Methods 47 (2009) 44–52
Contents lists available at ScienceDirect
Methods j o u r n a l h o m e p a g e : w w w . e l s e v i e r. c o m / l o c a t e / y m e t h
Recombinant bacterial RNA polymerase: Preparation and applications Konstantin Kuznedelov a,*, Konstantin Severinov a,b,c a
Waksman Institute, Rutgers, The State University of New Jersey, 190 Frelinghuysen Road, Piscataway, NJ 08854, USA Institute of Molecular Genetics, Moscow, Russian Federation c Institute of Gene Biology, Russian Academy of Sciences, Moscow, Russian Federation b
a r t i c l e
i n f o
Article history: Accepted 8 October 2008 Available online 21 October 2008 Keywords: RNA polymerase Transcription Bacterial In vitro reconstitution Co-overexpression Aquifex aeolicus
a b s t r a c t Availability of DNA-dependent RNA polymerase from vario us bacteria is a key for setting up specific in vitro transcription systems necessary for understanding species-specific transcription regulation. We describe here two main strategies for recombinant RNA polymerase preparation—through in vitro reconstitution and heter ologous co-overproduction in Escherichia coli. Both strategies can be used for preparation of large amounts of RNA polymerases from any bacteria for which sequences of rpo (RNA polymerase) genes are known. © 2008 Elsevier Inc. All rights reserved.
1. Introduction DNA-dependent RNA polymerase (RNAP) is the key enzyme of gene expression. RNAP is a multisubunit, multifunctional molec ular machine, whose functions are regulated by various cellu lar factors in response to environmental cues. Despite the very high degree of evolutionary conservation at and around the cat alytic center, RNAP from different organisms, including different eubacteria, exhibit species-specific properties such as specificity of promoter recognition or ability to interact with and respond to transcription factors. Thus, understanding transcription regulation in a particular bacterium often means that an in vitro transcription system with cognate RNAP needs to be set up. Bacterial RNAP consists of a catal ytically proficient core enzyme (subunit composition a2bb9x, molecular weight »300–400 kDa) and a specificity subunit r. A complex of core with r is called the holoenzyme and is able to specifically recognize promoters. Dif ferent r factors direct the core to different groups of promoters. The smallest RNAP subunit, x (»10 kDa) is the only subunit that is dispensable for most in vivo and in vitro functions of the enzyme. While robust procedures suitable for RNAP purification from most bacteria have been developed, their application is hampered by dif ficulties associated with cultivating, let alone growing large vol umes of bacterial cultures needed for biochemical fractionation. Bacterial RNAP is unique in that it can be efficiently reconstituted in vitro from isolated recombinant subunits, or prepared by heter olog ous co-overexpression of subunits in surrogate hosts, opening way for functional and structural analyses of transcription mecha * Corresponding author. Fax: +1 732 445 5735. E-mail address: kuznedel[email protected]
man.rutgers.edu (K. Kuznedelov). 1046-2023/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2008.10.007
nism and regulation in hard-to-cultivate microorganisms. In addi tion, both the in vitro assembly and co-overexpression approaches allow preparing RNAP mutants lacking essential functions. This is very important for reverse genetics analysis of RNAP, which is an essential enzyme, meaning that classical genetic approaches are limited to analysis of viable mutations. Below, we present detailed protocols for preparation of recombinant RNAP by in vitro assem bly and co-overexpression. 2. In vitro reconstitution of recombinant bacterial RNAP A general scheme of in vitro reconstitution of bacterial RNAP is presented in Fig. 1 and is described in detail in the following sec tions. 2.1. Expression of recombinant RNAP subunits in Escherichia coli Plasmids of the pET series (Novagen) have been successfully used as vectors for expression of rpo genes from various bacterial sources. Plasmids with rpo genes are transformed into the E. coli BL21(DE3) cells and transformants are plated on solid medium containing appropriate antibiotics. In PET-based plasmids, the target protein overproduction is induced by the addition of IPTG (usually 1 mM). It is very important to find optimal conditions for overproduction for each RNAP subunit. In our experience, vary ing growth temperature, IPTG concentration, culture OD600 at the time of induction, and cultivation time after the induction allows high- to medium levels of expression of RNAP subunits from various sources, including Gram-negative bacteria (E. coli, Xan thomonas oryzae, and Francisella tularensis), Gram-positive bac teria (Bacillus subtilis, Bacillus cereus), and several thermophilic
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
