Broad range shuttle vector construction and promoter evaluation for the use of Lactobacillus plantarum WCFS1 as a microbial engineering platform

Abstract

As the field of synthetic biology grows, efforts to deploy complex genetic circuits in nonlaboratory strains of bacteria will continue to be a focus of research laboratories. Members of the Lactobacillus genus are good targets for synthetic biology research as several species are already used in many foods and as probiotics. Additionally, Lactobacilli offer a relatively safe vehicle for microbiological treatment of various health issues considering these commensals are often minor constituents of the gut microbial community and maintain allochthonous behavior. In order to generate a foundation for engineering, we developed a shuttle vector for subcloning in Escherichia coli and used it to characterize the transcriptional and translational activities of a number of promoters native to Lactobacillus plantarum WCFS1. Additionally, we demonstrated the use of this vector system in multiple Lactobacillus species, and provided examples of non-native promoter recognition by both L. plantarum and E. coli strains that might allow a shortcut assessment of circuit outputs. A variety of promoter activities were observed covering a range of protein expression levels peaking at various times throughout growth, and subsequent directed mutations were demonstrated and suggested to further increase the degree of output tuning. We believe these data show the potential for L. plantarum WCFS1 to be used as a nontraditional synthetic biology chassis and provide evidence that our system can be transitioned to other probiotic Lactobacillus species as well.

1. Introduction

The lactic acid bacteria (LAB) are a diverse group of Gram-positive microorganisms of general interest for industrial applications based on their fermentative abilities. The genus Lactobacillus has received increasing attention recently due to the versatility of its members, as different species are widely studied for practical applications based on their status as Generally Regarded as Safe food-grade microbes that occupy a variety of environments relevant to human health. Lactobacillus species such as various probiotic strains of Lactobacillus acidophilus, Lactobacillus reuteri, Lactobacillus rhamnosus, Lactobacillus johnsonii and Lactobacillus plantarum are known to be beneficial to human health through their activity in the gastrointestinal tract, the urogenital tract and the oral cavity (1). Due to the applicability of LAB cultures in industrial and health settings, considerable work has been devoted to the development (2) and modification (3–7) of inducible heterologous protein expression systems in Lactobacillus species for various purposes including the production and secretion of eukaryotic immune proteins (8). Most of the developed systems utilize regulatory genes and promoters involved in the production of the class II bacteriocin sakacin from L. sakei (2), which was adapted from the concept of the Nisin-inducible systems developed in another LAB Lactococcus lactis (9).

The specific method of optimal protein expression may vary depending on the particular application of the LAB strain employed. Despite the advantage of overcoming potential host toxicity, overexpression from inducible systems may not be desirable for fermentation processes or in the intestinal environment (10). Investigators have since identified options for protein expression using endogenous promoters in certain Lactobacillus species (10–13), with some examples of promoters that function in multiple species (11, 14, 15) that might offer alternatives to inducible systems. Given the current state of synthetic biology, however, more expansive heterologous expression mechanisms may be required to more precisely address the problems of the era.

Synthetic biology has attracted a growing interest in recent years and consequently has produced numerous phenomenal circuit concepts that have been demonstrated in highly characterized organisms. Simple transcriptional units have been demonstrated upon their combination to function as a variety of logic gates such as AND, NOT, NAND (16), NOR, OR, XNOR and XOR gates (17), as well as more complicated functions such as temporal oscillation (18), binary computation (19) and edge detection (20). The outstanding accomplishments exemplified by these circuits hint at the potential for synthetic biology solutions given the availability of adequately characterized parts (18, 21). The broad application of such molecular engineering, however, requires a push for such circuits to function outside of the bounds of Escherichia coli or S. cerevisiae lab strains and in more environmentally relevant organisms, but this transition requires the assessment of circuit parts in the individual target chassis such as those reviewed by Gyulev et al. (22) for Clostridium and Tang et al. (23) for Comamonas. In light of the work required to assess the function of transcriptional and translational machinery in specific organisms, investigators have looked into gene circuit components that can easily transfer between species, genus and kingdom in order to expedite the design and implementation of functional circuits (24).

The versatility of Lactobacillus species to act as probiotics in the human microbiome, to thrive on a variety of carbohydrate sources, and to participate in industrial applications make them a desirable chassis for the development of synthetic biology solutions. The diversity of the Lactobacillus genus, however, has provoked the discussion of reclassification recently (25) and is demonstrated by inconsistent plasmid compatibility amongst its members (26) that could be due to the prevalence of uncharacterized cryptic plasmids within different species (27). As such, plasmids containing replicons viable in multiple species are indispensable tools for synthetic biology in Lactobacillus. Commonly used theta-replication origins such as ColE1 or pMB1 found in pUC19 are popular in plasmids used in E. coli but do not replicate in some Lactobacillus species (28). While holding the potential for higher copy number and segregational stability, theta-replication plasmids have the downside of being traditionally narrow in host range (27). For example, theta-replicating origins such as the 256rep have been demonstrated to work in a number of LAB within the pSIP vector series; however, these plasmids required the incorporation of the pUC(pGEM)ori to function as a shuttle vector in E. coli (2). The host range of the pSIP system was later improved with the incorporation of the SH71 replicon, but a second origin was still required for replication in E. coli (7). Numerous other LAB shuttle vectors containing theta-replicating origins have been developed, all of which require multiple origins of replication to function in both the host and cloning strains (13, 15, 28–31). On the other hand, rolling circle replication (RCR) plasmids are characterized as quite the opposite of theta-replication plasmids, with typically broad-range function and relatively poorer segregational stability. This concept was demonstrated by the characterization of an array of RCR origins from Bacillus subtilis pE194, L. lactis pWV01, Lactobacillus buchneri pCD034-1 and L. buchneri pCD034-2 and their activity within the organisms L. plantarum CD033 and L. buchneri CD034 (26). The pWV01 origin from L. lactis ssp. cremoris (27) contains its own replication proteins and has been successfully shown to replicate in E. coli, Lactobacillus gasseri and L. acidophilus NCFM in addition to the organisms listed above (11, 26). While replication origins have been identified that allow for replication in both LAB and E. coli, maintaining selective pressure on organisms with a single resistance marker can be more challenging based on the high amounts of erythromycin required for E. coli (2, 7, 11) or the use of antibiotic resistance markers such as Cam^R (26), Kan^R (32) or Amp^R that are already prevalent in LAB genomes (27, 33, 34).

In this study, we aimed to establish a broad host-range shuttle vector for use in Lactobacillus species, namely L. plantarum WCFS1, in addition to E. coli 10β. A system comprised of the pWV01 origin and dual antibiotic selection markers for use in E. coli and Lactobacillus spp. was used to characterize the transcriptional and translational activities of constitutive promoters from the genes encoding pantothenate kinase (coaA), fructose-bisphosphate aldolase (fba), pyruvate kinase (pyk), pyruvate oxidase (pox2), glyceraldehyde-3-phosphate dehydrogenase (gapB), elongation factor Tu (tuf), elongation factor G (fusA1) and a putative TypA elongation factor (typA) of L. plantarum WCFS1. The observed promoter strengths and temporal variability in addition to the promoter elements involved in posttranscriptional regulation comprise a variety of parts for consideration in the potential construction of genetic circuits in L. plantarum. We further show that similar promoters native to other Lactobacillus species are also active in L. plantarum WCFS1 and E. coli 10β, indicating the potential for both the construction of circuits functional in multiple species as well as convenient preliminary screening of before deployment in Lactobacillus.

