Recombinant bacteriophage LysKB317 endolysin mitigates Lactobacillus infection of corn mash fermentations

Background Commercial ethanol fermentation facilities traditionally rely on antibiotics for bacterial contamination control. Here we demonstrate an alternative approach to treat contamination using a novel peptidoglycan hydrolase (LysKB317) isolated from a bacteriophage, EcoSau. This endolysin was specially selected against Lactobacillus strains that were isolated as contaminants from a fuel ethanol plant. The LysKB317 gene was recombinantly expressed in Escherichia coli as a 33 kDa purified enzyme. Results In turbidity reduction assays, the recombinant enzyme was subjected to a panel of 32 bacterial strains and was active against 28 bacterial strains representing 1 species of Acetobacter, 8 species of Lactobacillus, 1 species of Pediococcus, 3 species of Streptococcus, and 1 species of Weissella. The activity of LysKB317 was optimal around pH 6, but it has broad activity and stability from pH 4.5–7.5 up to at least 48 h. Maximum activity was observed at 50 °C up to at least 72 h. In addition, LysKB317 was stable in 30% ethanol up to at least 72 h. In experimentally infected corn mash fermentations, 1 µM endolysin reduced bacterial load by 3-log fold change, while 0.01 µM reduced bacteria by 2-log fold change. Concentration of fermentation products (ethanol, residual glucose, lactic acid, and acetic acids) for infected cultures treated with ≥ 0.01 µM LysKB317 was similar to uncontaminated controls. Conclusion Exogenously added LysKB317 endolysin is functional in conditions typically found in fuel ethanol fermentations tanks and may be developed as an alternative to antibiotics for contamination control during fuel ethanol fermentations.


Background
Endolysins are peptidoglycan hydrolase enzymes (also known as phage lysins) produced by bacteriophages to enzymatically degrade host bacterium cell wall from within to release progeny virions at the end of lytic multiplication cycle [1]. Due to endolysins' antibacterial activity, they are considered potential alternatives to antibiotics [2,3].
The fuel ethanol industry in the United States has experienced a tremendous growth over past decade from 110 plants (6.5 billion gallons per year) in 2007 to 200 plants in 2017 with production capacity approaching 16 billion gallons per year [4,5]. An estimated production capacity for fuel ethanol would need to reach 60 billion gallons per year by 2030 to meet the proposed US Energy Independence and Security Act (EISA) of 2007, Renewable Fuel Standard (RFS) mandates, and goal set by the environmental protection agency (EPA) and other states to increase higher blend of ethanol in gasoline [6][7][8].
However, commercial ethanol fuel facilities rarely perform fermentations under aseptic conditions [9,10]. Fermentation tanks in ethanol production are constantly contaminated with a wide variety of microbes that can cause chronic and acute contaminations in commercial biorefineries [11]. These strains cause both chronic and acute infections in commercial biorefineries and can significantly reduce the level of ethanol production [4,9,10,12,13]. Potential sources of microbial contamination (bacteria, fungi, and wild yeast) can be found in raw materials such as corn, corn mash, and process water, although through the liquefaction process they appeared to be inactivated [9,[13][14][15]. Acute contamination often occurs unpredictably and can lead to a costly shutdown of facilities [16]. It is generally believed that lactic acid bacteria (LAB), and predominantly species of Lactobacillus, are the primary bacterial contaminants found in fuel ethanol fermentation facilities [17,18]. In addition to competition of nutrients and substrates with fermenting yeast, bacterial contaminants produced undesirable byproducts such as acetic and lactic acids can inhibit yeast growth [9,19]. The presence of Lactobacillus can cause "stuck fermentations" and decrease yields of ethanol production as Lactobacillus spp. compete for resources and negatively impact the health of Saccharomyces sp. [4,[16][17][18]20]. The solution to combat contamination in the United States has traditionally relied on the usage of antibiotics such as erythromycin, penicillin, and virginiamycin [21]. Concerns over long-term excessive usage of antibiotics are believed to contribute to the emergence of antibiotic resistant bacteria and remains controversial in the ethanol industry [22,23]. Alternative strategies such as the deployment of all-natural proteinaceous antimicrobial control agents such as endolysin are warranted.
In this study, we described the application of a novel recombinant peptidoglycan hydrolase (endolysin) LysKB317 derived from Lactobacillus bacteriophage vB_LfeS_EcoSau (abbreviated as EcoSau); isolated from commercial sauerkraut) to inhibit the growth of lactic acid bacteria known to contaminate ethanol fermentation facilities [24]. This endolysin derived from EcoSau was designated to LysKB317 in honor of Dr. Kenneth Bischoff. We demonstrated the effectiveness of exogenously added endolysin LysKB317 with predicted GH25 muramidase activity to a panel of Gram-positive bacterial species such as Lactobacilli. LysKB317 showed a robust antibacterial activity against eight species of Lactobacillus, including those that are problematic in the fuel ethanol industry. In addition, LysKB317 confirmed some activity against bacterial species such as Acetobacter pomorum, Pediococcus spp. Streptococcus spp., and Weissella confusa isolated from commercial biorefineries.
We determined the activity profile of LysKB317 under fermentation conditions, and with extended exposure to various pH, temperature, and percent ethanol. The robustness of LysKB317 was demonstrated in an experimentally infected corn mash fermentation to treat against Lactobacillus fermentum contamination and restored yield of ethanol fermentation by S. cerevisiae. Overall results showed the potential of endolysin LysKB317 as an alternative to conventional antibiotics to control contamination for the fuel ethanol industry.

