The Quorum-sensing Protein TraR of Agrobacterium tumefaciens is Susceptible to Intrinsic and TraM-mediated Proteolytic Instability

Summary

TraR of Agrobacterium tumefaciens is a LuxR-type transcription factor that regulates genes required for replication and conjugation of the tumor-inducing (Ti) plasmid. TraR binds the pheromone 3-oxo-octanoylhomoserine lactone (OOHL) and requires this molecule for folding into a protease-resistant, soluble conformation. Even after binding to OOHL, TraR is degraded at readily detectable rates. Here we show that the N-terminal domain of TraR, which binds OOHL, is more resistant to degradation than the full length protein, suggesting that sites on the C-terminal DNA binding domain (TraR(171-234)) enhance protein turnover. A fusion between GFP and TraR-(171-234) was poorly fluorescent, and truncations of this fusion protein allowed us to identify residues in this domain that contribute to protein degradation. TraR activity was previously shown to be inhibited by the antiactivator TraM. These proteins form 2:2 complexes that fail to bind DNA sequences. Here we show that TraM sharply decreased the accumulation of TraR in whole cells, indicating that TraM facilitates proteolysis of TraR. The TraM component of these complexes is spared from proteolysis, and could therefore act catalytically.

Introduction

The cellular abundance of many proteins can be greatly influenced by their degradation by cellular proteases (Gottesman, 2003). In eukaryotic cells, protein degradation generally occurs after a protein is polyubiquitinated by a ubiquitin ligase (Kriegenburg et al., 2012). These covalently modified proteins are processively reduced to short peptides within the 26S proteasome. Most bacteria are thought not to covalently modify proteins to mark them for proteolysis (although an exception is well documented in Mycobacterium tuberculosis (Burns & Darwin, 2010)). Bacteria can nonetheless use a number of different cytoplasmic proteases to selectively degrade proteins that are unfolded, misfolded, aggregated, or tagged with the SsrA peptide (Chandu & Nandi, 2004). In general terms, hydrophobic residues that are exposed to solvent are thought to act as a proteolysis signal (Gottesman, 2003). This can serve to target denatured or misfolded proteins, as correctly folded cytoplasmic proteins generally have hydrophobic interiors and hydrophilic exteriors. These proteases are multisubunit complexes containing a substrate binding component that binds and unfolds the target protein at the expense of ATP, and a protease component whose active site lies within a hollow core of the complexes (Baker & Sauer, 2006). There are five known cytoplasmic, ATP-dependent proteases in E. coli: ClpXP, ClpAP, HslUV, Lon, and FtsH (Langklotz et al., 2006).

Many proteins are required only at particular times of the cell cycle or in response to particular environmental stimuli and are inactivated by proteolysis when no longer needed (Chandu & Nandi, 2004). At least some of these proteins have evolved features such as solvent-exposed hydrophobic residues that enhance protease susceptibility. These sequences are frequently, though not always, found at the amino or carboxyl terminus of a protein. A well-studied example of this is the CtrA transcription factor of Caulobacter crescentus. CtrA accumulates during the G1 phase, regulates almost 100 promoters, and then is degraded by the ClpXP protease (Jenal, 2009). CtrA has two alanine residues at its C-terminus that provide a target for ClpXP (Domian et al., 1997). Other examples include CcrM of C. cresentus, the SulA, UmuC, UmuD, and MazE proteins of E. coli, and many others (Chandu & Nandi, 2004).

TraR of A. tumefaciens may provide another example of a protein with features that enhance proteolysis. TraR is a LuxR-type quorum sensing transcription factor that binds an autoinducer-type pheromone (3-oxo-octanoylhomoserine lactone, OOHL) (Zhu & Winans, 1999). TraR-OOHL complexes activate expression of Ti plasmid genes required for vegetative replication and for conjugative transfer by binding to specific sites called tra boxes that lie upstream of these promoters (Fuqua & Winans, 1996, Fuqua & Winans, 1994, Li & Farrand, 2000, Pappas & Winans, 2003). TraR requires OOHL, not only for activity, but also for solubility and for folding into a protease-resistant form (Zhu & Winans, 1999, Zhu & Winans, 2001). In the absence of OOHL, apo-TraR is degraded by the Clp and Lon proteases, with a half-life of just 2–3 minutes (Zhu & Winans, 2001).