should be preheated to 37 oC. The culture is grown at 37 °C with vigorous agitation until OD600 reaches 0.4–0.8 (usually 2–3 h), induced with 1 mM IPTG and allowed to grow further (usually for additional 2–5 h; the optimal time of induction time could be determined during the pilot experiments). A 0.1–0.5 ml aliquot of the culture is withdrawn before the induction and stored at room temperature in an Eppendorf tube. When the induction is com plete, another aliquot of the culture is withdrawn. Cells from both culture aliquots are collected by centrifugation in a microcentri fuge, the supernatant is removed, and the pellet is resuspended in 20 ll H2O and equal volume of Laemmli loading buffer is added. Samples are boiled for 5 min and analyzed by SDS–PAGE to deter mine recombinant RNAP subunits expression levels. If desired lev els of overproduction are detected, cells from induced cultures are collected by centrifugation (4000g, 10 min, 4 °C), the medium is removed and cell pellet drained and stored at ¡80 °C until further use. 2.2. Inclusion bodies preparation and purific ation of His-tagged RNAP subunits from cell extracts Escher ichia coli BL21(DE3) cells transformed with rpo gene expression plasmids usually overproduce individual core RNAP subunits at high-level upon induction and these RNAP subunits often form inclusion bodies. Nevertheless, it is very important to determine the localization (cytoplasmic versus inclusion bodies) of recombinant protein when overproducing an RNAP subunit from a new source. This is best done by performing a small-scale trial induction in »10 ml of LB medium. It is safer to perform local ization trials by growing cell cultures in tiny flasks rather than tubes, since induction conditions often do not scale properly from tubes to flasks. Induced cells are collected in an Eppendorf tube, resuspended in 400–500 ll of lysis buffer (see Table 1) and lysed by several 5- to 10-s sonication bursts using a microtip with 1-min rests between the bursts. The tube shall be kept in a water-ice bath during the sonication. An aliquot of lysed cells is next removed and cell debris and inclusion bodies are collected by a 2–5 min centri fugation in a refrigerated microcentrifuge. An aliquot of the cleared cell lysate is removed, the rest of the supernatant discarded and the pellet is resuspended in an initial volume of the lysis buffer
Table 1 Solutions and buffers. Fig. 1. Preparation of bacterial RNAP core enzyme by in vitro reconstitution. The sequence of steps involved in RNAP core enzyme reconstitution is schematically presented. The x subunit can be omitted from reconstitution reaction with minimal effects on most RNAP functions. See text for details.
rganisms (Thermus aquaticus and Aquifex aeolicus). Should prob o lems with expression be encountered, various procedures such as altering the ratio of isoaccepting tRNAs in the expression host can be used to attempt to increase the yield. Useful information about optimization of induction can be found in the PET system manual from Novagen (http://www.emdbiosciences.com/docs/docs/PROT/ TB055.pdf). Once the induction conditions have been optimized in pilot experiments, large-scale induction is performed. Only freshly transformed cells (grown at 37 oC for no more than 12 h) should be used. Cells are collected from plates by scraping with a micro biological loop and are carefully resuspended in »1 ml of liquid LB medium in a sterile Eppendorf tube. The resulting suspension is used to inoculate liquid cultures (a plate containing several hundred »1 mm colonies is suf ficient for inoculation of 1 l of LB medium). The medium should contain appropriate antibiotics and
40 mM Tris–HCl, pH 7.9; 300 mM KCl; 10 mM EDTA
40 mM Tris–HCl, pH 7.9; 100 mM NaCl; 10 mM EDTA
40 mM Tris–HCl, pH 7.9; 200 mM KCl; 50% glycerol; 1 mM EDTA; 1 mM 2-ME 6 M guanidine–HCl; 50 mM Tris–HCl, pH 7.9; 10 mM MgCl2; 10 lM ZnCl2; 10% glycerol; 1 mM EDTA 50 mM Tris–HCl pH 7.9; 200 mM KCl; 10 mM MgCl2; 10 lM ZnCl2; 10% glycerol; 1 mM EDTA 40 mM Tris–HCl, pH 7.9; 5% glycerol; 1 mM EDTA 20 mM Tris–HCl, pH 7.9; (or Hepes); 500 mM NaCl; 5% glycerol 40 mM Tris–HCl, pH 8.4; 40 mM KCl; 10 mM MgCl2
Before use supplemented with 15 mM 2-ME and 0.1 mM PMSF Before use supplemented with 15 mM 2-ME and 0.1 mM PMSF Used for dialysis
TGE buffer Start buffer
Before use supplemented with 10 mM DTT
Before use supplemented with 1 mM DTT or 2-ME (see text) Before use supplemented with 1 mM 2-ME For metal chelate affinity chromatography
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
with a brief burst of sonication. An aliquot of the resulting homo geneous suspension is removed. The three aliquots (crude lysate, cleared lysate, and pellet suspension) are next analyzed by SDS– PAGE and the amount of the target protein in soluble and insoluble fractions is estimated. Incomplete cell lysis can lead to errors in estimates of insoluble protein. The easiest way to determine that lysis is complete is to monitor the presence of the E. coli b and b9 RNAP subunit bands in the inclusion bodies fraction. These two subunits form very characteristic closely spaced bands with appar ent electrophoretic mobilities of 150 and 155 kDa. They shall be absent in the inclusion bodies sample (obviously, this method is not applicable when an overexpressed subunit matches the size of E. coli b and or b9). The following procedure can be used for large-scale prepara tion of RNAP subunit containing lysates or inclusion bodies. Cells (frozen or freshly collected) are resuspended in lysis buffer with 15 mM b-mercaptoethanol (2-ME) and 0.1 mM PMSF (5 ml of buffer per 1 g of bacterial paste). A homogeneous cell suspension is prepared using a glass rod or by vigorous pipetting through a glass or plastic pipette. The cell suspension is transferred into a glass beaker or a centrifuge tube. Cells are disrupted by 1-min bursts of sonication using maximal power setting for a tip used (varies depending on the total volume of cell suspension) at a 50% duty cycle. During sonication, the vessel with cell suspension is kept in a water-ice bath. The suspension is left for 1 min in a water-ice bath to cool down, and then the sonic ation step is repeated (4–5 times with 1 min rests between each sonication give