The cDNA encoding the full-length D. melanogaster Met protein (Q9VYW2) was provided by T. Godlewski (Prof. Thomas G. Wilson, Ohio State University). The MetC sequence (509–716 aa) was amplified using polymerase chain reaction (PCR) and the primers forward 5′ att aat tcc atg gCG GGC CGG CAA AAG GTG 3′ and reverse 5′ a tta agc ggc cgc TCA TCG CAG CGT GTC 3′, thus introducing a restriction site for NcoI and NotI, respectively (underlined). The obtained PCR product was digested with suitable enzymes and cloned into a pETH-SUMO vector (EMBL, Germany) in the reading frame with 6 × His and SUMO tags. The presence of the insert in the vector was confirmed by sequencing.

The BL21(DE3) E. coli cells (Novagen) were transformed with a pETH-SUMO vector containing 6 × His-SUMO-MetC cDNA and plated on lysogeny broth (LB) agar containing 50 μg/ml of kanamycin. After overnight incubation at 37°C, a single colony was used to inoculate 20 ml of the LB medium containing 50 μg/ml of kanamycin. The bacterial culture was incubated overnight in a rotary shaker (220 rpm) at 37°C. Then, 10 ml of the overnight starting culture was used to inoculate 500 ml of the N-5052 medium with 100 μg/ml of kanamycin (for unlabeled samples and samples labeled with N¹⁵; in the case of N¹⁵ and C¹³ labeling, minimal M9 medium was used). The cell cultures were incubated at 37°C until the optical density (OD₆₀₀) reached 1.5. Then, incubation was continued at 20°C for 15 h (in the case of the M9 medium, IPTG, in a final concentration of 0.2 mM, was added). Finally, the cells were harvested by centrifugation at 4,000 × g (10 min, 4°C), resuspended in lysis buffer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 20 μg/ml DNase I and RNase A, 0.2 mg/ml PMSF, and 1 mM β-mercaptoethanol (βME)), and disrupted by sonication (10 min, 60%, 30 s/30 s). After lysate centrifugation (20,000 rpm, 1 h, at 4°C), the soluble fraction containing 6 × His-SUMO-MetC was collected. Immobilized metal affinity chromatography (IMAC) on the Ni²⁺-NTA resin was chosen as the first purification step. The unbound proteins were washed away with lysis buffer, and the 6 × His-SUMO-MetC protein was digested with SUMO hydrolase dtUD1 (EMBL, Germany) (Weeks et al., 2007) on the column. Finally, tag-less MetC was collected as a flow-through, concentrated to 2 ml on the Amicon Ultra Centrifugal Filter (Millipore) with a cutoff of 10 kDa, and loaded on a Superdex 200 10/300 GL column (Amersham Biosciences) connected to ÄKTA avant (GE Healthcare). For MetC NMR analysis, 20 mM Bis-Tris pH 6.5, 150 mM NaCl, and 5 mM βME were used at the buffer. Samples of the purified MetC protein were collected for further analysis.

Cloning and the expression of GceC and the Ftz-F1 LBD were described in detail previously (Kolonko et al., 2020). The cDNA encoding the D. melanogaster Gce protein (UniProt accession number: Q9VXW7) was provided by Prof. Thomas G. Wilson (Ohio State University) and used as a template for PCR. The amplified GceC fragment (661–959 aa) was cloned into the pET-M11 vector. Next, the cDNA encoding the Ftz-F1 LDB (P33244, 786–1027 aa) was synthesized (Gene Art Thermo Fisher Scientific) and cloned into the pET-15b vector. Both vectors introduced a 6 × His tag at the protein N-terminus.

GceC and Ftz-F1 LBD expression procedures were analogous to the procedure described for MetC. The obtained cell pellets were resuspended in lysis buffer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 20 μg/ml DNase I and RNase A, 0.2 mg/ml PMSF, and 1 mM β-mercaptoethanol (βME)) and disrupted by sonication (10 min, 60%, 30 s/30 s), followed by centrifuging at 20,000 rpm for 1 h at 4°C.

In the case of GceC, the protein was present in the sediment fraction, so the protein was purified under denaturing conditions. After denaturation with 6 M GdmCl₂ and renaturation by 100 x dilution in refolding buffer (20 mM Tris-HCl pH = 8.0, 7.5% glycerol, and 5 mM DTT), the protein was bound to the Ni²⁺-NTA resin, eluted with 0.35 M imidazole, and loaded on the Superdex 200 10/300 GL column (Amersham Biosciences) equilibrated with GceC NMR buffer (PBS pH = 7.4, 2 mM DTT). Samples containing the purified GceC protein were collected.