TraR-OOHL complexes, while far more stable than apo-TraR, are nonetheless degraded at readily detectable rates, having a half-life of approximately 35 minutes in A. tumefaciens (Zhu & Winans, 2001). This turnover could play an important role in turning off TraR activity, especially given the fact that TraR appears to bind OOHL with an extremely high affinity and slow dissociation (Zhu & Winans, 1999). For this reason, dilution of cells to a low population density might not suffice to turn off expression of target genes, and proteolysis of the activator could play a central role in down-regulating the tra/rep regulon. A. tumefaciens C58 genome encodes three ClpP homologs and one homolog each of HslV, Lon, and FtsH.

TraR-OOHL complexes are inhibited by two antiactivator proteins, TrlR and TraM. TrlR is extremely similar to TraR but is truncated and lacks a DNA binding domain. TrlR can form heterodimers with TraR that contain just one DNA binding domain rather than two and are therefore unable to bind DNA (Chai et al., 2001, Oger et al., 1998, Zhu & Winans, 1998). TraM is a small protein that forms a 2:2 complex with TraR. A TraR-TraM complex from Rhizobium sp. NGR234 was visualized by X-ray crystallography (Chen et al., 2007). In this co-crystal, two TraM subunits contact a dimer of TraR in such a way that TraM forces the two CTDs apart, blocking their ability to bind to tra box DNA (Chen et al., 2007). The TraM binding site of TraR does not overlap the DNA binding site (Zheng et al., 2012). TraR activates the transcription of the traM gene, creating an interesting negative feedback loop (Fuqua et al., 1995, Hwang et al., 1995). Null mutations in traM show elevated levels of tra gene expression, OOHL production, and conjugation.

In this study we show that the DNA binding domain of TraR contains sites that decrease protein stability, and have mapped these sites. Little is known about the ultimate fate of TraM-TraR complexes. In this study we also show that TraR is degraded in the presence of TraM. Point mutations in TraR that are resistant to TraM-mediated inhibition were found to be resistant to proteolysis. In contrast, the concentration of TraM in whole cells was unaffected by TraR. We conclude that TraM targets TraR to proteolysis and that it itself is spared from proteolysis.

Results

Several amino acids in the TraR DNA binding domain decrease protein stability

In a previous study, the surface of TraR-CTD was mutagenized in a screen for positive control mutants (White & Winans, 2005). As might be expected, many of the mutants were strongly impaired for accumulation. Perhaps more surprisingly, several mutants showed an accumulation greater than wild type. These data suggested that the CTD of the wild type protein contains sites that destabilize the protein. To test this hypothesis, we performed pulse chase experiments on seven of these mutants, conducted in E. coli and using the T7 RNA polymerase to drive protein expression. The E. coli strain is BL21/DE3, which is naturally defective in the Lon protease (saiSree et al., 2001), so the turnover that we observed is presumably due primarily to the Clp protease (Zhu & Winans, 2001). Rifampicin was added to block the synthesis of all proteins other than TraR. The half-life of wild type TraR in this system was 212 minutes in the presence of OOHL and 3 minutes in its absence, in reasonable agreement with previous data (Pinto & Winans, 2009). All mutants showed a half-life that was 150% to 300% greater than wild type (Fig. 1). We conclude that the wild type residues at these positions in some way predispose the protein to proteolysis.

Figure 1.