satisfactory results). The lysate is next centrifuged at 18,000g for 1 h at 4 °C. Such long centrifugation step is only warranted if an overproduced RNAP subunit is found to be soluble. A 5- to 10-min centrifugation is suf ficient to pellet inclusion bodies. The supernatant is either discarded (if overproduced protein is mostly located in inclusion bodies) or, if it contains overproduced protein, transferred into a 50 ml screw-cap polypropylene tube. To the supernatant, dry finely grained (NH4)2SO4 is added to 60% saturation and dissolved com pletely by gentle mixing. The tube is left on ice for at least 1 h for protein pellet to form. Proteins precipitated with (NH4)2SO4 can be stored at +4 °C indefinitely (in the form of suspension in (NH4)2SO4containing mother solution) until further use. A pellet containing inclusion bodies is further washed by repeated cycles of sonication (as described above) and centrifuga tion, first in the lysis buffer, then in the same buffer containing 0.2% Na–deoxycholate. The final pellet is resuspended in a small volume of lysis buffer (2 ml of buffer per 1 g of starting bacterial paste) and aliquots of 0.2 ml are placed in individual Eppendorf tubes. The tubes are centrifuged for 5 min in a microcentrifuge, supernatants are discarded and pellets are stored at ¡80 °C until use. Purification of solub le untagged RNAP subunits from cell extracts can be complicated since at least large subunits (b and b9) are not properly folded and do not behave well during chro matographic steps. Therefore, if overproduced subunit is found soluble, it is best to affinity tag it by genetically fusing to a termi nal hexahistidine tag and purify it by metal chelate affinity chro matography. In our experience, affinity tags at either terminus of RNAP subunits from various sources have minimal effects on most RNAP functions. If a tag is of concern, it can be removed by treat ment with specific proteases, since most PET vectors incorporate a protease recognition site between the affinity purification tag and the target protein, Collect the (NH4)2SO4 precipitate (see above) by 20-min cen trifugation at 15,000g, 4 °C. Carefully discard the supernatant and drain the pellet by placing the centrifuge tube upside down on a paper towel. Wipe the remaining drops of the supernatant from the centrifuge tube walls using Kimwipes (avoid touching the pellet). Dissolve pellet in 5 ml of start buffer (Table 1) contain ing 5 mM imidazole (avoid creating bubbles when dissolving the
pellet). For better binding of the target protein, it is sometimes worth using Hepes instead of Tris–HCl (Tris–HCl tends to reduce protein binding by metal affinity columns). Remove undissolved material by 30-min centrifugation at 15,000g, 4 °C. Transfer the supernatant into a fresh screw-cap polypropylene tube and repeat the centrifugation step. Filter the supernatant through a 0.45 lm Nylon Membrane acrylic filter (PALL, Life Sciences) attached to a syringe and load onto a 1 ml Hi-Trap chelating column (GE Health care) or equivalent charged with Ni2+ according to manufacturer’s instructions and equilibrated in the start buffer. The column can be either attached to a liquid chromatography system or (an easier and faster way) all steps can be performed manually with a syringe. Wash the column with 5 ml of start buffer with 20 mM imidazole (up to 60 mM of imidazole can be used at the wash step; the con centration of imidazole at the wash step should be determined experimentally for each RNAP subunit: for example RNAP a sub unit tends to bind to the column stronger because of the bidentate nature of the interaction—since a is a dimer at native conditions). Because of considerable variability between metal ion chelating resins from different sources and even between batches from the same source, optimal binding and elution conditions should be adjusted each time a new column is used. In general, the higher the concentration of imidazole at the wash step the purer the final preparation of the protein. Elute the protein with start buffer containing 100 mM imidazole collecting 1 ml fractions (sometimes, up to 500 mM imidazole is required for complete elution, this should be determined experimentally). Analyze all fractions by SDS–PAGE and pool those containing the target subunit, which should be 70–90% pure. Concentrate pooled fractions by ultrafiltration using a Microsep (PALL, Life Science) centrifugal device (or an analogous one) with the appropriate molecular weight cut-off to a final protein concentration of no less than 1 mg/ml, dialyze against storage buffer (see Table 1) and store at ¡20 °C. Note that dialysis in 50% glycerol buffer further concen trates the protein. It is sometimes necessary to purify His-tagged RNAP subunits at denaturing conditions (for example, to decrease the background of E. coli RNAP subunits that may associate with heterologously over produced RNAP subunit). Purification is carried out the same way as described above but 6 M guanidine–HCl is added to all buffers. Fractions containing pure protein are pooled and dialyzed against reconstitution buffer (see Table 1) with 5 mM 2-ME. Precipitate formed during the dialysis is removed by 30-min centrifugation at 15,000g, 4 °C, soluble protein is concentrated and stored as above. 