Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G. Search for Related Content PubMed PubMed Citation Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G. Agricola Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, May 2009, p. 2705-2711, Vol. 75, No. 9
0099-2240/09/$08.00+0     doi:10.1128/AEM.01888-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Mutagenesis of the Bacterial RNA Polymerase Alpha Subunit for Improvement of Complex Phenotypes

Daniel Klein-Marcuschamer, Christine Nicole S. Santos, Huimin Yu, and Gregory Stephanopoulos^*

Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Received 13 August 2008/ Accepted 20 February 2009

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Combinatorial or random methods for strain engineering have been extensively used for the improvement of multigenic phenotypes and other traits for which the underlying mechanism is not fully understood. Although the preferred method has traditionally been mutagenesis and selection, our laboratory has successfully used mutant transcription factors, which direct the RNA polymerase (RNAP) during transcription, to engineer complex phenotypes in microbial cells. Here, we show that it is also possible to impart new phenotypes by altering the RNAP core enzyme itself, in particular through mutagenesis of the alpha subunit of the bacterial polymerase. We present the use of this tool for improving tolerance of Escherichia coli to butanol and other solvents and for increasing the titers of two commercially relevant products, L-tyrosine and hyaluronic acid. In addition, we explore the underlying physiological changes that give rise to the solvent-tolerant mutant.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Strain improvement strategies make use of a variety of methods for engineering and isolating microbial variants with desired traits. These techniques often fall under two general categories: "rational" approaches involve the targeted alteration of genetic information (42, 43), while "random" methods are typically based on the creation of mutant libraries containing nondirected changes in genotype with subsequent screening for phenotypes of interest (11). These strain improvement techniques have made possible the commercial-scale production of a variety of compounds, such as amino acids (18, 23), antibiotics and other pharmaceuticals (44, 49), and precursors for organic synthesis and commodity chemicals (8, 28). Because naturally occurring strains rarely perform optimally under ideal process conditions (7, 33, 34) and because the number of compounds that are synthesized microbiologically is projected to increase substantially (29), strain improvement programs are becoming a recurrent theme of industrial and academic research. Progress in this area is particularly important for the production of commodity chemicals and fuels, since profit margins are small and thus even minor gains in efficiency are economically significant (28).

Random approaches for strain engineering have been extensively employed for the improvement of complex or poorly understood phenotypes, such as metabolite overproduction or tolerance to toxic compounds (33, 34, 39, 48). Chemical or physical mutagenesis of the genome followed by screening has been a traditional means of improving several phenotypes, particularly for increasing antibiotic production in the pharmaceutical industry (10, 11). More recently, another approach called transcriptional engineering has been successfully used to improve tolerance and production phenotypes by reprogramming the transcriptome (1, 22, 33a). Because there are several ways in which cells regulate the process of transcription, we were interested in finding new targets for transcriptional engineering. A good target, when mutated, should generate richly diverse populations of strains from which new phenotypes can be selected (22).

Previous transcriptional engineering studies with Escherichia coli have been based on mutagenesis of artificial transcription factors, such as zinc finger proteins, as well as sigma factors (mainly sigma factor D [^D]) of the RNA polymerase (RNAP) (1, 3). The RNAP is a multisubunit complex that directs and regulates transcription in prokaryotes (the subunit composition of the core enzyme is ₂ββ'). The sigma-less RNAP is capable of all steps of transcription except initiation, which requires binding to one of the seven sigma factors (6). Because of their direct role in promoter recognition and binding, sigma factors have generally been regarded as interesting targets for mutagenesis (1, 22). However, although sigma factors are often recognized as the primary determinants of promoter specificity, the alpha subunit of RNAP has also been implicated in promoter recognition, interacting not only with DNA but also with many activator and repressor proteins mainly through its carboxy-terminal domain (C-terminal one-third of the alpha subunit, denoted CTD) (6, 9, 13, 37). Because it is part of the core enzyme, it also has the potential to change the promoter affinity of the RNAP regardless of which of the seven sigma factors is bound to it (6, 19, 20). Therefore, we hypothesized that the alpha subunit, encoded by the gene rpoA, could also be a good target for transcriptional engineering.