Proteolytic stability of TraR variants with mutations in the carboxyl-terminal domain. The half-life of each mutant was calculated from the pulse-chase experiments shown on the left. Each of these mutants was previously shown using western immunoblots to accumulate to concentrations greater than wild type protein (White & Winans, 2005), and the accumulation of each mutant, expressed as a percentage of wild type, was determined in that study. Data are reprentative of experiments carried out three times.

TraR(1-170) is more abundant and more stable than full length TraR

If TraR-CTD contains proteolysis signals, then removal of this domain could increase the stability and abundance of the remainder of the protein. This was tested in two ways. First, we expressed full length TraR and TraR(1-170) in E. coli by using the T7 promoter, and performed pulse chase experiments using radiolabeled methionine. Full length TraR was degraded with a half-life of approximately 120 min (Fig. 2A). In contrast, TraR(1-170) was not detectably degraded over a period of 4 hours. We also attempted to measure the half-life of TraR(171-234) but were unable to detect this fragment even at early time points (data not shown), presumably due to its instability.

Figure 2.

The N-terminal domain of TraR is more stable than the full-length TraR. A) The half-life of full length TraR (circles) and TraR(1-170) (triangles) using strains KYC55(pHC016) and KYC55(pYC108), respectively were determined by pulse-labeling chase experiments. B) OOHL sequestration assays of KYC55(pYC107), KYC55(pYC108) and KYC55(pHC016) showed that the N-terminal domain sequestered more OOHL than full length TraR, consistent with its longer halflife. AU: arbitrary units. Sequestration data are the average of three experiments with the standard errors indicated.

We also measured the sequestration of exogenous OOHL by cultures of cells that express full length TraR, TraR(1-170), or TraR(171-234). We cultured strains containing these proteins in the presence of OOHL and quantified the bound OOHL using a bioassay (Zhu et al., 1998). As expected, the strain that expressed TraR(171-234) did not sequester OOHL above background levels (Fig. 2B), while the strains expressing full length TraR or TraR(1-170) sequestered readily detectable amounts. Of these, the strain expressing TraR(1-170) sequestered approximately 2.5 times more OOHL than the strain expressing full length TraR (Fig. 2B). These data support the hypothesis that TraR(1-170) accumulates to higher levels than full length TraR.

Identification of TraR(170-234) residues that enhance proteolysis

To determine whether TraR(170-234) could destabilize an unrelated protein, we constructed a fusion between GFP and TraR(170-234), and expressed the fusion protein in E. coli using a Ptac promoter. A strain expressing this fusion showed approximately 4-fold lower levels of fluorescence than a strain expressing a GFP control (Fig. 3A), indicating that TraR(170-234) increased proteolysis of the fusion protein or impaired its folding into a mature, fluorescent form.

Figure 3.

The C-terminal domain of TraR, when fused to GFP, decreased GFP fluorescence, and a genetic screen was carried out to identify mutants that restore fluoresence. A) E. coli colonies expressing either GFP (left) or GFP-TraR(171-234) (right). B) A fusion between GFP and TraR(170-234), showing the positions and frequency of stop codon mutations that restored GFP fluorescence.

The gene encoding GFP-TraR(170-234) was mutagenized randomly using PCR amplification and screened for variants showing increased levels of fluorescence. We had hoped to isolate missense mutations that would resist proteolysis. To our initial surprise, all mutants isolated from this screen had nonsense codons within the TraR portion of the fusion protein (Fig. 3B). We recovered stop codon mutations at only 5 positions, one each at codons 173 and 177 (numbering from the start codon of full length TraR), and five stop codons at each of codons 181, 183, and 184. Each mutation was caused by a single nucleotide change, and all of them lay near the junction of the fusion protein (Fig. 3B). The fact that nonsense mutations were never recovered at any codon downstream of codon 184 provided suggestive evidence that a protease recognition motif lay close to that codon. Codons 184–187 encode the hydrophobic residues Trp-Ile-Ala-Val. A stop codon at position 188 would therefore create a C-terminus with four hydrophobic residues that might destabilize the protein, while a stop codon at position 184 would lack these residues and might therefore be more stable and show greater fluorescence.

