Construction of a Mycobacterium smegmatis Promoter Library for Therapeutic and Environmental Applications

TL;DR

This paper creates a reliable library of genetic promoters in Mycobacterium smegmatis to support its use in synthetic biology for medical and environmental purposes.

Contribution

The study introduces a dual-channel reporter system to characterize 18 M. smegmatis promoters, reducing variability in expression measurements.

Findings

A dual-channel reporter system was used to identify 18 M. smegmatis promoters with consistent expression levels.

The promoter library and reporter plasmid provide a standardized tool for future synthetic biology applications in M. smegmatis.

The system reduces extrinsic noise from culturing conditions and reporter gene choices.

Abstract

Mycobacterium smegmatis is a nonpathogenic
species of soil-dwelling mycobacteria that shows promise as a synthetic
biology chassis with both clinical and environmental applications.
The development of a nonmodel chassis requires a library of regulatory
genetic elements that cover a range of expression levels. Currently,
most studied M. smegmatis promoters
are characterized using single-channel reporter cassettes, which are
vulnerable to extrinsic noise introduced by different culturing conditions,
initial cell metabolic states, and reporter genes of choice. For constructing
predictable and reliable circuits in M. smegmatis, this study systematically identified and analyzed 18 M. smegmatis promoters by using a dual-channel reporter
system across different environments. Here, we show a well-characterized
promoter library and a standardizable reporter plasmid construct that
will allow future investigators to easily assess additional promoter
elements, promoting future use of M. smegmatis as an effective and field-deployable chassis.

Figures

Promoter Characterization Pipeline. (A) Library Construction: pSUM9.01 - pSUM9.19 are promoter reporter plasmids constructed using 3G assembly. Each promoter reporter plasmid contains a control TU that encompasses the promoter AB_004 upstream of an sfGFP coding sequence and a test TU that encompasses a test promoter upstream of an mCherry coding sequence. The negative control plasmid pSUM.tts has a terminator in place of a test promoter. (B) Measurements under Different Conditions: M. smegmatis MC2155 transformed with promoter reporter plasmids were grown in three different conditions: standard 7H9 media, nutrient-limiting 7H9 media, and 7H9 media supplemented with soil-extracted soluble organic matter (SESOM). OD600, green fluorescence, and red fluorescence measurements were taken every two h over ninety-eight h using a Synergy H1Microplate Reader. (C) Data Preprocessing: Promoter characterization was conducted either over a 16 h exponential phase period or an early stationary phase time point. Autofluorescence of M. smegmatis was used to normalize control TU green fluorescence outputs. M. smegmatis transformed with pSUM.tts was used to normalize the test TU red fluorescence outputs. Locally Estimated Scatterplot Smoothing (LOESS) models were built to normalize the TU outputs for each experimental run. After normalization, arbitrary units (AU) of fluorescence were converted to Molecules of Equivalent Fluorescein (MEFL) or Molecules of Equivalent Sulforhodamine 101 (MESR) using standard curves. (D) The slopes of the linear models fit by the control TUs’ or the test TUs’ fluorescence outputs with respect to OD600 were denoted as the control TU outputs (m control TU) or the test TU outputs (m test TU). The ratio between m test TU and m control TU was denoted as the ratiometric characteristic (α) of the test promoters. The ratios between the ratiometric characteristics of the test promoters and the control construct pSUM9.04 with reference promoter AB_004-Psmyc were calculated to obtain relative promoter strengths (ρ). n = 6. For a full description of the experimental procedures, refer to the Methods section.

Characterization of Test Promoters in Standard 7H9Media. (A) Average growth curve of M. smegmatis transformed with reporter plasmids. n = 108, ribbon = ± SD. The horizontal box plot displays the distribution of the initial time points of the exponential growth phase. n = 108. (B) Controls the TU outputs of all test promoter constructs. Green fluorescence data points were normalized using untransformed M. smegmatis. Normalized control TU outputs exhibit differences across test promoter constructs (Kruskal–Wallis rank-sum test, p < 0.001). n = 6. (C) Heat map of test TU outputs and ratiometric characteristics across test promoters (y axis) with six biological replicates (x axis). Both metrics were normalized by the mean of AB_004’s corresponding metric. Six data points of negative values were omitted after the log transformation of the color scale (five data points of AB_007 and one data point of AB_002). (D) Comparison of the coefficient of variations (CV) of promoter characteristic measurements using single-channel control TU outputs, single-channel test TU outputs, and dual-channel ratiometric characteristics. CV is defined as SD/mean. Each point represents a different promoter construct. The distribution of ratiometric characteristics displays a CV lower than both single-channel outputs across constructs. The Kruskal–Wallis rank-sum test, p = 0.0054 for control TU output vs ratiometric characteristic, *p < 0.001 for test TU output vs ratiometric characteristic. n = 6 per promoter construct. (E) Ratiometric characteristics of all test promoters. Reference promoter AB_004-Psmyc is highlighted in yellow. Six data points of negative values were omitted after the log transformation of y axis (five data points of AB_007 and one data point of AB_002). Each point represents one replicate. n = 6. (F) Relative promoter strength obtained in this study did not correlate with relative RPKM reported in the literature. Dashed lines represent the linear relationship between the relative promoter strength and relative RPKM on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R 2 = 0.0945. (G) Test promoters’ relative promoter strengths obtained from the exponential growth phase and the early stationary phase generally align with each other. Dashed lines represent the linear relationship between the relative promoter strengths obtained from the two growth phases on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R 2 = 0.9166.

