparably low regulation throughout all growth phases compared to paralogues, the NRC, at >1,300,000, is the highest for all identified alcohol and aldehyde dehydrogenases (Fig. 1). Tenfold lower transcription levels were detected for the enzymes encoded by Cspa_c04380, Cspa_c32270, and Cspa_c41390 (Fig. 1). All other potential aldehyde and alcohol dehydrogenases show at least 100-fold lower transcription levels than Cspa_c56880 (Bld) (Fig. 1).

Characterization of deletion and overexpression strains.

Based on the BLASTp or the transcriptome results, genes pta (Cspa_c13010) (encoding phosphotransacetylase) and/or bld (Cspa_c56880) (encoding butyraldehyde dehydrogenase) were deleted to construct butyrate-producing strains. The deletions were carried out in frame with the first and last 45 bp of the coding region remaining in the chromosome to prevent polar effects on the transcription of the respective operons. The resulting strains, C. saccharoperbutylacetonicum Δpta, C. saccharoperbutylacetonicum Δbld, and C. saccharoperbutylacetonicum ΔbldΔpta, were characterized in a growth experiment (Fig. 2). Throughout the growth experiment, the pH dropped and increased again for C. saccharoperbutylacetonicum Δpta and C. saccharoperbutylacetonicum wild type, whereas it dropped without a subsequent increase for C. saccharoperbutylacetonicum Δbld and C. saccharoperbutylacetonicum ΔbldΔpta. Furthermore, strains C. saccharoperbutylacetonicum Δbld and C. saccharoperbutylacetonicum ΔbldΔpta produced meager amounts of the solvents acetone, butanol, and ethanol (up to 37 mM, compared to 189 mM for C. saccharoperbutylacetonicum wild type) and high butyrate levels (up to 46 mM, compared to 10 mM for C. saccharoperbutylacetonicum wild type) (Fig. 2). Moreover, C. saccharoperbutylacetonicum ΔbldΔpta produced very low acetate levels (up to 3 mM, compared to 21 mM for C. saccharoperbutylacetonicum wild type) (Fig. 2). C. saccharoperbutylacetonicum Δpta produced higher solvent (up to 151 mM acetone, butanol, and ethanol) and lower acid (up to 8 mM acetate and 9 mM butyrate) levels than C. saccharoperbutylacetonicum Δbld and C. saccharoperbutylacetonicum ΔbldΔpta but lower solvent and acetate levels than C. saccharoperbutylacetonicum wild type (Fig. 2).

FIG 2.

Growth experiment using different C. saccharoperbutylacetonicum deletion strains. Bacterial growth (OD₆₀₀), pH, and concentrations of the substrates and products throughout the experiment are shown. Open gray diamonds, wild type C. saccharoperbutylacetonicum; orange triangles, C. saccharoperbutylacetonicum Δpta; inverted red triangles, C. saccharoperbutylacetonicum Δbld; black-and-white circles, C. saccharoperbutylacetonicum ΔbldΔpta. Error bars represent standard deviations (n = 3). The colored asterisks mark significant differences between the strain represented by the curve and the strain represented by the color of the asterisk (P < 0.05 by one-way analysis of variance [ANOVA]). (Modified from reference 81.)

The next objective was to channel the carbon flux further toward butyrate. For this purpose, the respective genes were cloned on plasmids under the control of the lactose-inducible promoter P_bgaL. According to the transcription patterns and transcript abundances, the following genes are responsible for butyrate formation: the bcs operon encoding crotonase (Crt), butyryl-CoA dehydrogenase (Bcd), electron-transferring protein subunits β and α (EtfB and EtfA), and 3-hydroxybutyryl-CoA dehydrogenase (Hbd) (Cspa_c04330 to Cspa_c04370) as well as the ptb-buk operon encoding phosphotransacetylase and butyrate kinase (Cspa_c02520 and Cspa_c02530). The resulting plasmids together with a vector control were used to transform C. saccharoperbutylacetonicum Δbld Δpta.

The resulting strains, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (vector control), C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_PB_P_bgaL] (ptb-buk operon), C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_PB_P_bgaL] (bcs operon), and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS-PB_P_bgaL] (bcs and ptb-buk operon), were characterized in a growth experiment (Fig. 3). All newly constructed strains grew to a low maximal OD₆₀₀ without an increase of the pH toward the end of the growth experiment compared to C. saccharoperbutylacetonicum wild type and C. saccharoperbutylacetonicum ΔbldΔpta. The acetone, butanol, ethanol, and acetate levels of all C. saccharoperbutylacetonicum ΔbldΔpta-based strains were decreased compared to those of C. saccharoperbutylacetonicum wild type, with induced strains C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS-PB_P_bgaL] producing the smallest amounts of the solvents acetone, butanol, and ethanol (85 mM and 96 mM, compared to 184 mM for C. saccharoperbutylacetonicum wild type). Butyrate levels were increased for all C. saccharoperbutylacetonicum ΔbldΔpta-based strains, with the induced strains C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS-PB_P_bgaL] producing the highest butyrate levels, 59 mM (13.14 mM butyrate·OD₆₀₀⁻¹ and 0.39 mM butyrate·mM glucose⁻¹) and 56 mM, respectively, compared to 10 mM (14.35 mM butyrate·OD₆₀₀⁻¹ and 0.36 mM butyrate·mM glucose⁻¹) for C. saccharoperbutylacetonicum wild type (Fig. 3). The identity of all strains was verified by 16S rRNA gene sequencing, amplification of the deletion regions, and restriction digestion of the respective plasmids.

FIG 3.

Growth experiment using C. saccharoperbutylacetonicum strains optimized for butyrate production. Bacterial growth (OD₆₀₀), pH, and concentrations of the substrates and products throughout the experiment are shown. Open gray diamonds, C. saccharoperbutylacetonicum wild type; black-and-white circles, C. saccharoperbutylacetonicum ΔbldΔpta; black squares, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (vector control); upward- and downward-pointing lilac triangles, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_PB_P_bgaL]; open and closed blue circles, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]; leftward- and rightward-pointing purple triangles, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS-PB_P_bgaL]; dashed lines with empty symbols, non-induced strains; straight lines with half-filled and filled symbols, induced strains; dashed-dotted red lines, time of induction with 20 mM lactose; upper lines in the lactate-lactose panel, lactose. Error bars represent standard deviations (n = 3). (Modified from reference 81.) The colored asterisks mark significant differences between the strain represented by the curve and the strain represented by the color of the asterisk, and empty asterisks represent the vector strain or noninduced strains (P < 0.05 by one-way ANOVA).