2. Materials and Methods

2.1 Bacterial strains and culturing

E. coli 10β cells (New England Biolabs) were maintained in Luria-Bertani (LB) media at 37°C in air with or without the addition of 50 µg ml⁻¹ ampicillin. Lactobacillus plantarum WCFS1 (ATCC BAA-793) and L. reuteri DSM 20016 (ATCC 53609) were obtained from the American Type Culture Collection and maintained in De Man, Rogosa and Sharpe (MRS) media at 37°C in air, 5% CO₂, or in an anaerobic environment generated by an Anaerobic Gas Generator bag (Thermofisher Scientific). Lactobacillus acidophilus NCFM was isolated from Respiratory Care Probiotic (SKU #733739029096, Now Foods, Bloomingdale, IL, USA) and confirmed via PCR amplification and sequencing of 16S ribosomal DNA (35). Lactobacillus acidophilus NCFM was maintained in 5% CO₂ or anaerobically as specified above using MRS media.

2.2 Plasmid construction

All plasmids were assembled using in vivo recombination (36). In brief, plasmid backbones and inserts were PCR amplified with Q5 Hotstart High-Fidelity polymerase (New England Biolabs) following the manufacturer’s protocol (1 ng DNA, 1X Q5 reaction buffer, 0.25 mM dNTPs, 0.5 µM forward and reverse primer, 0.02 U µl⁻¹ Q5 Hotstart polymerase and water to 5 µl) to generate 25 bp overlap regions. After initial amplification of seventeen cycles, products were diluted 1:100 in water and amplification was repeated for seventeen cycles more before adding 0.5 µl each of backbone and insert reactions to 50 µl Chemically Competent 10β cells (New England Biolabs) for heat shock transformation. Transformants were selected by plating 10β cells on LB agar with 50 µg ml⁻¹ ampicillin (LB/amp).

Plasmid pTRK-892 (Addgene) containing the erythromycin resistance gene ermC, the pWV01 replicon, and the gusA1 reporter was modified by the addition of an ampicillin resistance gene from pUC19 as well as the deletion of the gusA1 reporter to generate the base plasmid pTRKa. The base pTRKa vector was used to generate promoter::reporter constructs by inserting superfold Green Fluorescent Protein (sfGFP) with select regions of genomic DNA isolated from L. plantarum WCFS1, L. reuteri DSM 20016 and L. acidophilus NCFM using the Promega Wizard Genomic DNA extraction kit (Promega). All constructs were verified by sequencing.

2.3 Electroporation

Electrocompetent Lactobacillus cells were made using a modified version of the protocol described elsewhere (37), inoculating MRS media with 2% overnight culture (v/v) and growing at 37°C shaking for 3 h, followed by gently pelleting cells at 5000 × g in 4°C and washing twice with cold water, once with cold 50 mM EDTA (pH 8.5), again with cold water, and then twice with cold electroporation solution (0.5 M sucrose, 10% glycerol) before resuspending cells in 1 ml cold electroporation solution and aliquoting to 90 µl on ice. Cells were electroporated following a protocol described elsewhere (38) with some modifications, wherein 10 ng plasmid DNA and 90 µl electrocompetent cells were combined and incubated on ice for 5 min, followed by pulsing 1 kV mm⁻¹ for 5 ms and immediate recovery with 1 ml MRS supplemented with 80 mM MgCl₂. Cells were grown 37°C for 24 h in recovery media and plated on MRS agar with 5 µg ml⁻¹ erythromycin (MRS/ery) to incubate in either 5% CO₂ or in an anaerobic chamber to select transformants.

2.4 Fluorescence measurements

Overnight cultures were prepared by the addition of a single colony to 3 ml MRS/ery and incubating 37°C shaking. The following day, 7 ml MRS/ery was inoculated with 1% (v/v) overnight culture and grown similarly, with 1 ml aliquots taken at predetermined time intervals to be frozen in −80°C after resuspension in 0.5 ml RNA Later (Qiagen), and 0.5 ml aliquots taken for measurements of optical density at 600 nm (OD₆₀₀) and sfGFP fluorescence after resuspension in 0.5 ml phosphate buffered saline (PBS). Fluorescence and OD₆₀₀ measurements were taken in triplicate using a Bio-Tek Synergy Neo2 plate reader (BioTek), where sfGFP fluorescence was measured using excitation and emission wavelengths of 485 and 510 nm, respectively. Fluorescence measurements were also obtained via Accuri C6 Flow Cytometer (BD Biosciences) using the Green Fluorescence filter applied to 10 000 cells. All data were analyzed by Student’s two-tailed paired t-test.

2.5 RT-qPCR

Cells frozen in RNA Later were thawed and incubated in 400 µL Lysis Solution (1 mg ml⁻¹ lysozyme, 50 mM EDTA, pH 8.5) for 30 min at 37°C. The lysis reaction was pelleted at 16 000 × g for 2 min and resuspended in 1 ml cold RNAzol RT (Molecular Research Center) for RNA extraction following the manufacturer’s protocol. Final RNA pellets were resuspended in 20 µL pure water (Sigma Aldrich) and quantified by Nanodrop (Thermofisher Scientific). Reverse Transcription was carried out with iScript RT Supermix (Bio-Rad) using 45 ng total RNA in a 20 µL reaction following the manufacturer’s protocol. Resulting cDNA was used for qPCR with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) following the manufacturer’s protocol. Analysis of qPCR was carried out with the comparative Ct method (39) using rpoB as a reference gene (40).

2.6 Plasmid stability

Longevity of plasmid retention was assessed by cultivation of transformed L. plantarum in the absence of erythromycin in both MRS broth and agar media in parallel with controls containing the antibiotic. A single colony was used to inoculate or streak a plate of specified MRS media and grown overnight at 37°C shaking (broth) or 5% CO₂ (plate), and the following morning 1 ml of cells were pelleted and then resuspended in PBS for OD₆₀₀ and fluorescence measurements. In addition, flow cytometry was used to partition cultures into transformed and cured subgroups based on the green fluorescence of 10 000 cells per sample in order to assess percent plasmid retention over time. Cultures were streaked to MRS/ery agar each day, and new overnight cultures were inoculated with 0.5% (v/v) from the previous day’s culture to repeat the analysis and track plasmid activity over time.