Characterization of Test Promoters in Nutrient-Limiting 7H9 Medium and 7H9 Supplemented with SESOM. (A) Average growth curves of M. smegmatis transformed with reporter plasmids in 7H9, NL-7H9, and Soil-7H9. n = 108 per growth condition, ribbon = ± SD. (B) Average OD600 when M. smegmatis culture exits the exponential growth phase in three different culturing conditions. M. smegmatis cultures in NL-7H9 enter the stationary phase at a lower OD600 than cultures in other culturing conditions. Wilcoxon rank-sum test, p < 0.0001 for 7H9 vs NL-7H9 and 7H9 vs Soil-7H9. (C) Fold-change of ratiometric characteristics of test promoters in NL-7H9 or Soil-7H9 relative to their ratiometric characteristics obtained from standard 7H9 cultures is, in general, conserved. Dashed lines represent linear relationships between average ratiometric characteristics of test promoters in NL-7H9 or Soil-7H9 and ratiometric characteristics in 7H9 on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R2 = 0.8576 for NL-7H9 vs 7H9, adjusted R 2 = 0.9698 for Soil-7H9 vs 7H9. (D) Two strong promoters, AB_001 and AB_009, have statistically nonsignificant relative promoter strengths across three culturing conditions. Wilcoxon rank-sum test, p > 0.05 for all comparisons. n = 6. (E) Five promoters, AB_012, AB_013, AB_015, AB_016, and AB_019, exhibit differential relative promoter strengths between 7H9 cultures and NL-7H9 cultures. n = 6. Wilcoxon rank-sum test, p < 0.0043 for AB_012, *p < 0.041 for AB_013, p < 0.0022 for AB_015, p < 0.0022 for AB_016, p < 0.0022 for AB_019. (F) Two promoters, AB_010 and AB_014, exhibit differential relative promoter strengths between 7H9 cultures and Soil-7H9 cultures. n = 6. Wilcoxon rank-sum test, p < 0.0087 for AB_010, p < 0.0043 for AB_014.

Full text

Introduction

Mycobacterium smegmatis is a fast-growing, nonpathogenic species that possesses multiple features that make it a particularly attractive chassis for synthetic biology. Since the introduction of a transformable M. smegmatis strain, namely M. smegmatis MC^2^155, ?,?

M. smegmatis has served as a model system for investigating the properties of pathogenic mycobacteria, including clinically important species such as Mycobacterium tuberculosis and nontuberculous mycobacteria pathogens such as Mycobacterium abscessus.? With advances in recombinant DNA technology and multiomics techniques, researchers have investigated M. smegmatis as a vector or a host for antituberculosis and antitumor vaccines, ?−? ? ? ? as well as phage therapy. ?−? ? ? Additionally, due to its natural sterol metabolic pathways and its resistance to stressful bioreactor conditions, M. smegmatis can be effectively utilized by the biomanufacturing industry for the biotransformation of high-value pharmaceutical sterol intermediates. ?,?

M. smegmatis has also been utilized as a model organism for characterizing metabolic pathways such as atmospheric carbon monoxide and hydrogen oxidation pathways. ?,? Additionally, due to its ubiquity across both aquatic and terrestrial biomes, ?,?,?

M. smegmatis has been probed as an environmental chassis capable of operating beyond the standard laboratory setting.?

A collection of quantitatively characterized gene regulatory elements covering a spectrum of strengths is fundamental for the implementation of a successful synthetic biology chassis. Because of its significance, a number of mycobacterial promoters ?−? ? ? and promoter collections ?,? have been identified, constructed, and characterized using single-channel reporter cassettes. However, such single-channel assays are vulnerable to changes in global variations, such as plasmid copy numbers, different culturing conditions, and the varying state of the initial inoculated cells, undermining both the reproducibility among different laboratories and the transferability of such measurements from laboratory to field conditions. ?,? Multiple approaches have been proposed to enhance the reliability and reproducibility of promoter characterizations, including the utilization of a reference promoter, ?,? a dual-channel reporter plasmid system? and multicolor fluorophore-calibrated fluorescence units.? Adopting these approaches for promoter characterization, we systematically identified and analyzed 18 M. smegmatis promoters in different environments to create a well-characterized promoter library as well as a generalizable backbone plasmid construct that will allow future investigators to easily assess additional promoter elements and employ M. smegmatis as an efficacious and field-deployable synthetic biology chassis.

