Global transcription engineering (gTME) was applied to E. coli, Lactobacillus plantarum, and Rhodococcus ruber TH by tailoring the housekeeping σ factor to investigate ethanol tolerance, hyaluronic acid production, lactic and inorganic acid tolerance, and acrylamide tolerance, respectively. The best σ factor mutants showed enhanced stress tolerance than the wild-type strains (Alper and Stephanopoulos, 2007; Klein-Marcuschamer and Stephanopoulos, 2008; Yu et al., 2008; Ma and Yu, 2012).

The introduction of exogenous σ factor into an industrial strain Actinoplanes teichomyceticus L-27 was recently demonstrated. For this strain, the exact principal sigma factors σ^HrdB are not well understood, but the actinomyces genera shares high similarity in σ^HrdB amino acid sequences. Thus, three σHrdB genes from A. missouriensis 431, Micromonospora aurantiaca ATCC27029, and Salinispora arenicola CNS-205 were rationally selected and engineered by random mutagenesis, DNA shuffling, and point mutation, and the resulting library transferred in A. teichomyceticus L-27. Screening yielded a recombinant strain with a twofold increase in teicoplanin production in a pilot scale fermentation (5.3 mg L⁻¹). The mutants showed significant diversity in the region 1.1 of σ^HrdB, apparently generated by DNA shuffling (Wösten, 1998; Campbell et al., 2002; Wang et al., 2014).

While σ^s mainly plays a significant role in the metabolism of E. coli during the stationary phase, it was reported that during the exponential growth phase, deletion of σ^s gene (rpoS) enhanced the tricarboxylic acid cycle and the glyoxylate shunt (Rahman et al., 2006). The rpoS gene was knocked out from a metabolically engineered E. coli for 1-propanol production, and the yield increased by over 100% to about 10 g L⁻¹ (Choi et al., 2012). Similarly, for an E. coli strain producing putrescine, a four carbon linear chain diamine with a wide range of industrial applications, the deletion of rpoS led to a 10% increase in the productivity in both batch and fed-batch cultures (Qian et al., 2009).

An engineered E. coli was constructed for improved isobutanol production by random mutagenesis and selection with an analog of valine (norvaline), which theoretically stimulated the production of the common precursor 2-ketoisovalerate. The mutant strain NV3 produced an improved isobutanol level of about 8 g L⁻¹ in comparison to the wild-type strain with 5 g L⁻¹ in 24 h. A truncation mutation was found in the rpoS gene of the mutant strain. Repair of this gene further elevated the yield of isobutanol to 21.2 g L⁻¹ in 99 h (Smith and Liao, 2011), likely because in this case σ^S was critical for isobutanol tolerance.

The 5′ untranslated region of rpoS mRNA forms an inhibitory loop, which blocks the ribosome binding site and represses the translation of rpoS. Three non-coding sRNAs (DsrA, RprA, and ArcA) can disrupt this loop and overexpression of these sRNAs increased acid tolerance supra-additively up to 8500-fold during active cell growth (Gaida et al., 2013). On the other hand, deletion of any of these sRNA decreased acid resistance (Bak et al., 2014). The sRNA GcvB also positively regulates the rpoS expression level. This sRNA was identified from a single gene knock out library for 79 sRNA in E. coli MG1655. The overexpression of gcvB caused a threefold increase in the rpoS translation. Thus, GcvB can be a potential candidate for modulating rpoS expression and for engineering acid tolerance (Jin et al., 2009).

A stress-induced system was designed to activate polyhydroxybutyrate (PHB) production by placing the synthesis genes under the control of the 5′ untranslated region and the promoter of rpoS in E. coli, which formed a so-called stress-induced region (SIR). The engineered strain produced PHB up to 85.8% of cell dry weight in a glucose medium without the addition of any additional inducer (Kang et al., 2008). PHB with shorter carbon chains and polyhydroxyalkanoate (PHA) with longer carbon chains are two types of well known bioplastics.

The stationary phase σ factor, SigE in cyanobacteria is a positive regulator of sugar catabolism. For the unicellular cyanobacterium, Synechocystis sp. PCC 6803, overexpression of sigE enhanced the transcription of genes for the oxidative pentose phosphate pathway and also for glycogen catabolism, and altered the metabolite levels (e.g., acetyl coA and citrate) of the TCA cycle (Osanai et al., 2011). Further studies revealed more pleiotropic effects. The cell size of sigE over-expressing strain increased likely due to an aberrant cell division. The hydrogen production was also increased under microoxic conditions. However, sigE overexpression caused reduced photosynthesis and respiration due to changes in several regulatory proteins. Regardless of these changes, the engineered strain showed normal growth and viability both under nitrogen replete and nitrogen-limiting conditions (Osanai et al., 2013a). In addition, several cyanobacteria including Synechocystis sp. PCC 6803 accumulate mostly PHB as carbon and energy source during unfavorable growth conditions. Under nitrogen-limiting conditions, sigE overexpression in Synechocystis 6803 increased glucose catabolism, which turned the metabolic pathway toward increased PHB production (Osanai et al., 2013b).