The Ftz-F1 LBD was present in the soluble fraction. IMAC on Ni²⁺-NTA was used as the first step of the purification procedure. The unbound proteins were washed away with wash buffer (20 mM Tris-HCl pH 7.5 and 100 mM NaCl), and the 6 × His-Ftz-F1 protein was eluted with 0.5 M imidazole. The collected fusion protein was digested overnight with thrombin (50 U for 10 mg of 6 × His-Ftz-F1, Sigma-Aldrich), concentrated to 2 ml on the Amicon Ultra Centrifugal Filter (Millipore) with a cutoff of 10 kDa, and loaded on the Superdex 200 10/300 GL column (Amersham Biosciences) equilibrated with a suitable buffer (Ftz-F1 LBD NMR buffer: 20 mM sodium phosphate pH 6.8, 50 mM NaCl, 1 mM TCEP, and 0.5 mM EDTA). Samples containing the purified Ftz-F1 LBD protein were collected.

To obtain the GceC protein in phosphorylated form (P-GceC), its overexpression was carried out in the presence of a protein kinase A catalytic subunit (PKA) with S/T phosphorylation activity (Plasmid #14921, Addgene) (Narayana et al., 1997). The BL21(DE3) E. coli cells (Novagen) were co-transformed with the pET-M11 plasmid encoding 6 × His-GceC and the pET-15b plasmid encoding PKA and plated on LB agar containing 50 μg/ml kanamycin and 50 μg/ml ampicillin (resistance to ampicillin was introduced with the pET-15b vector). The following protein expression and purification steps were the same as for the non-phosphorylated GceC control since the PKA was present only in the soluble fraction.

The full-length 14-3-3 human isoform ζ, cDNA (P63104) used for tests was provided by Dr. Marek Orłowski (Wroclaw University of Science and Technology) and amplified (PCR) with the primers, forward 5′ aaa acc atg gcc ATG GAT AAA AAT 3′ and reverse 5′ aaa agc ggc cgc TTA ATT TTC CCC TCC 3’, thus introducing restriction sites for NcoI and NotI, respectively (underlined). The digested product was cloned into a pET-M11 vector in a reading frame with a 6 × His tag. The correctness of the obtained pET-M11-6 × His-14-3-3 vector was confirmed by sequencing.

The protein expression procedure was the same as for MetC. The obtained cell pellet was resuspended in lysis buffer and disrupted by sonication (10 min, 60%, 30 s/30 s), followed by centrifugation (20,000 rpm, 1 h, at 4°C). The soluble fraction containing 6 × His-14-3-3 was loaded on the Ni²⁺-NTA resin. Next, 6 × His-14-3-3 was eluted with the buffer supplemented with 0.5 M imidazole, concentrated to 2 ml on the Amicon Ultra Centrifugal Filter (Millipore) with a cutoff of 10 kDa, and loaded on the Superdex 75 10/300 GL column (Amersham Biosciences) equilibrated with the appropriate buffer. Fractions containing purified 14-3-3 were collected.

The lyophilized 19-residue MetC and GceC peptides (Met^PEP, LERIVLYLIENLQKSVDSA; Gce^PEP, LEKAVLRLIQNLQKSGENG) representing the novel NR-box were synthesized (PSL GmbH, Heidelberg). The product purity determined by NMR was >98%. The peptides were finally dissolved in appropriate buffers.

Samples collected during protein expression and purification were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (12% polyacrylamide gels developed in a Tris/glycine system (Laemmli, 1970)). The Precision Plus Protein™ Standards Weight Marker (Bio-Rad) was used as a molecular mass (MM) protein standard. The gels were stained with SimplyBlue™ SafeStain (Invitrogen).

The NMR spectra of MetC (for resonance assignment and chemical shift perturbation experiment), as well as the NMR spectra of Ftz-F1 LBD (for resonance assignment transfer), were collected using the Bruker Avance III HD spectrometer with a proton frequency of 800 MHz, equipped with a 5 mm TCI-HCN cryogenically cooled probe. GceC, as well as Ftz-F1 LBD spectra for the chemical shift perturbation experiment, was measured in the absence and presence of selected partners (Ftz-F1 or 14-3-3 for GceC and MetC, GceC, Met^PEP, or Gce^PEP for the Ftz-F1 LBD) using the Bruker Avance III HD spectrometer with a proton frequency of 600 MHz, equipped with a 3 mm cryoprobe and automatic sample changer (SampleJet). All the protein samples for NMR measurements were prepared in the appropriate buffer (20 mM Bis-Tris pH 6.5, 150 mM NaCl, and 5 mM βME for MetC; PBS pH 7.4 and 2 mM DTT for GceC; and 20 mM sodium phosphate pH 6.8, 50 mM NaCl, 1 mM TCEP, and 0.5 mM EDTA for Ftz-F1 LBD), with the addition of 10% D₂O to provide a lock signal, and measured at defined temperatures (18°C for MetC and GceC, and 29°C for Ftz-F1 LBD). The conditions of MetC measurements were optimized to slow down the chemical exchange of amide protons, which, under the initial conditions, led to the disappearance of many peaks in the ¹⁵N-HSQC spectrum. For Ftz-F1, the conditions were set identically to the ones in the initial resonance assignment (Daffern et al., 2019). The conditions of GceC were optimized to avoid protein precipitation. The interaction and titration analysis were performed using ¹⁵N-HSQC experiments.