In this paper, we explore the use of rpoA mutant libraries generated by error-prone PCR for eliciting three industrially relevant phenotypes: improved butanol tolerance and increased production of the small molecule L-tyrosine and the biopolymer hyaluronic acid (HA). Butanol has recently emerged as an attractive biofuel alternative to ethanol (12, 46), and as such, the problem of mitigating butanol toxicity has been an important concern for large-scale production (2). The aromatic amino acid L-tyrosine serves as a precursor for several pharmaceutical and industrial compounds, from novel polymers and adhesives to biocosmetics, thus making it an interesting target for production in E. coli (27). Finally, HA possesses several health and cosmetic applications (14), and its synthesis in E. coli has recently been reported (47). Each of these complex phenotypes is likely influenced by several genetic factors, thus making these systems particularly well suited for transcriptional engineering studies.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Bacterial strains and cultivation conditions.
L-Tyrosine production experiments used the parental strain E. coli K-12 pheA tyrR::P_LTET-O1 tyrA^fbr aroG^fbr lacZ::P_LTET-O1 tyrA^fbr aroG^fbr/pTrcmelA^mut1 (where the superscript fbr indicates feedback resistance) (26, 27; C. N. S. Santos, unpublished data). These were performed at 37°C with 225-rpm orbital shaking in 50-ml MOPS (morpholinepropanesulfonic acid) minimal medium (Teknova) cultures supplemented with 5 g/liter glucose and an additional 4 g/liter NH₄Cl. Butanol experiments were carried out with DH5 (Invitrogen, Carlsbad, CA), using LB medium (Difco BD, Franklin Lakes, USA), following previously established practices (1). Chloramphenicol at 34 µg/ml was used as needed. HA production experiments were completed with recombinant E. coli Top10/pMBAD-sseABC in LB liquid medium supplemented with Mg²⁺, using L-arabinose as an inducer as previously reported (47).

Library construction.
The native rpoA gene was amplified from genomic DNA using Phusion DNA polymerase (NEB) and cloned into the ApaLI and XmaI sites of the multicloning site of pHACm (1), using NEB restriction enzymes. The correct insert was verified by sequencing, and strains transformed with this plasmid are denoted "wild type" throughout the study. Error-prone PCR was carried out with the same primers using the GeneMorph II kit (Stratagene, La Jolla, CA), and the mutation frequency was varied by changing the initial amount of target DNA from 700 ng, 250 ng, and 25 ng for rpoA*L (low), rpoA*M (medium), and rpoA*H (high), respectively. After ligation with Fast-link ligase (Epicentre, Madison, WI), the libraries were transformed into DH10B cells (Invitrogen, Carlsbad, CA), plated in LB agar, and pooled together after overnight growth. The plasmids were recovered with a miniprep (Qiagen, Valencia, CA) and used to retransform the three host strains. The original size of the library was approximately 10⁵.

The saturation mutagenesis library for rpoA14 was constructed with the QuikChange Multi site-directed mutagenesis kit (Stratagene, La Jolla, CA) by designing primers according to the manufacturer's instructions with degenerate bases to substitute for the codons corresponding to V257 and L281.

Library screening. (i) Butanol tolerance.
DH5 cells transformed with the libraries (5 x 10⁶ colonies were obtained) were pooled together, and cultured on LB liquid medium at 37°C with shaking for 2 h before placing them under selection conditions. Selection for butanol tolerance was carried out in screw-cap shake flasks in 0.9% (vol/vol) butanol. The culture was grown overnight and used to reinoculate fresh selection medium. This process was repeated twice before plating cells to test individual colonies. For verification of the phenotype, plasmids were isolated, reintroduced into a clean background, grown overnight in LB, and inoculated to an optical density of 0.1 in 5 ml of medium in the presence of different concentrations of 1-butanol, 2-butanol, 3-pentanol, or 1-pentanol. Tubes were sealed with Parafilm to avoid evaporation of alcohols.

(ii) L-Tyrosine production.
Libraries of L-tyrosine production mutants were constructed by transforming the parental strain E. coli K-12 pheA tyrR::P_LTET-O1 tyrA^fbr aroG^fbr lacZ::P_LTET-O1 tyrA^fbr aroG^fbr/pTrcmelA^mut1 (C. N. S. Santos, unpublished data) with the rpoA libraries. A total of 7.5 x 10⁵ viable colonies were obtained and subsequently screened for L-tyrosine production as described previously (40).

Saturation mutagenesis libraries for rpoA14 were similarly transformed into an rpoA14 mutant lacking its rpoA plasmid. The plasmid was cured from the strain following four rounds of subculturing in LB medium.

(iii) HA production.
Mutant libraries of HA-producing strain E. coli Top10/(pMBAD-sseABC, pHACm-rpoA) transformed with the rpoA libraries were first screened by a translucent colony phenotype and then selected by alcian blue staining as described previously (48). E. coli Top10/(pMBAD-sseABC) was used as a control and, in total, 76 translucent colonies in the library were stained by alcian blue.

Analytical methods. (i) Determination of pH_i.
The intracellular pH (pH_i) was monitored by expressing a pH-responsive green fluorescent protein (GFP) (a gift from G. Miesenbock and J. Rothman) (30). The response of the GFP variant was monitored as the excitation ratio at 395 nm to 475 nm, and the emission was measured at 530 nm. A standard curve was constructed by resuspending DH5 cells in various buffers as described previously (21) without a carbon source for 45 min at 37°C. For pH_i difference quantification, cells were grown in LB and resuspended in potassium phosphate buffer (50 mM, extracellular pH [pH_e] = 4.7) with 0.5% glucose as an energy source in the presence and absence of butanol.

