PCID2 Subunit of the Drosophila TREX-2 Complex Has Two RNA-Binding Regions

Drosophila PCID2 is a subunit of the TREX-2 mRNA nuclear export complex. Although the complex has long been studied in eukaryotes, it is still unclear how TREX-2 interacts with mRNA in multicellular organisms. Here, the interaction between Drosophila PCID2 and the ras2 RNA was studied by EMSA. We show that the C-terminal region of the WH domain of PCID2 specifically binds the 3′-noncoding region of the ras2 RNA. While the same region of PCID2 interacts with the Xmas-2 subunit of the TREX-2 complex, PCID2 interacts with RNA independently of Xmas-2. An additional RNA-binding region (M region) was identified in the N-terminal part of the PCI domain and found to bind RNA nonspecifically. Point mutations of evolutionarily conserved amino acid residues in this region completely abolish the PCID2–RNA interaction, while a deletion of the C-terminal domain only partly decreases it. Thus, the specific interaction of PCID2 with RNA requires nonspecific PCID2–RNA binding.

Drosophila Xmas-2/yeast Sac3 is a scaffold TREX-2 subunit, around which dPCID2/ yThp1 and dENY2/ySus1 are organized [7,16,24,25]. Additionally, it has been found that Sem1, a small, negatively charged protein, is associated with TREX-2 and stabilizes the complex in yeast [7,26]. PCID2 is a subunit responsible for the TREX-2-mRNA interaction [7]. Its major part is the PCI domain, which is a bipartite fold that consists of a structurally conserved C-terminal winged-helix (WH) domain and a more divergent N-terminal HD domain and is based on a stack of α-helices [7,25]. The WH domain is a common nucleic acid-binding motif shown to interact with dsDNA, ssDNA, and RNA [7,[26][27][28].
Crystallization of yeast Thp1-Sem1 and Sac3 has shown that Sem1 is associated with the HD domain of Thp1, while Thp1 is associated with Sac3 through its C-terminal WH domain [7]. The WH domains of Sac3 and Thp1 form a common, positively charged surface, which mediates the interaction of the complex with nucleic acids [7]. In EMSA experiments, the individual components of the complex (Thp1-Sem1 and Sac3 250-563) do not interact with RNA, but efficiently bind RNA after being preincubated together [7]. However, an earlier study has demonstrated that Thp1 is capable of nucleic acid binding in vitro regardless of the other proteins of the complex [29]. While it has been demonstrated that human PCID2 is associated with Sem1 in a way similar to that of their yeast homologues [7], the interactions of Drosophila PCID2 and the whole Drosophila TREX-2 complex with mRNA has not been studied yet.
In particular, it is unclear whether TREX-2 binds a specific region of mRNA and which regions of PCID2 are responsible for the interaction with RNA. Another unknown issue is whether the TREX-2 subunits dPCID2 and dXmas-2 interact with RNA independently of each other.
The interaction of Drosophila PCID2 with mRNA was studied in this work. We showed that a C-terminal fragment present in a WD domain-homologous RNA-interacting fragment of yeast Thp1 specifically interacts with the 3 -noncoding region of the ras64B (ras2) mRNA. The fragment is also necessary for the interaction with Xmas-2, but Xmas-2 is not required for the PCID2 interaction with the ras2 mRNA in vitro. An additional RNA-binding region, which interacts with RNA nonspecifically, is in the N-terminal part of the PCI domain. Point mutations of evolutionarily conserved amino acids in this region totally abolish the PCID2-RNA interaction. Thus, PCID2 binds with RNA independently of the other TREX-2 subunits via its two domains, and a distorted interaction with RNA in the N-terminal part of the PCI domain disturbs the RNA binding with the C-terminal domain of the protein.
The recombinant GST-or His-tagged proteins were expressed in E. coli BL21 cells at 20 • C for 24 h. Cells were collected by centrifugation and stored at −70 • C. The recombinant proteins with the GST-epitope tag were purified from cell lysates by binding to Glutathione SepharoseTM 4 fast flow (Cytiva, Marlborough, MA, USA) and subsequently eluted by displacement with glutathione according to the manufacturer's recommendations (GE Healthcare). The recombinant proteins with the His-epitope tag were purified from cell lysates by binding to Ni-SepharoseTM (GE) and subsequently eluted by increasing the imidazole concentration according to the manufacturer's recommendations (GE Healthcare).