The resonance assignment of MetC was performed using a uniformly isotopically labeled (¹³C and ¹⁵N) sample with a concentration of 450 μM. A set of multidimensional experiments was utilized, including 3D HNCO, 5D HN(CA)CONH (Kazimierczuk et al., 2010), 5D (HACA)CON(CA)CONH (Zawadzka-Kazimierczuk et al., 2012b), and 5D HabCabCONH (Kazimierczuk et al., 2010). All these experiments were acquired using nonuniform sampling of the evolution time space. They were processed using the signal separation algorithm for 3D (Stanek and Koźmiński, 2010) and 5D (Kosiński et al., 2017) data. For sparse processing of 5D datasets (Kazimierczuk et al., 2010), 3D HNCO was used as a basis spectrum. Spectra were analyzed using NMRFAM-SPARKY software (Lee et al., 2015). The resonance assignment was supported by the TSAR (Zawadzka-Kazimierczuk, 2012a; Piai et al., 2016). The experimental parameters are given in Supplementary Table S2. The MetC chemical shifts were deposited in the BMRB (Hoch et al., 2023), under accession number 51720.

The chemical shift perturbation experiments of MetC upon adding Ftz-F1 LBD were performed using the ¹⁵N-HSQC technique. The titration scheme was adapted from the study by Sekhar et al. (2017); the two initial samples were prepared (sample A with 150 μM MetC concentration and with no Ftz-F1 and sample B with 150 μM MetC concentration and 500 μM Ftz-F1 LBD concentration). The Ftz-F1 LBD concentration was changed by cross-mixing (simultaneously exchanging equal amounts of the solution between the samples) so that the sample volumes and MetC concentration remained constant. The measurement was performed for the following concentrations of the Ftz-F1 LBD: 0, 40, 82, 122, 199, 301, and 500 μM. The dissociation constant was calculated according to the published protocol (Williamson, 2013).

For the Ftz-F1 LBD, the resonance assignment (of the peaks showing changes in the position or intensity on binding to the Met^PEP and Gce^PEP peptides) was transferred from the study by Daffern et al. (2019). The transfer was supported with 3D HNCO (Kay et al., 1990), HN(CA)CO (Clubb et al., 1992), HN(CO)CA (Bax and Ikura, 1991), HNCA (Kay et al., 1990), and CBCA(CO)NH (Grzesiekt and Bax, 1992) spectra. The sample used for these experiments was uniformly ¹³C,¹⁵N-labeled, and its concentration was 450 μM. The 3D HNCO experiment was recorded using nonuniform sampling of the evolution time space, while all other experiments were performed conventionally. The experimental parameters are given in Supplementary Table S3.

The chemical shift perturbation experiments of GceC upon adding Ftz-F1 LBD or 14-3-3 and FTZ-F1 LBD upon adding MetC, GceC, Met^PEP, or Gce^PEP were performed using the ¹⁵N-HSQC technique. The 100 μM ¹⁵N-labeled GceC spectra were recorded in the absence and presence of the equimolar amount of unlabeled Ftz-F1 LBD or 14-3-3. The 150 μM ¹⁵N-labeled FTZ-F1 LBD spectra were recorded in the absence and presence of an equimolar quantity of unlabeled MetC, GceC, Met^PEP, or Gce^PEP. Finally, the peptide concentration was used up to the solubility limit (approximately 3 mM for Met^PEP and 2 mM for Gce^PEP).

The binding of Met^PEP to the Ftz-F1 LBD was monitored using a Nano Isothermal Titration Calorimeter (ITC) (TA Waters, USA) at 25°C with a cell volume of 1 ml. All experiments were performed in duplicate, in 20 mM HEPES buffer, pH 6.8, 150 mM NaCl, and 1 mM TCEP. The titrate (Ftz-F1 LBD) concentration was 0.035 mM, whereas the titrant (Met^PEP) concentration was 1.06 mM. After system equilibration, 20 injections of the titrant were made into the reaction cell in 12.5 µl increments at 300 s intervals with stirring at 200 rpm. To determine the heats of titrant dilution, control titrations were performed by identical injections of Met^PEP to the buffer. The net reaction heat was calculated by subtracting the dilution heat from the corresponding total heat of the reaction. The titration data were analyzed using NanoAnalyze (version 3.3.0), NITPIC (version 2.0.7) (Scheuermann and Brautigam, 2015), and SEDPHAT (version 15.2b) (Houtman et al., 2007). Data were fitted to an independent model A + B ⇌ AB using the global fitting mode. The error estimates for the fitting results were produced by Monte Carlo analysis with 1,000 iterations and confidence levels of 0.7.