In an effort to more finely dissect this possible protease recognition site, we used PCR amplification to construct the fusions GFP-TraR(170-184), GFP-TraR(170-185), GFP-TraR(170-186) and GFP-TraR(170-187). GFP-TraR(170-184) and GFP-TraR(170-185) had relatively high levels of fluorescence (Fig. 4). In contrast, GFP-TraR(170-186) and GFP-TraR(170-187) were far less fluorescent, suggesting that Ala186 and Val187 can function as a protease recognition motif (PRM), at least when they lie at the C-terminus of the fusion. We then mutated the gene encoding GFP-TraR(170-187) to make its C-terminus more hydrophilic. We individually altered A186 and V187 to glutamate residues using site-directed mutagenesis. Both mutations increased fluorescence levels (Fig. 4). We note that the mutation V187E in an otherwise wild type TraR protein enhanced accumulation and stability (Fig. 1). These findings suggest that V187 may be part of a protease recognition motif.

Figure 4.

Fluorescence Intensity of GFP-TraR(171-234) variants bearing truncations within the TraR portion of the fusion protein. A truncation that removed residues 185–234 restored fluorescence, while a truncation that removed residues 188–234 did not, suggesting that a protease recognition motif lies at least partly in this region. Fluorescence arbitrary units are averages of triplicate cultures with standard errors indicated in parentheses.

TraM enhances TraR proteolysis

It was previously shown that TraR can bind to the antiactivator TraM and that the resulting complexes are unable to bind DNA (Chen et al., 2007, Qin et al., 2007, Luo et al., 2000, Swiderska et al., 2001). However, the ultimate fate of the TraR-TraM protein complexes in whole cells was not known. To determine whether either of these proteins was degraded, we constructed two plasmids: pEC501, which constitutively expresses native TraR from a Ptet promoter, and pEC508, which expresses a Plac-flag-traM epitope-tagged fusion, and also contains lacI^Q, so that expression of the fusion can be controlled using IPTG. The abundance of TraR protein was controlled by varying the concentrations of OOHL, as apo-TraR is rapidly degraded (Zhu & Winans, 2001). These plasmids were introduced into A. tumefaciens strain NTL4, which lacks a Ti plasmid, cultured to late log phase in the presence of varying concentrations of IPTG and OOHL, and analyzed for TraR and Flag-TraM contents by western immunoblotting.

In experiments to detect TraR degradation, our goal was to express TraM in mole excess over TraR, and therefore we provided OOHL at a limiting concentration (0.1 nM), just high enough to allow detection. In contrast, IPTG was added at various concentrations, including saturating levels. As expected, the abundance of TraM was proportional to IPTG concentration (Fig. 5A). TraM was not detected after growth in the presence of 0, 40, or 80 µM IPTG, and expression was saturated by growth in 1280 µM. Significantly, the abundance of TraM and TraR were inversely correlated. At high levels of TraM, TraR was undetectable. At 80 µM IPTG, TraM was not detected but TraR was less abundant than at 0 or 40 µM IPTG, suggesting that sufficient TraM was synthesized to impact TraR accumulation. In control experiments, the addition of IPTG to cells lacking traM did not alter TraR abundance (data not shown). Also, TraR did not accumulate in the absence of OOHL regardless of TraM levels, in agreement with previous studies that apo-TraR is rapidly degraded even in strains lacking TraM (Zhu & Winans, 1999, Zhu & Winans, 2001). We conclude that TraM decreases the abundance of TraR.

Figure 5.