Results and Discussion

Construction of a Dual-Channel Promoter Reporter Plasmid Library

To enable reproducible measurements that could be compared across varying laboratory and environmental conditions, we designed a dual-channel fluorescence plasmid system to measure promoter activity in M. smegmatis (FigureA). This system contains two divergent transcriptional units (TU), namely, a test transcriptional unit (test TU) and a control transcriptional unit (control TU) separated by a bidirectional terminator. Both TUs are flanked by unique nucleotide sequences (UNS), which simplifies cloning procedures for future applications.? The test TU contains the test promoter to be analyzed, followed by a ribosomal binding site (RBS),? the mCherry coding region,? and the bidirectional ttsbiB terminator.? The control TU contains the reference promoter P_smyc_, ?,? a strong constitutive promoter followed by the same RBS sequence used for the test TU, the sfGFP coding region? and the bidirectional ttsbiB terminator. The control TU enables normalization of the output of the test TU, reducing extrinsic variations introduced by factors such as different culturing conditions, metabolic burden imposed by the plasmid, or differences in assay equipment, thus allowing the accurate assessment of the intrinsic characteristics of the test promoters across varying conditions and platforms.?

*Promoter Characterization Pipeline. (A) Library Construction: pSUM9.01

• pSUM9.19 are promoter reporter plasmids constructed using 3G assembly. Each promoter reporter plasmid contains a control TU that encompasses the promoter AB_004 upstream of an sfGFP coding sequence and a test TU that encompasses a test promoter upstream of an mCherry coding sequence. The negative control plasmid pSUM.tts has a terminator in place of a test promoter. (B) Measurements under Different Conditions: M. smegmatis MC2155 transformed with promoter reporter plasmids were grown in three different conditions: standard 7H9 media, nutrient-limiting 7H9 media, and 7H9 media supplemented with soil-extracted soluble organic matter (SESOM). OD600, green fluorescence, and red fluorescence measurements were taken every two h over ninety-eight h using a Synergy H1Microplate Reader. (C) Data Preprocessing: Promoter characterization was conducted either over a 16 h exponential phase period or an early stationary phase time point. Autofluorescence of M. smegmatis was used to normalize control TU green fluorescence outputs. M. smegmatis transformed with pSUM.tts was used to normalize the test TU red fluorescence outputs. Locally Estimated Scatterplot Smoothing (LOESS) models were built to normalize the TU outputs for each experimental run. After normalization, arbitrary units (AU) of fluorescence were converted to Molecules of Equivalent Fluorescein (MEFL) or Molecules of Equivalent Sulforhodamine 101 (MESR) using standard curves. (D) The slopes of the linear models fit by the control TUs’ or the test TUs’ fluorescence outputs with respect to OD600 were denoted as the control TU outputs (m control TU) or the test TU outputs (m test TU). The ratio between m test TU and m control TU was denoted as the ratiometric characteristic (α) of the test promoters. The ratios between the ratiometric characteristics of the test promoters and the control construct pSUM9.04 with reference promoter AB_004-Psmyc were calculated to obtain relative promoter strengths (ρ). n = 6. For a full description of the experimental procedures, refer to the Methods section.*

A total of 18 M. smegmatis promoters, covering a wide range of promoter strengths, were selected as test promoters for analysis in this study (Table S1). Five promoters, AB_001 (rpsL),? AB_002 (ftsZ P2),? AB_003 (MSMEG_5228),? AB_004 (P_smyc_) and AB_006 (P_wmyc_)? were selected as benchmarks from promoters previously characterized using single-channel assays in the literature. The other 13 promoters (AB_007–AB_019), not previously characterized, were identified from two published M. smegmatis transcriptome analyses. ?,? Using average reads per kilobase per million mapped reads (RPKM) values obtained with exponential phase cells (sampled after 16 h of growth) reported? as a proxy for promoter strength, we selected promoters ranged from a relative RPKM (average RPKM normalized to the average RPKM of the reference promoter AB_004) of 0.004 for promoter AB_006 to 2.513 for promoter AB_012 (the MSMEG_3050 promoter) in hope for selecting a collection of promoters that cover a wide dynamic range of promoter strengths. Only transcriptional start sites (TSS) that were in agreement between both transcriptome studies were included. The 50 base pairs upstream of the TSS were then used as the test promoters.

