Bioluminescence Resonance Energy Transfer System for MeasuringDynamic Protein-Protein Interactions in Bacteria

Boyu Cui,a Yao Wang,a Yunhong Song,a Tietao Wang,a Changfu Li,a Yahong Wei,a Zhao-Qing Luo,b Xihui Shena

State Key Laboratory of Crop Stress Biology for Arid Areas and College of Life Sciences, Northwest A&F University, Yangling, Shaanxi, Chinaa; Department of BiologicalSciences, Purdue University, West Lafayette, Indiana, USAb

B.C. and Y.W. contributed equally to this article.

ABSTRACT Protein-protein interactions are important for virtually every biological process, and a number of elegant approacheshave been designed to detect and evaluate such interactions. However, few of these methods allow the detection of dynamic andreal-time protein-protein interactions in bacteria. Here we describe a bioluminescence resonance energy transfer (BRET) systembased on the bacterial luciferase LuxAB. We found that enhanced yellow fluorescent protein (eYFP) accepts the emission fromLuxAB and emits yellow fluorescence. Importantly, BRET occurred when LuxAB and eYFP were fused, respectively, to the inter-acting protein pair FlgM and FliA. Furthermore, we observed sirolimus (i.e., rapamycin)-inducible interactions between FRBand FKBP12 and a dose-dependent abolishment of such interactions by FK506, the ligand of FKBP12. Using this system, weshowed that osmotic stress or low pH efficiently induced multimerization of the regulatory protein OmpR and that the multim-erization induced by low pH can be reversed by a neutralizing agent, further indicating the usefulness of this system in the mea-surement of dynamic interactions. This method can be adapted to analyze dynamic protein-protein interactions and the impor-tance of such interactions in bacterial processes such as development and pathogenicity.

IMPORTANCE Real-time measurement of protein-protein interactions in prokaryotes is highly desirable for determining the rolesof protein complex in the development or virulence of bacteria, but methods that allow such measurement are not available.Here we describe the development of a bioluminescence resonance energy transfer (BRET) technology that meets this need. Theuse of endogenous excitation light in this strategy circumvents the requirement for the sophisticated instrument demanded bystandard fluorescence resonance energy transfer (FRET). Furthermore, because the LuxAB substrate decanal is membrane per-meable, the assay can be performed without lysing the bacterial cells, thus allowing the detection of protein-protein interactionsin live bacterial cells. This BRET system added another useful tool to address important questions in microbiological studies.

Received 9 March 2014 Accepted 10 April 2014 Published 20 May 2014

Citation Cui B, Wang Y, Song Y, Wang T, Li C, Wei Y, Luo Z-Q, Shen X. 2014. Bioluminescence resonance energy transfer system for measuring dynamic protein-proteininteractions in bacteria. mBio 5(3):e01050-14. doi:10.1128/mBio.01050-14.

Invited Editor Ilya Manukhov, State Research Institute of Genetics and Selection of Industrial Microorganisms (GosNIIGenetika) Editor Stephen Winans, Cornell University

Copyright © 2014 Cui et al. This is an open-access article distributed under the terms of the Creative Commons Attribution-Noncommercial-ShareAlike 3.0 Unported license,which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

Address correspondence to Xihui Shen, xihuishen@nwsuaf.edu.cn, or Zhao-Qing Luo, luoz@purdue.edu.

Protein complexes of multiple components govern essentiallyall biological processes, and a number of elegantly designed

methods have been developed to identify interacting proteins andto evaluate such interactions (1–3). Successful utilization of thesemethods has led to the identification of numerous interacting pro-teins and the elucidation of many important biological processes(4, 5). Two-hybrid systems represent one class of such methods.These systems, depending upon the activity of the reporter pro-teins used as the readout, can be divided into two categories. In thefirst category, exemplified by the yeast two-hybrid system, a tran-scriptional factor was split into two halves; each has lost the orig-inal activity in the activation of transcription. Fusion of interact-ing proteins to each of the fragments restores such activity bybringing these two fragments into physical proximity (1). Thesecond category involves in the use of specific enzymes whoseenzymatic activity can be restored when the split fragments arebrought together by interacting proteins (6). For example, be-cause of its simplicity in experimental manipulation, the cyclic

AMP synthase (Cya)-based system has been widely used in thestudy of protein-protein interaction in bacteria (7).

Whereas two-hybrid systems offer the advantage of screeningthrough numerous candidate proteins from complex libraries po-tentially representing every protein of a given organism, they arenot feasible for measuring real-time dynamic binding. This limi-tation was overcome by several protein fragment complementa-tion assays (PCA) (8) and by fluorescence and bioluminescenceresonance energy transfer (FRET and BRET, respectively) (9, 10).In PCA, inactive fragments of reporter proteins, such as the greenfluorescent protein (GFP) and light-emitting luciferase, werefused to the proteins of interest and the binding of these proteinsrestores their activity (11, 12). In FRET and BRET, energy gener-ated by either a fluorescent protein or by a light-emitting lu-ciferase is transferred to an acceptor fluorophore (13, 14). Suchtransfer occurs when the molecules are in close proximity, gener-ally within 10 to 100 Å, depending on the extent of spectral overlap(15), leading to the emission of a different fluorescence signal (13,

RESEARCH ARTICLE

May/June 2014 Volume 5 Issue 3 e01050-14 ® mbio.asm.org 1

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

16, 17). These methods allow direct visualization of the interac-tions in the native organism and thus can reveal the cellular com-partment within which the interactions occur (18).

BRET assays have been widely used to analyze protein-proteininteractions in eukaryotic cells, particularly in experiments thatmandate real-time dynamic measurement of the bindings (9, 18–20). For example, by using this method, Perroy et al. successfullymeasured the kinetics of ubiquitination of �-arrestin 2 (21). How-ever, similar applications are limited in bacterial cells, probablydue to factors such as lower sensitivity caused by suboptimal ex-pression of the luciferase of eukaryotic origin. Indeed, althoughthe use of firefly Hotaria parvula luciferase (HpL) and Gaussiaprinceps luciferase (GLuc) has been reported (22–24), the 35-kDaRenilla luciferase (Rluc) is the most commonly used light-emitting luciferase in BRET (25).

We hypothesized that a BRET assay based on a luciferase ofbacterial origin may have the potential to overcome the limita-tions associated with the existing systems and allow the develop-ment of methods more suitable for measuring interactions in pro-karyotic cells. To this end, we established a BRET system using theLuxAB from Photorhabdus luminescens (26) and enhanced yellowfluorescent protein (eYFP) (27). We found that the spectrum ofLuxAB emission is sufficiently broad to excite fluorescence ofeYFP but not that of enhanced GFP (eGFP) or mOrange, a fluo-rescent protein from Discosoma sp. (28). Using this system, weobserved interactions between FlgM and FliA, two proteins in theflagellar regulon; we also examined sirolimus (i.e., rapamycin)-dependent, FK506-titratable dynamic interactions between FRB,the sirolimus-binding domain of mTOR (29, 30), and the FK506-binding protein FKBP12 (31). This system was further used toevaluate the impact of distinct environmental stimuli on the mul-timerization of OmpR, the transcriptional regulatory protein in-volved in sensing osmolarity/pH changes (32–35). Our results es-tablished LuxAB-based BRET as a potentially useful tool indissecting dynamic protein-protein interactions in bacterial cells.