Electrophoretic Mobility Shift Assay (EMSA)
Purified recombinant proteins corresponding to the full-size PCID2; truncated PCID2; mutated PCID2 variants; and the N, M, and C parts of PCID2 were incubated with the radiolabeled RNA fragments in 20 µL of a binding buffer, which contained 25 mM Tris-HCl, pH 8.0, 100 mM NaCl, 0.5% Triton X-100, 5% glycerol, 1 mM EDTA, 1 mM DTT, 25x PIC, and 40x RiboLock (Thermo Scientific, Waltham, MA, USA), at 4 • C for 1 h. After binding was completed, a PAGE loading buffer was added to the samples, and the samples were applied to a native 5% polyacrylamide gel. Electrophoretic separation was carried out in a 0.5x TBE buffer at 150 V at 4 • C for 80 min. The radioactive signal was detected using a Cyclone Storage Phosphor Screen device. The signal from each lane (area) was quantitated using the ImageJ program. Direct titration reactions were plotted as bound RNA fraction versus protein concentration, and K D was determined according to [30,31]. Immunoprecipitation was performed as described [32], using extracts pretreated with DNase I (Thermo Scientific, Waltham, MA, USA, 0.6 U/mL), RNase (Thermo Scientific, 10 U/mL), and antibodies against the FLAG and HA epitopes.

Statistical Analysis
Error bars represent SD from at least three independent experiments. All data are presented as mean ± SD, and Student's t-test was used to compare the control and treatment groups. An asterisk (*) indicates statistical significance with p-value < 0.05; (**) indicates statistical significance with p-value < 0.01.

Biomolecular Resources
The predicted Drosophila PCID2 structure (AF ID: AF-Q9VTL1-F1) was provided by the AlphaFold Protein Structure Database (ebi.ac.uk, accessed on 28 October 2022), which has been created in partnership with EMBL's European Bioinformatics Institute (EMBL-EBI) [33,34]. The PCID2 structure was visualized using UCSF ChimeraX software, which has been developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases [35]. The electrostatic properties of the protein surface were computed using the APBS web service (poissonboltzmann.org, accessed on 23 April 2023) [36].