Lytic activity of purified endolysin LysKB317 confirmed
We were able to express and purify the phage lytic protein in recombinant E. coli as the N-terminus 6 × Histagged LysKB317 ( Fig. 1a; GenBank accession number AIY32273.1). The endolysin consists of a glucohydrolase family 25 (GH25) muramidase-superfamily domain and a cell wall binding SH3b homologue domain (Fig. 1a). Based on blast search and homologous to protein sequences analyses with known functions, the predicted muramidase activity in LysKB317 is thought to cleave β-(1,4)-glycosidic bond of the peptidoglycan N-acetylglucosamine-N-acetylmuramic acid (NAG-NAM) linkages ( Fig. 1b; [25,26]) The SDS-PAGE and western blot analysis were performed on the nickel-NTA column purified protein, which produced a single prominent band for LysKB317 with the predicted molecular mass of 33.8 kDa ( Fig. 2a and Additional file 1: Figure S1). Spot plate assay (using MRS agar plate incorporated 1 mL of live L. fermentum 0315-25 (OD 600 = 0.8) in 0.7% soft top agar) demonstrated exolytic activity after spotting of 5 µL LysKB317 (either expressed whole cell lysate supernatant or purified LysKB317; Fig. 3). Under visual observation, the zone of clearing in both whole cell lysate supernatant or purified LysKB317 samples were significantly more pronounced compared to those of LysA (minimum activity against L. fermentum 0315-25; [4]), and lysozyme (positive control) confirming the exolytic activity of the enzyme. Zymogram analysis was performed with copolymerized L. fermentum 0605-B44 into the gel matrix. Single translucent bands in the same size region as the predicted LysKB317 were clearly visible for whole cell lysate, soluble fraction, and purified enzyme of LysKB317 from the expression host E. coli (Fig. 2b).

The LysKB317 lyses a range of Lactobacillus species
Purified endolysin LysKB317 was tested using a turbidity reduction assay against several bacterial species (Table 1) isolated from commercial fuel ethanol fermentation plant (Fig. 4). LysKB317 had a strong lytic activity (> 100 OD 600 /min/µM enzyme; Fig. 4) against all of the L. fermentum strains tested in the panel (Table 1). When other lactic acid bacteria species were included in the tests (Fig. 4), approximately 87% of the Lactobacillus spp. tested from the panel were susceptible to LysKB317 (Fig. 4). Furthermore, the endolysin showed lytic activity against bacterial species other than Lactobacilli such as Weissella confuse, Acetobacter pomorum, Pediococcus acidilactici, Staphylococcus lugdunensis and two species of Streptococcus (Fig. 4). However, bacterial species including Enterococcus faecium, L. amylovorus, and L. brevis presented minimum to none exolytic activity by the LysKB317.

LysKB317 endolysin remains active in the fermentation environment
To examine the enzymatic activity of endolysin LysKB317 under typical fuel ethanol fermentation conditions, the enzyme was tested under a range of pH, temperature, and ethanol concentration over time using turbidity reduction assays. The optimal pH for LysKB317 was achieved at pH 6 and it was stable for up to at least 48 h (Fig. 5). In addition, the enzyme was functionally stable in a range of pH 4.5-7.5 up to at least 48 h, but the lytic activity of the endolysin at pH 4 was compromised. Thermostability of the enzyme was observed from 4 °C to 50 °C for at least 72 h (Fig. 6). At 60 °C, thermal stability of the enzyme started to deteriorate after 41 h of incubation and the lytic activity was abolished by 72 h. Minimal to no lytic activity was observed at 95 °C regardless of the time LysKB317 was incubated (Fig. 6). The presence of ethanol at or below 5%, did not have a significant impact on the lytic activity of the endolysin regardless of incubation time (0-72 h; Fig. 7). LysKB317 remained active upon exposure of ethanol concentration up to 30%, although activity was approximately 45-54% less than samples without added ethanol (Fig. 7).