2.7 Microscopy

Cultures of L. plantarum containing pTRKa-9G were grown as described above overnight in 1 ml MRS/ery. Cells were pelleted and washed twice with 1× PBS before resuspending in 400 µl PBS containing 20 µg ml⁻¹ 4ʹ,6-diamidino-2-phenylindole (DAPI, Thermofisher Scientific) for 5 min. Cells were then washed further as above and resuspended in 0.5× volume with PBS, where 10 µl cell suspension was pipetted onto 20 × 20 mm glass coverslips (VWR, Radnor, PA, USA). The cover glass was inverted onto precleaned microscope slides and sealed with acrylic nail polish. Confocal and widefield images were taken with a Zeiss AxioObserver.Z1/7 LSM 800 Airyscan confocal microscope with Plan-Apochromat 63×/1.40 Oil DIC M27 or 40×/1.3 Oil DIC (UV) VIS-IR M27 objectives. For confocal images, DAPI and GFP fluorescence was excited at 353 nm: 0.20% laser power and 488 nm: 0.20% laser power, respectively. The emission spectra of DAPI was collected with 400–490 nm filters and detected with the LSM 800 GaAsP Pmt-1 detector. GFP emission spectra was collected with 490–550 nm filters and detected with the LSM 800 Airyscan detector. Each image was taken with a 1.03 µs pixel dwell, 5.06 s scan time per frame with 4× averaging. Images are maximum intensity projections composed of 0.27 µm optical sections over a 1.89 µm Z-stack interval in a 20.28 × 20.28 µm field of view. For widefield images, GFP fluorescence and brightfield signals were collected over a 322.8 × 322.8 µm field of view. GFP fluorescence was excited with 475 nm LED-module using 455–483 nm filters. Emission spectra were collected using 499–539 nm filters over a 150 ms exposure by a Hamamatsu ORCA-Flash 4.0 C13440 camera. Images were recorded using the Zeiss Zen Blue imaging software (Carl Zeiss, LLC, Thornwood, NY, USA).

2.8 Materials availability

All DNA constructs will be provided upon request following the completion of a Materials Transfer Agreement and any other documentation that may be required.

3. Results and discussion

3.1 Plasmid construction

Shuttle vectors used in various LAB have been previously characterized (9) and demonstrated to replicate in certain Lactobacillus species (11) based on the use of the pWV01 RCR origin. The previously mentioned group had constructed a series of vectors (pTRK) derived from the pGK12 plasmid containing this origin and an erythromycin resistance gene ermC, which were capable of being maintained in E. coli as well as the Lactobacillus species studied. We found, however, that the maintenance of this vector using E. coli was unreliable at best, for common lab strains of E. coli have no problem growing in LB media concentrations of up to 2 mg ml⁻¹ erythromycin regardless of whether they contain these plasmids (data not shown). In order to simplify culturing requirements, we decided to modify the pTRK plasmid with the ampicillin resistance cassette from pUC19 that has been shown to function in E. coli, therefore generating the plasmid series pTRKa (Figure 1A, Table 1). As an alternative to more traditional but time-consuming β-glucuronidase reporter assays typically carried out with Lactobacillus species, we sought the use of the highly stable superfolder green fluorescent protein (sfGFP) reporter (41). The incorporation of this quickly folding fluorescent reporter allowed the immediate visualization of promoter activity without the technical error involved in traditional gusA assays that require partial lysis of cells, the addition of substrate, and prolonged incubation in order to visualize enzyme activity via formation of a colorimetric product. The establishment of the pTRKa system as a viable shuttle vector with a versatile and convenient reporter system gave us a platform with which to screen promoter activity.

Figure 1.

Plasmid design for this study. The pTRKa plasmid (A) was created to contain erythromycin and ampicillin resistance genes for the evaluation of promoter activity in controlling expression of superfolder GFP. Derivatives of the pTRKa plasmid include pTRKa4 with a truncated version of the pWV01 origin (B) and pTRKa3 containing the ColE1 origin (C). All plasmid sizes given correspond to versions containing the tuf promoter region from L. plantarum WCFS1 (−9G).

Table 1.

Plasmids used in this study

Plasmid	Description	References
pUC19	AmpR, ColE1 origin
pET28	KanR, ColE1 origin
pTRK-892	EryR, pWV01 origin, gusA1	(11)
pTRKa	pTRK-892 containing AmpR from pUC19 and ΔgusA1	This study
pTRKa-3G	pTRKa with sfGFP controlled by coaA promoter from L. plantarum WCFS1	This study
pTRKa-3.1G	pTRKa-3G with 10 bp added between the RBS of the coaA promoter and the sfGFP ORF	This study
pTRKa-5G	pTRKa with sfGFP controlled by fba promoter from L. plantarum WCFS1	This study
pTRKa-5.1G	pTRKa-5G with 10 bp added between the RBS of the fba promoter and the sfGFP ORF	This study
pTRKa-7G	pTRKa with sfGFP controlled by pyk promoter from L. plantarum WCFS1	This study
pTRKa-8G	pTRKa with sfGFP controlled by gapA promoter from L. plantarum WCFS1	This study
pTRKa-9G	pTRKa with sfGFP controlled by tuf promoter from L. plantarum WCFS1	This study
pTRKa-15G	pTRKa with sfGFP controlled by pox2 promoter from L. plantarum WCFS1	This study
pTRKa-16G	pTRKa with sfGFP controlled by fusA1 promoter from L. plantarum WCFS1	This study
pTRKa-19G	pTRKa with sfGFP controlled by typA promoter from L. plantarum WCFS1	This study
pTRKa-9Ga	pTRKa with sfGFP controlled by tuf promoter from L. acidophilus NCFM	This study
pTRKa-9Gr	pTRKa with sfGFP controlled by tuf promoter from L. reuteri DSM 20016	This study
pTRKa-9Gr-R	pTRKa with sfGFP controlled by reverse tuf promoter from L. reuteri DSM 20016	This study
pTRKa3-9G	pTRKa-9G with ColE1 origin from pET28 in place of pWV01 origin	This study
pTRKa4-9G	pTRKa-9G with pWV01 truncated as described in (32)	This study

The identification of constitutive promoters in Lactobacillus species has previously been undertaken based on the generation of a consensus sequence from rRNA promoters (12) as well as a more data-driven technique of basing construct designs on previously obtained microarray data (11, 42). We wanted to employ a combination of both ideas, and selected constitutive promoters displaying a variety of activities based on the rationale of focusing on genes important to Lactobacillus survival in addition to those from previously published proteomic data (43). As previous work had been done to confirm the constitutive activity of the phosphoglucomutase (pgm) promoter in L. acidophilus NCFM, it seemed prudent to assess the activity of promoters controlling important glycolytic enzymes from L. plantarum WCFS1. As such, promoters from the genes encoding fructose bisphosphate aldolase (fba), glyceralde-3-phosphate dehydrogenase (gapB) and pyruvate kinase (pyk) were selected based on their roles in this integral metabolic pathway, as well as that of pyruvate oxidase (pox2) based on its role in acetate production (44). Considering the indispensable nature of Coenzyme A synthesis throughout all life (45), the promoter governing expression of the pathway’s initial enzyme pantothenate kinase (coaA) was assessed. Another protein important for cell survival is Elongation Factor Tu encoded by the tuf gene, and its appearance as a top identified protein from a recent proteomic analysis of Lactobacillus cultures (43) prompted us to look further into the activity of its promoter. Also included were the promoters of genes encoding Elongation Factor G (fusA1) based on a similar requirement for translation and ribosomal recycling (46), and a TypA class protein (typA) for its relatively unknown function in translation (47, 48).

