A tryptophan-based assay method to search regulatory compounds for TCTP

Transcriptionally controlled tumor protein (TCTP) is a highly conserved protein performing a large number of cellular functions by binding with various partner proteins. The importance of its roles in many diseases requires an assay method to find regulatory compounds. However, the molecular characteristics of TCTP made it difficult to search for chemicals interacting with it. In this study, a tryptophan-based assay method was designed and Y151W mutant TCTP was constructed to search binding chemicals. Since there is no tryptophan in the native sequence of TCTP, the incorporation of tryptophan in the Y151W mutant was very effective to establish the method. A flavonoid library was employed to the assay with the method. With the native and Y151W mutant TCTPs, three flavonoids such as morin, myricetin and isobavachalcone have been found to interact with TCTP. Combined with native gel electrophoresis, the binding region of isobavachalcone was suggested to be the flexible loop of TCTP. This approach can be easily applicable to find binding compounds of proteins with similar molecular characteristics of TCTP.


Introduction
Transcriptionally controlled tumor protein (TCTP) is a highly conserved protein with fundamental cellular functions. It exerts numerous functions by interacting with various partner proteins in many core cell biological processes [1][2][3]. Especially it demonstrates cytoprotective action in responses to a variety of cellular stresses. In addition, TCTP diversifies its functions depending on oligomeric states that would be triggered by cellular proteolytic digestion or oxidative environment [4]. HRF (Histamine Releasing Factor) is a representative dimeric form activating mast cells and basophils by interacting with a subset of IgE molecules. Paradoxically, the essential cytoprotective function of TCTP also causes it connected to various cancers [5].
Intriguingly, anti-apoptotic activity, promotion of mitosis and anticancer drug resistance of TCTP are all related to the survival of cancer cells. Therefore, if regulatory chemicals of TCTP could be found, various cellular functions mediated by TCTP can be controlled.
To find those compounds, a convenient in-vitro method to assay molecular functions of TCTP should be present. Unfortunately, no direct in-vitro method correlated to molecular functions of TCTP has been reported. One of the reasons is that TCTP is not a catalytic enzyme.
Since high-throughput approaches to find out hit chemicals absolutely require easy in-vitro assay methods, TCTP is not applicable. Another reason is the absence of a prominent binding site on the surface of TCTP. Since large cavities on surfaces of proteins increase the binding probability of chemicals and release stabilization energy during interaction with partner molecules, calorimetric techniques can be used to monitor the interaction phenomena. However, there are only two small cavities on the surface of TCTP such as Cavity-1 comprised of H2, β7 and the loop between β8 and β9 and Cavity-2 lined up with H3 and β-strands β8, β9, β10 [6].
Their cavity volumes are around ~360 A 3 . The small cavity volume implies that common calorimetric methods may be hard to be applied due to the limitation of detecting small binding or stabilization energy by instruments. At last, spectroscopic methods are also inapplicable to TCTP due to the absence of tryptophan residues or chromophores.
In order to overcome the above obstruction and to find regulatory chemicals for TCTP, a tryptophan-based assay has been designed. Since there is no tryptophan residue in TCTP, one tryptophan residue was introduced to TCTP by mutation. Subsequently, the binding of chemicals to TCTP could be detected with a general spectrometry method. Since flavonoids have molecular characteristics of both hydrophilicity and hydrophobicity, a flavonoid library was applied to prove the usefulness of the method together with finding binding compounds quickly.