LysKB317 reduces Lactobacillus in a model fermentation flask
As described previously [4], we emulated fermentations using corn mash solids to test the effects on LysKB317 (Fig. 8). In experimentally infected corn mash fermentations, the addition of endolysin at 1 µM reduced bacterial load by approximately 3-log fold over time (black circle) compared to the challenged control fermentation (gray triangle), which rose above 9-log CFU/mL. Uninfected corn mash fermentations (negative control) and LysKB317-treated fermentations without infection (negative control) did not have detectable bacterial load over 3-log CFU/mL (limit of detection) were not included in the graph.

Fermentation products of infected but LysKB317 treated were similar to those of uninfected controls
In our model bacterial infected flask yeast fermentation runs, the highest concentration of LysKB317 (10,000 nM ≡ 330 μg/mL) reduced bacterial load by 4-log fold CFU/ mL change, while 100 nM (3.3 μg/mL) was able to reduce bacteria load by approximately 2-log fold ( Table 2). The bacterial L. fermentum fermentation byproducts, such as lactic acid and acetic acid, which are known to inhibit S. cerevisiae and reduce ethanol yields were reduced significantly with the addition of LysKB317. Lactic acid was reduced more than 20% from 19.8 g/L to 15.4 g/L, while acetic acid decreased by more than 70% from 3.6 g/L to 1.0 g/L. The glucose utilization by S. cerevisiae after LysKB317 treatment (10,000 nM) in L. fermentum infected corn mash compared to uninfected corn mash had an over 98% improvement (from 38.9 g/mL glucose prior to treatment down to 0.7 g/mL glucose after endolysin treatment). End concentration of ethanol after fermentation had increased to 21.3 g/mL (~ 22% increase). The LysKB317-treated flask fermentation resulted comparable levels of glucose utilization and ethanol production when compared with that of uninfected flask fermentation controls.  Table 1), purified LysA2 (5 µL; negative control; [54]), 20 µg/mL lysozyme (positive control), and 5 µL of MRS broth (negative control) were spotted on to plate and allow to air dry before incubating at 37 °C until zone of clearance can be visualized

Discussion
Bacterial contamination is inevitable during the propagation and fermentation processing of fuel ethanol production [12]. Mitigating bacterial contamination using antibiotics and adjustment of fermentation process, such as pH or temperature, have been used to control infection [23]. In addition, commercially available chemical-based products, such as hop acids and chlorine dioxide, have shown some success [27,28]. However, there still is a need to improve the current technology by finding alternatives to control bacterial contamination in these types of biorefining processes. In the United States, ethanol production accounts for one of the largest industrial uses of  antibiotics consumption [23,29]. Prolonged excessive usage of antibiotics to treat bacterial contamination has raised concerns on the contribution to antimicrobial resistance [19,21,23]. Furthermore, it has been demonstrated that low concentration of biologically active antibiotic such as virginiamycin can persist in distilled grain coproducts when used in ethanol production facilities [30]. Low concentrations of bioactive antibiotics could potentially present a selection pressure resulting in anthropogenic influences that may contribute to bacterial resistance [31][32][33]. To date, rare accounts of resistance to bacterial peptidoglycan lytic enzymes have been reported, which makes it an effective and desirable alternative treatment to antibiotics [3,34]. Our goal in this study was to demonstrate that purified endolysin LysKB317 could be a useful tool to mitigate bacterial contamination for the bioethanol industry [9,14,21]. The putative endolysin gene LysKB317 was first identified from EcoSau bacteriophage isolated from commercial sauerkraut [24]. The application of purified endolysin LysKB317 has demonstrated a high lytic activity against numerous Gram-positive LAB including several Lactobacillus species such as L. fermentum, which have previously been shown to negatively impact the rate of fermentation and often lead to stuck fermentations [19,24,35].