PCID2 Specifically Interacts with a Fragment of the 3 -Noncoding Region of the ras2 mRNA Independently of Xmas-2
A direct interaction of Drosophila PCID2 with RNA was studied by EMSA. The ras2 (ras64B) mRNA was used in experiments. The choice was explained by our previous finding that Xmas-2 and PCID2 bound the ras2 mRNA in RNA immunoprecipitation experiments and the RNA pull-down assay [14,19]. The ras2 mRNA sequence was divided into five approximately equal fragments ( Figure 1A). ras2 fr1 corresponded to the 5 -noncoding region, ras2 fr2 and ras2 fr3 to the coding region, and ras2 fr4 and ras2 fr5 corresponded to the 3 -noncoding region of the ras2 mRNA. The respective radiolabeled RNA fragments were synthesized using T3 and T7 RNA polymerases and [α-32P]UTP. A sequence encoding the full-size PCID2 was cloned in pGEX 5.1, and the protein was expressed in a bacterial system. Each RNA fragment was tested for binding with the full-size PCID2 fused with the GST epitope. ras2 fr4, which corresponded to the 3 -noncoding region of the ras2 mRNA, was found to interact with PCID2 in EMSA experiments ( Figure 1B).
To determine the binding site in the ras2 mRNA more precisely, ras2 fr4 was divided into three parts ( Figure 1C). Only the ras2 4_2 fragment was found to interact with PCID2 in EMSA experiments ( Figure 1D). The specificity of the interaction was confirmed in a competition assay (Figure 2A,B). ras2 4_1 (left panel) and ras2 4_2 (right panel) fragments were used as cold competitors. A competitor was added to the reaction in 0.5-to 250-fold excess over the hot 4_2 fragment in a gradient manner. Figure 2B shows the level of binding of GST-PCID2 with hot ras2 fragment 4_2. The interaction of the hot 4_2 fragment with PCID2 decreased gradually with a minimal increment in the presence of cold fragment 4_1, while cold fragment 4_2 competed with its hot counterpart and already decreased its interaction with the protein when used in a minor excess ( Figure 2B).
To better visualize the specific interaction of PCID2 with fragment 4_2 of the ras2 mRNA, supershift EMSA was carried out with anti-PCID2 antibodies ( Figure 2C). The experiment showed that a supershift, a larger antibody-protein-RNA complex ( Figure 2C, lane 3), formed in addition to the protein-RNA complex ( Figure 2C, lane 2) when antibodies against PCID2 were added. Western blotting was performed to verify the result ( Figure 2D). Membrane hybridization with the anti-PCID2 antibodies also detected the formation of the antibody-protein-RNA complex ( Figure 2D, lane 3). To obtain quantitative information on PCID2 binding specificity, we determined the apparent K D of RNA fragment 4_2. Apparent K D values of about 37 nM were observed for ras2 fragment 4_2, indicating high-affinity binding to PCID2 ( Figure S1).
Thus, PCID2 directly binds RNA and specifically interacts with a sequence located in the 3 -noncoding region of the ras2 mRNA. The experiments demonstrated additionally that PCID2 interacts with RNA without being associated with other TREX-2 subunits, such as Xmas-2, which has been implicated in the binding in yeast [7].    PCID2 and its homologs are structurally similar; their evolutionarily conserved C-terminal region (the PCI domain) includes the highly structurally conserved WH domain [7,26]. The C-terminal part of the WH domain of the yeast PCID2 homolog has been shown to interact with the yeast Xmas-2 homolog to form a platform for interactions with an RNA molecule. Certain amino acid sequences were identified as responsible for the interaction of the protein with RNA within the TREX-2 complex [7].
We studied the role that the homologous C-terminal region of PCID2 plays in its interaction with RNA. We constructed a PCID2 devoid of the 35 C-terminal amino acid residues homologous to those shown to be involved in interactions in yeast (PCID2 1-360 ). The full-size PCID2 and PCID2 1-360 , each fused with the GST epitope, were tested for binding fragment 4_2 of the ras2 mRNA by EMSA ( Figure 3A). PCID2 1-360 far less efficiently interacted with the RNA fragment as compared with the full-size PCID2 ( Figure 3B), indicating that the C-terminal domain plays an important role in the interaction with RNA. However, it should be noted that the interaction with RNA was not completely abolished. The finding indicates that another PCID2 region is probably also involved in interacting with the ras2 mRNA.

The C-Terminal Fragment of the PCI domain Interacts with the 3 -Noncoding Region of the ras2 mRNA
A role in interacting with Xmas-2 was additionally demonstrated for the C-terminal domain of PCID2 in our experiments. The interaction of the truncated PCID2 (PCID2 1-360 ) with Xmas-2 was studied in D. melanogaster S2 cells. PCID2 1-360 was fused with the HA epitope, Xmas-2 was fused with the FLAG epitope, and the two proteins were coexpressed in S2 cells. The results of protein coimmunoprecipitation from a cell lysate are shown in Figure 3C. PCID2 1-360 did not interact with Xmas-2 in the experiment ( Figure 3C, right panel). At the same time, the full-size PCID2 and Xmas-2 interacted with each other ( Figure 3C, left panel). Thus, the domain utilized by D. melanogaster PCID2 to interact with Xmas-2 is the same that the domain utilized by its yeast homolog [7]. Still, the PCID2 interaction with the specific RNA fragment was shown to be independent of Xmas-2 in our experiments.