Eighteen different reported plasmids (pSUM9.XX) were cloned and confirmed via whole plasmid sequencing (FigureA). OD600, green fluorescence, and red fluorescent measurements were then collected from M. smegmatis transformed with these constructs using a Synergy H1 Microplate Reader (FigureB). Data from either a 16-h exponential growth period? or a single stationary phase time point ?,? were extracted from our time-series data for promoter strength characterization (FigureC). Autofluorescence of M. smegmatis was used to normalize the green fluorescence output of the control TU. For normalizing the red fluorescence output of the test TU, we generated a negative control plasmid (pSUM9.tts) by inserting the bidirectional terminator ttsbiB in place of a test promoter in the test TU. M. smegmatis transformed with pSUM9.tts does exhibit a low basal level of red fluorescence, possibly due to unintentional leakage or noise, but nonetheless higher than untransformed M. smegmatis (Figure S1). Thus, pSUM9.tts was included in each plate run as a negative control for the test TU background fluorescence normalization. Unique LOESS regression models,? giving background green (M. smegmatis no-plasmid control) and red fluorescence (M. smegmatis with pSUM9.tts) as a function of OD600, were built for each plate run (Figure S2). ?,? Normalized fluorescence outputs were then converted from arbitrary units (AU) to Molecules of Equivalent Fluorescein (MEFL) or Molecules of Equivalent Sulforhodamine 101 (MESR) using standard curves (FigureC).? For calculating the ratiometric characteristics, similar to the Rudge et al.? method, the rates of change of fluorescence outputs with respect to OD600 during exponential phase were calculated and denoted as the m test TU and m contol TU. The ratio of m test TU and m contol TU was denoted as the unitless ratiometric characteristic α. The ratio of the α_test promoter_ and α_reference promoter (AB_004‑Psmyc)_ was then denoted as the relative promoter strength of the corresponding test promoter (FigureD).

Dual-Channel Characterization Allows Capture of Intrinsic Promoter

Properties

Overall, even with stringent normalization, there is a presence of noise and variation across our data set, indicated by the varying timing of transitions between growth phases and overall ranging growth patterns across constructs (FigureA). Additionally, control TU outputs (m_contol TU_) differ across different promoter reporter plasmids (FigureB), showing the necessity of the internal normalization using a dual-channel characterization system. Computing the ratiometric characteristics reduces the variation across biological replicates (FigureC), dissipating outlier data points’ influences, and significantly reduces the coefficient of variation (CV) of promoter measurements compared to single-channel outputs (FigureD, Table S2 and S3). These results indicate that our promoter characterization pipeline reduces extrinsic variations rooted in both technical limitations ?,? and the inherent stochasticity in gene expression,? allowing us to capture the intrinsic properties of test promoters.

*Characterization of Test Promoters in Standard 7H9Media. (A) Average growth curve of M. smegmatis transformed with reporter plasmids. n = 108, ribbon = ± SD. The horizontal box plot displays the distribution of the initial time points of the exponential growth phase. n = 108. (B) Controls the TU outputs of all test promoter constructs. Green fluorescence data points were normalized using untransformed M. smegmatis. Normalized control TU outputs exhibit differences across test promoter constructs (Kruskal–Wallis rank-sum test, p < 0.001). n = 6. (C) Heat map of test TU outputs and ratiometric characteristics across test promoters (y axis) with six biological replicates (x axis). Both metrics were normalized by the mean of AB_004’s corresponding metric. Six data points of negative values were omitted after the log transformation of the color scale (five data points of AB_007 and one data point of AB_002). (D) Comparison of the coefficient of variations (CV) of promoter characteristic measurements using single-channel control TU outputs, single-channel test TU outputs, and dual-channel ratiometric characteristics. CV is defined as SD/mean. Each point represents a different promoter construct. The distribution of ratiometric characteristics displays a CV lower than both single-channel outputs across constructs. The Kruskal–Wallis rank-sum test, **p = 0.0054 for control TU output vs ratiometric characteristic, **p < 0.001 for test TU output vs ratiometric characteristic. n = 6 per promoter construct. (E) Ratiometric characteristics of all test promoters. Reference promoter AB_004-Psmyc is highlighted in yellow. Six data points of negative values were omitted after the log transformation of y axis (five data points of AB_007 and one data point of AB_002). Each point represents one replicate. n = 6. (F) Relative promoter strength obtained in this study did not correlate with relative RPKM reported in the literature. Dashed lines represent the linear relationship between the relative promoter strength and relative RPKM on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R 2 = 0.0945. (G) Test promoters’ relative promoter strengths obtained from the exponential growth phase and the early stationary phase generally align with each other. Dashed lines represent the linear relationship between the relative promoter strengths obtained from the two growth phases on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R 2 = 0.9166.