(ii) Quantification of L-tyrosine.
Cell-free culture supernatants were filtered through 0.2-µm PTFE membrane syringe filters (VWR International) and used for high-pressure liquid chromatography analysis with a Waters 2690 separations module connected with a Waters 996 photodiode array detector (Waters) set to a wavelength of 278 nm. Samples were separated on a Waters Resolve C₁₈ column with 0.1% (vol/vol) trifluoroacetic acid in water (solvent A) and 0.1% (vol/vol) trifluoroacetic acid in acetonitrile (solvent B) as the mobile phase. The following gradient was used at a flow rate of 1 ml/min: 0 min, 95% solvent A plus 5% solvent B; 8 min, 20% solvent A plus 80% solvent B; 10 min, 80% solvent A plus 20% solvent B; and 11 min, 95% solvent A plus 5% solvent B.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Butanol tolerance. (i) Isolation of a mutant with improved growth on butanol.
The rpoA libraries were first tested for their ability to improve tolerance to 0.9% butanol through repeated subculturing in LB medium. After several mutants were isolated and reintroduced into the parental strain, a single mutant (denoted L33) was selected for further characterization. This mutant grew faster (data not shown) and exhibited a higher accumulated cell mass (Fig. 1) than the control in the presence of butanol.

View larger version (14K): [in this window] [in a new window]

FIG. 1. Final optical densities of batch cultures of DH5 cells containing either the wild-type or the L33 mutant of rpoA. Growth experiments were conducted in several different alcohol solvents; the first number of each label indicates the carbon atom at which the hydroxyl group is attached, and the second number represents the total number of carbon atoms in the molecule (e.g., 2-C4 is isobutanol). The concentration of each alcohol used is indicated in parentheses (vol/vol). Error bars indicate standard deviations.

(ii) Tolerance of the isolated mutant to other alcohols.
It is generally believed that butanol toxicity results from fluidization and disordering of membrane lipids and the consequent leakage of ions through it (41). In particular, dissipation of the transmembrane pH gradient has an energy-uncoupling effect (15), similar to that caused by weak acids at low external pH (36, 45). We therefore hypothesized that if the presence of a mutated rpoA gene negated the effects of butanol through a decrease in membrane fluidity or a related response, such a mutant would also exhibit resistance to other solvents known to act via similar mechanisms. We tested the tolerance of the mutant to other alcohols possessing desirable biofuel properties and found that the strain performed better than the control in all cases (Fig. 1).

(iii) Response to ion leakage and energy uncoupling.
In order to test the ability of the mutant to cope with ion leakage and energy uncoupling, we measured the pH_i in the presence of a weak acid at low pH_e with and without added butanol. Cells were acid shocked by resuspending them in potassium phosphate buffer at a pH_e of 4.7. By comparing the pH_is of wild-type and mutant cells in the presence of butanol to that of the control with no added butanol, we were able to test the capacity of the strains to maintain the pH_i when in contact with the alcohol. The results are displayed in Fig. 2. A negative pH difference implies that the pH_i of the strain in the presence of butanol is lower than that without butanol, which is the expected result for the wild type (as shown). In contrast, throughout the experiment L33 maintains a higher pH_i than both the wild-type with the same amount of butanol and the control strain with no butanol (thus, the pH difference is positive). These observations suggest that the mutant L33 copes with the stress either by reducing the fluidity of the membrane, by increasing the rate at which protons are forced out of the cell, or by a combination of both actions. Because the mutant does not show improved growth compared to the wild type in the absence of butanol or other solvents, the phenotype does not seem to be merely a result of a general growth advantage.

View larger version (15K): [in this window] [in a new window]

FIG. 2. Abilities of the mutant and wild-type strains to maintain their pHi in the presence of butanol. The y axis shows the pHi difference between cells resuspended in buffer (pHe = 4.7) with the indicated amount of butanol (vol/vol, %) and cells resuspended in buffer with no butanol. Error bars are not shown, but the coefficient of variation of the pH measurements was on average 0.4%.

(iv) Sequence analysis of mutant rpoA.
Upon analyzing the rpoA plasmid isolated from mutant L33, we found a single nucleotide mutation which replaced amino acid E244 with a stop codon, thus resulting in a truncated protein that lacks the CTD (Fig. 3). Interestingly, similar rpoA mutants have been isolated in the past and have been widely studied. A protein lacking the CTD is capable of being assembled into the RNAP and carrying out transcription; however, it does not respond to signals in the DNA or from protein effectors (16, 17, 25, 31, 37). For example, alpha subunits lacking the CTD do not react to the strong activating signals from the UP regions of rrnB promoters (35, 37) or the cyclic AMP (cAMP)-cAMP receptor protein complex (17). The changes associated with these interactions would be significant even in the absence of other effects, considering that the products of rrnB promoters make up a major fraction of the total RNA in the cell (35) and that more than 100 loci are activated by cAMP receptor protein (5, 25). It is important to note that because the chromosomal copy of rpoA remains intact, the observed phenotype likely arises from the combined action of the truncated and nontruncated forms of the protein.