Controlled overexpression of TraM leads to decreases in TraR accumulation. Plasmid pEC508 has an IPTG-inducible Plac-Flag-traM fusion, while pEC501 expresses traR constitutively. A. Western immunoblots of NTL4(pEC501)(pEC508) cells cultured in the presence of 0.1 nM OOHL with different concentrations of IPTG (indicated in the top of the figure). Increasing concentrations of Flag-TraM are associated with decreases in TraR abundance. B. Flag-TraM caused decreased OOHL sequestration by TraR. Data are representative of experiments carried out three times.

We did similar experiments using higher concentrations of OOHL, which enhanced the abundance of TraR. In the presence of 1 nM OOHL, ITPG caused a small decrease in TraR abundance, while in the presence of 10 nM or 100 nM OOHL, IPTG had no detectable effect (Table 1). These data indicate that high levels of TraR can overcome the destabilizing effects of TraM, confirming earlier studies (Hwang et al., 1999, Hwang et al., 1995). We also used the same strain to measure the sequestration of OOHL in the presence of 0.1 nM OOHL and the presence or absence of 1 mM IPTG. The strain expressing both TraR and TraM sequestered only 5% as much OOHL per cell as the cell expressing only TraR (data not shown). These data confirm that TraM directs the destruction of TraR.

Table 1.

TraR accumulation in the presence of increasing amounts of TraM¹

	TraR Abundance
IPTG (µM)	0	40	80	160	320	640	1280	2560
OOHL 0.1 nM	(100)2	98	71	26	14	11	< 10	< 10
OOHL 1.0 nM	(100)	96	102	98	87	72	69	63
OOHL 10 nM	(100)	NT3	NT	NT	NT	106	107	101
OOHL 100 nM	(100)	NT	NT	NT	NT	104	104	107

¹A. tumefaciens strain NT1(pEC501)(pEC508) contains a constitutive Ptet-traR fusion and an IPTG-regulated Plac-flag-traM fusioin. OOHL and IPTG were added at the indicated concentrations to stabilize TraR against proteolysis and to express Flag-TraM, respectively.

²TraR accumulation was assessed by Western immunoblots and normalized to the abundance in the absence of IPTG (second column). All band intensities were also normalized to cross reacting material.

³not tested.

Previous studies have shown that TraR mutations L182F, A195T, and A195V interfere with antiactivation by TraM (Qin et al., 2007, Luo et al., 2000). The abundance of TraR proteins containing these mutations was measured in the presence and absence of Flag-TraM. TraM was detected only in the presence of IPTG, as expected (Fig. 6A). Also as expected, the abundance of wild type TraR was severely decreased by IPTG (compare Fig. 6B, lanes 1 and 2). In contrast, TraR(L182F), TraR(A195T) and TraR(A195V) were still readily detected in the presence of IPTG (Fig. 6B). The accumulation of TraR was not completely restored by these mutations, consistent with earlier findings that these single mutations in TraR do not disrupt completely the interaction between TraM and TraR (Qin et al., 2007).

Figure 6.

TraR mutations that were previously shown to resist inhibition by TraM are resistant to TraM-mediated degradation. A. Western immunoblots of NTL4(pEC501)(pEC508) cells cultured in the presence and absence of IPTG, which induces expression of the Plac-Flag-traM fusion. B. Western immunoblots that TraR mutations that were previously shown to resist inhibition by TraM also resist TraM-mediated proteolysis. Accumulation of wild type and each mutant was normalized to the band intensity of the same TraR protein in the absence of IPTG. All band intensities were also normalized using the intensity of cross-reacting material. Data are representative of three experiments.

TraM levels do not decrease with increasing amounts of active TraR

Data in the previous section suggest that TraM targets TraR for proteolysis. We wanted to determine whether the converse is also true, that is, whether TraR can target TraM for proteolysis. We cultured the same strain expressing Ptet-traR and Plac-Flag-traM in the presence of a low concentration of IPTG (160 µM). This concentration of IPTG was sufficient for detection of Flag-TraM (Fig. 7) and low enough not to affect the concentration of TraR at high OOHL levels (Table 1). The media were also supplemented with 0, 1, 10, 100, or 1000 nM OOHL in order to vary the concentrations of TraR. As expected, increasing amounts of OOHL caused increasing abundance of TraR (Fig. 7). However, the abundance of TraM was constant in all cultures. These results indicate that TraM was not degraded even in the presence of excess TraR.