Eighteen promoters were characterized with ratiometric characteristics (p) ranging approximately 2.5 orders of magnitude (FigureE). The characterization of the five previously characterized promoters (AB_001–AB_004, AB_006) from the literature generally aligns, in relative terms, with the assessment of promoter activity in this study. AB_001 was reported to be a strong promoter? and was characterized by this study to possess the highest promoter strength among all promoters tested. Using single-channel TUs, AB_004 was characterized to be driving a 16-fold stronger expression of GFPuv compared to the promoter AB_006.? This relationship was also observed in this study. The average promoter activity calculated for AB_004 is 15.04 times higher than that of AB_006 (FigureE and Table S3). AB_003 was characterized by Uhía et al.? to be a TetR family regulator-repressible promoter located upstream of the gene encoding a 3-b-hydroxysteroid dehydrogenase involved in the cholesterol catabolic pathway. In this study, the relative strength of AB_003 was measured to be 0.041, a relatively weak promoter, potentially due to the absence of cholesterol in the lab culture conditions of M. smegmatis used in this study.

While there is general agreement between the promoter strengths of the previously characterized promoters described above and our assessment of promoter strength, it is notable that the promoter strengths quantified in our study did not typically correlate with the relative RPKM values reported in the literature? (FiguresF and S3). For example, based on the relative RPKM, five promoters were reported to be stronger than our reference promoter AB_004. However, we found that only one of these five promoters (AB_001) displayed a relative promoter strength greater than 1. For example, AB_012 had a relative RPKM of 2.513, the second highest among test promoters; however, in our study, promoter AB_012 had a relative promoter strength of 0.256. Similarly, AB_010 (the MSEMG_1919 promoter) was reported to be a strong promoter according to its relative RPKM (1.184) but was characterized as a relatively weak promoter in our study.

On the other hand, AB_014 (the MSMEG_3976 promoter) was reported to be a weak promoter (relative RPKM of 0.015), while AB_018 (the MSMEG_5180 promoter) was reported to be a medium-strong promoter (relative RPKM of 0.732). However, our study indicates exactly the reverse, with AB_014 exhibiting significantly higher expression levels than AB_018. These discrepancies suggest that the RPKM values from RNA-seq experiments are not suitable for predicting the strength of the corresponding promoter but are rather only appropriate for generating preliminary hypotheses to be tested. This could be due to various post-transcriptional modifications affecting TUs’ outputs, which RPKM does not take into account,? exclusion or truncation of upstream regulatory elements during promoter sequence extraction,? or the different culturing conditions the cells encounter across different experimental procedures and laboratories. Such discrepancies were also noted during other promoter characterization studies with M. smegmatis,? Methylotuvimicrobium buryatense,? Rhodobacter sphaeroides,? Staphylococcus aureus,? and Streptomyces albus,? highlighting the necessity of stringent wet lab experimental validations similar to this study.

Notably, AB_007 displayed a negative average ratiometric characteristic of −0.018 due to pSUM9.07’s test TU mCherry expression being indistinguishable from the background red fluorescence of the negative control plasmid pSUM.tts (Table S3). This correlates to AB_007’s relative RPKM? of 0.005, the second lowest among our test promoters. This suggests one of the shortcomings of fluorescence protein-based assayshaving a high detection limit unsuitable for characterizing weak promoters.? Downstream analysis of AB_007 was still carried out due to the non-negative ratiometric characteristics displayed by some biological replicates.

Recognizing the merits of promoters that are stable across growth phases for both biomanufacturing purposes? and field-deployable applications,? we utilized our data set to characterize the test promoters’ transcriptional activity in the early stationary phase. Using the method similar to Yu et al.,? we calculated the ratiometric characteristic of test promoters over a single time point during the stationary phase (Table S4). We chose to not use the Rudge et al.? method here because of its assumption of a linear relationship between the cells’ fluorescent protein production rate and growth rate no longer holds as cells transition to the stationary phase. Test promoters’ ratiometric characteristics obtained from the exponential and stationary phases are generally consistent (FigureG). Notably, AB_010 shifted from being the 13th strongest promoter during the exponential phase to being the ninth during the stationary phase (Figure S4). AB_010, the MSMEG_1919 promoter, is located upstream of the MSMEG_1919 gene encoding a WhiB-like transcriptional factor.? WhiB-like family proteins are intrinsically disordered transcriptional factors widely distributed in the Actinobacteria phylum and actinobacteriophages, and typically act as regulators in response to oxidative and nitrosative stresses.? Currently, there is no evidence in literature to support WhiB-like transcriptional factors’ expression is upregulated as M. smegmatis transits to the stationary phase, but there is data showing that whcB, a Corynebacterium glutamicum whiB homologue, had a 3-fold expression in the stationary phase over the exponential phase.?