2.3. RNAP reconstitution from individual subunits RNAP reconstitution procedure developed for E. coli RNAP and described in detail elsewhere  can be successfully used for reconstitution of RNAP from other bacterial species. Below, we present a reconstitution protocol based on this procedure with some modifications. Prior to in vitro reconstitution, inclusion bodies are solubilized in denaturation buffer (Table 1) with 10 mM DTT (added imme diately before use). For more efficient solubilization, the inclu sion bodies pellet from a single aliquot (above) is resuspended by pipetting and the suspension is left for 30 min on ice. Undissolved material is removed by 30-min centrifugation in a microcentrifuge (4 °C). The supernatant is transferred into a fresh tube and protein concentration is determined using the Bradford assay with BSA as a standard. RNAP subunits deteriorate after solubilization in the dena turation buffer and should be used shortly after solubilization. RNAP subunits are mixed in a molar ratio of 2:8:4 (a:b:b9), the total protein concentration is adjusted to 0.5 mg/ml with dena turation buffer containing 10 mM DTT and the mixture is dialyzed
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
(at 4 °C) for 16 h against two changes of 250 volumes of reconsti tution buffer containing 10 mM 2-ME. Any precipitate formed dur ing dialysis is removed by centrifugation. To the supernatant, one molar equivalent of the primary RNAP r subunit in storage buffer is added and the mixture is incubated for 1 h at 30 °C. This “ther moactivation” step allows the holoenzyme to form and seems to generally increase the yield of assembled active enzyme. The addi tion of r step can be omitted, especially if RNAP core is prepared. However, total yields (and specific activity of resultant RNAP) become considerably lower, The resulting RNAP preparations can be used directly in tran scription assays or stored under (NH4)2SO4 (65% saturation) until further use. RNAP in reconstitution mixtures is not very stable and loses activity with time, even if stored in the presence of 50% glyc erol at ¡20 °C. Stable pure RNAP can be obtained by subsequent Superose-6 gel-filtration and Mono-Q or Resource-Q ion-exchange chromatographic purification steps. In vitro reconstitution conditions described above always yield a mixture of RNAP subassemblies and core in different ratios; some times, especially when working with RNAP mutants, no assembled enzyme is produced (see, for example, Ref. ). The Superose-6 step separates assembled enzyme from assembly intermediates and unassembled subunits and allows one to determine whether assembly of recombinant RNAP occurred in the first place, judging by the appearance of characteristic chromatographic peaks. The pellet obtained after (NH4)2SO4 precipitation of RNAP reconstitu tion mixture is collected by centrifugation, drained thoroughly and dissolved in 0.25 ml of TGE buffer (see Table 1) with 1 mM 2-ME. Undissolved material is removed by several centrifugation steps (until no visible pellet is formed) and the supernatant is loaded on a Superose-6 10/30 HR (GE Healthcare) column attached to an FPLC and equilibrated in the TGE buffer containing 200 mM NaCl and 1 mM 2-ME. The chromatography is conducted in the same buffer (0.4 ml/min flow rate); 1 ml fractions are collected and analyzed by SDS–PAGE. Fractions containing RNAP (and the a2b subassembly) usually form a sharp peak that elutes after a peak of aggregates (elutes in void volume) and a diffuse peak of unassembled large RNAP subunits. The Mono-Q step separates the a2b subassembly, RNAP core, and holoenzyme from each other and also removes excess r and con taminating ribonucleases. Superose-6 fractions containing RNAP are pooled, diluted 2-fold with TGE buffer, and loaded (1 ml/min flow rate) onto a 1 ml Mono-Q column equilibrated with TGE con taining 100 mM NaCl and attached to FPLC. The column is washed with 10 ml of the same buffer and bound proteins are eluted with a linear gradient of NaCl (from 200 to 400 mM NaCl) in TGE over the course of 35 min. The column is next washed with TGE contain ing 1 M NaCl. During Mono-Q chromatography, RNAP core and the holoenzyme elute separately (RNAP core first, in a series of closely spaced and poorly separ ated peaks that probab ly represent dif ferent conformation of the core, followed by a sharp holoenzyme peak). The r subunit elutes last. RNAP core and holoenzyme con taining fractions are identified by SDS–PAGE, pooled, made 50% with glycerol and stored at ¡20 °C. Alternatively, RNAP in pooled fractions can be concentrated »4-fold by dialyzing against the storage buffer. If the amount of RNAP is low, fractions can be con centrated on a centrifugal device such as Microsep 100 K Omega (PALL, Life Science) followed by the addition of glycerol to 50%. The concentration of RNAP is determined using the Bradford assay with BSA as a standard or, preferably, determined spectrophoto metrically using the calculated extinction coef ficient at 280 nm. At concentrations of 1 mg/ml or more, thus purified RNAP samples remain active for many years when stored at ¡20 °C. If any RNAP subunit is affinity tagged, the enzyme can be batchpurified by metal chelate affinity chromatography . To this end, reconstitution mixtures after dialysis and thermoactivation (above)