Figure 7.

Western blots of Flag-TraM and Flag-TraR in strains cultured with limiting amounts of IPTG in order to limit the expression of Flag-TraM, and a variety of concentrations of OOHL to vary the concentration of Flag-TraR. Even high concentrations of Flag-TraR did not decrease the abundance of Flag-TraM, indicating that the TraM portion of TraM-TraR complexes is stable to proteolysis.

Discussion

This study has shown that the DNA binding domain of TraR plays a critical role in both intrinsic and TraM-mediated turnover of the protein. It contains several residues mutations in which enhance stability, suggesting that the wild type residues destabilize the protein, and also contains a binding site for TraM, which inactivates the protein (Chen et al., 2007) and ultimately directs TraR to the cytoplasmic proteolytic machinery (this study).

In earlier studies, the surface of TraR was systematically mutated in a search for residues required for positive control of transcription (Costa et al., 2009, White & Winans, 2005), as these residues probably make direct contact with RNA polymerase. Of the 60 mutations constructed at 38 positions in the C-terminal domain of TraR, 35 accumulated to levels 50% or less that of wild type, while nine accumulated to levels 125% greater than wild type (White & Winans, 2005). In a similar study of the TraR-NTD, 117 mutations were made at 103 positions. Only eight of these mutants impaired accumulation by 50% or less, and only two caused over-accumulation of 125% or more (Costa et al., 2009). These data provide further evidence that the TraR-CTD plays a significant role in proteolysis, and that many point mutations can influence this activity.

Evidence using GFP-TraR fusions suggested that TraR residues W184, I185, A186, and V187 could destabilize TraR, possibly in conjunction with neighboring residues (Fig. 4). Of these, residues I185 and A186 are sequestered by the opposite subunit of a TraR dimer (Vannini et al., 2002, Zhang et al., 2002). They would therefore be protected from proteases in TraR dimers but exposed in monomers. Residues W184 and V187 are exposed to solvent in TraR dimers. However, both residues are essential for positive control, suggesting that they make direct contact with RNA polymerase (Costa et al., 2009, Qin et al., 2009, White & Winans, 2005). They might therefore be protected by RNA polymerase from the proteolysis machinery. The stability of TraR may also be impacted by its binding to DNA. Of the seven mutants tested, three (K201A, K208A, and R210A) are completely defective in DNA binding, while a fourth (F216W) is almost completely defective. Structural studies indicate that all four wild type residues are in close proximity with DNA. This suggests that these destabilizing residues would be protected by DNA from proteolysis. Taken together, these data suggest that TraR monomers may present several protease-sensitive amino acids, while TraR dimers would present fewer residues, TraR-DNA complexes would present even fewer, and that TraR-DNA-RNA polymerase ternary complexes would present fewest of all.

We also have examined the fate of complexes containing TraR and TraM. The finding that TraM blocks the accumulation of TraR could be interpreted to mean that TraM causes a general increase in proteolysis. However, we also found that three TraR mutants (L182F, A195V and A195T) that are resistant to the effects of TraM in transcription assays (Qin et al., 2007, Luo et al., 2000) are also resistant to TraM in proteolysis assays. TraM still shows residual activity in both assays. Furthermore, TraR residues L182 and A195 are in close proximity to TraM in co-crystals (Chen et al., 2007). These findings argue strongly that TraM destabilizes TraR via direct protein-protein interactions.