Promoter Characterization in Differential Culturing Conditions

To assess both the reliability of the dual-channel system and the variability of the test promoters’ activity across different culturing conditions, we then characterized this collection of promoters with two additional growth media that mimic the scarcity of essential nutrients found in soil environments (7H9, 0.02% glycerol, w/o albumin–dextrose (AD) supplement (NL-7H9)) and the presence of metabolites distinct from commonly used lab compounds such as glucose and yeast extract (7H9, 0.2% glycerol, supplemented with AD supplement and soil-extracted soluble organic matter (SESOM) (Soil-7H9)) ?,? (FiguresB and ?A). Notably, M. smegmatis transformed with test constructs displays different growth profiles across the three culturing conditions. In general, M. smegmatis grown in NL-7H9 halts exponential growth at lower OD600 values compared to others (FigureB). Normalized relative strengths of each test promoter, combining measurements from all growth conditions, displayed lower coefficients of variation (CVs) compared to the outputs of single-channel test TUs, indicating that the ratiometric promoter characteristics are less affected by the extrinsic variations introduced by differential growth conditions than single-channel fluorescence outputs (Figure S5). We observed that across growth conditions fold changes of ratiometric characteristics of test promoters measured in NL-7H9 or in Soil-7H9 with respect to their corresponding ratiometric characteristics measured in 7H9 are moderately conserved (FigureC), suggesting that promoter measurements were heavily influenced by universal factors dependent on culturing conditions. Additionally, we observed that test promoters generally exhibit higher levels of noise under nutrient-limiting conditions, which matches Sureka et al.?’s results that, in M. smegmatis, an increase in noise in basal gene expression contributes to cells’ transition to alternative transcriptional landscapes in response to stress. After normalizing the ratiometric characteristics of test promoters using the mean ratiometric characteristic of the reference promoter AB_004, the majority of the test promoters displayed a consistent relative strength across different culturing conditions (Table S5), suggesting the necessity of normalizing against a reference promoter, which aids in screening out media-dependent global factors.?

*Characterization of Test Promoters in Nutrient-Limiting 7H9 Medium and 7H9 Supplemented with SESOM. (A) Average growth curves of M. smegmatis transformed with reporter plasmids in 7H9, NL-7H9, and Soil-7H9. n = 108 per growth condition, ribbon = ± SD. (B) Average OD600 when M. smegmatis culture exits the exponential growth phase in three different culturing conditions. M. smegmatis cultures in NL-7H9 enter the stationary phase at a lower OD600 than cultures in other culturing conditions. Wilcoxon rank-sum test, ****p < 0.0001 for 7H9 vs NL-7H9 and 7H9 vs Soil-7H9. (C) Fold-change of ratiometric characteristics of test promoters in NL-7H9 or Soil-7H9 relative to their ratiometric characteristics obtained from standard 7H9 cultures is, in general, conserved. Dashed lines represent linear relationships between average ratiometric characteristics of test promoters in NL-7H9 or Soil-7H9 and ratiometric characteristics in 7H9 on a log–log scale. AB_007 is omitted due to its nonpositive average ratiometric characteristic. Adjusted R2 = 0.8576 for NL-7H9 vs 7H9, adjusted R 2 = 0.9698 for Soil-7H9 vs 7H9. (D) Two strong promoters, AB_001 and AB_009, have statistically nonsignificant relative promoter strengths across three culturing conditions. Wilcoxon rank-sum test, p > 0.05 for all comparisons. n = 6. (E) Five promoters, AB_012, AB_013, AB_015, AB_016, and AB_019, exhibit differential relative promoter strengths between 7H9 cultures and NL-7H9 cultures. n = 6. Wilcoxon rank-sum test, **p < 0.0043 for AB_012, *p < 0.041 for AB_013, **p < 0.0022 for AB_015, **p < 0.0022 for AB_016, **p < 0.0022 for AB_019. (F) Two promoters, AB_010 and AB_014, exhibit differential relative promoter strengths between 7H9 cultures and Soil-7H9 cultures. n = 6. Wilcoxon rank-sum test, **p < 0.0087 for AB_010, *p < 0.0043 for AB_014.