Characterization of strains optimized for butyrate production.

The strain C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] was further characterized by transcriptome analysis during the exponential and early stationary growth phases. The results of the growth experiment with analyses of substrate consumption and product formation are shown in Fig. S4. Figure 4 shows the evaluated transcriptomic data for C. saccharoperbutylacetonicum wild type, C. saccharoperbutylacetonicum ΔbldΔpta, and non-induced and induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]. The data show that the transcription of the sol operon and the pta-ack operon is not influenced throughout the growth experiment compared to the wild type, except for the deleted genes bld and pta (Cpa_c56880 and Cspa_c13010) (Fig. 4). In addition, the overexpression of the bcs operon occurs only in induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], and the overexpression is clearly detectable during the exponential and early stationary growth phases (Fig. 4). Analysis of the transcriptomic data for all previously identified aldehyde and alcohol dehydrogenases revealed that the highest detectable expression level after the loss of bld (Cspa_c56880) is measured for Cspa_c37180, increasing up to 5.6-log₂ fold in C. saccharoperbutylacetonicum ΔbldΔpta-based strains. The NRC measured for Cspa_c37180 is equally as high as that for bld (Cspa_c56880) in C. saccharoperbutylacetonicum wild type (Fig. 4).

FIG 4.

Transcriptomic data from the optimized butyrate producer C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]. Data are shown as log₂ fold changes compared to C. saccharoperbutylacetonicum wild type and as normalized read counts determined using DeSeq2 (52). Black-and-white-striped bars, C. saccharoperbutylacetonicum ΔbldΔpta; closed light-blue bars, non-induced, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]; closed dark-blue bars, induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], open gray bars, C. saccharoperbutylacetonicum wild type; open black bars, C. saccharoperbutylacetonicum ΔbldΔpta; open light-blue bars, non-induced, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] open dark-blue bars, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], induced. Error bars represent standard deviations (n = 3). Sampling during the exponential/early stationary growth phase are shown (OD₆₀₀ of C. saccharoperbutylacetonicum wild type, 3.25/6.93; OD₆₀₀ of C. saccharoperbutylacetonicum ΔbldΔpta, 3.82/5.35; OD₆₀₀ of non-induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], 1.66/2.65; OD₆₀₀ of induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], 1.58/2.99).

Fermentations using C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] and controls.

C. saccharoperbutylacetonicum wild type, C. saccharoperbutylacetonicum Δbld Δpta [pMTL83151], and C. saccharoperbutylacetonicum ΔbldΔpta[pMTL83151_BCS_P_bgaL] (only with induction) strains were used in fermentations with controlled pH in 1-L fermentors on glucose and Excello. Figure 5 shows the fermentation results with glucose as the substrate. Glucose was depleted by all strains at the end of the growth experiment. GC analysis revealed sharp increases and decreases for acetone, butanol, and ethanol for C. saccharoperbutylacetonicum wild type. C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] yielded very low acetone, butanol, and acetate levels but yielded ethanol levels similar to those of C. saccharoperbutylacetonicum wild type (Fig. 5). The butyrate levels of C. saccharoperbutylacetonicum ΔbldΔpta-based strains were drastically higher than those of C. saccharoperbutylacetonicum wild type. C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] produced higher butyrate levels (133.3 mM [21.02 mM butyrate·OD₆₀₀⁻¹ and 0.63 mM butyrate·mM glucose⁻¹]) than C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (118.4 mM [16.06 mM butyrate·OD₆₀₀⁻¹ and 0.49 mM butyrate·mM glucose⁻¹]), and butyrate production did not reach a plateau at the end of the fermentation (Fig. 5). The concentration of the inductor lactose decreased during the fermentation. Furthermore, all strains produced approximately the same amounts of hydrogen (Table 1). The produced carbon dioxide decreased from C. saccharoperbutylacetonicum wild type to C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] (Table 1). C. saccharoperbutylacetonicum wild type produced twice as much methane as C. saccharoperbutylacetonicum ΔbldΔpta-based strains (Table 1). Moreover, approximately 10-fold higher levels of ethanol were detected in the headspace for C. saccharoperbutylacetonicum wild type than for C. saccharoperbutylacetonicum ΔbldΔpta-based strains (Table 1). In conclusion, the ethanol levels detected in the liquid samples and from the headspace revealed that C. saccharoperbutylacetonicum ΔbldΔpta-based strains produced only 35% of the ethanol produced by C. saccharoperbutylacetonicum wild type (Table 1; Fig. S5). The identity of all strains was verified by restriction digestion of the plasmid if present, 16S rRNA gene sequencing, and amplification of the deletion regions.

FIG 5.

Controlled fermentation on glucose using C. saccharoperbutylacetonicum strains optimized for butyrate production. Bacterial growth (OD₆₀₀), pH, and concentrations of the substrates and products throughout the experiment are shown. Open gray diamonds, C. saccharoperbutylacetonicum wild type; black squares, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (vector control); blue circles, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]; dashed-dotted blue line, time of induction with 20 mM lactose; upper line in the lactate-lactose panel, lactose. Error bars represent standard deviations (n = 3). The colored asterisks mark significant differences between the strain represented by the curve and the strain represented by the color of the asterisk (P < 0.05 by one-way ANOVA).

TABLE 1.

Additional data on growth experiments with C. saccharoperbutylacetonicum strains grown in batch and pH-controlled fermentations^a

Fermentation and strain	Mean product concn (mM) ± SD
	HBu	HAc	Sol	H2	CO2	CH4	EtOH (g)
Batch, glucose
C. saccharoperbutylacetonicum wild type (n = 3)	10.0 ± 0.4	20.7 ± 4.0	187.2 ± 3.6
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (n = 3)	44.2 ± 2.3	11.7 ± 0.6	152.0 ± 2.9
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_PbgaL],  induced (n = 3)	59.0 ± 4.4	7.1 ± 0.4	85.4 ± 6.6
Controlled, glucose
C. saccharoperbutylacetonicum wild type (n = 1)	39.1	31.8	71.4	7.8	137.2	5.8	10.8
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (n = 3)	118.4 ± 12.2	11.1 ± 0.5	8.0 ± 0.7	7.9 ± 0.6	106.8 ± 16.4	2.8 ± 0.2	1.1 ± 0.1
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_PbgaL],  induced (n = 3)	133.2 ± 2.6	12.6 ± 0.8	8.8 ± 0.4	7.9 ± 0.1	91.4 ± 3.8	2.5 ± 0.0	1.0 ± 0.1
Controlled, Excello
C. saccharoperbutylacetonicum wild type (n = 1)	38.6	48.3	134.6	6.4	132.8	11.0	22.8
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (n = 1)	162.5	21.6	16.2	5.7	57.3	4.4	2.6
C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_PbgaL], induced (n = 2)	132.2 ± 1.4	23.1 ± 4.4	11.7 ± 3.6	4.0 ± 0.1	25.2 ± 3.7	3.8 ± 0.0	1.4 ± 0.1

^aThe product concentrations depicted are the maximum concentrations of the respective products or the accumulated stripped gas or solvent. HBu, butyrate; HAc, acetate; Sol, total solvents (i. e. acetone, butanol, and ethanol); EtOH (g), stripped ethanol.