3.2 Translational promoter activity

The translational activity resulting from mRNA derived from the selected promoters was assessed by fluorescence measurements at 4, 7 and 24 h postinoculation to represent early-, mid- and late-log phase/early stationary phase of aerobic growth in MRS media. The tested promoters displayed activities spanning approximately three orders of magnitude from 10² to 10⁵ fluorescence units per 10 000 cells at each observed timepoint via flow cytometry (Figure 2A). While all were constitutive in nature, the activity of some promoters appeared to be influenced by the progression through log phase toward stationary phase of growth. The highest sfGFP protein yield of the studied promoters belonged to the tuf and typA promoters (pTRKa-9G and pTRKa-19G, respectively). Considering the importance of Ef-Tu to protein synthesis, it is no surprise that the translational activity related to this promoter would be maintained at a high level throughout active growth phases. The overall trend of the tuf promoter activity was characterized by high activity throughout growth with the low point occurring during mid-log phase at 7 h. This low point was mild and determined to only be 1.6-fold below levels observed at 4 h and 1.2-fold lower than measurements at 24 h (P < 0.05). Therefore the highest activity displayed by the tuf promoter occurred during early-log phase. Previous investigators have assessed tuf promoter activity in L. plantarum CD033 using a different origin of replication and showed peak fluorescence between 16 and 24 h (10), but considering differences in experimental format it is hard to compare activity at specific times between our studies. Contrary to the previous study, however, we observed a relatively stable amount of fluorescence from our cultures throughout growth, which could be due to the differing methods of detection as well.

Figure 2.

Promoter activity was assessed by fluorescence measurements (A) and RT-qPCR (B) of cultures during early-log phase (4 h), mid-log phase (7 h) and the late-log phase/stationary phase transition (24 h). Experiments were carried out in triplicate and values represent mean values of fluorescence obtained via flow cytometry of 10 000 cells (A) or mean expression values from triplicate technical replicates of RT-qPCR (n = 9) targeting sfGFP mRNA produced from respective promoters. mRNA expression values are normalized to the lowest value of typA at 24 h. Error bars represent standard deviation.

The peak activity we observed from the tuf promoter at 4 h was considered the standard to which other promoter activities would be compared throughout experiments, where High Activity was defined as <10-fold lower, Medium Activity as >10-fold lower and Low Activity as >100-fold lower, comparatively. Fluorescence measurements of wild-type (WT) L. plantarum cells were also taken at each timepoint and showed a low level of fluorescence above what might have been expected as a background (Figure 2A). Before continuing with further promoter characterization, images of wild-type L. plantarum cells and pTRKa-9G transformants were taken individually using both confocal (Figure 3) and widefield microscopy (Supplementary Figure S1) to establish that fluorescence readings of wild-type cultures were background measurements and distinctly different than any signal related to sfGFP.

Figure 3.

Confocal microscopy of WT L. plantarum cultures or L. plantarum transformed with pTRKa-9G that produced sfGFP under the control of the tuf promoter. Scale bar represents 2 µm.

The analysis of the typA promoter showed an opposite trend than what was observed with the tuf promoter, where mid-log phase was determined to be the high point of activity rather than the low point (Figure 2A). The overall levels of GFP were comparable to those observed for the tuf promoter, with 2.2- to 5.7-fold lower protein production throughout growth (P < 0.05), establishing the typA promoter as High Activity. While the GFP produced by the typA promoter was similar for early- and mid-log phases, a 2-fold reduction was noted during the mid- to late-log phase transition (P < 0.05). The functions of TypA-type GTPase proteins are relatively unknown outside of their shared domains with other elongation factors and subsequent putative roles in translation (48). These proteins are not widespread throughout prokaryotes, but there have been studies linking them with stress responses in Proteobacteria and Firmicutes that note their functions to vary among species (47). Given the reduction in protein production seen during the late growth phase in our experiments, the typA promoter may be reduced in activity related to the stationary phase transition (49).

The fusA1 promoter (pTRKa-16G) displayed much lower activity than the tuf and typA promoters, with up to 80-fold and 30-fold less observed protein (P < 0.05), respectively (Figure 2A). sfGFP production from the fusA1 promoter did not change significantly from 4 to 7 h, but then increased nearly 7-fold during the subsequent log to stationary phase transition (P < 0.05). Due to this alteration of activity, the fusA1 promoter could be characterized as a promoter with transitioning Low to Medium Activity throughout growth. While Ef-G is equally important for protein synthesis as Ef-Tu, there is an implied complexity in the regulation of these two genes as the fusA1 promoter activity was >70-fold lower than that of the tuf promoter during early- and mid-log phase and 14-fold lower during stationary phase (P < 0.05). The Ef-G protein typically exists in the str operon with ribosomal protein genes (50), but L. plantarum WCFS1 contains two of such annotated proteins (51). The gene fusA2 (lp_1027) is located within the str operon, but fusA1 (lp_0076), whose promoter is studied here and shares homology with fusA2, is located in a cluster of uncharacterized genes putatively involved in transport. While still annotated as Ef-G, fusA1 likely encodes an Ef-G duplication that functions as a GTPase in transport. Given this putative activity of fusA1, it is understandable that the levels of protein originating from its promoter are so different from those of the tuf promoter.

The coaA promoter was also chosen based on its importance for survival, as it controls the expression of the indispensable pantothenate kinase (PanK) enzyme involved in Coenzyme A synthesis. The activity of this promoter (pTRKa-3G) was interesting in that it showed the highest activity in the early-log phase timepoints and subsequently reduced throughout growth (Figure 2A). Compared with early-log phase of growth, the coaA promoter produced roughly 15-fold and 20-fold less GFP at 7 and 24 h (P < 0.05), respectively. The coaA promoter was one of three showing a trend of reduced activity over time, but it displayed the largest decreases in protein production of all investigated promoters, with the activity at the latest timepoint drawing close to previously determined background fluorescence levels of WT cultures. At its highest level the coaA promoter was observed to produce 21-fold less protein the tuf promoter (P < 0.05), and at its lowest level it produced 401-fold less (P < 0.05). The temporal variation represented a 1.3 Log change of activity, and established this promoter as transitioning Medium to Low Activity. CoA synthesis has been studied extensively in bacteria (52–56), and the coaA gene encodes PanK as the initial rate-limiting enzyme of this pathway. Investigations in E. coli have determined that PanK activity is tightly regulated by the state of CoA within the cell, where free reduced CoA was a powerful downregulator of pantothenate phosphorylation as a form of negative feedback (57). The ratio of free reduced CoA to its thioesters has also been noted to increase based on shifts from glucose as a carbon source, and this also likely coincides with late growth phase transitions. It is therefore possible that this proposed late growth phase transition-induced feedback extends from enzyme inactivation to translational repression as well.

The glycolytic gene promoters, while variable among different tested candidates, proved to be the most stable promoters over the course of experiments. Activity of the fba promoter (pTRKa-5G) decreased over time, albeit slowly with a 1.4-fold decrease from early- to mid-log phase (P < 0.05) followed by an insignificant change during the mid-log to stationary phase transition (Figure 2A). The protein production by the fba promoter at its peak was only 8-fold lower than that of the tuf promoter (P < 0.05), and at its lowest was merely 17-fold lower (P < 0.05). Due to the change in activity over time being minor (only 2-fold change throughout the entire experiment), the promoter was determined to be Stable and at the higher end of the Medium Activity group. The fba gene encodes the enzyme fructose bisphosphate aldolase that is responsible for the cleavage of the six carbon fructose bisphosphate into the two phosphorylated trioses dihydroxyacetone phosphate (DHAP) and glyceraldehyde-3-phosphate (G3P) in glycolysis. As one of the higher-expressed proteins identified by proteomic studies on L. plantarum (43, 58), the activity of fructose bisphosphate aldolase is a major regulatory step in glycolysis, for intracellular concentrations of fructose bisphosphate are used to tune the outcome of glycolysis in nutrient-rich and nutrient-deprived situations (59). Based on the reliance of the cell on the activity of this enzyme to determine the fate of such a central metabolic pathway, it is understandable that protein production from its promoter would be steady.