Purified LysKB317 demonstrated lytic activity against most Lactobacillus species
Based on Pfam protein domain prediction, LysKB317 has a predicted peptidoglycan hydrolase similar to a glycoside hydrolase family 25 LysA-like domain active site (a muramidase) and a bacterial SH3b-like cell wall binding domain ( Fig. 1a; [36]). A panel of 32 commonly found bacterial contaminant strains at ethanol fermentation facilities were tested (Table 1), and 26 (81% effective rate) strains were lysed by LysKB317. As a muramidase, LysKB317 is thought to cleave Gram-positive bacterial cell wall that shared similar peptidoglycan backbone. As a potential method to treat ethanol fermentation contaminants (e.g., L. fermentum), differences in the makeup of peptidoglycan chemotypes could have minimal impact on the catalytic activity of the endolysine to treat infection. Interestingly, in turbidity reduction assay, the highest (L. fermentum) and the lowest (L. amylovorus) lytic activities were all from Lactobacillus species (Fig. 4). Not all peptidoglycan chemotypes of Lactobacillus spp. were equally sensitive to LysKB317 (Table 3; [37][38][39][40][41]) as seen with L. fermentum and L. mucosae. Differences in affinity of the SH3b cell wall binding domain and/ or accessibility of the LysKB317 could affect target based on differences in strain specific cell wall surface moieties [42][43][44][45] cause interferences to the predicted cleavage site (Fig. 1b). More research is needed to determine the substrate specificity in L. fermentum by LysKB317 compared to L. amylovorus.

Endolysin LysKB317 exhibited a robust and stable characteristic under physical conditions of fermentation
Conditions typically found in fuel ethanol fermentation facility fermentation tank can have a temperature ranging

Table 2 Treatment model for bacterial load and fermentation products of experimentally infected ethanol fermentations treated with exogenously added LysKB317 endolysin
Cultures of S. cerevisiae grown on corn mash feedstock were challenged with 10 6 CFU/mL of L. fermentum 0315-25 [27], and treated with the indicated concentration of recombinant LysKB317 endolysin. The control culture was not challenged with L. fermentum 0315-25. After 72 h incubation, viable L. fermentum was determined by enumeration on MRS agar plates, and the fermentation broth was analyzed by HPLC for the following fermentation products: ethanol, residual glucose, lactic acid, and acetic acid

LysKB317 is effective under small-scale corn mash fermentation conditions
Exogenous addition of purified LysKB317 alone was sufficient and successful in treating and controlling infected corn mash matrix (Fig. 8). Effective treatment seen in 50 mL Erlenmeyer flasks ( Table 2) was encouraging in control L. fermentum bacterial load. The effectiveness in controlling bacterial contamination is also reflected upon the level of acetic and lactic acid in reducing the byproducts and restored ethanol production. Alternative methods to increase production of the lysin could be beneficial and will reduce cost at the industrial scale. Nevertheless, current method of exogenous addition of purified LysKB317 alone was able to control L. fermentum contamination and restore healthy fermentation characteristics.

Conclusion
Bacteriophage-derived lytic endolysin enzyme such as LysKB317 is a strong candidate of antimicrobial control against LAB contamination in fuel ethanol fermentations.
LysKB317 demonstrated the ability to lyse L. fermentum at pH, temperature, and ethanol concentrations similar to conditions found during fuel ethanol fermentations by at least two-log fold change in small-scale corn-mash fermentation. These qualities make LysKB317 an excellent candidate for antimicrobial control for use in biofuel fermentations.

Bacterial and yeast strains and culture conditions
Wildtype bacterial strains were isolated from a Midwestern dry-grind fuel ethanol plant and selected from a previous screen [15]. Unless otherwise stated, all bacterial strains described here (Table 1) were grown in its respective culture media. Escherichia coli strains in Miller's LB (LB broth) medium (Difco Laboratories, Inc.). When used, ampicillin (Amp; Sigma-Aldrich, Inc.) at 100 µg/ mL or kanamycin (Kan; Sigma-Aldrich, Inc.) at 50 µg/mL was added to LB media when required. Here we acknowledge newly reclassification and genera naming of some Lactobacillus spp. listed in this study (e.g., Lactobacillus fermentum as Limosilactobacillus fermentum and Lactobacillus mucosae as Limosilactobacillus mucosae) [48]. For consistency, older species names are being used here. Lactobacillus spp. and Weissella. were grown in Lactobacilli MRS (MRS broth) medium (Difco Laboratories, Inc.). Acetobacter and Pediococcus were grown in rapid lemonade spoilage organism broth (RLS broth; Sigma-Aldrich).

Western blot analysis
Purified N-terminus His-tagged LysKB317 recombinant protein described above was separated by SDS-PAGE as previously described. A Trans-Blot turbo transfer system (Bio-Rad) was used for protein transfer onto a low-fluorescence polyvinylidene difluoride (PVDF) membrane with 0.2 µm pore size (Bio-Rad). Protein electrophoresis transfer was verified using Ponceau S staining (Cell Signaling Technology, Inc.). Nonspecific binding was blocked by 3% bovine serum albumin (BSA) in 1 × tris-buffered saline (TBS) containing 0.1% Tween-20 (Sigma-Aldrich).
Mouse anti-His-tag antibody conjugated to DyLight 488 was applied and incubated at 4 °C overnight (1:1,000; Thermo Fisher Scientific). Fluorescent band signals were detected using a ChemiDoc XRS + imaging system (Bio-Rad).  Strains that exhibited zones of clearance were deemed susceptible to LysKB317. As controls, 5 µL of MRS broth served as the negative control and 5 µL of 20 µg/mL purified endolysin LysA and lysozyme separately served as the positive control [50].