The fermentations of C. saccharoperbutylacetonicum wild type, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151], and C. saccharoperbutylacetonicum ΔbldΔpta[pMTL83151_BCS_P_bgaL] using Excello, with C. saccharoperbutylacetonicum wild type grown on glucose as a control, are shown in Fig. 6. This showed behavior similar to that of the strain in Fig. 5. C. saccharoperbutylacetonicum wild type grown on Excello produced higher solvent (acetone, butanol, and ethanol) and acetate levels but lower butyrate levels than with growth on glucose. Most of the sugars contained in Excello, i.e., glucose, mannose, galactose, xylose, fructose, and arabinose, were depleted by all strains at the end of the growth experiment (Fig. S6). C. saccharoperbutylacetonicum ΔbldΔpta-based strains produced very small amounts of solvents but high acid concentrations compared to fermentation on glucose (Fig. 5 and 6). The acid production of C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] reached a plateau at the end of the fermentation (162.5 mM butyrate [33.50 mM butyrate·OD₆₀₀⁻¹ and 0.67 mM butyrate·mM sugar⁻¹]), whereas the acid production of C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] continued to rise steeply, ending with 132.22 mM butyrate (24.76 mM butyrate·OD₆₀₀⁻¹ and 0.60 mM butyrate·mM sugar⁻¹) (Fig. 6). The identity of all strains was verified by amplification of the deletion regions, 16S rRNA gene sequencing, and restriction digestion of the plasmid.

FIG 6.

Controlled fermentation on Excello using C. saccharoperbutylacetonicum strains optimized for butyrate production. Bacterial growth (OD₆₀₀), pH, and concentrations of the main substrate (i.e., glucose) and products throughout the experiment are shown. Open gray diamonds, C. saccharoperbutylacetonicum wild type (glucose) (n = 1); gray-and-white diamonds, C. saccharoperbutylacetonicum wild type (Excello) (n = 1); black squares, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (vector control) (n = 1); blue circles, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] (n = 2) (error bars represent standard deviations); dashed-dotted blue lines, time of induction with 20 mM lactose and inoculation of wild type C. saccharoperbutylacetonicum.

The levels of hydrogen and methane production were highest for C. saccharoperbutylacetonicum wild type and lowest for C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] (Table 1). C. saccharoperbutylacetonicum wild type produced the highest carbon dioxide levels, whereas C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] produced the lowest carbon dioxide levels (Table 1). The levels of ethanol are comparable for the C. saccharoperbutylacetonicum wild type, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151], and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] strains in the samples taken during fermentation (Fig. 6). Analysis of the headspace revealed an increase in ethanol formation by C. saccharoperbutylacetonicum wild type grown on Excello, in contrast to growth on glucose (Fig. S7). C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] produced lower ethanol levels than C. saccharoperbutylacetonicum wild type, with smaller amounts stripped from the fermentation broth (Fig. S7).

DISCUSSION

According to BLASTp and transcriptome analyses, deletion targets were chosen (Fig. 1; see also Fig. S1 and S3 in the supplemental material). The transcript levels of Cspa_c13010 (pta), coding for phosphotransacetylase, were highest after the butyrate peak and, hence, before the acetate peak (Fig. 1; Fig. S3). In addition, pta (Cspa_c13010) was previously deleted, leading to a decrease in acetate production (49). Cspa_c56880 (bld), coding for butyraldehyde dehydrogenase, was chosen for deletion because it exhibited the highest NRC, at 1,300,000, and its transcript level was high after the acetate peak at the verge of and during the stationary growth phase (Fig. 1; Fig. S3). Additionally, the overexpression of the sol operon containing bld had previously resulted in increases in acetone, butanol, and ethanol (53), and the deletion of bld resulted in a decrease in solvent formation (42).

The resulting single- and double-deletion strains C. saccharoperbutylacetonicum Δpta, C. saccharoperbutylacetonicum Δbld, and C. saccharoperbutylacetonicum ΔbldΔpta were characterized. Analysis showed that C. saccharoperbutylacetonicum Δpta was still able to produce acetate (Fig. 2). This effect was previously described for C. acetobutylicum as well as for C. saccharoperbutylacetonicum and was postulated to be a result of the broad substrate spectrum of the phosphotransbutyrylase (38, 49, 54–56). Since the gene encoding the phosphotransbutyrylase (ptb) was not deleted in this strain, this is probably also true for C. saccharoperbutylacetonicum Δpta. Total solvent formation consisting of acetone, butanol, and ethanol of C. saccharoperbutylacetonicum Δbld and C. saccharoperbutylacetonicum ΔbldΔpta was decreased to 17% and 19% of the C. saccharoperbutylacetonicum wild type levels, respectively (Fig. 2), showing results similar to those described previously by Herman and coworkers for their deletion strain (42). This suggests that all other identified aldehyde and alcohol dehydrogenases together are responsible or can compensate for approximately 20% of the alcohol formation (Fig. 2). However, the growth experiment employing strains further optimized for butyrate formation revealed differences in this respect (Fig. 3). C. saccharoperbutylacetonicum ΔbldΔpta produced 19% of the total solvent level of C. saccharoperbutylacetonicum wild type in the first growth experiment (Fig. 2), whereas it produced 73% of the total solvent level of C. saccharoperbutylacetonicum wild type in the second growth experiment (Fig. 3). This suggests that the bacteria are able to compensate for the loss of bld, coding for butyraldehyde dehydrogenase, by altering the expression of other genes encoding aldehyde and alcohol dehydrogenases (38, 54, 57). The same is valid for acetate formation. In the first growth experiment, C. saccharoperbutylacetonicum Δpta formed 36% and C. saccharoperbutylacetonicum ΔbldΔpta formed 12% of the maximal acetate level of wild type C. saccharoperbutylacetonicum (Fig. 2), whereas in the second growth experiment, 43% of the acetate level of C. saccharoperbutylacetonicum wild type was formed by C. saccharoperbutylacetonicum ΔbldΔpta (Fig. 3). Since it was previously described that phosphotransbutyrylase (encoded by ptb) can compensate partially for the loss of phosphotransacetylase (encoded by pta), this compensation seems to be expanded by subculturing and adaptation for growth experiments (56, 57).