The gapB gene encodes the enzyme glyceraldehyde-3-phosphate dehydrogenase that natively catalyzes the next step in glycolysis. The promoter of gapB (pTRKa-8G) was shown to produce up to 3-fold more protein than the fba promoter throughout growth. Peak activity for this promoter occurred during early-log phase, but the overall variation only reached a 1.24-fold difference during mid-log phase (P < 0.05) with no significant changes thereafter. Based on the 5-fold difference in activity from the tuf promoter (P < 0.05), the gapB promoter could be considered Stable and High Activity. As previous investigators have noted GapB to be high in abundance in proteomic studies of L. plantarum (43, 58), the high activity of this promoter along with that of the fba promoter comes as no surprise. While some investigators have shown GapB in Staphylococcus aureus to be regulated dependent on glucose concentrations (60), others have used the gapB gene as a highly stable reference gene for Reverse Transcriptase - quantitative Polymerase Chain Reaction (RT-qPCR) (61). In S. aureus, the two enzymes GapA and GapB are responsible for opposite reactions in glycolysis and gluconeogenesis, respectively. Therefore, the regulation of each based on glucose availability serves as an efficient energy conservation strategy. L. plantarum, on the other hand, only has one copy of this enzyme (GapB) that functions in both pathways. The observed glucose regulatory mechanisms in S. aureus therefore likely do not exist in L. plantarum, and steady-state concentrations of protein are likely required in order for the cell to adjust to the culture’s changing environment throughout growth.

The pyk gene encoding pyruvate kinase is natively involved in a reaction four steps downstream from that of GapB in glycolysis. The promoter for pyk (pTRKa-7G) was observed to decrease in activity over time (P < 0.05 for each time) but not nearly to the degree observed with the coaA promoter. Similar to observations with the fba promoter, the pyk promoter only decreased a maximum of 2.2-fold throughout the experiment, and could similarly be considered stable. Its overall activity was below the fba promoter, with between 3- and 4-fold less protein produced throughout experiments (P < 0.05). The pyk promoter was therefore determined to be Stable and Medium Activity. Like the fba gene, pyk encodes an enzyme that is an important control point for the fate of glycolysis. Responsible for the ATP generating conversion of phosphoenol pyruvate (PEP) to pyruvic acid, the pyruvate kinase enzyme has been shown to be one of the lower abundance glycolytic proteins in L. plantarum (58). Lactobacillus casei has been shown to coexpress both pyruvate kinase and phosphofructokinase, and studies have suggested this expression to increase in the presence of phosphotransfer system sugars such as glucose (62, 63). While there is enough intergenic space between pyk and pfk in L. plantarum, it is unknown whether the predicted promoter sequence from that region is actually a functional promoter. Based on previous observations that mRNA from the pyk gene decreases with pH (64), it is possible that the reduction of pH with the progression through different growth phases is responsible for the noted decrease in protein concentration by this promoter.

The final promoter assessed from the glycolytic pathway was that of the gene pox2 (pTRKa-15G). Unlike the other glycolytic promoters, peak activity from the pox2 promoter was observed during mid-log phase. Protein production at 7 h was determined to be 1.5-fold higher than at 4 h (P < 0.05) and 3-fold higher than at 24 h (P < 0.05). This variation was observed to be lower than previously mentioned coaA and fusA1 promoters, but was still the most variable of the glycolytic promoters analyzed. Despite this relative variation, it could still be considered a Stable Medium Activity promoter. The pox2 gene encodes the enzyme pyruvate oxidase that catalyzes the conversion of pyruvate to acetyl-phosphate, which can subsequently be used by alternate enzymes to generate CoA or to produce acetate and ATP. Investigators have noted Pox activity to be repressed when L. plantarum is grown anaerobically on glucose, as cells prefer to produce lactate over acetate (44). While protein production from the pox2 promoter is not characterized as low as stated above, the observed 3-fold decrease in protein from 7 to 24 h may be similarly attributed to the propensity for lactate production over acetate in late growth phases, as the two pathways of acetate and lactate production diverge one step upstream from Pox.

The assessed promoters were observed to produce fluorescence ranging three orders of magnitude with varying temporal patterns and expression levels throughout growth. While the lowest protein levels observed were expressed by the coaA promoter at 24 h, its activity was still quite high at the 4 h timepoint. In fact, all studied promoters but one were observed to be near the same level of activity as the coaA promoter or higher at 4 h. We were therefore curious whether some degree of directed mutation could be employed in order to lower the observed activity of different promoters in order to reach a lower range during early-log phase. Previous work had shown the effects of the spacer region length between the ribosomal binding site of a promoter and the start codon of the ORF to have a peculiar effect on translation (10). We therefore wanted to alter these spacer regions in some of our characterized promoters to tune expression to the desired low level. The coaA promoter spacer region was therefore increased from four to fourteen nucleotides by the addition of the sequence CACAGCCACC just upstream of the start codon to sfGFP (pTRKa-3.1G), and the same was done with the fba promoter to increase the spacer from eight to eighteen nucleotides (pTRKa-5.1G) yielding the coaA* and fba* promoters, respectively (Figure 4A). The alterations to the coaA promoter reduced activity 34-fold during early-log phase (P < 0.05), and this reduction brought observed fluorescence near WT levels (Figure 4B). The changes to the fba promoter, however, reduced activity 56-fold at 4 h (P < 0.05) to levels that were notably above background at 7 h (P < 0.05) via flow cytometry (Figure 4B). The protein production observed from the fba* promoter was stable similar to activity seen with the fba promoter, where fluorescence levels only changed by 10% throughout the experiment. With these specific changes, the fba* promoter was diminished in classification from Stable High Activity to Stable Low Activity, and displayed the desired Low Activity during early-log phase.

Figure 4.

Promoter mutation to reduce levels of protein expression at the 4 h mark was carried out following the scheme showing the addition of 10 nts directly upstream of the start codon (A). Fluorescence measurements were taken from cultures at different phases of growth by flow cytometry of 10 000 cells, where bars represent the mean of biological replicates and error bars represent standard deviation (n = 3).