are combined with the appropriate amount Ni2+–NTA agarose (Qiagen) equilibrated with 50 mM Tris–HCl pH 7.9, 0.5 mM EDTA, 5% glycerol. After »30 min binding with gentle agitation, agarose beads are washed three times with the same buffer containing 5 mM imidazole, and bound protein is eluted with the buffer con taining 150 mM imidazole. The samples are concentrated (and the buffer changed to one without imidazole) using a centrifug al ultra filtration device and the protein is stored as above. Thus purified RNAP is suitable for most in vitro transcription assays but tends to deteriorate upon long-term storage (storing under (NH4)2SO4 helps to avoid this problem). 3. Preparation of recombinant RNAP by co-overexpression in E. coli The first successful preparation of functional RNAP by co-expres sion in E. coli of a heterologous set of rpo genes was achieved for T. aquaticus RNAP , the first RNAP for which a high-resolution structure became available . This experimental system was later improved and used for preparation of milligram amounts of sev eral structure-based recombinant T. aquaticus RNAP mutants [5–8]. Since then, enzymes from several other sources were prepared by this method. Of note here is the preparation of recombinant E. coli RNAP harboring a large deletion in an evolutionarily variable region of the b’ subunit  and the preparation of crystallizable recombinant T. aquaticus RNAP . In the former case, the mutant enzyme could not have been prepared by in vitro reconstitution, possibly reflecting the role of protein chaperones in the process of RNAP assembly in vivo. In the latter example, co-overexpres sion turned out to be the only way to obtain recombinant RNAP forming diffraction quality crystals. Again, the superior qualities of recombinant enzyme obtained by co-overexpression (as com pared to one prepared by in vitro reconstitution) may either reflect the presence of chaperones (which, however, must be species nonspecific) or result from special conditions of in vivo assembly such as co-translational folding, ionic conditions etc. For every RNAP obtained by co-overexpression, a rather labori ous procedure of creating a large PET-based plasmid with rpo genes expressed from individual T7 RNAP promoters or jointly expressed from a single promoter as part of an artificial operon needs to be accomplished. The particular strategy depends very much on the sequence and availability of restriction sites in rpo genes of a microorganism under study. In general, PET-based plasmids used for individual expression of rpo genes for in vitro reconstitution serve as a useful point of departure. Various arrangements of rpo genes in the final co-overexpression plasmids have led to success ful preparation of recombinant RNAP, suggesting that gene order is not particularly important. Since co-overexpression strategy for Thermus RNAP is adequately described in published literature [3,6,7], below, we present the application of the same strategy to obtain and initially characterize RNAP from hyperthermophilic Gram-negative eubacterium A. aeolicus. Currently, co-overexpres sion strategy is most successful when preparing RNAP from ther mophilic organisms (or more generally, RNAPs that substantially differ from E. coli RNAP in terms of their physical stability), since clean separation of the target enzyme from contaminating E. coli RNAP and, more importantly, from interspecies hybrid enzymes becomes possible. Escher ichia coli mutants defective in RNAP assembly due to mutations in all genes coding for RNAP core sub units (except for x) have been described [10,11]. It is likely that the use of such mutant strains or their derivatives as hosts for co-over expression will allow the use of co-overexpression strategy for preparation of microbial RNAPs whose physical properties resem ble those of the E. coli RNAP. Use of multiple orthogonal affinity tags positioned on different recombinant RNAP subunits should also be helpful to decrease contamination by hybrid enzymes.
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
3.1. Cloning Aquifex aeolicus rpo genes in E. coli expression and co-expression plasmids Primers for PCR amplification of A. aeolicus rpoA, rpoB, rpoC, rpoZ, and rpoD genes were designed using available A. aeolicus genome sequence data . The primers allowed the cloning of each of amplified A. aeolicus rpo genes in pET series plasmids between the NdeI (or NcoI) and EcoRI (or BamHI or NotI) sites of the polylinker. Plasmids pET11-AaeA, pET28-AaeB, pET28-AaeC, pET28-AaeZ, and pET11-AaeD and pET28-AaeD overexpressing, respectively, untagged A. aeolicus RNAP a and b subunits, C-ter minally hexahistidine-tagged b9, untagged x, and untagged and N-terminally hexahistidine-tagged principal r subunit were con structed using routine cloning methods. The A. aeolicus rpo expres sion plasmid set provides a source of individual RNAP subunits for in vitro reconstitution experiments.
Plasmid pET28-AaeABZC, co-overexpressing A. aeolicus rpoA, rpoB, rpoC, and rpoZ genes, was created according to a scheme pre sented in Fig. 2. First, two intermediate plasmids each containing two genes—rpoA and rpoB (pET28-AaeAB), and rpoC + rpoZ (pET28AaeCZ)—were constructed by (1) inserting the rpoA cassette from pET11-AaeA into pET28-AaeB and (2) inserting the rpoZ cassette from pET28-AaeZ into pET28-AaeC. The pET28-AaeABZC plasmid was obtained by inserting the rpoA–rpoB cassette from pET28AaeAB into the pET28-AaeCZ plasmid. pET28-AaeABZC contains four genes in the following sequence: rpoA, rpoB, rpoZ, and rpoC; each rpo gene is preceded by T7 RNAP promoter. The rpoC gene is fused to the C-terminal hexahistidine tag and is followed by T7 transcription terminator. Previous work in several systems sug gests that C-terminal tagging of b9 has no effect on RNAP activ ity and allows affinity purification of RNAP and immobilization of functional transcription complexes. Table 2 lists A. aeolicus rpo
Fig. 2. Construction of a plasmid co-expressing A. aeolicus rpo genes. Steps involved in creation of A. aeolicus rpo genes co-overexpressing plasmids are schematically pre sented. See text for more details.