Genes located in the bcs (i.e., crotonase, butyryl-CoA dehydrogenase, electron-transferring protein subunits β and α, and 3-hydroxybutyryl-CoA dehydrogenase) and ptb-buk (i.e., phosphotransbutyrylase and butyrate kinase) operons were chosen for overexpression to channel the metabolic flux further toward butyrate formation. The strains further optimized for butyrate production were characterized in a growth experiment with and without the induction of gene expression (Fig. 3). Only the induced strains harboring additional copies of the bcs operon, namely, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], showed an increase in butyrate production compared to C. saccharoperbutylacetonicum ΔbldΔpta (Fig. 3). From this, it was concluded that only the genes located in the bcs operon can channel metabolic flux toward butyrate or that one of the encoded proteins is the bottleneck in butyrate formation.

A second transcriptomic analysis was carried out to test for the successful overexpression of the bcs operon and potential genetic compensation (Fig. 4; Fig. S4). With these data, it was shown that the deletion of bld and pta (Cspa_c56880 and Cspa_c13010) was successful and did not result in expression changes in the remaining operons throughout the exponential and early stationary growth phases of C. saccharoperbutylacetonicum ΔbldΔpta and non-induced and induced C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL]. Investigation of the transcription levels of the identified aldehyde and alcohol dehydrogenases revealed that the aldehyde-alcohol dehydrogenase (encoded by adhE [Cspa_c37180]) increased up to the levels of bld in C. saccharoperbutylacetonicum wild type (Fig. 4). This result suggests compensation for the loss of bld by adhE, subsequently allowing the detected alcohol formation (57). This compensational mechanism was not characterized previously for C. saccharoperbutylacetonicum.

C. saccharoperbutylacetonicum wild type, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151], and the best butyrate producer, C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], were cultivated on glucose with pH control (Fig. 5). This resulted in increases in butyrate levels of up to 125% by C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] compared to the results of the batch experiment (Fig. 3 and 5). The fermentation of the strains on Excello resulted in the highest butyrate levels for C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (Fig. 6).

The batch experiment revealed an increase of 340% in the butyrate levels of C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] compared to C. saccharoperbutylacetonicum wild type and a significant increase of up to 30% for C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] compared to C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] (Fig. 3). These results could not be confirmed by fermentation on glucose since the increase in butyrate production by C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] compared to C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] is not significant (Fig. 5). The increase of 280% in butyrate production by both optimized strains compared to C. saccharoperbutylacetonicum wild type is statistically significant (Fig. 5). Growth on Excello showed that this substrate leads to higher acetate and ethanol levels and lower butyrate levels for C. saccharoperbutylacetonicum wild type than with growth on glucose (Fig. 6). C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] grown on Excello showed an increase of 500% compared to C. saccharoperbutylacetonicum wild type (Fig. 6). In contrast to the growth experiment (Fig. 3), the optimized strain C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] did not perform better than C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] in fermentations on glucose and Excello (Fig. 5 and 6). The highest absolute butyrate levels detected during fermentation were 132 mM and 133 mM for C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] grown on glucose and Excello and 162 mM for C. saccharoperbutylacetonicum ΔbldΔpta grown on Excello (Fig. 5 and 6). These values are lower than the values reported for native producers such as Clostridium tyrobutyricum or Clostridium butyricum (58, 59), whereas the yields of the optimized strains in this study are similar to those of the natural butyrate producers, with very few by-products (59). The lack of differences in C. saccharoperbutylacetonicum ΔbldΔpta and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] probably results from the pH control. This could keep the expression of the bcs operon active so that additional transcription originating from the plasmid (∼10-fold increase according to Fig. 4) does not matter for butyrate production. The constant pH could be the reason for the constant NADH/NAD⁺ ratio inside the cells, resulting in the continuous activation of the promoter of the bcs operon by the transcription factor Rex (60–62). Detection of gas stripping was established only for ethanol but not for acetone and butanol (Fig. 5 and 6; Fig. S5 and S7). Yao and coworkers cultivated C. saccharoperbutylacetonicum on combinations of different sugars in batch fermentations and showed that the patterns of solvent production, i.e., acetone, butanol, and ethanol, differed for different combinations (18). Ezeji and coworkers established gas stripping for the purification of the solvents acetone, ethanol, and butanol from fermentation broth (63). They measured various specificities for the different compounds, which are affected by the different concentrations of the respective compounds and cell density (63). Therefore, no statement can be made regarding the absolute levels of acetone and butanol during fermentation.

Hydrogen and carbon dioxide formation did not differ for C. saccharoperbutylacetonicum wild type between growth on glucose and Excello, while it in fact differed for C. saccharoperbutylacetonicum ΔbldΔpta-based strains (Table 1). This difference might be due to the additional sugars and carbon sources contained in Excello that are metabolized via different pathways, i.e., the glycolysis, pentose phosphate, and phosphoketolase pathways (64). Furthermore, twice as much methane is formed during fermentation on Excello as on glucose (Table 1). Methane formation by C. saccharoperbutylacetonicum can be explained by a reductive decarboxylation mechanism resulting from pyruvic phosphoroclasm, as was described previously for Clostridium pasteurianum (65, 66). The smaller amounts formed by the strains based on C. saccharoperbutylacetonicum ΔbldΔpta could be due to the modifications made in the genome or gene products located on the plasmid backbone as C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151] and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] produced comparable amounts of methane (Table 1). Further optimization of butyrate production may be possible by integrating a second copy of the bcs operon or an exchange of the native promoter region upstream of the bcs operon in the genome of C. saccharoperbutylacetonicum ΔbldΔpta (2). Furthermore, other genes could be deleted to prevent compensation for the formation of acetone, butanol, and ethanol and to channel metabolic flux further toward butyrate, e.g., spo0A, encoding the global regulator Spo0A (47, 57).

MATERIALS AND METHODS

Bacterial strains and cultivation conditions.