RESULTSExcitation of eYFP by LuxAB. To establish a BRET system basedon the bacterial luciferase LuxAB, we first examined the ability ofits emission to excite several receptor fluorophores. To this end,we chose to test GFP, the monomeric fluorophore mOrange, de-rived from the coral red fluorescent protein DsRed, and eYFP, theGFP mutant that emits yellow fluorescence. Based on their pho-tophysical properties, a significant spectral overlap only exists be-tween LuxAB and eYFP (Fig. 1A), which potentially can be ex-ploited for a BRET assay in which LuxAB and eYFP will be broughtinto proximity by interacting proteins, respectively, fused to thefluorophores (Fig. 1B and C). On the other hand, due to the lack ofsignificant spectral overlap between LuxAB and GFP or mOrange,BRET is not expected when either of these two proteins is used asthe receptor fluorophore.

We thus made constructs that allow lac promoter (Plac)-dependent expression of each of these three fluorophores fused tothe carboxyl end of LuxAB (Fig. 2A). As expected, the light pro-duced by Escherichia coli cells expressing LuxAB became detect-able around 430 nm and peaked at 490 nm (Fig. 2B to D). Cellsexpressing GFP, mOrange, or eYFP also emitted fluorescence sig-nals with their respective spectra (Fig. 2B to D). Furthermore, theemission spectrum of cells expressing LuxAB-GFP or LuxAB-mOrange strain was indistinguishable from that of LuxAB itself

(Fig. 2B and C), indicating that resonance energy transfer did notoccur. Importantly, in cells expressing LuxAB-eYFP, in additionto the peak at 490 nm, another peak was detected at 527 nm, awavelength that is identical to the emission of eYFP itself(Fig. 2D). This excitation is consistent with the significant spectraloverlap between LuxAB and eYFP (Fig. 1A). Furthermore, thesignal intensity of this peak was comparable to that emitted byeYFP, suggesting that the excitation signal from LuxAB is able totrigger the fluorescence of eYFP almost at the maximal level. Theinability of LuxAB to excite GFP or mOrange is consistent with thefact that no spectral overlap exists between LuxAB and each ofthese fluorophores, which have major excitation peaks at 395 nm(36) and 548 nm (28), respectively (Fig. 2B and C). Because theluciferase of Photorhabdus luminescens is not related to eYFP fromAequorea, intrinsic interactions between LuxAB and eYFP are veryunlikely. Thus, we chose these two fluorophores to establish aBRET assay for analyzing protein-protein interactions.

The sensitivity of the LuxAB-based BRET depends upon theorientation of protein fusion. To provide the proof-of-conceptevidence for this method, we examined the interactions betweenthe alternative sigma factor FliA specific for the flagellar regulon(37) and the anti-sigma factor FlgM (38, 39). FliA and FlgM werefused to the carboxyl or amino terminus of LuxAB and eYFP,respectively (Fig. 3A). BRET signals indicative of interactions be-tween FlgM and FliA were detected when FliA was fused to thecarboxyl end of LuxAB irrespective of the orientation of the eYFP-FlgM fusion (Fig. 3B and C). Although the eYFP-FlgM fusion wasexpressed at a high level (see Fig. S1 in the supplemental material),its BRET signal was considerably lower than that of the FlgM-eYFP fusion (Fig. 3C). Similar results were obtained when theratio of emission at 527 nm to that at 490 nm (Y:B ratio) wasmeasured (Fig. 3D). On the other hand, luminescence signalscould not be detected when FliA was fused to the N-terminal endof LuxAB, and no further experiments were performed with thisconstruct. Thus, the orientation of the fusion is important for thesensitivity or sometimes even for the functionality of LuxAB-based BRET to work.

Dynamic interactions between FRB and FKBP12. To deter-mine the usefulness of the LuxAB-based BRET system, we first setout to analyze the interactions of two proteins known to regulat-ably interact with each other. To this end, we chose FRB, thesirolimus-binding domain of mTOR and its binding partner,FKBP12, a 12-kDa protein that also binds the immunosuppres-sant FK506 with high affinity (30). This interacting protein pairoffers at least two advantages in the evaluation of the LuxAB-basedBRET. First, their interaction depends on sirolimus (40) (Fig. 4A),a property that allows us to examine whether intrinsic associationbetween the LuxAB and eYFP fusions occurs. Second, the FRB-FKBP12 complex can be abolished by FK506 (31, 41), which po-tentially will enable us to determine the reversibility of the inter-actions in real time.

To ensure that these proteins are properly expressed in bacte-rial cells, we first codon optimized the mouse cDNA of bothFKBP12 and FRB in accordance with the codon preference ofeubacteria, such as E. coli (see Fig. S2 in the supplemental mate-rial). To determine the interactions between these two proteins,we used the ratio of emission at 527 nm to that at 490 nm (Y:Bratio) to measure the intensity of BRET or the strength of theinteraction (19). E. coli cells expressing both LuxAB-FKBP12 andFRB-eYFP were measured for BRET under different conditions.

Cui et al.

2 ® mbio.asm.org May/June 2014 Volume 5 Issue 3 e01050-14

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

In samples receiving the solvent dimethyl sulfoxide (DMSO), noBRET signal was detected, and the Y:B ratio was constantly atapproximately 0.55 (Fig. 4B) over the experimental duration. Insamples receiving 100 nM sirolimus, BRET signals became detect-able within 2 min and plateaued 10 min after addition of theinducer (Fig. 4B). Interestingly, when FK506, the inhibitory com-petitor for FKBP12 was added together with sirolimus to the bac-terial cultures, the emission of BRET signal was arrested, which isconsistent with the fact that FK506 competes for the binding siteof sirolimus, leading to reduced interactions between FKBP12 andFRB (Fig. 4B). The sirolimus-dependent BRET signals betweenthis protein pair indicate that no intrinsic association occurs be-tween LuxAB and eYFP. To further explore the usefulness of thismethod, we sought to determine the inhibitory constant of FK506against sirolimus by measuring its effects on the Y:B ratio at dif-ferent concentrations. The inhibitory effect of FK506 became de-tectable at 0.001 nM and reached the maximal at 10 �M (Fig. 4C).Calculations based on these results revealed that in our system, the50% inhibitory constant (IC50) of FK506 was 0.37 nM (Fig. 4C).These results indicate that the LuxAB-based BRET method allowsthe measurement of dynamic protein-protein interactions.