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
Table 2 Aquifex aeolicus rpo genes and expression plasmids. Gene, plasmid
rpoA rpoB rpoC rpoD rpoZ pET11-AaeA pET28-AaeB pET28-AaeC pET11-AaeD pET28-AaeD pET28-AaeZ pET28-AaeAB pET28-AaeZC pET28-AaeABZC
954 4407 4725 1728 231 6272 9679 9964 7368 7060 5496 10784 10315 15954
a b b’ r x a b b9-His6-COOH r NH2-His6-r x a, b b9-His6-COOH, x a,b,b9-His6-COOH, x
genes PCR-amplified from genomic DNA and plasmids containing these genes. 3.2. Prepar ation of A. aeolicus RNAP core enzyme To purify A. aeolicus RNAP core from E. coli cells co-expressing A. aeolicus rpo genes, BL21-CodonPlus (DE3)-RIL (Stratag ene) cells harboring pET28-AaeABZC were used to inoculate 1 l of LB contain ing 25 lg/ml kanamycin and were grown at 37 °C for 18–20 h with vigorous shaking. Cells (»4–5 g of wet biomass) were collected by centrifugation (10 min at 4000g, 4 °C) and resuspended in 40 ml of grinding buffer (Table 1) containing 0.1 mM PMSF and 15 mM 2-ME. Cells were lysed by sonication and the debris was removed
by centrifugation (30 min, 18,000g, 4 °C). The cleared cell lysate was transferred into a 50 ml screw-cap polypropylene centrifuge tube and incubated at 80 °C for 30 min with occasional mixing. Massive pellet that was formed during this stage was removed by centrifugation (30 min, 18,000g, 4 °C) and discarded. The superna tant was loaded on a 5 ml Heparin Hi-Trap column (GE Healthcare) equilibrated in TGE buffer (Table 1) containing 100 mM NaCl and 1 mM 2-ME. After loading, the column was washed with TGE buffer containing 300 mM NaCl, and RNAP was step-eluted in »7–10 ml of TGE buffer containing 600 mM NaCl. Typically thermophile results of SDS–PAGE analysis of fractions up to the 1 M NaCl hep arin column elution are shown in Fig. 3A. The 1 M NaCl heparin column fraction was loaded on a 5 ml Hi-Trap chelating column (GE Healthcare) charged with Ni2+ and equilibrated in start buffer (Table 1) containing 5 mM imidazole. The column was washed with the same buffer supplemented with 20 mM imidazole, and A. aeolicus RNAP core was eluted with 100 mM imidazole in the same buffer. Fractions containing A. aeolicus RNAP core were pooled and precipitated overnight with powdered (NH4)2SO4 (0.3 g/ml) in a cold room. Ammonium sulfate pellet was collected by centrifuga tion in a 15 ml polypropylene centrifuge tube, drained thoroughly, dissolved in 250 lL of TGE buffer and loaded on a Superose-6 HR 10/30 column (GE Healthcare) attached to an FPLC and equili brated in TGE containing 200 mM NaCl and 1 mM 2-ME. A. aeolicus RNAP eluted in two peaks. (Fig. 3B) The first peak contained mate rial strongly absorbing at 260 nm and probably represented RNAP complexes with nucleic acids and was discarded. The second peak contained pure (no less than 95% pure) RNAP core. This peak was collected, glycerol was added to the final concentration of 50% and the enzyme was stored at ¡20 °C.
Fig. 3. Typical purification example of recombinant thermophile RNAP core enzyme from E. coli cells co-overexpressing T. aquaticus rpo genes. (A) SDS–PAGE analysis of pro teins in cell extract of induced co-overexpressing cells (L), in a supernatant after heat-treatment (H), in a flow-through from a heparin-agarose column (FT) and in fractions from the column eluted with 0.3, 0.6, and 1.0 M NaCl in a buffer. A Coomassie-stained gel is shown. (B) Superose-6 fractionation of T. aquaticus RNAP core enzyme. Chromato graphic elution profile, SDS–PAGE of indicated fractions are shown (lane labeled 0 contains material loaded on the column). Material from the peak eluting early (E) and late (L) was pooled and analyzed spectrophotometrically. The resultant spectra are shown on the righthand side of the figure.