It is worth mentioning that in previous studies (Greb-Markiewicz et al., 2011; Greb-Markiewicz et al., 2015) the Gce isoform (provided by Prof. T Wilson and T. Godlewski, Ohio State University) comprising 689 aa was used. Later, this isoform was substituted by a new isoform containing an additional N-terminus 270 aa long (UniProt accession number Q9VXW7). However, according to Baumann (2010a), both Gce isoforms perform the same functions, and the first 270 amino acid residues do not affect the protein activity. In addition, no NLSs/NESs were predicted to influence localization in this region. Hence, for expression studies in mammalian cells, to compare results with those of previous studies, we also used the short isoform of GCE in this study.

cDNA fragments corresponding to Met, Gce, MetC, and GceC, with a C-terminally introduced FLAG peptide sequence (DYKDDDDK), were cloned into the multiple cloning site (MCS) of the pEYFP-C1 vector (Clontech) (pEYFP-Met-FLAG, pEYFP-Gce-FLAG, pEYFP-MetC-FLAG, and pEYFP-GceC-FLAG) with EcoRI and SalI restriction enzymes. The cDNA of the Ftz-F1 LBD and 14-3-3 was cloned into the pECFP-C1 vector (Clontech) (pECFP-Ftz-F1 LBD and pECFP-14-3-3, respectively) with XhoI and BamHI restriction enzymes. All the primer sequences are presented in Supplementary Table S4. All constructs were confirmed by DNA sequencing.

All experiments were performed as described previously (Kolonko et al., 2020). In short, African green monkey kidney fibroblast COS-7 cells (ATCC CRL-1651) were grown in Ø6 cm plates at 37°C 24 h prior to transfection (Xfect Transfection Reagent Takara Bio) with 9 μg of the appropriate vector encoding the selected protein, or with 6 μg of each from two selected vectors in the case of co-transfection. pEYFP-C1 and pECFP-C1 vectors were used as controls. After transfection, the culture was further incubated for 24 h and used for fluorescence imaging.

All experiments were performed as described previously (Kolonko et al., 2020). After 24 h of incubation, the cells were lysed (lysis buffer: 25 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 1 mM EDTA, and 5% glycerol), transferred to a microcentrifuge tube, and centrifuged at 13,000 × g for 15 min. Importantly, in the case of 14-3-3 interaction assays dependent on partner phosphorylation, we used phosphatase inhibitors, 5 μM sodium orthovanadate, 10 μM sodium molybdate, 10 μM PMSF, and cOmplete™ EDTA-free Protease Inhibitor (cat. No. 11836170001, Roche) in lysis buffer. The soluble fractions were incubated on ice with 20 μl of EZview™ Red ANTI-FLAG M2 Affinity Gel (Sigma-Aldrich) for 2 h. The unbound proteins were washed away, and the gel was incubated with the elution buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 100 μg/ml FLAG peptide) for 30 min. The eluted fractions were collected for Western blot analysis.

All experiments were performed as described previously (Kolonko et al., 2020). Briefly, samples obtained during the pull-down assay were separated by SDS-PAGE using 12% gels and transferred to the Whatman Protran nitrocellulose transfer membrane (Protran BA85, Schleicher & Schuell Pure, Sigma-Aldrich). The membranes were blocked with milk powder (Milchpulver, blotting grade, Roth), then incubated overnight with specific primary anti-GFP polyclonal antibodies (1:5,000, cat. no. SAB2702197, Sigma-Aldrich) or anti-14-3-3 antibodies (1:5,000, cat. no. AB9748-I, Millipore), followed by the incubation with secondary horse anti-mouse antibodies (for anti-GFP, 1:10,000, cat. no. PI-2000, Vector Laboratories) or goat (for anti-14-3-3, 1:10,000, Sigma-Aldrich, cat. no. A 6154) anti-mouse antibodies coupled to horseradish peroxidase. Signals were detected using the SuperSignal™ West Pico PLUS Substrate Chemiluminescent kit (Thermo Scientific™). Finally, the membranes were exposed to Kodak BioLight film. The MM of each fusion protein was calculated based on aa sequences using the ProtParam server (Gasteiger, 2019) (YFP-Met-FLAG 107 kDa, YFP-Gce-FLAG 105 kDa, YFP-MetC-FLAG 51.4 kDa, YFP-GceC-FLAG 64 kDa, CFP-Ftz-F1 LBD 55 kDa, and CFP-14-3-3 57 kDa).

The experiment was performed as described previously (Kolonko et al., 2020). Briefly, fluorescence microscopy was performed on an Olympus IX71 microscope with a CFP or YFP filter 24 h after transfection. Cells transfected with empty pEYFP-C1 and pECFP-C1 vectors were used as controls.