Overall, our characterization pipeline, which includes normalization steps using an internal control TU and a reference promoter construct, enables us to capture the intrinsic characteristics of the test promoters across the culturing conditions. Two strong promoters, AB_001 and AB_009, displayed consistent promoter dynamics across all growth conditions (FigureD), making them ideal candidate parts for constructing synthetic systems intended to be deployed into soil systems. Both AB_001 and AB_009 are located upstream of genes encoding ribosomal proteins (RPs),? suggesting that RP-expressing promoters could be a reservoir for discovering strong constitutive promoters that maintain consistent expression levels when experiencing different conditions.? Multiple sequence alignment of the six strongest promoters characterized in this study revealed a clear −10 box matching the mycobacterial sigma factor SigA motif from previous studies, ?,? providing empirical insights for the rational design of additional strong mycobacterial promoters (Figure S6). Comparing constructs grown in 7H9 and NL-7H9, 5 promoters (AB_012, AB_013, AB_015 (the MSMEG_4272 promoter), AB_016 (the MSMEG_4276 promoter), and AB_019 (the MSMEG_6182 promoter)) displayed a lower activity that was statistically significant (FiguresE and S7). Similarly, 2 out of 18 promoters (AB_010, AB_014 (the MSMEG_3976 promoter)) displayed significantly lower promoter activity when comparing cultures grown in standard 7H9 and Soil-7H9 (FiguresF and S7). The majority of promoters that exhibit differential activity possess low-to-medium strength under standard laboratory conditions and could potentially be regulated promoters that are further repressed under nutrient starvation or the presence of antimicrobial metabolites in soil. ?,? These differentially expressed promoters could be useful for monitoring bacterial stress levels and controlling gene expression in response to stress level fluctuations.

Conclusion

In this study, we constructed a modular system for synthetic biologists to use for accurate M. smegmatis promoter characterization under varying conditions (Figures S8 and S9, Sequences S1 and S2). We constructed 18 promoter-reporter plasmids (Sequences S3–S20) and provided an assessment of 18 M. smegmatis promoters under laboratory and soil-mimicking environments, where we identified promoters suitable for an array of different circuit-design applications. This study lays the foundation for both the characterization and development of M. smegmatis as a fieldable synthetic biology chassis and the construction of field-deployable M. smegmatis systems.

Methods

Bacterial Strains and Growth Conditions

Both Escherichia coli strain 5α (NEB) and DH5α (Thermo Fisher) were used as chassis for plasmid assembly. For growing E. coli liquid culture, Luria–Bertani (LB) medium was used. Cells were placed in incubation at 37 °C with shaking at 250 rpm for 14–16 h until saturation. For plating, LB plates with a 1.5% agar were used. When necessary, antibiotics were added to the medium at the following final concentrations: kanamycin, 50 μg/mL; chloramphenicol, 25 μg/mL. Mycobacterium smegmatis MC^2^155 (ATCC no. 700084) was used for promoter characterization. For starting liquid cultures, M. smegmatis was grown at 37 °C and 250 rpm in Middlebrook 7H9 broth supplemented with 0.2% glycerol, 10% albumin–dextrose supplement, and 3–4 beads per tube to prevent cell clumping. Cultures were grown for 36 h to an optical density measured at 600 nm (OD600) of approximately 0.8–1.0 and diluted 100-fold into either standard 7H9 broth (0.2% glycerol, 10% albumin–dextrose supplement), nutrient-limiting 7H9 broth (0.02% glycerol, w/o albumin–dextrose supplement), or 7H9 supplemented with soil-extracted soluble organic matter (SESOM). SESOM was prepared as described by Valian et al.? For preparing 2X SESOM, 200 g of air-dried soil was suspended in 500 mL of 20 mM MOPS buffer with shaking at 200 rpm at 37 °C for 1 h. Liquid extract was then filtered through coffee filter paper and a 0.22 μM vacuum filtration system (Thermo Scientific). 2X SESOM was then mixed with 2X 7H9 broth with a 1:1 ratio to obtain the 1X 7H9 supplemented with SESOM. For plating, 7H9 plates with 1.5% agar were used. For all M. smegmatis cultures, antibiotics were added to the medium at the following final concentrations: carbenicillin, 50 μg/mL; cycloheximide, 10 μg/mL. When necessary, kanamycin (35 μg/mL) was supplemented into the medium for plasmid-transformant selection.

Physical Parts Isolation

Fragments comprising individual parts were either isolated from M. smegmatis MC^2^155 genomic DNA or available plasmids (Addgene) using PCR, or ordered as linear double-stranded gene blocks (gBlock, Integrated DNA Technologies) (Tables S6 and S7). All promoter fragments are flanked with BsaI cut sites to facilitate downstream seamless insertion into the backbone destination vectors using the Golden Gate assembly. Individual parts were inserted into modified pSB1C3? vectors and stored in E. coli DH5α. All parts were confirmed by Sanger sequencing.

Plasmid Assembly

The overall strategy for plasmid assembly for this study is the 3G assembly described by Halleran et al.? Individual parts were assembled into transcriptional units via the Golden Gate assembly. The constructed transcriptional units were then inserted into a linearized pSUM-lacZ-gfp backbone? via Gibson assembly. At each step, the assembly reaction products were transformed into E. coli DH5α. Transformation products were plated on LB agar plates with the corresponding antibiotics for transformant selection. Colony PCR was used to confirm the successful assembly. Confirmed colonies were then inoculated into the LB liquid medium. All constructs were then extracted and sequence confirmed (Plasmidsaurus, Louisville, KY).