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
3.3. Aquifex aeolicus r factor preparation
3.5. Transcription by holoenzymes with interspecies r replacements
The pET28-AaeD plasmid expressing N-terminally hexa histidine-tagged A. aeolicus r factor was transformed in E. coli BL21(DE3) cells. Fresh transformants were inocul ated in 1 l LB containing 25 lg/ml kanamycin and grown at 37 °C. 0.5 mM IPTG was added when the culture OD600 reached 0.7. After 6 h of growth with vigorous agitation at 37 °C, cells were harvested and cell pellet (»1.5 g) was resuspended in 10 ml of start buffer (Table 1) containing 1 mM EDTA, 5 mM 2-ME, and 0.1 mM PMSF. Cells were lysed by sonication, cell debris removed by two cen trifugation steps (30 min, 15,000g, 4 °C each) and the lysate vol ume was adjusted to 20 ml with the same buffer. The lysate was transferred into a 50 ml screw-cap polypropylene centrifuge tube and incubated for 30 min at 80 °C with occasional mixing. The massive pellet was removed by low-speed centrifugation (20 min, 15,000g, 4 °C) and discarded. Then supernatant was fil tered through Acrodisc 25 mm Syringe Filter with 0.45 lm Nylon Membrane (PALL, Life Sciences) and loaded onto a 1 ml Hi-Trap chelating column (GE Healthcare) charged with Ni2+ according to manufacturer’s instructions and equilibrated in start buffer containing 5 mM imidazole. The column was washed with the same buffer supplemented with 20 mM imidazole, and A. aeoli cus sigma factor was eluted with 100 mM imidazole in the same buffer. Fractions containing A. aeolicus r were pooled and con centrated on a centrifugal filter Amicon Ultra (Millipore), dia lyzed against storage buffer (Table 1) and stored at ¡20 °C.
When hexahistidine-tagged A. aeolicus r was purified from E. coli extracts without the heat-treatment step, large amounts of E. coli RNAP core were always present in the 100 mM imidazole fraction, suggesting that hybrid holoenzymes form with high efficiency. No E. coli RNAP core is observed when recombinant T. aquaticus r is puri fied. Previous work also suggested that while E. coli core was some what active with T. aquaticus r, the reverse combination was fully inactive . We used abortive transcription initiat ion to check activ ity of hybrid holoenzymes containing RNAP core and r subunits from A. aeolicus, E. coli, and T. aquaticus. As can be seen from Fig. 4, the A. aeolicus RNAP core was only active with its cognate r factor. In con trast, RNAP core from E. coli was active with all three r factors tested, with activity decreasing in the following order: E. coli r > A. aeolicus r > >T. aquaticus r, supporting the idea that A. aeolicus r binds E. coli core better than T. aquaticus r does. Of the two T. aquaticus RNAP core-based hybrids, the one containing E. coli r was inactive, while the one containing A. aeolicus r exhibited good activity levels.
3.4. In vitro transcription Transcription assays developed for E. coli RNAP were used to test activity of A. aeolicus RNAP with a single modification: the reaction temperature was raised to 70 °C. Standard abortive initi ation reactions contained, in 10 ll of transcription buffer (Table 1), 50 nM A. aeolicus RNAP core enzyme and 100 nM of recombinant A. aeolicus r. Reactions were pre-incubated for 10 min at 70 °C, followed by the addition of 100 nM T7 A1 promoter-containing DNA fragment (¡54 + 92) and additional 10- to 15-min incuba tion at 70 °C. Abortive transcription was initiated by the addition 100 lM CpA, 5 lM UTP, and 5 lCi a-[32P]UTP (3000 Ci/mmol), and allowed to proceed for 10 min at 70 °C. Run-off transcription was initiated by the addition 100 lM CpA (or 0.5 mM ATP), 5 lCi a-[32P]UTP (3000 Ci/mmol) and 10 lM NTPs, and allowed to pro ceed for 10 min at 70 °C. Reactions were terminated by the addi tion of equal volume of formamide-containing loading buffer and analyzed by denaturing gel-electrophoresis (8 M urea, 20% poly acrylamide) and autoradiography.
3.6. Antibiotic sensitivity of A. aeolicus RNAP Despite the high degree of evolutionary conservation, bacterial RNAPs exhibit differential sensitivity to antibiotics. This property must always be kept in mind when comparing structural data from available Thermus RNAP structures with genetic data obtained in other systems, as erroneous interpretations become possible. We determined the effect of two transcription inhibitors, microcin J (McJ) and rifampicin (Rif) on in vitro activity of A. aeolicus RNAP (McJ does not bind to and has no effect on Thermus RNAP; Ther mus RNAP behaves as a strong rifampicin-resistant mutant of E. coli RNAP). With McJ, we hoped that if robust McJ inhibition of A. aeolicus RNAP, a potentially crystallizable enzyme, is observed, then structural basis of transcription inhibition could be investi gated. This expectation was not fulfilled, unfortunately (Fig. 5). At 0.1 mg/ml, McJ had little, if any, effect on A. aeolicus RNAP tran scription either in an abortive initiation or in a steady-state runoff transcription assay. Transcription by the E. coli enzyme at these conditions was strongly inhibited, as expected. The transcription template used in the experiment of Fig. 5 also contained an intrin sic transcription terminator k tR2 between the transcription start site and the end of the transcription unit. As can be seen, the effi ciency of transcription termination (defined as the ratio of the amount of radioactivity in the band of terminated transcript to the sum or radioactivity in the terminated and run-off transcript bands) by the A. aeolicus RNAP was much lower than that of the E. coli enzyme. Similar observations were previously made for
Fig. 4. In vitro transcription by interspecies RNAP hybrids. The results of abortive initiation from a T7 A1 promoter-containing DNA fragment by holoenzymes reconstituted from indicated core enzymes and r subunits are presented.