Multimerization of OmpR measured by the LuxAB-basedBRET. Many important bacterial regulatory proteins function bybinding to promoters of their target genes as multimers uponreceiving proper environmental stimuli. Because of the lag causedby transcription, translation, and the final maturation of the pro-tein produced by conventional reporter genes such as lacZ and gfp,it has been difficult to determine the kinetics of the response of theregulatory proteins. To test the usefulness of the LuxAB-basedBRET in analyzing protein multimerization in response to envi-ronmental cues, we chose OmpR, the widely distributed bacterialregulatory protein that responds to signals such as changes in os-molarity and pH (34, 35, 42). The interaction between OmpRmolecules became detectable within 5 min after the bacterial cul-tures were treated with either osmolarity shock or low pH (Fig. 5Aand B). Importantly, such multimerization did not occur in amutant in which Asp 55, the recipient residue for EnvZ-mediatedphosphorylation, was mutated (Fig. 5A and B). Moreover, no in-teraction was detected between wild-type OmpR and the D55Amutant (Fig. 5B), indicating that phosphorylation of both mono-mers is required for the formation of the dimer. Taken together,these results indicate that the formation of active OmpR dimer

FIG 1 Emission spectra of LuxAB and eYFP and schematics of fusion constructs for the LuxAB-based BRET assay. (A) The bioluminescence spectra illustratethe emission spectra of LuxAB and the enhanced yellow fluorescent acceptor protein. The luciferase substrate used in this case was decanal. Spectral resolutionof the system is marked on the x axis. (B) When testing proteins X and Y do not interact, eYFP cannot be excited by the emission from LuxAB, and thus no BREToccurs. (C) The binding between X= and Y= brings LuxAB and eYFP to close proximity, which allows the transfer of the energy to the receptor fluorophore andthe emission of fluorescence signals.

Real-Time Dynamic Protein Interactions in Bacteria

May/June 2014 Volume 5 Issue 3 e01050-14 ® mbio.asm.org 3

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

depends on the proper signaling cascade that leads to its post-translational modification.

We next examined the kinetics of OmpR multimerization inresponse to changes in environmental cues by neutralizing theacidic medium at different time points. A culture in which OmpRmultimerization had been established by low pH was split intoseveral subcultures, and NaOH was added to one of the subcul-tures at different times. When NaOH was added 5 min after acid-mediated induction, a time point when the multimerization hadjust started to occur, little dimerization was detected (Fig. 5C),which was consistent with the fact that activation of OmpR did notoccur in a neutral-pH medium. After the dimer has formed (10 or15 min after low-pH induction), neutralization of the mediumbrought down the strength of BRET signals within 10 to 15 min(Fig. 5C). Because the half-life of P-OmpR is 1 to 2 h (43) and theupstream regulator EnvZ can function as a kinase or phosphataseupon sensing distinct environmental cues (44), the reduction inBRET signals likely was attributed to the disassembly of estab-lished dimers under unfavorable conditions.

To correlate the kinetics of OmpR multimerization with thetranscription of genes under its control, we measured the expres-sion of a bioluminescence reporter fused to PompC, an OmpR-regulated promoter (33). Bioluminescence did not become de-

tectable 20 min after the osmolarity of the medium was disrupted(Fig. 5D). The time required for transcription, translation, andother events necessary for the production of the measurable prod-uct (light) apparently accounted for the longer time needed todetect gene expression. These results further validate the useful-ness of the LuxAB-based BRET system in the study of dynamicprotein-protein interactions.

DISCUSSION

Many elegantly designed genetic methods, such as two-hybrid sys-tems (1), PCA (8), and bimolecular fluorescence complementa-tion (BiFC) (45), have been developed to measure protein-proteininteractions. However, in prokaryotes, most, if not all of thesemethods only offer the detection of steady-state interactions (46,47). Here we described the development of a bioluminescence-based BRET system that for the first time allows real-time mea-surement of protein-protein interactions in live bacterial cells. Inthis system, the excitation light source is the bacterial LuxAB, andthe receptor fluorophore is eYFP. This strategy has several advan-tages. First, the endogenous excitation light circumvents the prob-lems associated with the need for an exogenous light source, whichoften requires a sophisticated instrument. Second, because theLuxAB substrate decanal is membrane permeable, the assay can be

FIG 2 BRET in E. coli cells expressing LuxAB and different receptor fluorophores. (A) Diagrams of the expression cassettes of Plac-LuxAB-eYFP. Plac, the lacpromoter region; Ter, transcription terminator region. The nucleotide sequence of the 15-amino-acid fusion linker between luxAB and eYFP is shown below thePlac-LuxAB-eYFP diagram. (B to D) Luminescence or fluorescence emission spectra of strains expressing LuxAB, mOrange, or eYFP or the fusion proteinsLuxAB-GFP (B), LuxAB-mOrange (C), and LuxAB-eYFP (D), respectively. Luminescence reactions were initiated by the addition of decanal to 1% in thecultures. All spectra were normalized by using the signals of suspensions of overnight cultures.

Cui et al.

4 ® mbio.asm.org May/June 2014 Volume 5 Issue 3 e01050-14

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

performed without lysing the bacterial cells. This feature makesexperiments aiming at addressing potentially reversible interac-tions possible. Third, with further modifications, such as havingthe protein fusions be expressed from the bacterial chromosomeby the cognate promoter, this system will allow the detection of aprotein interaction close to the natural condition.

The analyses of interactions of two protein pairs with distinctfeatures indicate that the BRET signals were generated by specificbinding of the testing proteins. First, the binding between FRB andFKBP12 occurred only when sirolimus, the molecule that directlylinks these two proteins, was added to the bacterial cultures. Sec-ond, no multimerization of OmpR was detected in the absence ofthe environmental cues important for its activation. Importantly,the use of FK506 to reverse sirolimus-mediated binding betweenFRB and FKBP allowed us to obtain an IC50 of 0.37 nM for thecompetitive inhibitor, which is comparable to values obtainedfrom other methods (30), further indicating the reliability of thismethod.