The protein outputs described above represent a variety of options to consider when designing constructs with specific constitutive expression patterns in mind. While all promoters are endogenous to the host L. plantarum, their expression activities span approximately three orders of magnitude with different temporal patterns. Stable protein production throughout growth can be achieved at various levels considering the tuf, gapB, typA and pyk promoters. High expression early in growth that tapers off to baseline levels afterwards is possible through the use of the coaA promoter. The use of the fusA1 promoter will allow relatively low expression throughout growth that spikes near the late-log phase/stationary phase transition. Expression tuning of certain promoters is also possible such as the reduction of pH to reduce the output of the pyk promoter, or the demonstrated promoter alterations to the fba promoter that reduce its early-log phase activity over an order of magnitude resulting in a lower activity version of this promoter. Due to the endogenous nature of these promoters to L. plantarum, they are not immune to cellular control mechanisms as orthogonal promoters might be. Therefore some of these promoters would be better suited for growth in microcosm experiments as shown above, while others could benefit from continuous culture where growth phase transitions could be avoided while nutrient levels and pH are maintained.

3.3 Transcriptional promoter activity

Promoters were checked for their transcriptional activities by RT-qPCR, and the resulting data were compared with protein production in order to hypothesize the degree of posttranscriptional regulation associated with each tested promoter. The tuf promoter was considered to have the highest activity at the protein level, and similar results were noted at the RNA level. The mRNA generated by this promoter was remarkably highest at the 24 h timepoint (P < 0.05) showing 7-fold more mRNA than at 4 h (Figure 2B), which was the timepoint of peak translational activity as described above. Based on the inconsistent trend of tuf promoter activity from RNA to protein levels, it seemed apparent that a degree of posttranscriptional regulation was preventing the translation of protein during the late-log to stationary phase transition. While the translation from the tuf promoter was considered to be relatively stable with less than a 2-fold difference in protein levels between different timepoints, the transcription from this promoter was observed to vary up to 13-fold (P < 0.05).

The coaA promoter showed similar results to those from the tuf promoter at the mRNA level, for both promoters had high activity in early-log phase at the protein level with increases in mRNA abundance at 24 h (P < 0.05) (Figure 2). Considering the lack of protein observed from the coaA promoter at 24 h, it was apparent that there was a strong translational repression occurring during the later growth phase. The activities of both the tuf and coaA promoters were interesting considering that the tuf promoter only produced 11-fold more mRNA at 4 h (P < 0.05), but resulted in 21-fold more fluorescence from protein production (P < 0.05). Furthermore, the tuf promoter only produced 11-fold more mRNA than the coaA promoter at 7 h (P < 0.05), but resulted in 194-fold more protein (P < 0.05). It is unclear whether this is the result of increased translation rates from the tuf promoter, reduced and tightly regulated translation from the coaA promoter, or a combination of both.

Transcription from the typA promoter was observed to decrease over time, unlike its previously observed protein production (Figure 2B). Levels of mRNA during the late-log to stationary phase transition at 24 h were shown to be 17-fold lower than those from early-log phase, and were also determined to be the lowest of all measured promoters. The observed changes in mRNA and protein from the typA promoter might indicate a downregulation of transcription but not necessarily of translation during this late growth transition. Such observations could indicate an increased mRNA stability due to untranslated regions (UTRs) associated with the mRNA transcribed by this promoter, for protein production was also shown to increase at 7 h while mRNA decreased. Overall, however, it is hard to draw strong conclusions from this mRNA data due to the observed variability.

The fusA1 promoter was shown to produce a stable amount of mRNA, with transcript abundance varying insignificantly throughout experiments. The highest transcriptional activity from fusA1 was observed at 7 h during mid-log phase, at 2.2-fold higher than early-log and 1.5-fold higher than late growth. Given the protein production increases and reduction in mRNA levels observed during late growth, it seems that there is a stability associated with transcripts produced from the fusA1 promoter similar to what was observed with the typA promoter. Furthermore, the increase in protein seen in late growth indicates a possible late growth induction characteristic that likely originates from the promoter’s UTR.

The glycolytic gene promoters on the protein level were all considered to be stable with varying levels of activity. On the mRNA level, however, they were quite different (Figure 2B). The pyk promoter showed a trend of producing insignificantly changing mRNA levels with significant decreases in protein production (P < 0.05). The fba promoter, however, was shown to significantly decrease mRNA abundance by 7.4-fold between early- and mid-log phase (P < 0.05), and then increase 13-fold during the late growth phase (P < 0.05). These changes are quite different from the steady 1.4-fold decrease in protein from 4 to 7 h (P < 0.05) and 7 to 24 h as previously described. The variation in mRNA levels from the fba promoter did not appear to affect translation, therefore reinforcing the promoter’s status as Stable with Medium Activity. While the fba promoter was apparently regulated by posttranscriptional means, in comparison it seemed that the pyk promoter was not. The gapB promoter was similar to the pyk promoter, in that the variation in mRNA levels over time were minute, much like the variation observed with protein production from this promoter. It seemed, therefore, that the gapB promoter was also not regulated by posttranscriptional means in a similar manner to the fba promoter. Finally, the pox2 promoter showed an increasing amount of mRNA throughout experiments that was not consistent with its protein production. The early- to mid-log phase levels of mRNA were shown to increase over 6-fold while protein only increased 1.5-fold (P < 0.05), and the mid-log to stationary phase times showed stable mRNA levels compared with protein levels that decreased 3-fold (P < 0.05). While the other glycolytic genes showed trends of protein levels relatively following mRNA levels, the pox2 promoter is an example where stable mRNA over time resulted in a decrease in translation. Considering the Pox enzyme functions as a branching point after glycolysis, its regulatory mechanisms are likely independent from those of the fba, pyk and gapB promoters. Based on the differences in activity from the pox2 promoter it seems the emphasis on regulation occurs more on the translation level.

While all promoters tested produced changes in the range of 500-fold difference of mRNA levels, the effect of these changes did not always correspond to similar changes on the protein level. Indeed, it is not a universally accepted norm that all transcriptional changes must be reflected on the protein level. Based on the construction of each tested promoter: fusion construct, the only variables related to translation rates would originate from the 5ʹ UTR of each respective promoter. Studies on the 5ʹUTR in E. coli showed that replacement of a UTR with one from a known labile gene such as ompA can increase mRNA half-life up to 4-fold higher based on intrinsic secondary structures (65–67). Transcriptomic studies have also shown that increased secondary structure of UTRs is negatively correlated with gene expression, and poorly structured regions upstream of the start codon can allow nonspecific ribosomal binding (68). We therefore wanted to look at the possibility of secondary structure formation from promoter UTRs, especially those that showed inconsistent trends with both transcription and translation. Considering that exact transcriptional start sites are not universally located within promoters and require experimental validation, we considered the 5ʹUTRs for our purposes to be between the −10 element of each promoter and the AUG start codon of the sfGFP mRNA (Table 2). The length of UTRs ranged from 39-nt (fusA1 promoter) to 228-nt (fba promoter), and each was capable of producing some sort of secondary structure as determined by mFold (69). The free energies associated with each secondary structure were calculated to range from −0.9 kCal mol⁻¹ (gapB promoter) to −40 kCal mol⁻¹ (fba promoter). The associated free energies allowed us to speculate the possibility of secondary structure formation, and taken together with trends observed from mRNA and protein production by each promoter the likelihood of structure-related posttranscriptional regulation could be proposed. The strongest examples include the pyk and gapB promoters that show relatively consistent mRNA and protein levels throughout growth, contain shorter UTRs, and have fewer predicted UTR secondary structures with smaller free energy calculations than the other studied promoters. Such data further implies a lack of UTR structure-related posttranscriptional regulation. Additionally, the tuf, fba and coaA promoters are examples with longer UTRs, more predicted UTR secondary structures with larger calculated free energies, and also show inconsistencies between mRNA and protein levels at certain times throughout growth. These observations, in contrast, do provide evidence of posttranscriptional regulation involving UTR structure.