Electroporation of Mycobacterium smegmatis

M. smegmatis electrocompetent cells were prepared and transformed similar to protocols described by Beggs et al.? Briefly, single M. smegmatis MC^2^155 colonies were picked, grown to an OD600 of 0.8–1.0, washed with 10% glycerol and harvested with centrifugation at 5000 rpm 4 times. 1 μg (maximum volume of 5 μL) of DNA was transformed into electrocompetent M. smegmatis using an eporator system (Eppendorf). For each electroporation, one pulse was delivered (time constant: 5.0 ms; field strength: 12 kV/cm). Cells were allowed to recover in supplemented 7H9 broth for 1 h and plated on 7H9 plates with kanamycin (35 μg/mL).

Microplate Reader Fluorescence Measurements

For each promoter construct, six M. smegmatis transformant colonies were picked and grown in supplemented 7H9 media with kanamycin (35 μg/mL) until an OD600 of 0.8–1.2 was reached. 50 μL of the culture were then diluted 100-fold into 5 mL of supplemented 7H9 media with kanamycin (35 μg/mL). For each measurement, 200 μL of each promoter reporter plasmid culture was loaded onto a 96-well plate. Blank media, untransformed M. smegmatis, M. smegmatis transformed with the negative control plasmid pSUM.tts, and a dilution series of fluorophore calibrants were also included for each plate for calibration and data normalization. A Synergy H1 Microplate Reader was used to measure OD600, green fluorescence (excitation wavelength: 488 nm; emission wavelength: 530 nm; bandwidth: 20 nm; gain: 65), and red fluorescence (excitation wavelength: 570 nm; emission wavelength: 620 nm; bandwidth: 20 nm; gain: 100). Readings were taken every two h over ninety-eight h.

Readout Data Normalization and Analysis

Arbitrary fluorescence units were converted into molecules of equivalent fluorophore calibrants (Molecules Equivalent of Fluorescein (MEFL) or Molecules Equivalent of Sulforhodamine-10 (MESR)).? For each plate run, two 2-fold serial dilutions of fluorescein (0.01 to 10.00 μM) and sulforhodamine-10 (0.001 to 2 μM) were included, and standard curves specific to each plate run were constructed. The green autofluorescence of untransformed M. smegmatis MC^2^155 was used to normalize the outputs of the control TUs, and the red fluorescence of M. smegmatis transformed with the negative control construct pSUM9.tts was used to normalize the outputs of the test TUs. Locally estimated scatterplot smoothing (LOESS) regression models ?,?,? were built to predict the two negative control strains’ TUs’ fluorescence outputs at different OD600 values. LOESS models were constructed with the loess function in RStudio.? Six biological replicates of both negative control strains were included for each plate run for constructing LOESS models specific to each plate. OD600 values were normalized by using the median OD600 of the appropriate media specific to each plate. The overall normalization process is as follows:

$$[eqn]$$

where

● F sample (t) = Fluorescence output at time point t

● F LOESS (OD sample (t)) = Negative control’s fluorescence output predicted by the corresponding LOESS regression model

● OD sample (t) = OD600 at time point t

● OD blank = OD600 of the corresponding media

The method from Rudge et al.? was used to calculate the promoter characteristics. For both TUs, the slope m of the fluorescence (normalized with LOESS model outputs) vs OD600 curves during the exponential growth phase was calculated. The exponential growth phase was defined as the eight consecutive time points that had the maximum rate of range in OD600 with respect to time. Linear models were fit using OD600 and fluorescence outputs. The slopes of linear models fitted using control TUs’ green fluorescence outputs with respect to OD600 were denoted as m controlTU. The slopes of linear models fitted using test TUs’ red fluorescence outputs with respect to OD600 were also denoted as m testTU. The ratio between m_testTU_ and m controlTU was denoted as the ratiometric characteristic (α) of a test promoter:

$$[eqn]$$

And to calculate the relative promoter strength (ρ):

$$[eqn]$$

For determining promoter strength at the stationary phase, similar to the approach described by Yu et al.,? the ratio between the normalized outputs of the two TUs at a single time point (6 h after the end of the 16-h exponential growth period) was calculated and normalized to the OD600 value of that time point. The early stationary phase promoter strength of AB_004 was used as the reference to obtain the relative promoter strengths of test promoters at the early stationary phase.

Data Analysis

The fluorescence levels were analyzed using either the Wilcoxon rank-sum test or the Kruskal–Wallis rank-sum test. Significance was set at p value <0.05. Data analyses were performed using R Statistical Software (v4.3.3; R Core Team 2024) and RStudio? with packages including ggbump,? ggplot2,? ggpubr,? and tidyverse.? Multiple sequence alignment was performed using Kalign? and visualized using Jalview.?

Supplementary Material