Our study of the multimerization of OmpR also revealed somenovel features associated with this important regulatory protein inresponse to environmental stimuli. First, in our system, multim-erization of OmpR became apparent within 5 min of changing the

environment to favor its activation. However, the expression ofthe target gene as measured by a bioluminescence reporter did notbecome detectable until 20 to 30 min after induction (Fig. 5D).Such delay may even be longer in the response of the cell to theenvironmental challenge because the proteins directly responsiblefor dealing with the stress will need to target to the right cellularlocation to fulfill their functions. Second, apo-OmpR is believedto mainly exist as a monomer (34, 35, 48): our results indicate thatphosphorylated OmpR cannot form multimers with apo-OmpR.Phosphorylation may allow OmpR to assume a conformation dis-tinctly different from that of apo-OmpR, which prevents the for-mation of heterodimers. Our results also suggested that changes inenvironmental cues could potentially lead to disassembly of themultimer. However, the mechanism of such disassembly remainsto be determined; it is possible that dephosphorylation ofp-OmpR by the phosphatase activity of EnvZ (44), the upstreamregulator of EnvZ, is attributed to the disassembly of the dimer.Alternatively, the reduction of the BRET signal upon the removalof the stimulus may be caused at least in part by the decay ofexisting proteins and the inability of the newly synthesized pro-teins to form multimers.

In addition to providing the proof-of-principle evidence for

FIG 3 Interactions between FlgM and FliA measured by the LuxAB/eYFP BRET system. (A) Schematics of constructs expressing relevant LuxAB or eYFPfusions. Proteins were fused both N terminal and C terminal of LuxAB and eYFP, respectively. With the exception of pLac::FliA-LuxAB, all fusions wereexpressed. (B and C) Luminescence spectra of strains harboring pLac::luxAB-fliA/pTac::eYFP-flgM and pLac::luxAB-fliA/pTac::eYFP-flgM. A strain harboringpLac::luxAB-FliA and pTac::eYFP was used as the negative control. (D) Ratios of 527 nm to 490 nm (Y:B ratios) from strains expressing indicated fusions. In allcases, luminescence production was initiated by the addition of decanal to 1% in the cultures.

Real-Time Dynamic Protein Interactions in Bacteria

May/June 2014 Volume 5 Issue 3 e01050-14 ® mbio.asm.org 5

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

the method, the analysis of binding between FliA and FlgM re-vealed that protein fusions of different orientations could affectthe sensitivity and potentially even the usefulness of the method.Reasons such as the conformation, abundance, and stability of thefusion proteins can account for such differences. This phenome-non should be of note in the use of this method to analyze inter-actions of novel protein pairs.

As indicated by the abolishment of the sirolimus-inducedFRB-FKBP interaction by FK506, this method can be adapted toestablish screenings aiming at identifying molecules capable ofinhibiting the binding of proteins involved in important biologi-cal processes. Clearly, modifications such as engineering the bac-terium of interest to produce the substrate for LuxAB may providemore accurate measurements for some experimental systems.Similarly, construction of bacterial strains expressing the relevantprotein fusions from the chromosome by their cognate promoterswill allow the study of the proteins of interest under conditionsmore closely resembling the natural conditions. Nevertheless, thedevelopment of a novel BRET system that allows real-time mea-surement of protein-protein interactions in live bacterial cells hasadded another useful tool to address important questions in thefield.

MATERIALS AND METHODSBacterial strains, plasmids, and growth conditions. The bacterial strainsand plasmids used in this study are listed in Table S1 in the supplementalmaterial. Strains were grown as described previously (39). Briefly, Esche-richia coli strains were cultured in Luria-Bertani (LB) medium containingappropriate antibiotics aerobically on a rotary shaker (180 rpm) or onagar plates at 37°C. Yersinia pseudotuberculosis strains were cultured inYersinia Luria-Bertani (YLB) broth at 26°C with appropriate antibioticswhen necessary. Antibiotics were added at the following concentrations:ampicillin, 100 �g/ml; kanamycin, 50 �g/ml; tetracycline, 10 �g/ml;chloramphenicol, 30 �g/ml.

DNA manipulation. General DNA manipulations and transforma-tions were performed using standard genetic and molecular techniques(49). Restriction enzyme digestion, ligation, and plasmid purificationwere done in accordance with the manufacturer’s instructions (TaKaRa,Dalian, China). PCR was performed with TaKaRa Taq or Pyrobest DNApolymerase (TaKaRa, Dalian, China). Plasmid DNA was isolated with theTiangen Biotech plasmid DNA Miniprep spin columns, and chromo-somal DNA was purified with the Tiangen Biotech bacterial genomicDNA purification kit (Tiangen Biotech, Beijing, China). DNA fragmentswere purified from agarose gels by using the EasyPure quick gel extractionkit (Transgen, Beijing, China). DNA sequencing and primer synthesiswere carried out by Sangon Biotech (Shanghai, China).

FIG 4 Interactions between FRB and FKBP12 measured by the LuxAB-based BRET. (A) Molecule model of sirolimus-induced FRB-FKBP12 associationdetected by BRET. (B) Sirolimus-induced FRB-FKBP12 association and its dissociation triggered by FK506. A 100 nM concentration of sirolimus (noted on thefigure by the common name “rapamycin”), 10 �M FK506, or the same volume of DMSO (negative control) was added together with 1% decanal to bacterial cells,and the BRET signals were recorded for 60 min. (C) Competitive inhibition coefficient of FK506 against sirolimus (IC50). Samples expressing FRB-eYFP andLuxAB-FKBP containing 100 nM sirolimus were treated with the indicated concentrations of FK506, and the BRET signals were measured to calculate the Y:Bratios. Similar results were obtained in at least three independent experiments done in triplicate, and results from one representative experiment are shown.

Cui et al.

6 ® mbio.asm.org May/June 2014 Volume 5 Issue 3 e01050-14

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

Plasmid construction. The plasmids and primers used in this studyare listed in Tables S1 and S2, respectively, in the supplemental material.To construct LuxAB-eYFP, LuxAB-GFP, and LuxAB-mOrange fusion ex-pression cassettes, the coding region of luxAB was amplified with primersLuxA-XbaI-F and LuxB-L15-BglII-R and cloned into XbaI/BamHI-digested pBluescript II KS� to give pBluescript-luxAB. A 15-amino-acidlinker (GGGSG)3 was fused in frame to the luxB open reading frame(ORF) by deleting the stop codon TAA and adding codons for the linkerafter the luxB coding region in primer LuxB-L15-BglII-R. The lac pro-moter was amplified with primers pLac-SacI-F and pLac-XbaI-R, digestedwith SacI/XbaI, and then ligated into pBluescript-luxAB, resulting in plas-mid pBluescript-pLac::luxAB (see Fig. S3 in the supplemental material).The eYFP coding region was amplified with primers eYFP-BamHI-F andeYFP-SalI-R, digested with BamHI/SalI, and inserted into similarly di-gested pBluescript-pLac::luxAB to produce pBluescript-pLac::luxAB-eYFP, in which the eYFP ORF is fused in frame with the luxB gene, spacedby the (GGGSG)3 linker. The plasmids pBluescript-pLac::luxAB-mOrangeand pBluescript-pLac::luxAB-gfp were constructed similarly, with theORFs of GFP and mOrange amplified with primer pairs GFP-BamHI-F/GFP-SalI-R and mOrange-BamHI-F/mOrange-SalI-R, by using the plas-mids pEGFP-N1 and pmOrange, respectively, as the template.