View larger version (12K): [in this window] [in a new window]

FIG. 3. Schematic mapping of the mutations in the RpoA protein. Gross features, such as the NTD and CTD, the NSH, and the four -helices of CTD (13), are indicated.

L-Tyrosine production. (i) Isolation of a mutant with improved L-tyrosine production.
We began by transforming the rpoA libraries into a preengineered L-tyrosine-producing strain and subjecting it to a high-throughput screen based on the synthesis of the black pigment melanin (40). From an initial library size of 7.5 x 10⁵ mutants, 30 of the darkest strains were selected by visual inspection and tested for L-tyrosine production in MOPS minimal medium. We were able to isolate a strain (denoted the rpoA14 strain) exhibiting a 91% increase in titer above the parental strain with a final concentration of 798 mg/liter L-tyrosine (Fig. 4).

View larger version (13K): [in this window] [in a new window]

FIG. 4. L-Tyrosine production by the parental strain (E. coli K-12 pheA tyrR::PLTET-O1 tyrAfbr aroGfbr lacZ::PLTET-O1 tyrAfbr aroGfbr) and strains harboring mutant rpoA genes. The rpoA14 strain was isolated from the error-prone PCR libraries, and the rpoA22 strain was selected from the saturation mutagenesis libraries. Error bars indicate standard deviations.

(ii) Sequence analysis of mutant rpoA.
Sequencing of the mutant rpoA gene revealed two amino acid changes, V257F and L281P (Fig. 3), located in the CTD near amino acids known to contact regulatory factors and the UP element (31). The first change occurred in the so-called "nonstandard helix" (NSH) of the CTD, while the other was located in one of the four -helices of the CTD (13). It is likely that both of these mutations alter the interaction of the CTD with its target proteins or sequences through changes in the CTD structure. In particular, the amino acid proline has been implicated in destabilizing -helices under some conditions (24). A change in helix conformation could indirectly affect the positions of the amino acids responsible for making contacts with the promoter, thus altering the affinity of the RNAP for some of its targets. Interestingly, the mutant rpoA gene was not successful in increasing L-tyrosine titers when reintroduced into a parental strain background (unpublished data), suggesting that the rpoA14 mutant had acquired an additional chromosomal mutation (or mutations) critical for L-tyrosine production. However, because the background strain by itself also did not exhibit high titers (unpublished data), the rpoA14 plasmid must act synergistically with the chromosomal mutation(s) in order for the phenotypic improvement to be observed. Although such an occurrence may ultimately limit transferability of the phenotype, we do believe that such a phenomenon is likely to be an exception rather than the rule.

(iii) Saturation mutagenesis of the rpoA14 plasmid.
In order to test whether other amino acid substitutions in V257 or L281 could further improve the production of L-tyrosine, we constructed a saturation mutagenesis library with mutations restricted to these two residues. The resulting library was transformed into the rpoA14 mutant devoid of its plasmid (generated by serial subculturing) and screened as before. After 39 of the darkest strains were selected and individually tested, one was chosen for further analysis. This mutant strain, denoted the rpoA22 strain, was found to have only a V257R change in protein sequence. As shown in Fig. 4, along with a modest increase in L-tyrosine titer (15%) compared to the rpoA14 strain, the mutant also exhibited a substantial increase in the rate of production (productivity). At 24 h, when the rpoA14 mutant was still indistinguishable from the parental strain, the rpoA22 mutant had already reached nearly 90% of its final titer. Therefore, this mutant offers an interesting opportunity for developing an industrial platform for the continuous production of L-tyrosine.

Isolation of a mutant with improved HA production.
As an additional phenotype to demonstrate the utility of our new approach, we chose to study HA production in E. coli. Using a previously reported high-throughput screening platform (48), we were able to screen an rpoA library and isolate several improved rpoA mutants. Figure 5 shows the HA concentrations relative to the control, quantified using the alcian blue method as described previously (48). It is clear from these results that significant phenotypic diversity was generated through the introduction of the rpoA library. Because phenotypic diversity is often linked to a greater probability of finding a desirable strain within a library (22), this observation serves as a qualitative indication that rpoA is a good target for transcriptional engineering. The rpoA-HA mutant (Fig. 5) was chosen for further study. A single mutation (L254Q) located in the NSH of the CTD (13) was revealed after sequencing the mutant rpoA gene on the plasmid (Fig. 3).

View larger version (45K): [in this window] [in a new window]

FIG. 5. Alcian blue quantification of HA production by selected mutants of E. coli Top10/(pMBAD-sseABC, pHACm-rpoA). The control strain (black) is Top10/pMBAD-sseABC, and the rpoA-HA mutant is indicated with a hatched bar. All samples were measured in duplicate. Error bars indicate standard deviations.