Table 2.

5ʹ untranslated region spacers

	−10 to RBS		RBS to AUG		−10 to AUG
ID	Sequence (5ʹ-3ʹ)	Length (nt)	Sequence (5ʹ-3ʹ)	Length (nt)	Sequence (5ʹ-3ʹ)	Length (nt)	Predicted Structures	kCal/mol range
coaA	TTGAATTTTAATCAGCGTTTGATAAGGGTTTCATGTTATAATCAACCCATTAATTTAGTAATGGTGGCAGACAGAAATCAT	81	TCAG	4	TTGAATTTTAATCAGCGTTTGATAAGGGTTTCATGTTATAATCAACCCATTAATTTAGTAATGGTGGCAGACAGAAATCATGGAAGATCAG	91	2	−14.5
fba	ATTCACAATTAATTATTGATTAAATTCGAAGCAACTTTTGGCTTGATCTGTAACTGATTTTATATAAATGGAAGGGTTTACAAACGGGCGGTTGTCAAATAAGCAAGAATTGACGGGGAATCGTTGGTTAATAAATATGACAAATCGGTTATCACTCCTATGTAAAATGTGCTAAAATGAATGTGATTCAAAAATCGTATCTCTTAATTAACCT	214	GTTATTTT	8	ATTCACAATTAATTATTGATTAAATTCGAAGCAACTTTTGGCTTGATCTGTAACTGATTTTATATAAATGGAAGGGTTTACAAACGGGCGGTTGTCAAATAAGCAAGAATTGACGGGGAATCGTTGGTTAATAAATATGACAAATCGGTTATCACTCCTATGTAAAATGTGCTAAAATGAATGTGATTCAAAAATCGTATCTCTTAATTAACCTAGGAGGGTTATTTT	228	8	−38 to -40
pyk	AAAAATGACGTTAATGTCTTTTTATAAAAACTTTCAA	37	AGATTTTTCTT	11	AAAAATGACGTTAATGTCTTTTTATAAAAACTTTCAAGGGAGAGATTTTTCTT	53	1	−6.8
gapB	TCCACA	6	AAATTCTAGT	10	TCCACAAGGAGGAAATTCTAGT	22	4	−0.9 to -1.9
gapBa	GAAGTTGCTTTTTAAAATAAAACTATTTTTATTTCCACA	39	AAATTCTAGT	10	GAAGTTGCTTTTTAAAATAAAACTATTTTTATTTCCACAAGGAGGAAATTCTAGT	55	1	−6.9
tuf	ACTAACTAAAGAATTGTTGAGACCATTTTGGCCTCGACGTTATTCTTGCGAAAATCAC	58	TTTCATTA	8	ACTAACTAAAGAATTGTTGAGACCATTTTGGCCTCGACGTTATTCTTGCGAAAATCACAGGAGGTTTCATTA	72	2	−16.4
tufa	ATTTAAGGATTCTCAGTGATGGGTGCGCGATTTGGCCTTTTCACTAGGATGTAGTATAATACTAACTAAAGAATTGTTGAGACCATTTTGGCCTCGACGTTATTCTTGCGAAAATCAC	118	TTTCATTA	8	ATTTAAGGATTCTCAGTGATGGGTGCGCGATTTGGCCTTTTCACTAGGATGTAGTATAATACTAACTAAAGAATTGTTGAGACCATTTTGGCCTCGACGTTATTCTTGCGAAAATCACAGGAGGTTTCATTA	132	6	−30.6 to -32.1
pox2	ACGTTTGTTAAATGAATTGAACCAAGCAGATAAATTGGTTCTAAGTCG	48	TGAATAGTG	9	ACGTTTGTTAAATGAATTGAACCAAGCAGATAAATTGGTTCTAAGTCGAAGGGATGAATAGTG	63	1	−12.3
fusA1	GTTAATTCCATTTATATTGA	20	TAACACATTTTCTA	14	GTTAATTCCATTTATATTGAGAGGAGTAACACATTTTCT	39	1	−5.7
typA	ATGTAAGATTTCTGCTGTCGAAGATAGTCAGAATTTTGAAGGATGA	46	TCAATT	6	ATGTAAGATTTCTGCTGTCGAAGATAGTCAGAATTTTGAAGGATGAAAGGAAGATCAATT	60	2	−9.7 to -9.9
coaAb	TTGAATTTTAATCAGCGTTTGATAAGGGTTTCATGTTATAATCAACCCATTAATTTAGTAATGGTGGCAGACAGAAATCAT	81	TCAGCACAGCCACC	14	TTGAATTTTAATCAGCGTTTGATAAGGGTTTCATGTTATAATCAACCCATTAATTTAGTAATGGTGGCAGACAGAAATCATGGAAGATCAGCACAGCCACC	101	2	−16.4 to -16.6
fbab	ATTCACAATTAATTATTGATTAAATTCGAAGCAACTTTTGGCTTGATCTGTAACTGATTTTATATAAATGGAAGGGTTTACAAACGGGCGGTTGTCAAATAAGCAAGAATTGACGGGGAATCGTTGGTTAATAAATATGACAAATCGGTTATCACTCCTATGTAAAATGTGCTAAAATGAATGTGATTCAAAAATCGTATCTCTTAATTAACCT	214	GTTATTTTCACAGCCACC	18	ATTCACAATTAATTATTGATTAAATTCGAAGCAACTTTTGGCTTGATCTGTAACTGATTTTATATAAATGGAAGGGTTTACAAACGGGCGGTTGTCAAATAAGCAAGAATTGACGGGGAATCGTTGGTTAATAAATATGACAAATCGGTTATCACTCCTATGTAAAATGTGCTAAAATGAATGTGATTCAAAAATCGTATCTCTTAATTAACCTAGGAGGGTTATTTTCACAGCCACC	238	6	−41.1 to -42.9

^aSecond predicted promoter upstream.

^bAltered promoters with extended regions between RBS and AUG.

The assessment of transcription provides evidence of promoters fitting a variety of activity patterns, both in terms of transcription level and of temporal activation. The typA and tuf promoters provide transcriptional activity over two orders of magnitude different at 24 h. The pyk and typA promoters display decreasing transcription over time. The pox2 promoter proved capable of increasing mRNA abundance throughout growth. The fba and gapB promoters showed decreases in mRNA during mid-log phase of different degrees, whereas the fusA1 promoter showed mRNA increases at that time. Taken together with consideration of the apparent posttranscriptional regulation observed with the coaA, fba and tuf promoters, as well as the lack of such regulation observed with the pyk and gapB promoters, we provide evidence of an array of potential activities that can be engineered based on the desired outcome. For example, it is plausible that the High Stable expression of the tuf promoter can be tuned to show an increase in protein production by swapping its UTR for that of the pyk promoter. Conversely, the High Stable expression of the gapB promoter can hypothetically be tuned to significantly decrease translation during late growth phases by replacing its UTR with the coaA UTR. We believe that we have provided convincing evidence that Lactobacillus plantarum has the great potential for new synthetic biology applications outside of those established by E. coli, not only for industrial use for the heterologous production of complex proteins, but for use in healthcare fields as an engineered probiotic capable of housing complex genetic circuits.