The tac promoter was amplified with primers pTac-SacI-F and pTac-

XbaI-R, digested with SacI/XbaI, and then ligated into pBluescript II KS�,resulting in pBluescript-pTac. An eYFP coding sequence with or withoutthe stop codon was PCR amplified using primers eYFP-EcoRI-F/eYFP-SalI-R and eYFP-SpeI-F/eYFP-SalI-R=, respectively, and then cloned intothe corresponding sites of plasmid pBluescript-pTac, to give pBluescript-pTac::�-eYFP and pBluescript-pTac::eYFP-�.

The lac promoter was amplified with primer pair pLac-SphI-F/pLac-SpeI-R, digested with SphI/SpeI, and ligated into similarly digestedpKT100 (see Fig. S3 in the supplemental material), to produce pKT100-pLac. The luxAB coding region was amplified with primer pair LuxA-SpeI-F/LuxB-L15-BamHI-R and LuxA-L15-XbaI-F/LuxB-KpnI-R andcloned into the SpeI/BamHI and XbaI/KpnI sites of the vector pKT100-pLac, respectively, resulting in plasmids pKT100-pLac::luxAB-� andpKT100-pLac::�-luxAB. For the plasmid pKT100-pLac::luxAB-�, a 15-amino-acid linker (GGGSG)3 was fused in frame to the luxB ORF byremoving the stop codon TAA. Finally, for plasmid pKT100-pLac::�-luxAB, the 15-amino-acid linker (GGGSG)3 was fused in-frame to theORF of luxA directly upstream of the start codon.

The coding regions of flgM and fliA were amplified from the genome ofY. pseudotuberculosis strain YpIII with primer pairs FliA-BamHI-F/FliA-SalI-R, FliA-BamHI-F/FliA-SalI-R-TAA, FlgM-BamHI-F/FlgM-SalI-R,and FlgM-BamHI-F/FlgM-SalI-R-TAA, respectively. PCR products of

FIG 5 Kinetics of OmpR dimerization under different conditions. (A) High-salt-concentration-induced dimerization of OmpR requires the phosphorylationsite D55. Salt was added to 1.5 M to cultures expressing the indicated testing protein fusions, and the BRET signals were recorded for an experimental durationof 60 min (B) Wild-type OmpR does not interact with its D55A mutant. The pH of cultures expressing the indicated protein fusion pairs was lowered to 4, andthe BRET signals were recorded for 60 min. Note that no BRET occurred in cultures expressing OmpR and its D55A mutant fused to the two BRET components,respectively. (C) Abolishment of the inducing condition led to reduction of dimerization. NaOH was added to cultures that had been induced by low pH for 5,10, or 15 min, and the BRET signals were measured at the indicated time intervals for 60 min. A culture that received an equal volume of fresh medium was usedas a control. (D) Kinetics of the PompC promoter activity in response to high salt. Salt was added to cells containing the PompC::LuxAB reporter, and the expressionof bioluminescence was monitored for 90 min. Note that it takes about 30 min to detect significant LuxAB activity.

Real-Time Dynamic Protein Interactions in Bacteria

May/June 2014 Volume 5 Issue 3 e01050-14 ® mbio.asm.org 7

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

flgM and fliA were digested with BamHI/SalI and inserted intopBluescript-pTac::�-eYFP, pBluescript-pTac::eYFP-�, pKT100-pLac::�-luxAB, and pKT100-pLac::luxAB-�, respectively, to produce plasmidspLac::luxAB-fliA, pLac::fliA-luxAB, pTac::eYFP-flgM, and pTac::flgM-eYFP. In the resulting plasmids, the fusion proteins FlgM and FliA werefused in frame to the upstream and downstream sequences of eYFP andLuxAB and were all transcribed from the lac promoter.

The cDNA encoding the mouse FK506 binding protein (FKBP) andthe 11-kDa FKBP12-rapamycin binding domain within FRAP (FRB) werecodon optimized according to Neurospora codon usage data from theKazusa DNA Research Institute (http://www.kazusa.or.jp/codon) andsynthesized by GenScript Co., Ltd. (Nanjing, China). For each residue, thecodon with the highest usage frequency in the bacterium was chosen. Theentire coding regions of FRB and FKBP were cloned into pBluescript-pTac::�-eYFP and pKT100-pLac::luxAB-�, respectively, to produce theplasmids pBluescript-pTac::FRB-eYFP and pKT100-pLac::luxAB-FKBP.

The ompR gene fragment was amplified with primers OmpR-F-BamHI and OmpR-R-SalI, using E. coli genomic DNA as the template.The PCR product was digested with BamHI and SalI and inserted intosimilarly digested pBluescript-pTac::�-eYFP and pKT100-pLac::luxAB-�, to give pBluescript-pTac::OmpR-eYFP and pKT100-pLac::luxAB-OmpR. Site-directed mutagenesis was performed according to theQuikChange protocol (Stratagene, La Jolla, CA) as described by the sup-plier with primers OmpR*-F and OmpR*-R, to construct pBluescript-pTac::OmpR(D55A)-eYFP and pKT100-pLac::luxAB-OmpR(D55A).

To construct the reporter plasmid pBluescript-POmpC::luxAB, theompC promoter was amplified with primer pair pOmpC-SacI-F/pOmpC-XbaI-R from E. coli genomic DNA. The PCR product was cloned intoSacI/XbaI-digested pBluescript-luxAB, which harbors an ampicillin resis-tance gene and a promoterless luxAB reporter gene. The integrity of theinsert in all constructs was confirmed by DNA sequencing.

Bioluminescence and fluorescence assay. E. coli strains were grownovernight in LB medium containing appropriate antibiotics at 37°C. Toeliminate the background fluorescence in LB medium, the E. coli cultureswere washed and resuspended in M9 minimal salts medium before assay.To detect eYFP fluorescence, 3 ml of E. coli cells resuspended in M9 me-dium (optical density at 600 nm [OD600] of 1.0) was transferred to a 4-mlcuvette. Fluorescence excitation and emission measurements were per-formed on an F-7000 fluorescence spectrophotometer with a 150-W xe-non lamp (Hitachi). eYFP fluorescence was excited at 510 nm, and itsemission was scanned from 515 nm to 600 nm. For the luminescenceassay, the lamp was shut off, and the cuvette was gently bubbled with air tosupply sufficient oxygen for the luciferase reaction after the addition of1% decanal (Aladdin). Bioluminescence emission spectra were collectedbetween the wavelengths 400 nm and 600 nm as described previously (14).