Testing rpoA plasmid specificity for different phenotypes.
Interestingly, although all of the amino acid substitutions isolated in this study were found in the CTD of the protein, none were located in residues that are known to contact DNA or protein regulators. This observation opens up the possibility that the isolated plasmids may act in a nonspecific fashion and may in fact lead to improvements in phenotypes for which they were not originally selected. For example, it is plausible to hypothesize that the rpoA14 allele may be a CTD loss-of-function variant and therefore may confer better growth in butanol (recall that the L33 allele contained a complete truncation of the CTD, as shown in Fig. 3). The single amino acid substitution in the rpoA-HA CTD may result in a similar loss-of-function mutation. To test the phenotypic specificity of the mutant plasmids, we transformed the rpoA14 and rpoA-HA plasmids into DH5 and measured growth in 0.9% n-butanol as before. Interestingly, neither showed an improvement comparable to that of the L33 mutant (Table 1), which implies that these mutant alleles have retained at least some CTD functionality.

View this table: [in this window] [in a new window]

TABLE 1. Phenotypic specificity of mutant plasmids

A similar experiment for L-tyrosine production was conducted by transforming the L33 and rpoA-HA plasmids into the background of the rpoA14 mutant strain. As shown in Table 1, all three plasmids conferred similar increases in L-tyrosine production, suggesting that small alterations in CTD function, perhaps through partial misfolding, can increase production in this background. Thus, it appears that an L-tyrosine production phenotype is not as discriminatory as a butanol tolerance phenotype and is able to accept several different rpoA alleles to yield the same gains in titer. This idea is further supported by the isolation of other plasmids from the saturation mutagenesis library (e.g., rpoA22) that also exhibited increases in L-tyrosine production. However, the fact that rpoA22 confers a significant improvement in productivity with respect to rpoA14 shows that specificity is observed in the phenotype of interest.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Earlier reports have suggested that multilocus traits can be manipulated by introducing altered transcription factors (1, 22, 33a). In this study, we demonstrate that mutant versions of rpoA which encode a subunit of the polymerase itself are capable of substantially altering the cellular phenotype, thus making it an attractive target for transcriptional engineering. Libraries of rpoA delivered improvements in three distinct phenotypes, i.e., butanol tolerance, L-tyrosine production, and HA accumulation, and we expect that this technique can be used for improving a broad spectrum of other interesting traits. This capacity is related to the function of the CTD, because all the mutations found in this study were mapped to this region. This domain of the protein has been implicated in contacting promoter DNA and protein activators and repressors (6, 9), suggesting that it is the regulatory function of the alpha subunit which gives rise to the pleiotropic alteration of the transcriptome that results in novel phenotypes. The NTD has also been shown to be a target for transcription regulation (32), but no mutations were found there in the present study.

Although the underlying mechanisms for the improvements were not fully explored, the specificity of the mutant plasmids for each phenotype was studied. We observed that although the improvement in butanol growth was specific for the mutant L33 plasmid, all three mutant plasmids (L33, rpoA14, and rpoA-HA) were able to confer higher L-tyrosine titers in the rpoA14 mutant background (but improved productivity was limited to rpoA22). This suggests that the specificity of the different rpoA variants for a particular phenotype is a complex phenomenon, and general conclusions cannot be offered.

In principle, all global regulators of transcription could be targets for reprogramming metabolism, but not all would be equally effective. In fact, our own (unpublished) work with other sigma factors (e.g., ^H and ^E) has produced less impressive results, despite supporting evidence that these factors could yield improved strains under the conditions tested. In some cases, even established mutagenesis targets, such as ^D, failed to deliver results for certain phenotypes of interest. For example, for the case of butanol tolerance, rpoD libraries were tried in parallel with those of rpoA, but no rpoD variants capable of conferring a growth advantage were isolated. In other cases, rpoA libraries led to comparable or better results than those found through rpoD mutagenesis. During our studies of HA production, rpoA mutants showed up to 60% improvement in titer (Fig. 4) compared to 40% in the case of rpoD as measured by the same method (48). Despite these results, we do not believe that this constitutes sufficient evidence for favoring one target or the other; instead, we have argued that a better measure of the usefulness of a library is related to the a priori probability of finding improved mutants (22), not to whether a particular target delivers a phenotype of interest or not. However, we do believe that mutagenesis of the alpha subunit may complement the use of sigma factors because it is permanently bound to the RNAP holoenzyme and has been shown to interact with most, if not all promoters (38), circumventing the fact that some transcriptional states may be hard to access by ^D in certain conditions (19, 20).