In the case of MetC, the previously developed protocol (expression from the pColdTM TF vector, Takara) (Kolonko et al., 2016) was not sufficient to obtain isotopically (¹⁵N and ¹³C) labeled samples. Thus, we used another vector, pET-SUMO (EMBL, Germany), introducing 6 × His and SUMO tags at the protein N-terminus. The SUMO tag acts both as a chaperonin and an initiator of protein folding and enhances the stability and solubility of target proteins (Jg et al., 2006). The expressed fusion protein was present in the soluble fraction, and the developed MetC purification procedure consisted of IMAC with a Ni²⁺-NTA resin, fusion protein digestion on a column with SUMO hydrolase (dTUd1, EMBL, Germany) (Weeks et al., 2007), and final purification combined with buffer exchange on the Superdex 200 10/300 GL column (Figure 1A, elution profile). As a result, the purified MetC was visible as a single band on the 12% SDS-PAGE gel (Figure 1B) with decreased electrophoretic mobility (approximately 29 kDa) compared to the MM of MetC (23.4 kDa) (Kolonko et al., 2016). Such behavior is a characteristic of IDPs (Tompa, 2002; Iakoucheva et al., 2009), binding less SDS due to the specific sequence composition.

GceC purification was described in detail previously (Kolonko et al., 2020). In short, the protein was insoluble under all tested conditions, so we decided to develop a purification procedure under denaturing conditions. After GceC denaturation with GdmCl₂, the protein was refolded by dilution and bound to the Ni²⁺-NTA resin since GceC was expressed in fusion with the 6 × His tag. Finally, SEC on the Superdex 200 10/300 GL column was used to remove small impurities. Purified GceC electrophoretic mobility was decreased similar to MetC: its electrophoretic mobility corresponded to the protein with an MM of 42 kDa instead of the expected 36 kDa. GceC belongs to IDPs (Kolonko et al., 2020).

The 14-3-3 protein interacts with its partners by binding phosphorylated S residues (Rajagopalan et al., 2008) (although binding of non-phosphorylated proteins is also possible (Masters et al. (1999)). Therefore, samples of GceC in the phosphorylated form (P-GceC) were prepared. Its purification procedure was identical to that developed for non-phosphorylated GceC. During SDS-PAGE analysis, in the P-GceC sample, two bands were observed (Figure 2A). The upper band, with decreased mobility, refers to a phosphorylated protein (P-GceC), while the lower band refers to non-phosphorylated GceC. When alkaline phosphatase was added to the sample, mainly, a lower band (GceC) was observed, which confirms the phosphorylation of GceC (Supplementary Figure S2). During expression, approximately half of the protein was phosphorylated.

The Ftz-F1 LBD purification procedure was described previously (Kolonko et al., 2020). In short, the 6 × His -FTZ-F1 LBD fusion protein was highly overexpressed and fully soluble. First, it was bound to the Ni²⁺-NTA resin, eluted with imidazole, and digested with thrombin overnight. Finally, it was purified from digesting enzymes and proteins co-purified on IMAC by SEC on the Superdex 75 10/300 GL column. The MM of LBD Ftz-F1 was calculated based on its aa sequence (28 kDa), and its apparent mass determinate from electrophoretic mobility overlapped (Figure 2B). The Ftz-F1 LBD is a globular domain.

The 14-3-3 human isoform ζ expressed from pET-M11 was fully soluble and highly overexpressed. IMAC on the Ni²⁺-NTA resin was used as the first step of purification. After elution, 14-3-3 was concentrated and purified by SEC on the Superdex 200 10/300 GL column, separating small impurities and changing the buffer. The apparent mass determined by SDS-PAGE (Figure 2B) is in agreement with the calculated MM (30 kDa when the tag is included) since 14-3-3 is a globular protein (Sluchanko and Gusev, 2012).

The native, active form of all purified proteins was controlled by determining their ability to interact (see the following section).

To determine functional differences between MetC and GceC, their ability to interact with both the Ftz-F1 LBD and 14-3-3 was analyzed by NMR spectroscopy. To determine aa residues directly involved in the interaction, spectrum assignment, and chemical shift perturbation analysis were performed. The MetC chemical shifts were deposited in the BMRB (Hoch et al., 2023) under the accession number 51720. The MetC resonance assignment percentage was 87.5%. We performed secondary structure prediction (neighbor-corrected Structural Propensity Calculator (ncSPC)) (Tamiola and Mulder, 2012)), which clearly showed the highly unstructured nature of the protein. The highest score (up to 0.3) in two regions (vicinity of residues 518 and 639 of full Met) indicated a slight propensity toward α-helix formation in these regions. Other parts remain mostly unstructured.

The 2D¹⁵N-HSQC spectrum of MetC, given in Supplementary Figure S3, shows features typical for IDPs, for which chemical environments of all the observed nuclei are averaged due to fast conformational changes, leading to a very narrow chemical shift range (Zuiderweg, 2002; Grudziąż et al., 2018); in the MetC case, it is 8–8.6 ppm in the proton dimension. Therefore, observed signals correlating amide proton and amide nitrogen frequencies strongly overlap. More dispersed signals (like 556E, 574E, 655V, and 716R) may correspond to aa residues involved in the formation of local and transient motifs of the secondary structure (like MoREs).