Detection of dynamic protein interactions by the LuxAB-basedBRET system. The plasmids pBluescript-pTac::OmpR-eYFP and pKT100-pLac::luxAB-OmpR were cotransformed into E. coli strain DH5�, and thetransformant was used to examine OmpR dimerization under differentconditions by growing the cells at 37°C in LB medium to an OD600 of 1.0.To detect the induced OmpR dimerization by high-osmolarity stress oracid stress, 3 ml E. coli cells resuspended in M9 medium (OD600 of 1.0) wastransferred to a 4-ml cuvette, and protein dimerization was initiated bythe addition of hydrochloric acid to pH 4.0 or NaCl to 1.5 M in thepresence of 1% decanal. Similarly, the sirolimus-induced FKBP-FRB in-teraction was detected by using E. coli DH5� cotransformed withpBluescript-pTac::FRB-eYFP and pKT100-pLac::luxAB-FKBP, and the in-teraction was initiated by adding 100 nM sirolimus (dissolved in DMSO)and 1% decanal. To obtain a competitive inhibitor coefficient of FK506 tosirolimus, eight groups of E. coli cells expressing FRB-eYFP and LuxAB-FKBP were treated with 100 nM sirolimus and incubated with differentconcentrations of FK506 for bioluminescence analysis. The BRET sensorfragments do not affect the dissociation constant for the FRB-sirolimus-FKBP complex and the inhibition constant for FK506 (41, 50). The BRET

assay was measured for indicated time durations (0 to 60 min) in live cellsafter being mixed and bubbled during the experiments (41).

BRET imaging of E. coli cultures. Cells of 1.5-ml overnight E. colicultures grown at 37°C in LB medium with appropriate antibiotics werewashed twice with M9 medium and resuspended in 1.5 ml M9 mediumcontaining 1% decanal. For imaging of the induced FKBP-FRB BRETsignal, 100 nM sirolimus was added into the suspension. Two hundredmicroliters of each suspension was immediately transferred to wells of adark microwell plate (Costar). The small volume of the sample allowedgood gas exchange, and so bubbling was not required. Six duplicates weremade for each strain, which were then divided into two groups. One groupwas visualized through a 480-nm (�5-nm) interference filter and theother through a 530-nm (�5-nm) filter. Images were captured with acooled charge-coupled device (CCD) camera of the ChemiDocXRS� gelimaging system (Bio-Rad).

LuxAB reporter fusion and bioluminescence assay. To determine thedynamics of OmpC transcription induced by high-osmolarity stress oracid stress, the reporter plasmid pBluescript-POmpC::luxAB was trans-formed into E. coli DH5� as a reporter strain. One hundred microliters ofa log-phase cell culture of E. coli reporter strain was transferred into a1.5-ml microcentrifuge tube. Luminescence reactions were initiated bythe addition of NaCl to a final concentration of 1.5 M or hydrochloric acid(stock concentration, 0.1 mol/liter) to a pH of 4.0 in the presence of 1%decanal. Luminescence activity was detected on a GloMax 20/20 lumi-nometer (Promega) as described previously (51). All experiments wereperformed independently at least three times, and the data shown arefrom one representative experiment done in triplicate.

Western blot analysis. E. coli cells expressing the BRET vectors, eYFPfusion proteins, were harvested and lysed with passive lysis buffer (Pro-mega) for 30 min on ice. Equal amounts of total protein were resolved by12% SDS-polyacrylamide gel and subsequently transferred onto polyvi-nylidene difluoride (PVDF) membranes. After blocking with 4% milk for2 h at room temperature, membranes were probed with anti-YFP mousemonoclonal antibody (1:2,000; Millipore, Billerica, MA) diluted in Tris-buffered saline–Tween (TBST) at 4°C overnight. The membrane waswashed five times with TBST and subsequently incubated with horserad-ish peroxidase (HRP)-conjugated secondary antibodies (ShanghaiGenomics) in TBST buffer for 1 h. The proteins were visualized by usingthe ECL Plus kit (GE Healthcare, United Kingdom) following the manu-facturer’s instructions.

SUPPLEMENTAL MATERIALSupplemental material for this article may be found at http://mbio.asm.org/lookup/suppl/doi:10.1128/mBio.01050-14/-/DCSupplemental.

Figure S1, PDF file, 0.1 MB.Figure S2, PDF file, 0.2 MB.Figure S3, PDF file, 0.3 MB.Table S1, PDF file, 0.2 MB.Table S2, PDF file, 0.1 MB.

ACKNOWLEDGMENTS

This work was supported by National Natural Science Foundation ofChina grant 31170121 to X.S. and grants 31170100 and 31370150 to Y.W.,Specialized Research Fund for the Doctoral Program of Higher Educationof China grant 20110204120023 to X.S., and NIH-NIAID grantK02AI085403 to Z.-Q.L.

REFERENCES1. Fields S, Song O. 1989. A novel genetic system to detect protein-protein

interactions. Nature 340:245–246. http://dx.doi.org/10.1038/340245a0.2. Dove SL, Hochschild A. 2004. A bacterial two-hybrid system based on

transcription activation. Methods Mol. Biol. 261:231–246. http://dx.doi.org/10.1385/1-59259-762-9:231.

3. Stynen B, Tournu H, Tavernier J, Van Dijck P. 2012. Diversity in geneticin vivo methods for protein-protein interaction studies: from the yeast

Cui et al.

8 ® mbio.asm.org May/June 2014 Volume 5 Issue 3 e01050-14

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

two-hybrid system to the mammalian split-luciferase system. Microbiol.Mol. Biol. Rev. 76:331–382. http://dx.doi.org/10.1128/MMBR.05021-11.

4. Wang Y, Cui T, Zhang C, Yang M, Huang Y, Li W, Zhang L, Gao C, HeY, Li Y, Huang F, Zeng J, Huang C, Yang Q, Tian Y, Zhao C, Chen H,Zhang H, He ZG. 2010. Global protein-protein interaction network inthe human pathogen Mycobacterium tuberculosis H37Rv. J. Proteome Res.9:6665– 6677. http://dx.doi.org/10.1021/pr100808n.

5. Morell M, Czihal P, Hoffmann R, Otvos L, Avilés FX, Ventura S. 2008.Monitoring the interference of protein-protein interactions in vivo bybimolecular fluorescence complementation: the DnaK case. Proteomics8:3433–3442. http://dx.doi.org/10.1002/pmic.200700739.

6. Remy I, Galarneau A, Michnick SW. 2002. Detection and visualization ofprotein interactions with protein fragment complementation assays.Methods Mol. Biol. 185:447– 459. http://dx.doi.org/10.1385/1-59259-241-4:447.

7. Karimova G, Pidoux J, Ullmann A, Ladant D. 1998. A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proc.Natl. Acad. Sci. U. S. A. 95:5752–5756. http://dx.doi.org/10.1073/pnas.95.10.5752.