Bacterial strains and plasmids used in this study are listed in Table 2. C. saccharoperbutylacetonicum strains were grown in clostridial growth medium (CGM), modified from methods described previously by Wiesenborn et al. (67). In brief, the medium was prepared by dissolving 252.3 mM glucose·H₂O, 5.5 mM KH₂PO₄, 4.3 mM K₂HPO₄, 15.1 mM (NH₄)₂SO₄, 2.9 mM MgSO₄·7H₂O, 0.059 mM MnSO₄·H₂O, 0.036 mM FeSO₄·7H₂O, 17.1 mM NaCl, 14.9 mM asparagine·H₂O, 0.5% (wt/vol) yeast extract, and 0.0044 mM resazurin in anaerobic water in an anaerobic cabinet with an N₂-H₂ (95:5) atmosphere. For solid medium, 1.5% (wt/vol) agar was added before autoclaving and after dissolving. Anaerobic water was prepared by boiling demineralized water for 20 min and cooling it in an ice bath under N₂ sparging. It was stored in an anaerobic cabinet until further use. Optimized synthetic medium (OMS) was used for growth experiments with C. saccharoperbutylacetonicum strains, modified according to methods described previously by Standfest (68). In brief, the medium contained 201.8 mM glucose·H₂O, 102.5 mM 2-(N-morpholino)ethanesulfonic acid, 4.04 mM KH₂PO₄, 2.41 mM K₂HPO₄, 15.1 mM (NH₄)₂SO₄, 0.81 MgSO₄·7H₂O, 0.059 mM MnSO₄·H₂O, 0.16 mM FeSO₄·7H₂O, 0.17 mM NaCl, 0.068 mM CaCl₂·2H₂O, 0.041 mM Na₂MoO₄·2H₂O, 0.015 mM p-aminobenzoic acid, 0.006 mM thiamine-HCl, 0.00041 mM biotin, and 0.0044 mM resazurin. All components were dissolved in demineralized water, and the pH was adjusted to 7.3 using a 10 M NaOH solution. Afterwards, the medium was boiled for 15 min, cooled in an ice bath while sparging with N₂, and transferred into an anaerobic cabinet, and the volume was adjusted using anaerobic water. The media were aliquoted into 50-mL aliquots in 125-mL serum bottles for growth experiments, transformations, and preculturing, whereas 5-mL aliquots were used for the picking of colonies or inoculation from glycerol stocks. The media were autoclaved for 15 min at 121°C with 2.2 × 10⁵ Pa pressure saturated with water vapor. Depending on the plasmid, the strains were cultured at 30°C with 75 μg·mL⁻¹ thiamphenicol or 40 μg·mL⁻¹ erythromycin or without antibiotics. Uncontrolled batch growth experiments were performed using 50-mL aliquots of OMS with 75 μg/mL^-1 thiamphenicol. A total of 20 mM lactose was added only to the strains to be induced, i.e., C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_PB_P_bgaL], C. saccharoperbutylacetonicum ΔbldΔpta[pMTL83151_BCS_P_bgaL], and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS-PB_P_bgaL].

TABLE 2.

Bacterial strains and plasmids used or constructed in this study

Strain or plasmid	Characteristic(s)	Source or reference
Strains
E. coli XL1-Blue MRF′	Δ(mcrA)183 Δ(mrr-hsdRMS-mcrBC)173 endA1 supE44 thi-1 recA1 gyrA6 relA1 lac [F− proAB lacIqZΔM15 Tn10 (Tetr)]	Agilent Technologies Inc. (Santa Clara, CA, USA)
E. coli DH5α	F− ϕ80lacZΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17(rK− mK−) phoA supE44 thi-1 gyrA96 relA1 λ−	Thermo Fisher Scientific Inc. (Waltham, MA, USA)
NEB Turbo competent E. coli	F− proA+B+ lacIqΔlacZM15 fhuA2 Δ(lac-proAB) glnV galK16 galE15 R(zgb-210::Tn10)Tets endA1 thi-1 Δ(hsdS-mcrB)5	New England BioLabs (Ipswich, MA, USA)
NEB 10-beta competent E. coli	Δ(ara-leu)7697 araD139 fhuA ΔlacX74 galK16 galE15 e14-ϕ80dlacZΔM15 recA1 relA1 endA1 nupG rpsL (Strr) rph spoT1 Δ(mrr-hsdRMS-mcrBC)	New England BioLabs (Ipswich, MA, USA)
C. saccharoperbutylacetonicum N1-4(HMT) (DSM 14923)	Type strain	DSMZa (Brunswick, Germany)
C. saccharoperbutylacetonicum Δbld	bld (Cspa_c56880) deleted	81
C. saccharoperbutylacetonicum Δpta	pta (Cspa_c13010) deleted	81
C. saccharoperbutylacetonicum ΔbldΔpta	bld (Cspa_c56880) and pta (Cspa_c13010) deleted	81
Plasmids
pMTL83151	catP ColE1 ori− lacZα pCB102 ori+ traJ	44
pMTL83251	catP ColE1 ori− lacZα pCB102 ori+ traJ	44
pMTL82154	catP ColE1 ori− lacZα pBP1 ori+ traJ	44
pMTL83151_ptaI	pMTL83151; homologous regions for deletion of pta (Cspa_c13010)	81
pMTL83151_sol-adhEI	pMTL83151; homologous regions for deletion of bld (Cspa_c56880)	81
pMTL82154a_ptaI	pMTL82154; homologous regions for deletion of pta (Cspa_c13010)	81
pMTL82154b_ptaI	pMTL82154; homologous regions for deletion of pta (Cspa_c13010)	81
pMTL83154_sol-adhEI	pMTL82154; homologous regions for deletion of bld (Cspa_c56880)	81
pMA-RQ-2pta	bla ColE1 ori−; direct repeats and spacers targeting pta (Cspa_c13010)	81
pMA-RQ-2sol-adhEI	bla ColE1 ori−; direct repeats and spacers targeting bld (Cspa_c56880)	81
pMTL83251a_2pta	pMTL83251; direct repeats and spacers targeting pta (Cspa_c13010)	81
pMTL83251b_2pta	pMTL83251; direct repeats and spacers targeting pta (Cspa_c13010)	81
pMTL83251_2sol-adhEI	pMTL83251; direct repeats and spacers targeting bld (Cspa_c56880)	81
pMTL83151_gusA_PbgaL	pMTL83151; gusA originating from E. coli, bgaR (activator) and promoter PbgaL originating from Clostridium perfringens, and terminator Tfdx originating from C. pasteurianum	81
pMTL83151_PB_PbgaL	pMTL83151; ptb-buk operon (Cspa_c02520–Cspa_c02530) originating from C. saccharoperbutylacetonicum, bgaR (activator) and promoter PbgaL originating from C. perfringens, and terminator Tfdx originating from C. pasteurianum	81
pMTL83151_BCS_PbgaL	pMTL83151; bcs operon (Cspa_c04330–Cspa_c04370) originating from C. saccharoperbutylacetonicum, bgaR (activator) and promoter PbgaL originating from C. perfringens, and terminator Tfdx originating from C. pasteurianum	81
pMTL83151_BCS-PB_PbgaL	pMTL83151; bcs operon (Cspa_c04330–Cspa_c04370) followed by ptb-buk operon (Cspa_c02520–Cspa_c02530) originating from C. saccharoperbutylacetonicum, bgaR (activator) and promoter PbgaL originating from C. perfringens, and terminator Tfdx originating from C. pasteurianum	81

^aDSMZ: Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (German Collection of Microorganisms and Cell Cultures).