The ¹⁵N-HSQC spectrum of MetC was used for the chemical shift perturbation experiment to define interactions between MetC and the Ftz-F1 LBD and determine aa residues affected by binding. The spectra were recorded in the absence and presence of an unlabeled Ftz-F1 LBD in increasing concentrations. Multiple signals show significant changes in intensity or are shifted (Figure 3A), which unequivocally indicates interactions. The greatest chemical shift changes were observed for residues 705S, 707L, 711L, 712T, 713S, 714T, 715L, and 716R (Figure 3A).

Analogous measurements were performed to analyze interactions between MetC and 14-3-3. In accordance with in silico analysis (Supplementary Table S1) (Greb-Markiewicz et al., 2015), in contrast to GceC, no specific shifts were observed (Figure 3B), which indicates no interactions. The result is consistent with the hypothesis based on our previous research (Greb-Markiewicz et al., 2015) and in silico analysis that Gce interaction with 14-3-3 depends on the GceC presenting 14-3-3-binding motifs.

The efficiency of GceC expression and purification was very low; therefore, it was not possible to obtain sufficiently concentrated samples to assign the observed signals to aa residues in the protein chain. However, samples suitable for 2D¹⁵N-HSQC measurements were prepared to verify if any changes in the reference spectrum in the presence of an equimolar amount of Ftz-F1 LBD or 14-3-3 can be observed.

The 2D¹⁵N-HSQC GceC spectrum, a characteristic of IDPs, was described previously (Kolonko et al., 2020). Specific signal shift perturbation in the presence of Ftz-F1 LBD was also presented (Kolonko et al., 2020). Here, we analyzed the ¹⁵N-labeled P-GceC spectrum in the presence of unlabeled 14-3-3 protein (Supplementary Figure S4). The phosphorylated form of GceC was used since 14-3-3 interacts mainly with phosphorylated S (Rajagopalan et al., 2008). The comparison of P-GceC spectra recorded in the absence and presence of 14-3-3 shows a change in intensity and shifting of many signals, which indicate an interaction between proteins. Importantly, the scheme of changes in the case of an interaction with the Ftz-F1 LBD and 14-3-3 differs: the shift perturbation depends on the partner being bound. Moreover, in the case of the interaction with the Ftz-F1 LBD, a slightly larger number of changes in the GceC spectrum was recorded (Kolonko et al., 2020), which may suggest a more robust and more stable interaction.

The interactions between full-length Ftz-F1 and Met/Gce were studied previously (Bernardo and Dubrovsky, 2012a). Here, we focused on the binding capacity of the Ftz-F1 LBD, comprising the AF2 motif, and MetC/GceC regions, consisting of the novel NR-box. In addition, we analyzed interactions of the Ftz-F1 LBD with Met^PEP and Gce^PEP, which were hypothetically directly involved and responsible for interactions with Ftz-F1 (Bernardo and Dubrovsky, 2012a).

The Ftz-F1 LBD 2D¹⁵N-HSQC spectrum was previously presented by Daffern et al. (2019). We recorded a very similar reference spectrum pattern; however, many unsystematic shifts of the resonance positions were observed, which made it impossible to transfer the assignment directly. To support the assignment transfer, a set of 3D NMR experiments was conducted, and the observed signals were reassigned to aa residues in the Ftz-F1 LBD chain. The Ftz-F1 LBD spectrum shows better peak dispersion characteristics compared to the spectra of IDPs; however, a large number of aa residues causes a substantial level of peak overlap in the middle part of the spectrum, even for this globular protein.

First, Ftz-F1 interaction capacity with Met^PEP and Gce^PEP was analyzed. In the presence of equimolar or double equimolar amounts of peptides, only subtle changes compared to the reference spectrum were observed. It indicated that the interaction of Ftz-F1 with both peptides is rather weak. Therefore, each peptide was added to the Ftz-F1 LBD sample up to its solubility limit (approximately 3 mM for Met^PEP and 2 mM for Gce^PEP). The recorded spectra present specific signal shifts (Figures 4A, B) that confirm peptides binding to the Ftz-F1 LBD. Importantly, the scheme of observed changes is similar for each Met^PEP and Gce^PEP (the same residues were affected, but the intensity of the effect is more pronounced on the former) and ensures that peptide sequences contain Ftz-F1-binding sites and are sufficient to form a stable interaction interface with the FTZ-F1 LBD in vitro. The solubility of Gce^PEP was decreased compared to that of Met^PEP, and thus, the observed shifts were smaller. Similar analyses were performed for the labeled Ftz-F1 LBD in the presence of equimolar amounts of MetC or GceC. Again, with the peptide-like pattern, many shift perturbations were noticed (Figure 4C, MetC). Importantly, in the case of MetC or GceC, the equimolar concertation was sufficient to observe specific shifts.