3.4 Plasmid stability and broad-range activity

The segregational stability of RCR plasmids is known to be comparably worse than their theta-replicating counterparts (27). Certainly the inability to be passed on to daughter cells in a selective environment can render a plasmid system useless for any application; however, RCR plasmids remain viable options for synthetic biology solutions. Various RCR plasmids have been shown to persist in different LAB species using nonselective media (26) dependent on the strain tested and the specific origin used. While there is evidence of plasmid viability in L. plantarum CD033, this organism only contains a single native plasmid (70) in comparison to L. plantarum WCFS1 that contains three (71). Due to these examples and the known strain diversity within the Lactobacillus genus (25) we wanted to assess the stability pTRKa plasmids containing various origins of replication in L. plantarum WCFS1.

We found that cultures containing plasmids with the pWV01 origin (pTRKa, Figure 1A), a truncated version of the pWV01 origin previously designed to improve replication efficiency (pTRKa4, Figure 1B) (32), and the ColE1 origin (pTRKa3, Figure 1C) were capable of replication and had varying stability in the absence of selective pressure (Figure 5). The growth yields of cultures carrying each plasmid varied, indicating that replication of pTRKa3 and pTRKa4 were taxing (Figure 5A), but growth of all cultures was observed to continue over time in the absence of selective pressure while continuing to produce fluorescence (Figure 5B). Furthermore, pTRKa persisted nearly twice as long in the absence of antibiotics compared with pTRKa3 and pTRKa4. On Day 11 of continuous subculturing when cultures with pTRKa3 and pTRKa4 showed <2% plasmid retention by flow cytometry analysis, samples were subcultured into selective media to recover the plasmid-containing population. At this point, however, only pTRKa-containing cultures were able to be recovered (Supplementary Figure S2). Therefore the pTRKa plasmid showed higher stability in L. plantarum WCFS1 than tested plasmids with other origins of replication, and showed a longer retention half-life than similar origins of replication tested in L. plantarum CD033 (26). Given the proposal for our characterized vector and promoters to be used for L. plantarum WCFS1 engineering, the observation of such plasmid retention is important as this presents the investigator with different options outside of genome engineering for designing a synthetic biology solution.

Figure 5.

(A) Fluorescence of L. plantarum WCFS1 cultures containing pTRKa vector with different origins of replication after 18 h growth at 37°C. (B) Plasmid retention in the absence of selective pressure of pTRKa plasmids containing various origins of replication in L. plantarum WCFS1 as measured by partitioning of cell subgroups by flow cytometry. Y-axis represents the percentage of cells maintaining plasmid with subculturing over time (X-axis).

Considering that the pWV01 origin has been shown to replicate in LAB such as L. lactis (72) and L. acidophilus NCFM (11), we wanted to test whether our pTRKa system containing this origin was transferable to other common Lactobacillus species such as L. reuteri DSM20016 that has previously been shown to replicate other common plasmids containing pAMβ1 (15) and ColE1 (73) origins. A similar pTRKa plasmid was constructed using the tuf promoter from L. reuteri to drive expression of sfGFP (pTRKa-9Gr). This plasmid was shown to successfully replicate in L. reuteri DSM20016 and produce greater amounts of sfGFP when grown anaerobically (P < 0.05) and lesser amounts when grown aerobically (P < 0.05) compared with L. plantarum WCFS1 grown in similar conditions (Supplementary Figure S3A), indicating that it can also serve as a vector for engineering L. reuteri DSM20016 in addition to L. acidophilus NCFM and L. plantarum WCFS1.

The tuf promoter has been demonstrated as highly active in some tested Lactobacillus strains (10, 14) in addition to what is shown above. Considering that the promoter sequences from different LAB sources vary, we were curious whether tuf promoters from other LAB species might function in L. plantarum WCFS1. We therefore constructed reporters for the tuf promoter from L. reuteri DSM20016 (pTRKa-9Gr) and L. acidophilus NCFM (pTRKa-9Ga) and found that they, too, can function in L. plantarum WCFS1 with slightly decreasing outputs, respectively (P < 0.05), that remain on the same order of magnitude activity to its native tuf promoter (Supplementary Figure S3B). Promoter activity was also noted in E. coli 10β subcloning cells, indicating the potential for preliminary screening of constructs, which might serve as a qualitative time-saver during vector construction. For example, when the direction of the tuf promoter from L. reuteri DSM20016 is reversed away from the sfGFP ORF, we noted >100-fold loss of activity in subcloning cells (P < 0.05, Supplementary Figure S3B).

Plasmids containing the Lactococcus-derived pWV01 origin have been considered broad range for quite some time (72). Given the diversity of LAB one cannot take for granted the fact that such broad-range vectors will work in every tested strain. A number of factors can determine whether a strain will take on a plasmid, for example stringent endogenous restriction enzyme systems (74) or the presence of native plasmids integral to survival that are within the same compatibility group. Therefore, the construction or acquisition of a broad-range vector may not result in positive transformants in every organism of choice, and strain specific testing must be done in order to insure plasmid compatibility and stability. Recent investigators have constructed shuttle vectors compatible with E. coli and target organisms such as L. casei (28), L. lactis (O’Sullivan and Klaenhammer, ³⁰) and L. plantarum WCFS1 (13), but these vectors house two origins of replication to satisfy the replicative needs of both species. Our pTRKa systems, on the other hand, contain a single origin of replication, and have been demonstrated to replicate in L. plantarum WCFS1, L. reuteri DSM20016 and E. coli. These results in conjunction with those from previous investigators (11) offer viable options for engineering multiple LAB species in addition to L. plantarum WCFS1.

4. Conclusions

Here, we describe the construction of a modified pWV01 origin-containing plasmid with dual antibiotic selection markers that allow straightforward subcloning in E. coli as well as transformation into different Lactobacillus species. To provide support for the use of L. plantarum WCFS1 as a potential synthetic biology chassis, we characterized the transcriptional and translational activity of eight promoters native to L. plantarum WCFS1 at distinct times during growth, and were able to identify a variety of expression ranges and patterns as well as potential solutions for tuning expression levels via UTR modification. We demonstrated plasmid stability of our pTRKa system in L. plantarum WCFS1 and suggest its use as a quick and temporary alternative to genomic engineering. Furthermore, the ability for E. coli to recognize the tested Lactobacillus promoters is a promising aspect of genetic circuit engineering with the potential for quick screening of simple protein outputs as a time-saver. The tools characterized in this study, in addition to the probiotic characteristics of the included species, together create the intriguing potential for the use of L. plantarum engineering as a synthetic biology solution for use in both industrial and healthcare settings.

Funding

Funding for this work was provided by Core funds provided the U.S. Naval Research Laboratory and by the Office of the Secretary of Defense funded Applied Research for the Advancement of S&T Priorities (ARAP) Synthetic Biology for Military-Relevant Environments (SBME) program.

All sequence data files are included in the Supplementary Data or can be requested from the corresponding author.

Conflict of interest statement. None declared.