8. Michnick SW. 2001. Exploring protein interactions by interaction-induced folding of proteins from complementary peptide fragments.Curr. Opin. Struct. Biol. 11:472– 477. http://dx.doi.org/10.1016/S0959-440X(00)00235-9.

9. James JR, Oliveira MI, Carmo AM, Iaboni A, Davis SJ. 2006. A rigorousexperimental framework for detecting protein oligomerization using bio-luminescence resonance energy transfer. Nat. Methods 3:1001–1006.http://dx.doi.org/10.1038/nmeth978.

10. Demarco IA, Periasamy A, Booker CF, Day RN. 2006. Monitoringdynamic protein interactions with photoquenching FRET. Nat. Methods3:519 –524. http://dx.doi.org/10.1038/nmeth889.

11. Hatzios SK, Ringgaard S, Davis BM, Waldor MK. 2012. Studies ofdynamic protein-protein interactions in bacteria using Renilla luciferasecomplementation are undermined by nonspecific enzyme inhibition.PLoS One 7:e43175. http://dx.doi.org/10.1371/journal.pone.0043175.

12. Magliery TJ, Wilson CG, Pan W, Mishler D, Ghosh I, Hamilton AD,Regan L. 2005. Detecting protein-protein interactions with a green fluo-rescent protein fragment reassembly trap: scope and mechanism. J. Am.Chem. Soc. 127:146 –157. http://dx.doi.org/10.1021/ja046699g.

13. Tramier M, Zahid M, Mevel JC, Masse MJ, Coppey-Moisan M. 2006.Sensitivity of CFP/YFP and GFP/mCherry pairs to donor photobleachingon FRET determination by fluorescence lifetime imaging microscopy inliving cells. Microsc. Res. Tech. 69:933–939. http://dx.doi.org/10.1002/jemt.20370.

14. Xu Y, Piston DW, Johnson CH. 1999. A bioluminescence resonanceenergy transfer (BRET) system: application to interacting circadian clockproteins. Proc. Natl. Acad. Sci. U. S. A. 96:151–156. http://dx.doi.org/10.1073/pnas.96.1.151.

15. Wu P, Brand L. 1994. Resonance energy transfer: methods and applica-tions. Anal. Biochem. 218:1–13. http://dx.doi.org/10.1006/abio.1994.1134.

16. Ke D, Tu SC. 2011. Activities, kinetics and emission spectra of bacterialluciferase-fluorescent protein fusion enzymes. Photochem. Photobiol. 87:1346 –1353. http://dx.doi.org/10.1111/j.1751-1097.2011.01001.x.

17. Breton B, Sauvageau E, Zhou J, Bonin H, Le Gouill C, Bouvier M. 2010.Multiplexing of multicolor bioluminescence resonance energy transfer.Biophys. J. 99:4037– 4046. http://dx.doi.org/10.1016/j.bpj.2010.10.025.

18. Xu X, Soutto M, Xie Q, Servick S, Subramanian C, von Arnim AG,Johnson CH. 2007. Imaging protein interactions with bioluminescenceresonance energy transfer (BRET) in plant and mammalian cells and tis-sues. Proc. Natl. Acad. Sci. U. S. A. 104:10264 –10269. http://dx.doi.org/10.1073/pnas.0701987104.

19. Subramanian C, Woo J, Cai X, Xu X, Servick S, Johnson CH, Neben-führ A, von Arnim AG. 2006. A suite of tools and application notes for invivo protein interaction assays using bioluminescence resonance energytransfer (BRET). Plant J. 48:138 –152. http://dx.doi.org/10.1111/j.1365-313X.2006.02851.x.

20. Pfleger KD, Eidne KA. 2006. Illuminating insights into protein-proteininteractions using bioluminescence resonance energy transfer (BRET).Nat. Methods 3:165–174. http://dx.doi.org/10.1038/nmeth841.

21. Perroy J, Pontier S, Charest PG, Aubry M, Bouvier M. 2004. Real-timemonitoring of ubiquitination in living cells by BRET. Nat. Methods1:203–208. http://dx.doi.org/10.1038/nmeth722.

22. Yamakawa Y, Ueda H, Kitayama A, Nagamune T. 2002. Rapid homo-

geneous immunoassay of peptides based on bioluminescence resonanceenergy transfer from firefly luciferase. J. Biosci. Bioeng. 93:537–542.http://dx.doi.org/10.1016/S1389-1723(02)80234-1.

23. Arai R, Nakagawa H, Kitayama A, Ueda H, Nagamune T. 2002. Detec-tion of protein-protein interaction by bioluminescence resonance energytransfer from firefly luciferase to red fluorescent protein. J. Biosci. Bioeng.94:362–364. http://dx.doi.org/10.1263/jbb.94.362.

24. Li F, Yu J, Zhang Z, Cui Z, Wang D, Wei H, Zhang XE. 2012. Bufferenhanced bioluminescence resonance energy transfer sensor based onGaussia luciferase for in vitro detection of protease. Anal. Chim. Acta724:104 –110. http://dx.doi.org/10.1016/j.aca.2012.02.047.

25. Shimizu TS, Delalez N, Pichler K, Berg HC. 2006. Monitoring bacterialchemotaxis by using bioluminescence resonance energy transfer: absenceof feedback from the flagellar motors. Proc. Natl. Acad. Sci. U. S. A. 103:2093–2097. http://dx.doi.org/10.1073/pnas.0510958103.

26. Winson MK, Swift S, Hill PJ, Sims CM, Griesmayr G, Bycroft BW,Williams P, Stewart GS. 1998. Engineering the luxCDABE genes fromPhotorhabdus luminescens to provide a bioluminescent reporter for con-stitutive and promoter probe plasmids and mini-Tn5 constructs. FEMSMicrobiol. Lett. 163:193–202. http://dx.doi.org/10.1111/j.1574-6968.1998.tb13045.x.

27. Srinivas S, Watanabe T, Lin CS, William CM, Tanabe Y, Jessell TM,Costantini F. 2001. Cre reporter strains produced by targeted insertion ofEYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1:4. http://dx.doi.org/10.1186/1471-213X-1-4.

28. Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE,Tsien RY. 2004. Improved monomeric red, orange and yellow fluorescentproteins derived from Discosoma sp. red fluorescent protein. Nat. Biotech-nol. 22:1567–1572. http://dx.doi.org/10.1038/nbt1037.

29. Banaszynski LA, Liu CW, Wandless TJ. 2005. Characterization of theFKBP · rapamycin · FRB ternary complex. J. Am. Chem. Soc. 127:4715– 4721. http://dx.doi.org/10.1021/ja043277y.