Although we were able to isolate improved mutants through rpoA mutagenesis alone, this technique can also be combined with more traditional strain improvement approaches. For example, the plasmid derived from L33, or similarly isolated mutants, could be expressed in an alcohol-synthesizing strain (2) to obtain a more robust platform for biofuel production. In the case of both HA and L-tyrosine, rational metabolic engineering strategies had been previously used for establishing and increasing titers (27, 47). Therefore, employing transcriptional engineering in the background of these preengineered strains can lead to more substantial increases in metabolite overproduction. It is interesting to note that the observed L-tyrosine-producing phenotype required both the background of the isolated strain and the presence of a mutant rpoA plasmid. Therefore, it is likely that this particular strain incurred additional mutations within the chromosome that act in concert with the mutant rpoA to enhance L-tyrosine production. Thus, rpoA mutagenesis can also act synergistically with natural variations introduced during normal replication processes.

The results presented here also open several possibilities for future strain improvement approaches through the use of transcriptional engineering. First, because each RNAP complex contains two alpha subunits, cooperation between two mutant alpha subunits could take place within the same RNAP or, alternatively, could result in several versions of RNAP within the same cell. Thus, the introduction of two (or more) mutant alpha subunits could potentially allow exploration of a larger phenotypic space. Second, since the alpha and sigma units regulate promoter preferences by different mechanisms, combining these mutants within the same strain could result in synergistic transcriptional responses unachievable by either subunit tested separately. Although such combinations significantly increase the number of possible libraries, the use of a previously developed phenotypic diversity metric (22) could aid in designing better ones, e.g., by increasing or decreasing the mutagenesis rate in these expanded libraries.

Put in context, the fact that complex responses were elicited by mutant alpha subunits reinforces the hypothesis that rapid phenotypic specialization can arise from global transcriptional changes (4, 22). Because these changes are introduced by mutations in single proteins, a main challenge becomes that of any other protein engineering project: the efficient exploration of the sequence space.

 
We thank J. E. Rothman and Gero Miesenböck for providing the pH-sensitive GFP.

D.K.-M. was partially supported by a CONACyT fellowship and C.N.S.S. was supported by a National Science Foundation Graduate Fellowship during the completion of this research.

 
* Corresponding author. Mailing address: Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139. Phone: (617) 253-4583. Fax: (617) 253-3122. E-mail: gregstep{at}mit.edu

Published ahead of print on 27 February 2009.