There Is an Additional RNA Interaction Region in PCID2
Because the deletion of the C-terminal domain of PCID2 did not fully abolish its interaction with RNA, additional RNA-binding sites were assumed for PCID2. To identify the respective regions, we engineered the N-PCID2, M-PCID2, and C-PCID2 constructs, in which the His epitope was fused with the approximately equal (about 130 amino acids) N-terminal, middle, and C-terminal fragments of PCID2, respectively. A fragmentation scheme is shown in Figure 4A. N-PCID2 corresponded to the N-terminal PCID2 part, which does not contain known domains. M-PCID2 included the N part of the PCI domain, while C-PCID2 corresponded to the C-terminal part of PCI domain, including the WH domain, which contains an RNA-binding motif. The PCID2 fragments were tested for interactions with ras2 RNA fragment 4_2 by EMSA ( Figure 4B). The C-PCID2 was found to interact with ras2 fragment 4_2 (K D~6 0 nM), which is in line with the results shown above ( Figure 3B). N-PCID2 did not virtually bind with RNA. We observed the strong interaction with the ras2 RNA for the M-PCID2 fragment (K D~1 1 nM). M-PCID2 formed several contacts with fragment 4_2 as could be seen from the migration pattern of the RNA-protein complex in gel. Thus, the N-terminal part of the PCI domain harbors an additional region responsible for RNA binding.
Then, we checked whether the interaction of M-PCID2 with RNA is specific. The PCID2 regions were tested for interactions with ras2 RNA fragment 4_1, which does not interact with the full-size PCID2, as reported above. N-PCID2 and C-PCID2 were not found to interact with fragment 4_1 ( Figure 4C). However, M-PCID2 bound the fragment with a considerable efficiency, with K D of about 9.2 nM ( Figure 4C). The M region of PCID2 (M-PCID2) was therefore concluded to nonspecifically bind with RNA ( Figure 4C).

Point Mutations of the M Domain Abolish PCID2 Binding with RNA
Then, we studied the role of the M region in the interaction of the full-size PCID2 with RNA. Conserved amino acid residues potentially responsible for the interaction with RNA were identified in the M domain by evolutionary sequence alignment of PCID2 homologs of various organisms. Two conserved residues, Arg191 and Arg216, were of particular interest according to the prediction of RNA binding sites with the online tools PRIdictor [37] and aaRNA [38]. Two PCID2 mutants fused with the GST epitope were constructed ( Figure 5A): Ala was substituted for Arg191 in one of them (PCID2 R191A ) and for Arg216 in the other (PCID2 R216A ). PCID2 R191A and PCID2 R216A did not bind ras2 RNA fragment 4_2 in contrast to the wild-type PCID2 ( Figure 5B). Thus, the mutations of the M domain fully distorted the interaction of the full-size PCID2 with RNA. Moreover, the results indicate that the C-terminal domain does not interact with RNA in the absence of RNA interactions with the M domain. One may assume that RNA binding with the M region of PCID2 is necessary for RNA binding with its C-terminal domain.  GST alone was added to the binding reaction at 125 nM. The complexes were analyzed by native 5% PAGE. Brackets indicate the major high-order gel-shifted complexes. Arrowheads show the migration of free RNAs. (C) [ 32 P] UTP-labelled, in vitro-transcribed ras2 mRNA fragment 4_1 (25 nM) was incubated with increasing amounts (0, 2.5, 5, 10, 15, 20, 25, and 50 nM) of purified N-PCID2-His (left panel), M-PCID2-His (middle panel), or C-PCID2-His (right panel). GST alone was added to the binding reaction at 125 nM. The complexes were analyzed by native 5% PAGE. A bracket indicates the major high-order gel-shifted complexes. Arrowheads show the migration of free RNAs.

Point Mutations of the M Domain Abolish PCID2 Binding with RNA
Then, we studied the role of the M region in the interaction of the full-size PCID2 with RNA. Conserved amino acid residues potentially responsible for the interaction with RNA were identified in the M domain by evolutionary sequence alignment of PCID2 homologs of various organisms. Two conserved residues, Arg191 and Arg216, were of particular interest according to the prediction of RNA binding sites with the online tools PRIdictor [37] and aaRNA [38]. Two PCID2 mutants fused with the GST epitope were constructed ( Figure 5A): Ala was substituted for Arg191 in one of them (PCID2 R191A ) and for Arg216 in the other (PCID2 R216A ). PCID2 R191A and PCID2 R216A did not bind ras2 RNA fragment 4_2 in contrast to the wild-type PCID2 ( Figure 5B). Thus, the mutations of the M domain fully distorted the interaction of the full-size PCID2 with RNA. Moreover, the results indicate that the C-terminal domain does not interact with RNA in the absence of RNA interactions with the M domain. One may assume that RNA binding with the M region of PCID2 is necessary for RNA binding with its C-terminal domain.