Results and Discussion
The original hypothesis of general tryptophan-based assay methods is based on the postulation that a change of fluorescence intensity would be detected if chemicals bind to proteins [7]. It is well known that the fluorescence intensity emitted from tryptophan is frequently perturbed by surrounding conditions [8]. In order to test whether the tryptophanbased assay method was feasible to search chemicals interacting with TCTP, one tryptophan residue was incorporated into TCTP. Since there is no tryptophan residue in the original human TCTP sequence, Tyr151 was mutated to Trp151. Tyr 151 was selected because the residue locates where its mutation was expected to influence minimal environmental perturbation. As expected, all the molecular behavior of native f-TCTP and Y151W f-TCTP was almost the same. ∆-TCTP and Y151W ∆-TCTP were also constructed to check if compounds interact with the flexible loop or not. The cell yields harvested for purification of Y151W f-TCTP and Y151W ∆-TCTP were 2.24 g and 3.7 g per 1000 ml of E. coli culture, respectively. The amount of finally purified proteins were 24.1 mg for the former and 10.8 mg for the latter. For storage and assay, both protein solutions were concentrated to 53.1 mg ml -1 and 43.2 mg ml -1 , respectively. For the assay, the final concentration was adjusted to 1 μM for both cases.
In order to find TCTP regulatory compounds, an in-house flavonoid library was built. Since most flavonoids show a high affinity to proteins due to the presence of hydrophobic aromatic rings and hydrophilic hydroxyl groups, their complementary chemical property endows broad affinities toward proteins. Therefore, a flavonoid library consisting of ten different scaffolds was constructed ( Figure 1). It contains three isoflavones, one isoflavane, five flavones, eight flavonols, four flavanols, five flavanones, two flavanonol, one prenylflavonoid, eight chalcones and two unclassified flavonoids (Supplementary Table 1). We applied the library to find flavonoids with affinity to two mutants, Y151W f-TCTP and Y151W ∆-TCTP. Using thirty-nine flavonoids, an inhibitory effect of each compound at 40 μM was tested. It has been reported that the flexible loop decreased the inflammatory cytokine activity of full dimeric TCTP (f-dTCTP) [9]. Therefore, using two mutants, we expected to find three different types of chemical compounds according to their binding affinity to; (1) f-TCTP (2) ∆-TCTP (3) f-TCTP and ∆-TCTP. If it is possible, this method can be used to find various regulatory compounds controlling TCTP related diseases.
The screening with flavonoids against the in-house flavonoid library was very successful.
Among thirty-nine, three flavonoids reduced the flavonoids intensity seriously. Those are morin, myricetin and isobavachalcone ( Figure 2). They reduced the fluorescence intensity of both mutants. The three flavonoids were further instigated to check whether they influence the cellular function of TCTP or not. Since TCTP is not a catalytic enzyme, the proof of regulatory functions of flavonoids was performed with TCTP constructs designed to cell experiments. In previous studies, N-terminus truncated TCTP can form dimers (∆N11-dTCTP) in cells and dimerization of TCTP is necessary for generating its cytokine-like activity [4]. It has also been reported that ∆N11-dTCTP stimulates the secretion of proinflammatory cytokines, such as IL-8 and GM-CSF in BEAS-2B cells [4,6]. Therefore, it is worthwhile to note that the overall architecture of f-dTCTP in-vitro is almost similar to ∆N11-dTCTP in-vivo. To verify the cellular function of three flavonoids, each chemical compound was analyzed for their inhibitory effect on ∆N11-dTCTP-induced cytokine release. Diosmin, not interacting with two mutant TCTPs was employed as negative control compounds. The secretion of IL-8 was significantly decreased in a dose-dependent manner when cells were treated with various concentrations of three flavonoids (Fig 3. A). Morin, myricetin and isobavachalcone reduced IL-8 secretion.
In order to independently identify the binding of three flavonoids on TCTP, native gel electrophoresis had been performed. Since TCTP presents as a monomer (f-TCTP) and dimer (f-dTCTP) depending on redox states of Cys172, DTT and tert-Butyl hydroperoxide (tBHP) were pre-treated before electrophoresis, respectively. Under the oxidizing condition, the monomeric and dimeric forms were present (Fig 4. A). Interestingly, three flavonoids, clearly In order to check whether the binding regions presumed by the above experiment includes the flexible loop of TCTP or not, native gel electrophoresis had been also performed with ∆-TCTP (Fig 4. B). Since ∆-TCTP also presents as a monomer (∆-TCTP) and dimer (∆-dTCTP) depending on redox conditions, electrophoresis was carried out under the same redox conditions as above. Under the oxidizing condition, monomer and dimer were present, too. The tryptophan-based assay method employed in this study was very successful. Specially, when combined with the native gel electrophoresis, the accuracy of tryptophan-based method could be confirmed independently. The capability of distinguishing interaction between the flexible loop and chemicals is also the merit of the combined method. Using this approach, a high-throughput method also can be applied to find TCTP regulatory compounds. We hope this method can be used to find binding chemicals against proteins which are none catalytic enzymes and have small molecular weight with small narrow cavities and no tryptophan residue.

3.2.Gene synthesis, protein expression and purification for the single mutant Y151W ∆-TCTP
In order to compare the molecular properties of Y151W f-TCTP, the Y151W mutation was also applied to the flexible loop domain (from Arg38 to Val66) deletion mutant, ∆-TCTP. This

3.3.Absorption spectroscopic studies based on the incorporated tryptophan
The fluorescence spectra of Y151W f-TCTP and Y151W ∆-TCTP with chemical compounds were investigated. The fluorescence measurements were recorded with a SpectraMax i3x Multi-mode microplate reader (Molecular Devices) at excitation and emission wavelengths of 280 nm and 310-500 nm, respectively. Both of them showed a strong fluorescence emission with a peak at 320 nm at the excitation wavelength of 280 nm, unlike native TCTP. In contrast, the chemical compounds were almost non-fluorescent under the same experiment condition.
Each 40 μM chemical compound was incubated with each mutant for 1 hour and the fluorescence intensity of each solution was measured.