Present address: Department of Chemical Engineering, Tsinghua University, Beijing 100084, China.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Alper, H., and G. Stephanopoulos. 2007. Global transcription machinery engineering: a new approach for improving cellular phenotype. Metab. Eng. 9:258-267.[CrossRef][Medline]
2
2. Atsumi, S., A. F. Cann, M. R. Connor, C. R. Shen, K. M. Smith, M. P. Brynildsen, K. J. Chou, T. Hanai, and J. C. Liao. 2007. Metabolic engineering of Escherichia coli for 1-butanol production. Metab. Eng. doi:10.1016/j.ymben.2007.08.003.[CrossRef]
3
3. Reference deleted.
4
4. Borneman, A. R., T. A. Gianoulis, Z. D. Zhang, H. Yu, J. Rozowsky, M. R. Seringhaus, L. Y. Wang, M. Gerstein, and M. Snyder. 2007. Divergence of transcription factor binding sites across related yeast species. Science 317:815-819.[Abstract/Free Full Text]
5
5. Botsford, J. L., and J. G. Harman. 1992. Cyclic AMP in prokaryotes. Microbiol. Rev. 56:100-122.[Abstract/Free Full Text]
6
6. Browning, D. F., and S. J. Busby. 2004. The regulation of bacterial transcription initiation. Nat. Rev. Microbiol. 2:57-65.[CrossRef][Medline]
7
7. Cakar, Z. P., U. O. Seker, C. Tamerler, M. Sonderegger, and U. Sauer. 2005. Evolutionary engineering of multiple-stress resistant Saccharomyces cerevisiae. FEMS Yeast Res. 5:569-578.[CrossRef][Medline]
8
8. Cameron, D. C., N. E. Altaras, M. L. Hoffman, and A. J. Shaw. 1998. Metabolic engineering of propanediol pathways. Biotechnol. Prog. 14:116-125.[CrossRef][Medline]
9
9. Dangi, B., A. M. Gronenborn, J. L. Rosner, and R. G. Martin. 2004. Versatility of the carboxy-terminal domain of the alpha subunit of RNA polymerase in transcriptional activation: use of the DNA contact site as a protein contact site for MarA. Mol. Microbiol. 54:45-59.[CrossRef][Medline]
10
10. Demain, A. L., and J. L. Adrio. 2008. Strain improvement for production of pharmaceuticals and other microbial metabolites by fermentation. Prog. Drug Res. 65:. 251:253-289.
11
11. Demain, A. L., and N. A. Solomon. 1986. Manual of industrial microbiology and biotechnology. American Society for Microbiology, Washington, DC.
12
12. Durre, P. 2007. Biobutanol: an attractive biofuel. Biotechnol. J. 2:1525-1534.[CrossRef][Medline]
13
13. Gaal, T., W. Ross, E. E. Blatter, H. Tang, X. Jia, V. V. Krishnan, N. Assa-Munt, R. H. Ebright, and R. L. Gourse. 1996. DNA-binding determinants of the alpha subunit of RNA polymerase: novel DNA-binding domain architecture. Genes Dev. 10:16-26.[Abstract/Free Full Text]
14
14. Goa, K. L., and P. Benfield. 1994. Hyaluronic acid. A review of its pharmacology and use as a surgical aid in ophthalmology, and its therapeutic potential in joint disease and wound healing. Drugs 47:536-566.[Medline]
15
15. Graca da Silveira, M., M. Vitoria San Romao, M. C. Loureiro-Dias, F. M. Rombouts, and T. Abee. 2002. Flow cytometric assessment of membrane integrity of ethanol-stressed Oenococcus oeni cells. Appl. Environ. Microbiol. 68:6087-6093.[Abstract/Free Full Text]
16
16. Igarashi, K., N. Fujita, and A. Ishihama. 1991. Identification of a subunit assembly domain in the alpha subunit of Escherichia coli RNA polymerase. J. Mol. Biol. 218:1-6.[CrossRef][Medline]
17
17. Igarashi, K., and A. Ishihama. 1991. Bipartite functional map of the E. coli RNA polymerase alpha subunit: involvement of the C-terminal region in transcription activation by cAMP-CRP. Cell 65:1015-1022.[CrossRef][Medline]
18
18. Ikeda, M. 2003. Amino acid production processes. Adv. Biochem. Eng. Biotechnol. 79:1-35.[Medline]
19
19. Ishihama, A. 2000. Functional modulation of Escherichia coli RNA polymerase. Annu. Rev. Microbiol. 54:499-518.[CrossRef][Medline]
20
20. Jishage, M., K. Kvint, V. Shingler, and T. Nystrom. 2002. Regulation of sigma factor competition by the alarmone ppGpp. Genes Dev. 16:1260-1270.[Abstract/Free Full Text]
21
21. Karagiannis, J., and P. G. Young. 2001. Intracellular pH homeostasis during cell-cycle progression and growth state transition in Schizosaccharomyces pombe. J. Cell Sci. 114:2929-2941.[Medline]
22
22. Klein-Marcuschamer, D., and G. Stephanopoulos. 2008. Assessing the potential of mutational strategies to elicit new phenotypes in industrial strains. Proc. Natl. Acad. Sci. USA 105:2319-2324.[Abstract/Free Full Text]
23
23. Koffas, M., and G. Stephanopoulos. 2005. Strain improvement by metabolic engineering: lysine production as a case study for systems biology. Curr. Opin. Biotechnol. 16:361-366.[CrossRef][Medline]
24
24. Li, S. C., N. K. Goto, K. A. Williams, and C. M. Deber. 1996. Alpha-helical, but not beta-sheet, propensity of proline is determined by peptide environment. Proc. Natl. Acad. Sci. USA 93:6676-6681.[Abstract/Free Full Text]
25
25. Lloyd, G. S., W. Niu, J. Tebbutt, R. H. Ebright, and S. J. Busby. 2002. Requirement for two copies of RNA polymerase alpha subunit C-terminal domain for synergistic transcription activation at complex bacterial promoters. Genes Dev. 16:2557-2565.[Abstract/Free Full Text]
26
26. Lutke-Eversloh, T., and G. Stephanopoulos. 2005. Feedback inhibition of chorismate mutase/prephenate dehydrogenase (TyrA) of Escherichia coli: generation and characterization of tyrosine-insensitive mutants. Appl. Environ. Microbiol. 71:7224-7228.[Abstract/Free Full Text]