The aa residues of Ftz-F1 experiencing a major shift were 1002L, 1003Y, 1004T, 1008A, 1009G, 1010S, 1011A, 1013T, 1024K, 1025R, and 1026K (localized in the Ftz-F1 LBD helix α12, AF2 sequence). Slightly smaller shifts were observed for 1006H, 1016L, 1017L, 1020M, and 1027G (also localized in the Ftz-F1 LBD helix α12, AF2 sequence). Additional changes are observed for 947R–948G and 801V+ (referring to the α3 helix and sequence connecting α8 and α9).

Isothermal titration calorimetry was used to confirm the binding of the novel NR-box represented by Met^PEP and Gce^PEP to the Ftz-F1 LBD and to determine the binding affinity. As shown in Supplementary Figure S5, the binding of Met^PEP to the Ftz-F1 LBD is represented as a decreasing exothermic process. The global data fitting to the 1:1 independent model uncovered micromolar affinity with a K _d value of 9.9 ± 0.5 µM and ΔH ^ITC of −1.1 ± 0.5 kcal ⋅ mol^-1. The calculated ΔG ^ITC and ΔS ^ITC values based on the fitted ones were −6.8 kcal⋅mol^-1 and 19.1 cal⋅mol^-1⋅K⁻¹, respectively (Padjasek et al., 2020). Since the solubility of Gce^PEP was very low, it was impossible to determine the K _d for this interaction due to experimental requirements.

To confirm the results of NMR studies and test interactions between full-length Met and Gce with partners, we performed the immunoprecipitation experiment using the ANTI-FLAG M2 Affinity Gel (Sigma-Aldrich) and COS-7 cells. The proteins were expressed as results of cell transfection with a single vector encoding CFP-Ftz-F1/LBD, CFP-14-3-3, YFP-Met-FLAG, YFP-Gce-FLAG, YFP-MetC-FLAG, or YFP-GceC-FLAG, or with two selected vectors (YFP-Met-FLAG and CFP-Ftz-F1 LBD or CFP-14-3-3; YFP-Gce-FLAG and CFP-Ftz-F1 LBD or CFP-14-3-3; YFP-MetC-FLAG and CFP-Ftz-F1 LBD or CFP-14-3-3; and YFP-GceC-FLAG and CFP-14-3-3) (YFP-GceC-FLAG in combination with the CFP-Ftz-F1 LBD was tested previously (Kolonko et al., 2020)), and pulled down from the cell lysates using the ANTI-FLAG M2 Affinity Gel. Western blotting was used to visualize the results. Importantly, a mix of proteases and inhibitors was used during cell lysis and precipitation experiments to prevent the loss of phosphate groups indispensable for the 14-3-3 interaction. CFP-Ftz-F1 LBD and CFP-14-3-3 were not tagged with the FLAG peptide and, therefore, could not bind to the ANTI-FLAG M2 Affinity Gel, providing a negative control—no band is visible in the elution fractions (Supplementary Figures S6A, B). Each YFP-Met-FLAG, YFP-Gce-FLAG, YFP-MetC-FLAG, and YFP-GceC-FLAG expressed alone was used as positive controls. The presence of the FLAG-tag at their C-termini allowed them to bind to the affinity gel. Each protein is observed as a single band at the expected area in the elution fractions (Supplementary Figures S6C, D). The additional bands observed in the elution fraction for MetC and GceC are nonspecific.

In contrast, the CFP-Ftz-F1 LBD and CFP-14-3-3 cannot bind to the resin and require interaction with the FLAG-labeled protein to be present in the eluted fraction. Hence, their detection with an anti-GFP antibody in the elution fraction clearly confirms the interaction. In the case of YFP-tagged Met-FLAG co-expressed with the CFP-Ftz-F1 LBD or CFP-14-3-3, only the band responding to YFP-Met-FLAG is detected in the elution fraction (Figures 5A, B). This result indicates that no interactions occurred. A similar result was obtained for YFP-MetC-FLAG co-expressed with CFP-14-3-3 (Figure 5D). Since the electrophoretic mobility of both proteins is highly similar, the results were confirmed by the detection of proteins with an anti-14-3-3 antibody (Figure 5E). CFP-tagged 14-3-3 is visible in the cell lysate (SUP), a fraction of proteins not bound to the resin (FT) and wash fraction (W1) but not in the fraction eluted from the resin. These results show that YFP-MetC-FLAG and CFP-14-3-3 do not interact. In the case of YFP-MetC-FLAG and CFP-Ftz-F1 LBD, two bands were observed in the elution fraction (Figure 5C), which indicate the interaction between proteins.