As described above, TraM accumulation is not affected by the presence or absence of TraR, indicating that complexes containing both proteins lead to the destruction of just TraR. There are other examples of proteins, often referred to as adaptors that deliver specific proteins to cellular proteases. Bacterial adaptor proteins include the E. coli proteins ClpS, RssB, and SspB (Flynn et al., 2004, Zhou et al., 2001b, Feng & Gierasch, 1998), and MecA and YpbH of B. subtilis (Nakano et al., 2002, Turgay et al., 1997).

Some adaptors, such as SspB, can modulate the degradation of several different proteins (Flynn et al., 2004). Our results suggest that TraM may work as an adaptor protein, that it not only inactivates TraR, but also delivers TraR to cellular proteases for degradation. It is far from clear why TraR benefits from a dedicated adaptor. It seems plausible that TraM synthesis, stability, or function may be responsive to some physiological or environmental signal. If so then expression of all genes regulated by TraR would be impacted by that signal.

Experimental Procedures

Bacterial strains, plasmids and oligonucleotides

Bacterial strains and plasmids used in this study are described in Table S1. Oligonucleotides used for PCR amplification and mutagenesis (Table S2) were obtained from Integrated DNA Technologies (Coralville, Iowa). A. tumefaciens strains were cultured in AT minimal medium at 28°C (Tempé et al., 1977), while E. coli strains were cultured in Luria broth (LB) or solid medium at 37°C (Miller, 1972). Synthetic OOHL was provided by A. Eberhard (Cornell University). Restriction enzymes were obtained from New England Biolabs. Spectinomycin and kanamycin were added to cultures of A. tumefaciens at final concentrations of 200 µg ml⁻¹. Spectinomycin, kanamycin, ampicillin, gentamycin and chloramphenicol were added to cultures of E. coli at final concentrations of 100 µg ml⁻¹, 100 µg ml⁻¹, 200 µg ml⁻¹, 15 µg ml⁻¹ and 10 µg ml⁻¹, respectively.

DNA manipulations

Recombinant DNA techniques were performed using standard procedures (Sambrook & Russell, 2001). Plasmid DNA was isolated from E. coli with QIAprep miniprep kits (Qiagen) for DNA sequence analysis. DNA sequences obtained by PCR were verified using automated DNA sequencing (The Cornell University Life Sciences Core Laboratories Center) and analyzed using the LaserGene program (DNASTAR). Plasmids were introduced into E. coli by transformation (Sambrook & Russell, 2001) and into A. tumefaciens by electroporation (Cangelosi et al., 1991). E. coli strain DH5á was used as a transformation recipient for all plasmid constructions.

Construction of plasmid pEC500 containing the Ptet and tetR was performed using transposon Tn10 as a template. A 783 bp fragment of Tn10 containing tetR, Ptet, operator regions, and 18 bp of tetA (three stop codons were added at the end of the tetA fragment) was amplified by PCR using the oligonucleotidesTn10F and Tn10R (Table S2). The resulting fragment was digested with EcoRI and SacI and cloned into the same restriction sites of pPZP200. The resulting plasmid is pEC500.

To construct the plasmid expressing TraR, the traR gene was PCR amplified with the use of plasmid pYC335 (Chai & Winans, 2004) as a template and oligonucleotides TraRF and TraRR (Table S2). The resulting DNA fragment was digested with KpnI and BamHI and cloned into pEC500 digested with the same enzymes, resulting in plasmid pEC501.

A plasmid expressing TraM was constructed by amplifying traM from the genome of A. tumefaciens strain R10 using oligonucleotides Flag-tag-TraMF and TraMR (Table S2). The resulting PCR fragment was digested using NdeI and BamHI and cloned into the same sites of pSRKGent (Khan et al., 2008) resulting in plasmid pEC508.

Random and site-directed mutagenesis

Random mutagenesis by error-prone PCR was performed using primers TraR-CTDF2 and TraR-CTDR3 (Table S2) and plasmid pYC350, which expresses GFP-TraR(171-234) as template. The PCR products and the plasmid pJBA113 were digested with PstI and HindIII and ligated. Transformants were screened for increased fluorescence under UV light. The candidates were sequenced using primer TraR-CTDR4 and fluorescence was quantified using Synergy HT Spectrophotometer (BioTek). Fluorescence intensity values were normalized against culture turbidity.