30. Chen J, Zheng XF, Brown EJ, Schreiber SL. 1995. Identification of an11-kDa FKBP12-rapamycin-binding domain within the 289-kDaFKBP12-rapamycin-associated protein and characterization of a criticalserine residue. Proc. Natl. Acad. Sci. U. S. A. 92:4947– 4951. http://dx.doi.org/10.1073/pnas.92.11.4947.

31. Liu J, Farmer JD, Jr, Lane WS, Friedman J, Weissman I, Schreiber SL.1991. Calcineurin is a common target of cyclophilin-cyclosporin A andFKBP-FK506 complexes. Cell 66:807– 815. http://dx.doi.org/10.1016/0092-8674(91)90124-H.

32. Zhang W, Wang Y, Song Y, Wang T, Xu S, Peng Z, Lin X, Zhang L,Shen X. 2013. A type VI secretion system regulated by OmpR in Yersiniapseudotuberculosis functions to maintain intracellular pH homeostasis.Environ. Microbiol. 15:557–569. http://dx.doi.org/10.1111/1462-2920.12005.

33. Forst S, Delgado J, Inouye M. 1989. Phosphorylation of OmpR by theosmosensor EnvZ modulates expression of the ompF and ompC genes inEscherichia coli. Proc. Natl. Acad. Sci. U. S. A. 86:6052– 6056. http://dx.doi.org/10.1073/pnas.86.16.6052.

34. Barbieri CM, Wu T, Stock AM. 2013. Comprehensive analysis of OmpRphosphorylation, dimerization, and DNA binding supports a canonicalmodel for activation. J. Mol. Biol. 425:1612–1626. http://dx.doi.org/10.1016/j.jmb.2013.02.003.

35. Gao H, Zhang Y, Han Y, Yang L, Liu X, Guo Z, Tan Y, Huang X, ZhouD, Yang R. 2011. Phenotypic and transcriptional analysis of the osmoticregulator OmpR in Yersinia pestis. BMC Microbiol. 11:39. http://dx.doi.org/10.1186/1471-2180-11-39.

36. Miller WG, Lindow SE. 1997. An improved GFP cloning cassette de-signed for prokaryotic transcriptional fusions. Gene 191:149 –153. http://dx.doi.org/10.1016/S0378-1119(97)00051-6.

37. Chilcott GS, Hughes KT. 2000. Coupling of flagellar gene expression toflagellar assembly in Salmonella enterica serovar Typhimurium and Esch-erichia coli. Microbiol. Mol. Biol. Rev. 64:694 –708. http://dx.doi.org/10.1128/MMBR.64.4.694-708.2000.

38. Aldridge P, Hughes KT. 2002. Regulation of flagellar assembly. Curr.Opin. Microbiol. 5:160 –165. http://dx.doi.org/10.1016/S1369-5274(02)00302-8.

39. Xu S, Peng Z, Cui B, Wang T, Song Y, Zhang L, Wei G, Wang Y, ShenX. 2014. FliS modulates FlgM activity by acting as a non-canonical chap-erone to control late flagellar gene expression, motility and biofilm forma-tion in Yersinia pseudotuberculosis. Environ. Microbiol. 16:1090 –1104.http://dx.doi.org/10.1111/1462-2920.12222.

Real-Time Dynamic Protein Interactions in Bacteria

May/June 2014 Volume 5 Issue 3 e01050-14 ® mbio.asm.org 9

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

40. Choi J, Chen J, Schreiber SL, Clardy J. 1996. Structure of the FKBP12-rapamycin complex interacting with the binding domain of human FRAP.Science 273:239 –242. http://dx.doi.org/10.1126/science.273.5272.239.

41. Remy I, Michnick SW. 2006. A highly sensitive protein-protein interac-tion assay based on Gaussia luciferase. Nat. Methods 3:977–979. http://dx.doi.org/10.1038/nmeth979.

42. Kenney LJ, Bauer MD, Silhavy TJ. 1995. Phosphorylation-dependentconformational changes in OmpR, an osmoregulatory DNA-binding pro-tein of Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 92:8866 – 8870. http://dx.doi.org/10.1073/pnas.92.19.8866.

43. Igo MM, Ninfa AJ, Stock JB, Silhavy TJ. 1989. Phosphorylation anddephosphorylation of a bacterial transcriptional activator by a transmem-brane receptor. Genes Dev. 3:1725–1734. http://dx.doi.org/10.1101/gad.3.11.1725.

44. Dutta R, Inouye M. 1996. Reverse phosphotransfer from OmpR to EnvZin a kinase�/phosphatase� mutant of EnvZ (EnvZ.N347D), a bifunc-tional signal transducer of Escherichia coli. J. Biol. Chem. 271:1424 –1429.http://dx.doi.org/10.1074/jbc.271.3.1424.

45. Hu CD, Chinenov Y, Kerppola TK. 2002. Visualization of interactionsamong bZIP and Rel family proteins in living cells using bimolecular flu-orescence complementation. Mol. Cell 9:789 –798. http://dx.doi.org/10.1016/S1097-2765(02)00496-3.

46. Kostecki JS, Li H, Turner RJ, DeLisa MP. 2010. Visualizing interactions

along the Escherichia coli twin-arginine translocation pathway using pro-tein fragment complementation. PLoS One 5:e9225. http://dx.doi.org/10.1371/journal.pone.0009225.

47. Sato T, Hanada M, Bodrug S, Irie S, Iwama N, Boise LH, ThompsonCB, Golemis E, Fong L, Wang HG, Reed JC. 1994. Interactions amongmembers of the Bcl-2 protein family analyzed with a yeast two-hybridsystem. Proc. Natl. Acad. Sci. U. S. A. 91:9238 –9242. http://dx.doi.org/10.1073/pnas.91.20.9238.

48. Mattison K, Kenney LJ. 2002. Phosphorylation alters the interaction ofthe response regulator OmpR with its sensor kinase EnvZ. J. Biol. Chem.277:11143–11148. http://dx.doi.org/10.1074/jbc.M111128200.

49. Sambrook J, Russell D. 2001. Molecular cloning: a laboratory manual,3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

50. Dragulescu-Andrasi A, Chan CT, De A, Massoud TF, Gambhir SS.2011. Bioluminescence resonance energy transfer (BRET) imaging ofprotein-protein interactions within deep tissues of living subjects. Proc.Natl. Acad. Sci. U. S. A. 108:12060 –12065. http://dx.doi.org/10.1073/pnas.1100923108.

51. Close DM, Patterson SS, Ripp S, Baek SJ, Sanseverino J, Sayler GS.2010. Autonomous bioluminescent expression of the bacterial luciferasegene cassette (lux) in a mammalian cell line. PLoS One 5:e12441. http://dx.doi.org/10.1371/journal.pone.0012441.

Cui et al.

10 ® mbio.asm.org May/June 2014 Volume 5 Issue 3 e01050-14

on Septem

ber 15, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from