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
Fig. 5. A. aeolicus RNAP is resistant to microcin J. Results of abortive and run-off transcription from a T7 A1-promoter template in the presence or in the absence of McJ are presented.
Fig. 6. Transcription inhibition by rifampicin. The results of steady-state transcription by the indicated enzymes in the presence of indicated concentrations of rifampicin are shown.
K. Kuznedelov, K. Severinov / Methods 47 (2009) 44–52
hermus RNAP, suggesting that thermophilic RNAPs may be less T effective in recognition of transcription terminators. Another compound tested was rifampicin, a classical inhibitor of bacterial RNAP. While most studies suggest that Rif is a sim ple steric inhibitor that blocks the extension of nascent transcripts beyond the length of 2–3 nucleotides , structural work with T. thermophilus RNAP led to more complex hypothetic al mechanisms, involving allosterical effects of Rif on the RNAP catal ytic center  (the conclusions of the latter paper were very recently seriously challenged and may not be valid, ). In the case of E. coli RNAP, Rif behaved in an expected way, inhibiting, at low concentrations, the production of run-off transcripts while strongly stimulating the production of abortive trimer CpApU (Fig. 6, note that in the gel system used here, the trimer has a lower mobility than the CpApUpC tetramer). In the case of T. aquatic us enzyme, increasing concentrations of Rif lead to progressively lower levels of run-off transcripts production (which however was never fully inhibited), with little or no effect on the CpApU production. The case of A. aeolicus RNAP presented yet another pattern of inhibition. Rif had little effect on run-off transcripts production, while clearly stimu lating abortive transcription. Such pattern may be consistent with reversible binding of Rif to the enzyme (high off rate). In contrast, in the case of T. aquaticus enzyme, the on rate of Rif binding may be low, but the binding itself is stable. Be that as it may, the observed differences suggest that the detailed mechanism of Rif action may be different in different systems, calling for caution when compar ing functional and structural data obtained with RNAPs from dif ferent origins. 4. Conclusions The protocols presented here were successfully used for RNAP preparation from various eubacteria. We believe that the same protocols can be used with minor modifi cation for preparation of RNAP from any bacteria. The availability of virtually unlimited amounts of bacterial RNAPs, coupled with modern proteomic tools
available for identific ation of RNAP-binding proteins and bioinfor matics tools for prediction of putative promoters shall allow iden tification of novel transcription regulatory mechanisms that are distinct from those operational in E. coli in the near future. Acknowledgments The project described above was supported by NIH RO1 Grants GM59295 and GM64503 and a Russian Academy of Sciences Pre sidium program grant in Cell and Molecular Biology to K.S. References  S. Borukhov, A. Goldfarb, Prot. Exp. Purif. 4 (1993) 503–511.  H. Tang, K. Severinov, A. Goldfarb, R.H. Ebright, Proc. Natl. Acad. Sci. USA 92 (1995) 4902–4906.  L. Minakhin, S. Nechaev, E.A. Campbell, K. Severinov, J. Bacteriol. 183 (2001) 71–76.  G. Zhang, E. Campbell, L. Minakhin, C. Richter, K. Severinov, S.A. Darst, Cell 98 (1999) 811–824.  K. Kuznedelov, N. Korzheva, A. Mustaev, K. Severinov, EMBO J. 21 (2002) 1369– 1378.  K. Kuznedelov, L. Minakhin, K. Severinov, Methods Enzymol. 370 (2003) 94– 108.  K. Kuznedelov, V. Lamour, G. Patikoglou, M. Chlenov, S.A. Darst, K. Severinov, J. Mol. Biol. 359 (2006) 110–121.  T. Naryshkina, K. Kuznedelov, K. Severinov, J. Mol. Biol. 361 (2006) 634–643.  I. Artsimovitch, V. Svetlov, K.S. Murakami, R. Landick, J. Biol. Chem. 278 (2003) 12344–12355.  E.C. Nedea, D. Markov, T. Naryshkina, K. Severinov, J. Bacteriol. 181 (1999) 2663–2665.  K. Igarashi, N. Fujita, A. Ishihama, Nucleic Acids Res. 18 (1990) 5945–5948.  G. Deckert, P.V. Warren, T. Gaasterland, W.G. Young, A.L. Lenox, D.E. Graham, R. Overbeek, M.A. Snead, M. Keller, M. Aujay, R. Huber, R.A. Feldman, J.M. Short, G.J. Olsen, R.V. Swanson, Nature 392 (1998) 353–358.  E. Campbell, N. Korzheva, A. Mustaev, K. Murakami, S. Nair, A. Goldfarb, S.A. Darst, Cell 104 (2001) 901–912.  I. Artsimovitch, M.N. Vassylyeva, D. Svetlov, V. Svetlov, A. Perederina, N. Igar ashi, N. Matsugaki, S. Wakatsuki, T.H. Tahirov, D.G. Vassylyev, Cell 122 (2005) 351–363.  A. Feklistov, V. Mekler, Q. Jiang, L.F. Westblade, H. Irschik, R. Jansen, A. Mus taev, S.A. Darst, R.H. Ebright, Proc. Natl. Acad. Sci. USA 105 (2008) 14820– 14825.