Avermectins that have potent anthelmintic and insecticidal properties are biosynthesized by Streptomyces avermitilis. About 47 ECF σ factors were found in the genome of S. avermitilis. The production of avermectin was negatively regulated by an ECF σ factor, Sig6 in S. avermitilis SAV663. Deletion of sig6 resulted in an increased avermectin yield of ~680 mg L⁻¹, while overexpression decreased the yield by 56–63%. The expression of a pathway-specific activator gene aveR was upregulated in the sig6 deletion mutant, resulting in the increased avermectin production (Jiang et al., 2011).

Alginate is biosynthesized as an exopolysaccharide by two bacterial genera, Pseudomonas and Azotobacter. As mentioned, the expression of alginate biosynthetic genes is regulated by the alternative σ factor AlgU. A mutation in the anti-σ factor MucA resulted in stabilization of AlgU in P. aeruginosa, thereby overproduction of alginate (Martin et al., 1993). An efficient method for alginate production in P. fluorescens SBW25 was subsequently demonstrated (Borgos et al., 2013) using an engineered MucA mutant with a deletion of the 37 C-terminal amino acids. This mutant was found to upregulate various genes including the TCA cycle, ribosomal and translational proteins, and downregulated the NADPH oxidizing cycle, when cells were grown on glycerol in chemostats under nitrogen limitation.

The putative σ factor Orf21 was engineered to study its role in the production of clavulanic acid (CA) in Streptomyces clavuligerus NRRL3585. Disruption of the orf21 gene decreased the CA production slightly, whereas overexpression of orf21 enhanced the production by 1.4-fold. RT-PCR analysis showed that Orf21 activated the synthesis genes ceas2 and cas2, and the activator gene ccaR, of the CA gene cluster (Jnawali et al., 2011).

Solvent tolerance and productivity are two important traits required for biofuel production. An adaptive evolution strategy was developed for improved ethanol tolerance and productivity in Thermoanaerobacter sp. X514. Long-term exposure to ethanol caused the ethanol sensitive wild-type to develop a low ethanol tolerance of 2% (X_I) and subsequently a higher tolerance of 6% (X_II). Genomic and transcriptomic analyses identified an iron containing alcohol dehydrogenase (ADH) and the ECF σ²⁴ as the key factors for improved tolerance for X_I and X_II, respectively. The adh over-expressing strain showed a 33% improvement in the ethanol productivity over the control, and a 31.8-fold enhanced growth under 1% ethanol. Under the same condition, ethanol tolerance was improved by 102-fold by over-expressing σ²⁴, with a 21% higher ethanol productivity than the control (Lin et al., 2013a).

Overexpression of the alternative σ factor σ⁵⁴ in E. coli successfully enhanced the heterologous expression of the oxytetracycline biosynthetic genes from a 32 kb type II oxytetracycline gene cluster. It was reasoned that this non-native cluster nonetheless contained σ⁵⁴ promoters. This provides useful clues for the production of other polyketides (Stevens et al., 2013).

So far, the σ factors have been largely manipulated with the more traditional genetic engineering approach. However, in recent years, several lines of studies using the synthetic biology approach have emerged. For example, an artificial sRNA was obtained from a rationally designed library that bound to the RBS region of the rpoS gene. This RNA alone or together with the sRNA DsrA can act to inhibit the translation of rpoS and convert the activator DsrA to a co-inhibitor (Jin et al., 2013). A theophylline dependent hybrid riboswitch was also created using the lactose inducible promoter for controlling the σ^S expression. When theophylline was added, the rpoS expression was repressed with increased acid sensitivity and increased acetate assimilation. The addition of IPTG restored the σ^S activity with enhanced acid survival and normal acetate assimilation (Jin and Huang, 2011).

Genome mining was used to identify an orthogonal set of 20 bacterial ECF σ factors and their cognate promoters. These σ factors and their corresponding anti-σ factors were further used to build synthetic transcriptional units in E. coli. Additional chimeric σ factors were obtained by swapping the −35 and −10 promoter binding domains from subgroups of ECFs. This study demonstrates the possibility of designing synthetic regulatory network within a cell by utilizing ECF σ factors (Rhodius et al., 2013).

Finally, a bisected version of the T7 polymerase was used to create orthogonal sigma factors. This split polymerase contained a larger catalytic core and a smaller segment for promoter recognition. By engineering the latter, different orthogonal promoter recognition fragments were generated (Segall-Shapiro and Voigt, 2013).