Site-directed mutagenesis of TraR was performed using synthetic overlap extension PCR (Sambrook & Russell, 2001). Plasmid pYC350 was used as a template and amplified using Taq polymerase High Fidelity (Invitrogen). The restriction sites for PstI and HindIII were used to introduce mutated DNA fragments into the wild type gene. The forward primer TraRCTDF1 and reverse primers ( R8, R9, R10, R11 and R12) were used to amplify TraR-CTD fragment with various substitutions (Table S2) using pYC335 as a template. The fluorescence of each TraR allele was measured using a Synergy HT Fluorometer from BioTek. Fluorescence intensity values were normalized against culture turbidity.

For the mutants TraR which are resistant to antiactivaton by TraM, a PCR product was generated by overlap extension PCR using Taq polymerase High Fidelity (Invitrogen). The flanking primers TraRF and TraRR were used in separate reactions with two different mutagenic primers that overlap at the mutation (L182FF, L182FR, A195TF, A195TR, A195VF and A195VR) (Table S2), using pEC501 as the template. These two PCR products are then combined and used as the template in a second round of PCR with primers TraRF and TraRR to generate the complete 747 bp fragment. The second set of PCR products generated were digested with KpnI and BamHI, and ligated to pEC501 digested with the same enzymes. Mutant sequences were confirmed by automated DNA sequencing.

TraR stability in E. coli

The measurement of TraR turnover was performed using BL21/DE3(pKP105), and BL21/DE3(pJZ358) to compare the stability of TraR-NTD and full length TraR. For the experiments with the mutants of TraR which are more stable than wild type TraR, the turnover of the protein was determined using strain BL21/DE3(pJZ358) (Zhu & Winans, 1999) or derivatives of pJZ358 carrying traR mutants. The experiments were performed as described previously (Zhu & Winans, 2001).

OOHL sequestration assay by TraR in whole cells

A. tumefaciens strains KYC55(pYC107), KYC55(pYC108), KYC55(pHC016) and NTL4(pEC501)(pEC508) were used for OOHL sequestration assays as previously described (Chai & Winans, 2004). All assays were performed twice with independent cultures.

Immunodetection of TraR

The abundance of each TraR and TraM was determined in A. tumefaciens by suspending centrifuged cells in 5% of their original volume in 1x cracking buffer (125 mM Tris pH 6.8, 4% SDS, 20% glycerol, 200 mM DTT, 0.02% bromophenol blue). Cells were disrupted by boiling for 5 min, cooling to −80°C and boiling for another 5 min. A fraction of each sample was size-fractionated using 12% SDS polyacrylamide gels, and electroblotted onto nitrocellulose membranes (BIORAD). The membranes were blocked using TBS (20 mM Tris pH 7.9, 500 mM NaCl, 0.05% Tween 20) supplemented with 5% skim milk. TraR was immunodetected in TBS with pre-adsorbed polyclonal anti-TraR rabbit antiserum (Chai and Winans, 2004) and goat anti-rabbit IgG conjugated with alkaline phosphatase (BIORAD) was used as the secondary antibody. TraM was immunodetected in TBS with monoclonal anti-Flag M2 antibody produced (Sigma-Aldrich). Goat anti-mouse IgG conjugated with alkaline phosphatase (Jackson ImmunoResearch Laboratories) was used as the secondary antibody. The membranes were stained with BCIP (5-bromo-4-chloro-3-indoyl phosphate p-toluidine salt) and NBT (p-nitro blue tetrazolium chloride) (BIORAD). Westerns were performed with fresh cell lysates for each strain at least three times. Data were analyzed using ImageJ (Rasband, 2004), and normalized against cross-reacting material in each lane.