All strains were revived using CGM and adapted in OMS before inoculation for growth experiments and fermentations. As a carbon source for uncontrolled batch growth experiments, 40 g·L⁻¹ glucose in OMS was used. Strains were revived from glycerol stocks in 5 mL CGM (1% [vol/vol] inoculum). After growth, 10% (vol/vol) was transferred to 5 mL OMS for the first adaptation step. For the second adaptation step, 1 to 10% (vol/vol) was transferred to 50 mL OMS. This culture was used for inoculation in the uncontrolled batch growth experiments. For pH-controlled fermentations, either 40 g·L⁻¹ glucose or Excello 95 adjusted according to the amount of glucose in Excello 95, giving a final glucose concentration of 40 g·L⁻¹ glucose, was used as the carbon source. The strains were revived from a frozen state using CGM (5 to 10% [vol/vol] inoculum). Once good growth was visible in the preculture (OD₆₀₀ of ∼3), 2% (vol/vol) was transferred to OMS containing either glucose or Excello 95. These cultures were used for inoculation in the pH-controlled fermentations. The fermentations were performed using DasGip parallel bioreactor systems for microbial applications (DasGip Plant MX44H; Eppendorf, Hamburg, Germany), with a total of 8 1-L reactors with a working volume of 700 mL each. The strains were cultivated at 30°C under anaerobic conditions. Anaerobicity was maintained by constant gassing with 100% nitrogen flow at 0.24 vvm (volume of nitrogen per volume of medium per minute) through sparging; the nitrogen quality was 6.0. Agitation was set at 150 rpm. The pH was controlled at 6.3 using 2 M NaOH, and the off-gas (CO₂, H₂, O₂, CH₄, and C₂H₅OH) was monitored by a Prima Pro mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) using Gasworks Prima Pro software. Foaming was controlled by adding 200 μL antifoam (10% antifoam 204; Sigma-Aldrich, St. Louis, MO, USA), as required, when foam started to build. Plasmid-borne gene transcription was induced at an OD₆₀₀ of approximately 0.25 by the addition of 20 mM lactose only to the strains to be induced, i.e., C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], and no thiamphenicol was added for the fermentations.

Escherichia coli strains were grown in Luria-Bertani (LB) medium, as described previously (69), for every purpose except for the transformation of chemically competent cells. Chemically competent E. coli cells were prepared according to methods described previously by Inoue et al. (70), using superoptimal broth medium with 10 μg·mL⁻¹ tetracycline, as described previously by Hanahan (71). Prepared E. coli cells were transformed using a modified method according to Inoue et al. (70). In brief, aliquots with chemically competent cells were thawed on ice, and 0.5 to 2 μg DNA was added. The cells were incubated on ice for 10 min, heat shocked at 42°C for 1 min, and cooled on ice for 10 min. Cells were regenerated by the addition of 800 μL LB medium and incubated at 37°C for 30 min under shaking conditions before plating on LB medium with 25 or 30 μg·mL⁻¹ chloramphenicol or 250 or 500 μg·mL⁻¹ erythromycin.

Construction of plasmids and recombinant C. saccharoperbutylacetonicum strains.

The construction of deletion strains and the transformation of C. saccharoperbutylacetonicum were performed as described previously by Brosseau et al. (72). All primers used in this study are listed in Table 3. The inserts of the plasmids pMTL83151_ptaI and pMTL83151_sol-adhEI were amplified from genomic DNA of C. saccharoperbutylacetonicum using CloneAmp HiFi PCR premix (TaKaRa Bio Inc., Kusatsu, Shiga, Japan). They were ligated with digested pMTL83151 (KpnI and SmaI; Thermo Fisher Scientific, Waltham, MA, USA) using the In-Fusion HD cloning kit (TaKaRa Bio Inc., Kusatsu, Shiga, Japan). The inserts of pMA-RQ-2pta and pMA-RQ-2sol-adhE were synthesized and subcloned by Thermo Fisher Scientific (Waltham, MA, USA). For the construction of deletion strains according to methods described previously by Atmadjaja et al. (47), the vectors, pMTL82154a, pMTL82154b, pMTL83251a, and pMTL83251b, were digested using KpnI and SmaI (pMTL83151 based and pMTL82154 based) or ZraI and XhoI (pMA-RQ based and pMTL83251 based) (New England BioLabs [NEB], Ipswich, MA, USA). The fragments were ligated using T4 DNA ligase (Promega Corporation, Madison, WI, USA). This resulted in the plasmids pMTL82154a_ptaI, pMTL82154a_sol-adhEI, pMTL82154b_ptaI, pMTL83251a-2pta, pMTL83251a-2sol-adhE, and pMTL83251b-2pta. These were used to construct the strains C. saccharoperbutylacetonicum Δpta, C. saccharoperbutylacetonicum Δbld, and C. saccharoperbutylacetonicum ΔbldΔpta, respectively.

TABLE 3.