Similarly, YFP-Gce-FLAG co-expressed with the CFP-Ftz-F1 LBD (Figure 6A) or CFP-14-3-3 (Figure 6B) and YFP-GceC-FLAG co-expressed with CFP-14-3-3 (Figure 6C) resulted in two bands observed in the fraction eluted from the ANTI-FLAG M2 Affinity Gel. Observed bands correspond to bands expected for the tested protein pair, which clearly confirms the interactions. A similar result was observed previously for YFP-GceC-FLAG paired with CFP-Ftz-F1 LBD (Kolonko et al., 2020).

Summarizing, we showed that the full-length Gce and GceC can interact with Ftz-F1 LBD and 14-3-3. Our results show that GceC is sufficient for interactions with Ftz-F1/LBD and 14-3-3. Interestingly, the full-length Met does not interact with any of these proteins under the same conditions, and MetC can interact only with the Ftz-F1 LBD. Importantly, we showed differentiation in the binding partners of Gce and Met, which can result in different functions of these proteins. In addition, the C-termini of Gce and Met seem to be factors responsible for this differentiation.

To continue interaction studies, we investigated whether interacting partners affect the subcellular localization of the tested proteins. We used fluorescence microscopy to observe the distribution of Met, MetC, Gce, and GceC in the presence and absence of the FTZ-F1 LBD and 14-3-3 in COS-7 cells and analogous to the pull-down method pEYFP-C1 and pECFP-C1 derivatives. The cells were analyzed by fluorescent microscopy 24 h after transfection of COS-7 cells. The cells transfected with a single vector were used as a control. The monitoring of protein overexpression and subcellular localization in living cells by fluorescent protein labeling is commonly used (Chalfie et al., 1994). Importantly, it was shown that GFP tagging does not affect the localization and activity of fused proteins (Chudakov et al., 2010). First, we performed a test for control proteins. As expected, cells transfected with pEYFP-C1 or pECFP-C1 presented ubiquitous fluorescence of the YFP or CFP proteins (Figures 7A, B). The expressed CFP-Ftz-F1 LBD was also distributed ubiquitously in the whole cell (Figure 7C), while CFP-14-3-3 was distributed only in the cytoplasm, with a substantial accumulation around the nuclei (Figure 7D). The expressed YFP-Met-FLAG localized predominantly in the nuclei (Figure 7E), while YFP-MetC-FLAG was in the cytoplasm with some accumulation around the nuclei (Figure 7F). Finally, the expressed YFP-Gce-FLAG fluorescence for most analyzed cells was observed in both compartments with a slight predominance of the nuclei, while some cells (less than 5%) presented a cytoplasmatic signal (Figure 7G). As expected, YFP-GceC-FLAG was observed exclusively in the nuclei (Figure 7H).

No shift was observed in the subcellular localization when YFP-Met-FLAG was co-expressed with the CFP-Ftz-F1 LBD (Figure 8A) or CFP-14-3-3 (Figure 8B) and when YFP-MetC-FLAG was co-expressed with CFP-14-3-3 (Figure 8D). In contrast, YFP-MetC-FLAG was co-expressed with the CFP-Ftz-F1 LBD, resulting in the shift of YFP-MetC-FLAG to nuclei, which, in turn, resulted in the almost homogenous distribution in both compartments of the cell, analogous to the distribution of the CFP-Ftz-F1 LBD (Figure 8C).

The co-expression of YFP-Gce-FLAG with the CFP-Ftz-F1 LBD shifted both proteins toward the nucleus (Figure 9). Similar results were observed previously for YFP-GceC-FLAG in the presence of the CFP-Ftz-F1 LBD. The CFP-Ftz-F1 LBD protein was shifted to the nucleus, where YFP-GceC-FLAG was present in the absence and presence of a partner (Kolonko et al., 2020).

Finally, we tested the impact of CFP-14-3-3 expression on the YFP-Gce-FLAG and YFP-GceC-FLAG subcellular localization. When co-expressed with cytoplasmic CFP-14-3-3, YFP-Gce-FLAG was also located in the cytoplasm (Figure 10A). In the case of YFP-GceC-FLAG, the expressed protein distribution was not ubiquitous. For approximately 50% of the analyzed cells co-transfected successfully, we observed the shift of the fluorescent signal from the nuclei to the cytoplasm (Figure 10B). Simultaneously, CFP-14-3-3 shifted toward the nuclei (Figure 10B).

We showed that a shift in the subcellular distribution of one or both partner proteins was observed for cells expressing Gce/GceC in the presence of FTZ-F1 LBD/14-3-3. In the case of MetC co-expressed with the FTZ-F1 LBD, we also observed the transfer in contrast to co-expression with 14-3-3, while the full-length Met did not present any translocation. Our results confirm that interacting partners affect the subcellular localization of the tested proteins and are consistent with results of pull-down experiments showing that full-length Met does not bind to the FTZ-F1 LBD or 14-3-3, while MetC can interact only with the FTZ-F1-LBD. The presented results allow us to document functional differences in Gce/GceC and Met/MetC behavior.