Discussion
In this work, the RNA-binding domains of Drosophila PCID2 were characterized. We showed that the association of D. melanogaster PCID2 with RNA may occur independently of the other components of the TREX-2 complex. The 34-residue C-terminal fragment of the PCID2 WH domain specifically interacts with the 3 -noncoding region of the ras2 mRNA. The fragment corresponds to a region of yeast Thp1, which has been shown to interact with Sac3, and together they form a platform for binding with RNA [7]. Our data show that the C-terminal fragment of PCID2 participates in the interaction with Xmas-2, a homolog of Sac3. However, in contrast to yeast Thp1, Drosophila PCID2 does not require Xmas-2 to interact with RNA. These findings agree with the published data that the yeast PCID2 homolog Thp1 may interact with RNA directly, without the involvement of other subunits of the complex [29]. The findings are consistent as well with the results of our previous study, which has shown that PCID2 occurs in the cytoplasm of S2 cells, where it is involved in cytoplasmic mRNA transport and is associated with mRNA independently of the other TREX-2 subunits [14].
We localized the ras2 mRNA nucleotide sequence to which PCID2 binds specifically. An association of Drosophila TREX-2 subunits with the ras2 mRNA has previously been demonstrated by coimmunoprecipitation [14,19]. In this work, we showed the direct binding of PCID2 with the ras2 RNA and localized the binding sites to 200 nt of the 3 -noncoding region of the ras2 mRNA. It is possible to assume that a certain secondary structure forms in the region and that the structure may occur in other mRNAs as well. The issue needs further investigation.
Here, we found the additional RNA-interacting region of PCID2. The region is in the N-terminal part of the PCI domain and binds RNA nonspecifically (M region). Based on the EMSA results, a lack of the C-terminal domain partly distorts the mRNA binding, while mutations of the middle part completely abolish the mRNA binding. The finding makes it possible to conclude that the interaction of the M region with RNA is necessary for the specific RNA binding with the C-terminal region. Primary RNA binding can be assumed to occur in the M region (because its binding with RNA is highly efficient) and to allow the C-terminal domain to specifically interact with the RNA (because its binding with RNA is less efficient). The M region of PCID2 is highly conserved in evolution (up to 91% homology in higher eukaryotes). A similar role is therefore possible to assume for the respective sequences of the PCID2 homologs of other organisms.
AlphaFold [33,34] and APBS [36] analyses of the M region structure and properties showed that the M region forms a positively charged groove, a possible surface for RNAinteracting protein domains ( Figure 6A). The highly evolutionarily conservative amino acids mutated in our experiments are critical for RNA-binding activity of the positively charged groove ( Figure 6B). The confidence of the PCID2 model is represented on Figure S2. The yeast Thp1 domain corresponding to the M region has been shown to interact with the Sem1p protein, and the Thp1-Sem1p complex does not interact with RNA [7]. It can be assumed that the role of Sem1 in the TREX-2 complex is to prevent PCID2 binding with mRNA and thus to regulate the TREX-2-mRNA interaction. The structure was accessed through the AlphaFold Protein Structure Database, which has been created in partnership with EMBL's European Bioinformatics Institute (EMBL-EBI). Electrostatics analysis was performed using APBS software. Molecular graphics were performed with UCSF ChimeraX, which has been developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from the National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases.

Supplementary Materials:
The following supporting information can be downloaded at: www.mdpi.com/xxx/s1, Figure S1: EMSA analysis of the Drosophila PCID2 interactions with the ras2 fragment 4_2; Figure S2: Confidence of the PCID2 model.  Database, which has been created in partnership with EMBL's European Bioinformatics Institute (EMBL-EBI). Electrostatics analysis was performed using APBS software. Molecular graphics were performed with UCSF ChimeraX, which has been developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from the National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.