Primers used in this study^a

Primer	Sequence	Restriction site(s)	Purpose(s)
bcs_rev	GCAGGCTTCTTATTTTTATGGCTAGCCTATTTAGAATAATCGTAGAATCCTTTTC	NheI	Cloning of bcs operon (pMTL83151_BCS_PbgaL)
PbgaL_bcs_fwd	AATGTATTGGGAGGGTAAACCTCGAGATGGAATTAAAGAATGTGATTC	XhoI, Eco88I	Cloning of bcs operon (pMTL83151_BCS_PbgaL, pMTL83151_BCS-PB_PbgaL)
pta_KO_down_fwd	ATCTGATAAAAATGTTGTTGCATTAACTG		Cloning of pMTL83151_ptaI
pta_KO_down_rev	GTCGACTCTAGAGGATCCCCCCCGGGCTAAAGGAGTGAATCCCATTG	SalI, XbaI, BamHI, Eco88I, SmaI
pta_KO_up_fwd	TACGAATTCGAGCTCGGTACGGTACCCCTCTTTGCATAAAATTACC	EcoRI, SacI, KpnI
pta_KO_up_rev	ATGCAACAACATTTTTATCAGATTGAGCTGC
sol-adhE_KO_down_fwd	GATTTAAAATTAGCAAGAAATTTTACAAGACAAAG		Cloning of pMTL83151_sol-adhEI
sol-adhE_KO_down_rev	GTCGACTCTAGAGGATCCCCCCCGGGAATCAATTAAGCCGCCTC	SalI, XbaI, BamHI, Eco88I, SmaI
sol-adhE_KO_up_fwd	TACGAATTCGAGCTCGGTACGGTACCATACACATTGTAATTCAAATTTGC	EcoRI, SacI, KpnI
sol-adhE_KO_up_rev	AATTTCTTGCTAATTTTAAATCTTTTGTTATAGAAACTAG
bcs-ptb_buk_fwd	ATAGGGATCCATGAGCAAAAACTTTGACG	BamHI	Cloning of ptb-buk operon (pMTL83151_BCS-PB_PbgaL)
ptb_buk_rev	TTTTGCTCATGGATCCCTATTTAGAATAATCGTAGAATCCTTTTC	BamHI
PB_rev	GCAGGCTTCTTATTTTTATGGCTAGCTTAATATACCTTAGCTTTTTCTTCC	NheI	Cloning of ptb-buk operon (pMTL83151_PB_PbgaL)
PbgaL_PB_fwd	AATGTATTGGGAGGGTAAACCTCGAGATGAGCAAAAACTTTGACG	XhoI, Eco88I
fD1	AGAGTTTGATGCTCAG		PCR, sequencing of 16S rRNA gene
rP2	ACGGCTACCTCGACTT
sol-adhE_fwd	AGGTCCGTCAGAATAGATT		PCR, sequencing of bld deletion
sol-adhE_rev	GTTTACGAGTTGGCCTTT
M13F	GTAAAACGACGGCCAG		PCR, sequencing of pMTL82154-based plasmids
M13R	CAGGAAACAGCTATGAC
MH202	TGTTTGCAAGCAGCAGATTACG		PCR, sequencing of pMTL83251-based plasmids
MH203	TCAGTTTCATCAAGCAATGAAACACG
pta_fwd	GCCATTTCAGCTCTCTTC		PCR, sequencing of pta deletion
pta_rev	TCTACTCCGCCCATTACT
pMTL83151_PoI_fwd	CGCTGTATCCATATGACCA		Sequencing of pMTL83151_PB_PbgaL, pMTL83151_BCS_PbgaL, pMTL83151_BCS-PB_PbgaL
pMTL83151_seq_gusA_PbgaL	CACAATTAGCAACACAGG		Sequencing of pMTL83151_PB_PbgaL, pMTL83151_BCS_PbgaL, pMTL83151_BCS-PB_PbgaL
seg_Tfdx2_rev	GGCTTGATGTGTTGGTAG
crt-adhE_fwd	ACTTGGAGCAGGAACTAT		Sequencing of bcs operon (pMTL83151_BCS_PbgaL, pMTL83151_BCS-PB_PbgaL)
seq_abr1_fwd	CTTCTTCAGTCATTTTTCCC
seq_abr2_fwd	TTGGCAAATACTCCTGGT
seq_abr3_fwd	GTCCGTCTTCTAATGCTC
seq_abr4_fwd	AAAGCTCCTTCTGCAATAC
seq_abr5_fwd	AGTTTGATCCTCTGTTGC
seq_buk_fwd	CCAGGATCAACATCAACAA		Sequencing of ptb-buk operon (pMTL83151_PB_PbgaL, pMTL83151_BCS-PB_PbgaL)
seq_ptb_fwd	ATTGCAGATCCTAAGAAAGC
seq_ptb_rev	GCTTTCTTAGGATCTGCAAT

^aIn the sequences, the motifs corresponding to the listed restriction enzymes are indicated with boldface type and underlining.

Inserts for plasmids for homologous overexpression were amplified from genomic DNA of C. saccharoperbutylacetonicum using Phusion green high-fidelity DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA). The plasmid pMTL83151_gusA_P_bgaL (73) was digested using NheI and XhoI (Thermo Fisher Scientific, Waltham, MA, USA), and the inserts harboring the genes of the bcs operon (crt, bcd, etfB, etfA, and hbd [Cspa_c04330 to Cspa_c04370]) (abbreviated BCS) and/or genes ptb and buk (Cspa_c02520 and Cspa_c02530) (abbreviated PB) were ligated using the In-Fusion HD cloning kit (TaKaRa Bio Inc., Kusatsu, Shiga, Japan) according to the manufacturer’s instructions. The finalized plasmids pMTL83151_PB_P_bgaL, pMTL83151_BCS_P_bgaL, pMTL83151_BCS-PB_P_bgaL, and pMTL83151 were used to transform C. saccharoperbutylacetonicum ΔbldΔpta as described previously by Atmadjaja et al. (47), resulting in strains C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151], C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_P_bgaL], C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL], and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL].

Determination of substrate consumption and product formation.

The growth of the bacterial cultures was measured by the determination of the optical density at a wavelength of 600 nm (OD₆₀₀) using an Amersham Biosciences Ultrospec 10 pro photometer (GE Healthcare, Chalfont St Giles, Buckinghamshire, UK). All further parameters were determined from 2-mL samples taken during the growth experiment. The samples were stored at −20°C until use. Analyses were carried out after thawing and centrifugation (12,000 × g for 30 min at 4°C).

The pH of the samples was measured using a semimicro-single-junction pH electrode (SM224; Avantor, Radnor, PA, USA) attached to a WTW pH 521 pH meter (Xylem Analytics Germany Sales GmbH & Co. KG, Weilheim, Germany).

The production of acetone, ethanol, butanol, acetate, and butyrate was determined using gas chromatography. Five hundred microliters of the prepared sample supernatant was mixed with 20 μL of 2 M HCl before capping the vials. External standards with concentrations ranging from 0.5 to 250 mM were prepared from stock solutions to a volume of 500 μL, and 20 μL 2 M HCl was added before capping the vials. Clarus 600 and Clarus 680 GC devices (PerkinElmer Corp., Waltham, MA, USA) were used, equipped with an Elite-FFAP column. A total of 0.5 μL of the standard or the sample was injected at 225°C with a split ratio of 60:1. The oven was set to 90°C, holding the temperature for 1 min, heating at 40°C·min⁻¹ until 250°C, and holding the temperature for 2 min. The mobile phase was H₂ with 2.25 mL·min⁻¹ as the flow rate. The flame ionization detector was set to 300°C, with a synthetic air (80% N₂–20% O₂) flow rate of 450 mL·min⁻¹ and an H₂ flow rate of 45 mL·min⁻¹.