27
27. Lutke-Eversloh, T., and G. Stephanopoulos. 2007. L-Tyrosine production by deregulated strains of Escherichia coli. Appl. Microbiol. Biotechnol. 75:103-110.[CrossRef][Medline]
28
28. Lynd, L. R., C. E. Wyman, and T. U. Gerngross. 1999. Biocommodity engineering. Biotechnol. Prog. 15:777-793.[CrossRef][Medline]
29
29. Maury, J., M. A. Asadollahi, K. Moller, A. Clark, and J. Nielsen. 2005. Microbial isoprenoid production: an example of green chemistry through metabolic engineering. Adv. Biochem. Eng. Biotechnol. 100:19-51.[Medline]
30
30. Miesenbock, G., D. A. De Angelis, and J. E. Rothman. 1998. Visualizing secretion and synaptic transmission with pH-sensitive green fluorescent proteins. Nature 394:192-195.[CrossRef][Medline]
31
31. Murakami, K., N. Fujita, and A. Ishihama. 1996. Transcription factor recognition surface on the RNA polymerase alpha subunit is involved in contact with the DNA enhancer element. EMBO J. 15:4358-4367.[Medline]
32
32. Niu, W., Y. Kim, G. Tau, T. Heyduk, and R. H. Ebright. 1996. Transcription activation at class II CAP-dependent promoters: two interactions between CAP and RNA polymerase. Cell 87:1123-1134.[CrossRef][Medline]
33
33. Parekh, S., V. A. Vinci, and R. J. Strobel. 2000. Improvement of microbial strains and fermentation processes. Appl. Microbiol. Biotechnol. 54:287-301.[CrossRef][Medline]
33
34. Park, K. S., D. K. Lee, H. Lee, Y. Lee, Y. S. Jang, Y. H. Kim, H. Y. Yang, S. I. Lee, W. Seol, and J. S. Kim. 2003. Phenotypic alteration of eukaryotic cells using randomized libraries of artificial transcription factors. Nat. Biotechnol. 21:1208-1214.[CrossRef][Medline]
34
35. Patnaik, R., S. Louie, V. Gavrilovic, K. Perry, W. P. Stemmer, C. M. Ryan, and S. del Cardayre. 2002. Genome shuffling of Lactobacillus for improved acid tolerance. Nat. Biotechnol. 20:707-712.[CrossRef][Medline]
35
36. Paul, B. J., W. Ross, T. Gaal, and R. L. Gourse. 2004. rRNA transcription in Escherichia coli. Annu. Rev. Genet. 38:749-770.[CrossRef][Medline]
36
37. Pieterse, B., R. J. Leer, F. H. Schuren, and M. J. van der Werf. 2005. Unravelling the multiple effects of lactic acid stress on Lactobacillus plantarum by transcription profiling. Microbiology 151:3881-3894.[Abstract/Free Full Text]
37
38. Ross, W., K. K. Gosink, J. Salomon, K. Igarashi, C. Zou, A. Ishihama, K. Severinov, and R. L. Gourse. 1993. A third recognition element in bacterial promoters: DNA binding by the alpha subunit of RNA polymerase. Science 262:1407-1413.[Abstract/Free Full Text]
38
39. Ross, W., and R. L. Gourse. 2005. Sequence-independent upstream DNA-CTD interactions strongly stimulate Escherichia coli RNA polymerase-lacUV5 promoter association. Proc. Natl. Acad. Sci. USA 102:291-296.[Abstract/Free Full Text]
39
40. Santos, C. N., and G. Stephanopoulos. 2008. Combinatorial engineering of microbes for optimizing cellular phenotype. Curr. Opin. Chem. Biol. 12:168-176.[CrossRef][Medline]
40
41. Santos, C. N., and G. Stephanopoulos. 2008. Melanin-based high-throughput screen for L-tyrosine production in Escherichia coli. Appl. Environ. Microbiol. 74:1190-1197.[Abstract/Free Full Text]
41
42. Sikkema, J., J. A. de Bont, and B. Poolman. 1995. Mechanisms of membrane toxicity of hydrocarbons. Microbiol. Rev. 59:201-222.[Abstract/Free Full Text]
42
43. Stephanopoulos, G., and J. Kelleher. 2001. How to make a superior cell. Science 292:2024-2025.[Free Full Text]
43
44. Stephanopoulos, G., and A. J. Sinskey. 1993. Metabolic engineering—methodologies and future prospects. Trends Biotechnol. 11:392-396.[CrossRef][Medline]
44
45. Thykaer, J., and J. Nielsen. 2003. Metabolic engineering of beta-lactam production. Metab. Eng. 5:56-69.[CrossRef][Medline]
45
46. Valli, M., M. Sauer, P. Branduardi, N. Borth, D. Porro, and D. Mattanovich. 2006. Improvement of lactic acid production in Saccharomyces cerevisiae by cell sorting for high intracellular pH. Appl. Environ. Microbiol. 72:5492-5499.[Abstract/Free Full Text]
46
47. Wackett, L. P. 2008. Biomass to fuels via microbial transformations. Curr. Opin. Chem. Biol. 12:187-193.[CrossRef][Medline]
47
48. Yu, H., and G. Stephanopoulos. 2008. Metabolic engineering of Escherichia coli for biosynthesis of hyaluronic acid. Metab. Eng. 10:24-32.[CrossRef][Medline]
48
49. Yu, H., K. Tyo, H. Alper, D. Klein-Marcuschamer, and G. Stephanopoulos. 2008. A high-throughput screen for hyaluronic acid accumulation in recombinant Escherichia coli transformed by libraries of engineered sigma factors. Biotechnol. Bioeng. 101:788-796.[CrossRef][Medline]
49
50. Zhang, Y. X., K. Perry, V. A. Vinci, K. Powell, W. P. Stemmer, and S. B. del Cardayre. 2002. Genome shuffling leads to rapid phenotypic improvement in bacteria. Nature 415:644-646.[CrossRef][Medline]

---

Applied and Environmental Microbiology, May 2009, p. 2705-2711, Vol. 75, No. 9
0099-2240/09/$08.00+0     doi:10.1128/AEM.01888-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G. Search for Related Content PubMed PubMed Citation Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G. Agricola Articles by Klein-Marcuschamer, D. Articles by Stephanopoulos, G.