Zymogram
Zymogram analysis was performed based on a previously described method with slight modification [4]. Briefly, L. fermentum 0315-25 cells (Table 1) were grown to mid-log phase in 50 mL MRS media and pelleted at 4,000×g for 15 min. Cells were washed with 10 mL of zymogram buffer (10 mM Tris, 150 mM NaCl, pH 7.5), harvested and resuspended in 300 µL zymogram buffer resulting in a final volume of approximately 600 µL. The purified LysKB317 protein described above, and protein standard (Precision Plus Protein All Blue standard; Bio-Rad) were run in parallel in two separate 15% SDS-PAGE gels. One gel contained 600 µL of resuspended L. fermentum 0315-25 cell (zymogram), and the other gel contained only 600 µL of buffer (negative control). Each of which was added prior to gel polymerization. Gels were electrophoresed for 1-2 h at 150 V until completion. SDS-PAGE gels were stained using LabSafe GEL Blue (G-Biosciences) and washed in deionized (DI) water for 1 h at room temperature. Additional de-staining incubation was done with gels submerged in de-staining buffer (50 mM Tris-HCl, 1% Triton X-114, pH 5.5) at room temperature with gentle swirling overnight or until translucent bands is clearly visible as described [49].

Turbidity reduction assay
Turbidity reduction assay was performed at 37 °C, unless otherwise stated, in Synergy 2 Microplate Reader (BioTek Inc.) with purified LysKB317 protein (described above) diluted in turbidity reduction assay buffer (300 mM NaCl, 30% (v/v) glycerol, 21 mM citric acid, 58 mM NaH 2 PO 4 , pH 5.5) to 1 μM concentration. Lactobacillus cultures (listed in Table 1) used in the turbidity reduction assay were prepared as previously described [4]. Briefly, bacterial cells were inoculated in 50 mL MRS media and grown to mid-log phase. Cells were washed in phosphate-buffered saline (PBS; pH 7.4, 30% glycerol) before being adjusted to an optical density (OD 600 nm ) = 2.0. Aliquots of 1 mL of cells were then centrifuged and pellet resuspended in 1 mL of turbidity reduction assay buffer. Each of the designated experimental wells of a 96-well microtiter plate (flat bottom; Falcon) contained 100 μL bacterial suspension and 100 μL of 1 μM endolysin. Wells containing bacterial cell suspension (100 μL) without endolysin (100 μL turbidity reduction assay buffer) were used to control the rate of autolysis of bacterial cells. Immediately upon addition of endolysin to bacterial suspension, absorbance readings (OD 600 ) were recorded every 30 s for 30 min. Treatment and control wells were run in triplicates. Specific actives were determined by (ΔmOD 600 nm /min/µM) described by Becker et al. [51].

Preparation of small-scale corn mash fermentation
This was done as described in Bischoff et al. and Roach et al. [4,35]. Briefly, the S. cerevisiae strain NRRL Y-2034 (Table 1) was grown overnight in YP broth supplemented with 5% (w/v) glucose at 32 °C with 200 rpm shaking. The infection L. fermentum strain 0315-25 (Table 1) was grown in static MRS media at 37 °C to mid-log phase (OD 600 nm = 0.4-0.6). Both yeast and bacteria cells were collected via centrifugation and inocula were resuspended in sterile phosphate-buffered saline (PBS; pH 7.4, Fisher Scientific) to OD 600 nm equivalent of 80 for yeast, and OD 600 nm equivalent of 8.0 for L. fermentum 0315-25. One OD 600 nm is approximately 6 × 10 7 CFU/ mL for yeast and 1 × 10 8 CFU/mL for bacteria. Corn mash (approximately 33% solids) was collected from a commercial dry-grind ethanol facility and stored at − 20 °C. Verification of aliquots of corn mash samples onto MRS agar did not detect transient bacteria in the mash (< 10 2 CFU/mL). In separate 50 mL Erlenmeyer flasks, 40 mL corn mash with ammonium sulfate (0.12%, w/v) and glucoamylase (20 μL of Optidex L-500; Genecor International Inc.) were dispensed.