Sugars and lactate were measured using high-performance liquid chromatography. Five hundred microliters of the sample supernatant was filled into vials. External standards with concentrations ranging from 0.5 to 250 mM were prepared from stock solutions to a volume of 500 μL. An Infinity 1260 device with a refraction index detector (RID) and a diode array detector (DAD) (35°C at 210 nm) (Agilent Technologies, Santa Clara, CA, USA) was used for analysis of the compounds. Twenty microliters of the sample was applied onto an organic acid resin column (CS [Chromatographie Service GmbH], Langerwehe, Germany) set to 40°C, and 5 mM H₂SO₄ at a flow rate of 0.7 mL·min⁻¹ was applied for 17 min for the detection of glucose, lactose, and lactate. All other sugars (mannose, galactose, xylose, fructose, arabinose, and cellobiose) were analyzed by applying 20 μL of the sample onto a Nucleogel sugar Pb column (Macherey-Nagel, Düren, Germany) set to 80°C, and water at a flow rate of 0.4 mL·min⁻¹ was applied for 30 min. All sugars were detected via the RID, and lactate was detected via the DAD.

Verification of strains.

Genomic DNA of C. saccharoperbutylacetonicum strains was isolated using the MasterPure Gram positive DNA purification kit (Epicentre, Madison, WI, USA) according to the manufacturer’s instructions. PCRs were performed using Phusion green high-fidelity DNA polymerase, Platinum SuperFi PCR master mix (Thermo Fisher Scientific, Waltham, MA, USA), CloneAmp HiFi PCR premix (TaKaRa Bio Inc., Kusatsu, Shiga, Japan), ReproFast proofreading polymerase (Genaxxon Bioscience GmbH, Ulm, Germany), and Phusion high-fidelity DNA polymerase (New England BioLabs, Ipswich, MA, USA) according to the respective manufacturer’s instructions. For the verification of strains during the construction process and after each growth experiment or fermentation, the respective deletions and the 16S rRNA gene were amplified and sent for sequencing. Primers for PCRs are shown in Table 3. Plasmids were verified by the transformation of isolated genomic DNA into E. coli, plasmid preparation using the Zyppy plasmid miniprep kit (Zymo Research, Freiburg, Germany) according to the manufacturer’s instructions, subsequent restriction digestion (Thermo Fisher Scientific, Waltham, MA, USA), and sequencing by Genewiz Inc. (South Plainfield, NJ, USA).

Transcriptomic data.

During a growth experiment, 1- and 2-mL samples were taken and centrifuged (13,000 × g for 1 min). The supernatant was discarded, and the pellet was stored immediately at −80°C until the end of the growth experiment. The RNeasy minikit (Qiagen NV, Hilden, Germany) was used for RNA isolation. The RNA integrity number was tested using an Agilent RNA 6000 Nano kit on an Agilent Bioanalyzer 2100 system (Agilent Technologies Inc., Waldbronn, Germany). Turbo DNase (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used for DNA removal, and a Ribo-Zero kit for bacteria (Illumina Inc., San Diego, CA, USA) was used for the reduction of rRNA in the samples under SRA accession numbers SRR16573869 to SRR16573880. The construction of strand-specific cDNA libraries was carried out with the NEBNext Ultra II directional RNA library preparation kit (New England BioLabs GmbH, Frankfurt, Germany). The size and quality of the libraries were checked using the Agilent high-sensitivity DNA kit on the Agilent Bioanalyzer 2100 system (Agilent Technologies Inc., Waldbronn, Germany). The Qubit double-stranded DNA (dsDNA) high-sensitivity assay DNA kit (Life Technologies GmbH, Darmstadt, Germany) was used for the determination of the concentration of the libraries. Sequencing of the samples under SRA accession numbers SRR16573869 to SRR16573880 was carried out using the HiSeq3000/4000 SR cluster kit and the HiSeq 3000/4000 SBS kit using the HiSeq4000 instrument from Illumina Inc. (San Diego, CA, USA). Analysis of the sequences was performed using Trimmomatic v-0.39 (74, 75), the Bowtie program (version 2) (76) using the previously published genome of C. saccharoperbutylacetonicum as a reference (15), featureCounts v.2.0.0 (77), R (version 3.6.1) (78), and DeSeq2 (version 1.26.0) (52). For the samples under SRA accession numbers SRR17835133 to SRR17835156, the Illumina Ribo-Zero plus rRNA depletion kit (Illumina Inc., San Diego, CA, USA) was used for rRNA depletion, and sequencing was performed using the NovaSeq 6000 machine (Illumina Inc., San Diego, CA, USA) with the NovaSeq 6000 SP reagent kit v1.5 (100 cycles) and the NovaSeq XP 2-Lane kit v1.5 in the paired-end mode with 2 by 50 cycles. Mapping against the reference genomes was performed with Salmon (v1.5.2) (78). As the mapping backbone, a file that contains all annotated transcripts excluding rRNA genes and the whole genome of C. saccharoperbutylacetonicum N1-4(HMT) (15) as a decoy was prepared with a k-mer size of 11. Decoy-aware mapping was done in selective-alignment mode with “–mimicBT2,” “–disableChainingHeuristic,” and “–recoverOrphans” flags as well as sequence and position bias correction. Values of 325 and 25 were used for –fldMean and –fldSD, respectively. The quant.sf files produced by Salmon were subsequently loaded into R (v4.0.3) (77) using the tximport package (v1.18.0) (79). DeSeq2 (v1.30.0) (52) was used for the normalization of the reads, and fold change shrinkages were also calculated with DeSeq2 and the apeglm package (v1.12.0) (80). Genes were considered differentially expressed when the adjusted P value was <0.05 and the log₂ fold change was equal to and greater or less than +2 and −2, respectively.

Data availability.

The raw reads of the 12 analyzed C. saccharoperbutylacetonicum wild type samples have been deposited in the SRA nucleotide archive under BioProject accession number PRJNA774278 and correspond to SRA accession numbers SRR16573869 to SRR16573880. The raw reads of the 16 analyzed samples regarding the optimized butyrate producers C. saccharoperbutylacetonicum ΔbldΔpta and C. saccharoperbutylacetonicum ΔbldΔpta [pMTL83151_BCS_P_bgaL] have been deposited in the SRA nucleotide archive under BioProject accession number PRJNA774278 and correspond to SRA accession numbers SRR17835133 to SRR17835156.