dZC3H3 is required for nuclear export of bulk poly(A) RNA and specific messages. (A) S2R+ cells depleted of dZC3H3 (bottom) or GFP (top) were fixed, permeabilized, and analyzed for RNA subcellular localization using Alexa Fluor 488–conjugated wheat germ agglutinin (green) to mark the nuclear rim and either a Cy3-OdT probe (left), HSP70-specific probes (middle), or an HSP83-specific probe (right; red). (B) Schematic of the dZC3H3 protein, its zinc finger domains (orange), composition of the v5-tagged deletion mutants (green), and sequences targeted by dsRNAs (red). (C) Overexpression of the C terminus of dZC3H3 (C-v5) partially rescues the poly(A) RNA nuclear export defect, whereas expression of the N terminus (N-v5) exacerbates the accumulation phenotype. The bar graph indicates the fraction of transfected cells displaying strong, weak, or no nuclear poly(A) RNA accumulation upon depletion of endogenous dZC3H3 or GFP and overexpression of various deletion mutants or GAPDH1-v5. (D) Representative images scored in C of S2R+ cells processed for poly(A) RNA localization (red) and immunofluorescence (v5; green) after being both depleted of endogenous dZC3H3 or GFP (labeled at left) and transfected with various v5-tagged deletion mutants (TFXN; labeled at right). Error bars indicate SDs of results between three experimental replicates. Bars, 5 µm.

dZC3H3's N terminus mediates a strong interaction with NXF1. (A) GFP-NXF1 interacts to different degrees by co-IP with all three v5-tagged dZC3H3 deletion mutants (Fig. 1) tested. (B) The bar graph illustrates relative interaction strengths of GFP-NXF1 with the v5-tagged deletion mutants as given by a ratio of IP signal to input signal for each reaction compared with that for full-length (FL) dZC3H3. INP, input; WB, Western blot. Error bars indicate SDs of results between three experimental replicates.

dZC3H3 interacts in a complex of export factors and with poly(A) RNA. (A) Silver-stained gel of final eluate samples from S2R+ cells that stably express TAP-dZC3H3 or TAP vector alone illustrate proteins that specifically purify with dZC3H3 (left) and not in control (right). The position of the TAP-tagged dZC3H3 bait is marked with a red dot. (B) Gene ontology analysis and domain searching reveals that 19 out of the 28 dZC3H3 interactors interact with or are predicted to interact with nucleic acids and 18 specifically with RNA. (C) Interactions of PABP2-v5 and swm-v5 with GFP-dZC3H3 are partially RNA dependent. PABP2-v5, swm-v5, or l(1)10Bb-v5 (as a negative control) were cotransfected with GFP-dZC3H3 and immunoprecipitated using v5-agarose resin in the presence or absence of RNase. (D) GFP-dZC3H3, swm-v5, NXF1, and PABP2 interact with poly(A) RNA. Nuclear poly(A) RNA was purified under nondenaturing conditions using oligo-dT cellulose resin from S2R+ cells transfected (TFXN) with GFP-dZC3H3, swm-v5, or GFP-v5 alone and analyzed by immunoblotting for associating proteins. INP, input; WB, Western blot.

Transcript hyperadenylation results upon dZC3H3 depletion. (A) Total poly(A) RNA from cells depleted of various export factors or GFP (control) was harvested, 3′ end labeled, and digested. The sizes of remaining poly(A) tails were analyzed by TBE Urea PAGE and autoradiography. Lengths of the adenylate (A) tails were assessed by comparison to an RNA marker at left. (B) The degree of hyperadenylation was determined by calculating the percentage of total poly(A) tail signal residing above 200 nucleotides in each sample. Error bars indicate SDs of results between two experimental replicates.

Function of ZC3H3 in mRNA nuclear export is conserved in humans. (A) U2OS cells were treated with control (ctrl) siRNAs or siRNAs targeting NXF1 or ZC3H3 for 3 d and fixed, permeabilized, hybridized with Cy3-OdT (left) to visualize poly(A) RNA, and stained for DNA (right). Images were taken using equivalent exposure times in the Cy3 channel (poly(A) RNA) to enable comparison of signal level between samples. Note the saturation of nuclear pixels in NXF1- and ZC3H3-depleted cells that results from the use of exposure times long enough to visualize poly(A) RNA signal in control cells. Poly(A) localization was further inspected in control cells (B), NXF1-depleted cells (C), and ZC3H3-depleted cells (D). (E) RPL32 mRNA and poly(A) RNA was visualized by hybridization of control, NXF1-, or ZC3H3- depleted cells with Cy3-RPL32 (red) and biotin-OdT (green) probes. Images within the white boxes are magnified at right. Bars, 20 µm.

ZC3H3 modulates mRNA polyadenylation downstream of PABPN1 activity. (A) U2OS cells were monitored for poly(A) RNA localization by hybridization with Cy3-OdT probe (left) as in Fig. 5 after depletion with either control siRNAs (ctrl) or siRNAs targeting ZC3H3, PABPN1, or both. Images acquired in the Cy3 channel (poly(A) RNA) were captured using equivalent exposure times to allow comparison of signal intensity between samples. (B) Quantification of the percentage of cells possessing nuclear poly(A) signal over background for cells treated as in A illustrates that codepletion of ZC3H3 and PABPN1 results in a nuclear accumulation phenotype indistinguishable from control. (C) Immunoblot demonstrating efficient depletion of siRNA-targeted factors in A. A nonspecific protein interaction occurs with anti-ZC3H3 antibodies (asterisk) in all samples, indicating equivalent protein loading between lanes. ZC3H3-11 is a second siRNA targeting ZC3H3 that was used to deplete cells. (D) A change in the localization of nuclear poly(A) RNA (Cy3-OdT, red) with respect to nuclear speckles (SC35, green) is resultant in ZC3H3-, PABPN1-, and ZC3H3/PABPN1-depleted samples as compared with control cells. (E) PABPN1 (green) colocalizes with poly(A) RNA (red) in both control and ZC3H3-depleted cells. Images in D and E were captured using confocal microscopy. Bars, 20 µm.

ZC3H3-dependent nuclear poly(A) RNA accumulations respond to transcriptional inhibition. Confocal images of U2OS cells that were treated with control (ctrl) siRNAs or siRNAs targeting ZC3H3 for 3 d and incubated with or without 50 ug/ml α-amanitin (Am) for 6 h before fixation, permeabilization, hybridization with Cy3-OdT probe, and immunostaining with anti-SC35 antibodies. Merged images within white boxes are enlarged at the top right to enable better resolution of features. Bar, 20 µm.

Model of potential roles of ZC3H3 in polyadenylation and nuclear export. RNA processing factors (light gray) are recruited to mRNAs during transcription by RNA polymerase II (red). During the processive stage of polyadenylation, CPSF (dark gray) and PABP2 (green) tether PAP (blue) to the elongating tail. ZC3H3 (orange) may aid in the recognition of tails of the proper length, trim the tail to the proper length, or release polyadenylation factors (such as PAP) after the proper length has been attained. This activity could additionally aid in the release of the poly(A) transcript from the transcription focus, thereby allowing it to proceed toward the nuclear periphery for export via interactions with export factors such as NXF1 (purple). For simplicity, ZC3H3 is shown to interact with the mRNP after the initiation of transcript polyadenylation. However, it is possible that this factor is recruited to transcripts before polyadenylation but functions at a later stage in mRNA biogenesis.