hnRNP K binds a core polypyrimidine element in the eukaryotic translation initiation factor 4E (eIF4E) promoter, and its regulation of eIF4E contributes to neoplastic transformation - PubMed (original) (raw)

hnRNP K binds a core polypyrimidine element in the eukaryotic translation initiation factor 4E (eIF4E) promoter, and its regulation of eIF4E contributes to neoplastic transformation

Mary Lynch et al. Mol Cell Biol. 2005 Aug.

Abstract

Translation initiation factor eukaryotic translation initiation factor 4E (eIF4E) plays a key role in regulation of cellular proliferation. Its effects on the m7GpppN mRNA cap are critical because overexpression of eIF4E transforms cells, and eIF4E function is rate-limiting for G1 passage. Although we identified eIF4E as a c-Myc target, little else is known about its transcriptional regulation. Previously, we described an element at position -25 (TTACCCCCCCTT) that was critical for eIF4E promoter function. Here we report that this sequence (named 4EBE, for eIF4E basal element) functions as a basal promoter element that binds hnRNP K. The 4EBE is sufficient to replace TATA sequences in a heterologous reporter construct. Interactions between 4EBE and upstream activator sites are position, distance, and sequence dependent. Using DNA affinity chromatography, we identified hnRNP K as a 4EBE-binding protein. Chromatin immunoprecipitation, siRNA interference, and hnRNP K overexpression demonstrate that hnRNP K can regulate eIF4E mRNA. Moreover, hnRNP K increased translation initiation, increased cell division, and promoted neoplastic transformation in an eIF4E-dependent manner. hnRNP K binds the TATA-binding protein, explaining how the 4EBE might replace TATA in the eIF4E promoter. hnRNP K is an unusually diverse regulator of multiple steps in growth regulation because it also directly regulates c-myc transcription, mRNA export, splicing, and translation initiation.

PubMed Disclaimer

Figures

FIG. 1.

Oligonucleotides used in EMSA experiments.

FIG. 2.

4EBE can replace the TATA box in the E1b promoter. (A) The indicated promoter constructs were transfected into HeLa cells and analyzed for CAT activity as described in Materials and Methods. The schematic diagram identifies pBLCAT3 (no element), E1bTATACAT (TATA), and 4EBE CAT (4EBE). The CAT reporter activity is represented as the fold activation compared to CAT alone. All values are reported as the mean increase ± the standard error for three replicas in this and the subsequent panels. (B) The diagrammed activator/basal element promoter constructs were transfected into HeLa cells and analyzed for CAT activity. Constructs are illustrated in the center and included plasmids containing two Sp1 sites upstream of CAT or 10 bases upstream of TATA, 4EBE, or 4EBE μ in the E1bCAT plasmid [(Sp1)2CAT, (Sp1)210 TATACAT, (Sp1)210 4EBE CAT, and (Sp1)210 4EBEmutCAT]. The CAT reporter activity is represented as the fold activation compared to the CAT alone value taken from panel A. (C) Representative experiment showing the effect of distance between Sp1 and TATA or Sp1 and 4EBE on transcriptional activity. Briefly, exponentially growing HeLa cells in 100-mm dishes were transiently transfected with 10 μg of CAT reporter plasmids containing two Sp1 sites 10 or 30 bases upstream of TATA or 4EBE [(Sp1)210 TATACAT, (Sp1)230TATACAT, (Sp1)210 4EBE CAT, and (Sp1)230 4EBE CAT]. (D) Representative experiment showing the effect of distance between the E-box element and TATA compared to the E-box element and 4EBE on transcriptional activity. Briefly, exponentially growing HeLa cells in 100 mm dishes were transiently transfected with 10 μg of CAT reporter plasmids containing four E-box elements 10 or 43 bases upstream of TATA or 4EBE [(E-box)410 TATACAT, (E-box)443 TATACAT, (E-box)410 4EBECAT, and (E-box)443 4EBE CAT], along with 2 μg of human growth hormone plasmid as an internal control. Transfected cells were harvested 48 h posttransfection and assayed for CAT activity in all experiments.

FIG. 3.

Purification of the 68-kDa protein binding to the 4EBE by DNA affinity chromatography identifies hnRNP K. (A) Hydrophobic exchange chromatography was performed as the initial step in protein purification. Fractions identifying the 68-kDa species binding to the 4EBE were identified by Southwestern analyses as described in Materials and Methods. Lane 1 contains a positive control HeLa nuclear extract (H), and lane 2 contains an aliquot of the HeLa extract (L) before loading on the column. (B) The second step in purification used anion exchange chromatography. Fractions identifying the 68-kDa species binding to the 4EBE were identified by Southwestern analyses as described in Materials and Methods. The positive control HeLa extract was again included for comparison (H). (C) Two sequential bindings to a DNA affinity chromatography matrix using a trimeric 4EBE sequence purified a 68-kDa band as demonstrated in this brilliant blue-stained acrylamide gel. Size markers (M, lane 1), an aliquot of the positive fraction from the hydrophobic exchange column (A, lane 2), an aliquot of the positive fraction from the monoS column (S, lane 3), and an aliquot of the DNA affinity wash (W, lane 5) are compared to the protein bound to the 4EBE-DNA affinity matrix (D) shown in lane 4. (D) The 68-kDa protein band was submitted for sequencing and the indicated amino acid sequences were identified. These correspond uniquely to hnRNP K. (E) Bacterially synthesized gst-hnRNP K was run in a standard Southwestern analysis, and the trimeric 4EBE probe bound to the gst-hnRNP K (H) but not to glutathione _S_-transferase alone (G). This band was confirmed to be hnRNP K by an immunoblot analysis with anti-hnRNP K. Finally, the gel was also stained with brilliant blue to demonstrate loading of the species tested for binding to the 4EBE.

FIG. 4.

Nucleotides in 4EBE sufficient and necessary for binding and transcription. (A) Diagram of 4EBE and comparison to promoter sites known to interact with hnRNP K (76) (12, 21, 51, 87, 95). hnRNP K binding sites are highlighted in promoter sequences previously shown to be regulated by hnRNP K. These are further compared to the 4EBE from human, rat, and mouse eIF4E promoters. (B) Schematic of mutations made in 4EBE to evaluate the function of the polypyrimidine tract shared between the 4EBE and known hnRNP K binding sites. (C) Representative EMSA depicting the nucleotides in 4EBE (5′-CTCTTACCCCCCCTT) sufficient for binding. Radioactively labeled 4EBE oligonucleotide shows two binding activities in an EMSA (lane 1 [arrows identify binding activity]). The binding to 4EBE is competed by adding 500-fold excess of cold wild-type 4EBE oligonucleotide (lane 2) or 500-fold excess of cold 4EBE mutant oligonucleotides μa, μb, μc, and μg (lanes 3, 4, 5, and 9, respectively). (D) Representative experiment correlating binding to 4EBE with transcriptional activity in transient-transfection assays in HeLa cells. Briefly, mutations in the 4EBE sequence (μd, μe, and μf) that did not show binding in EMSA were generated in the native 4E promoter, as described in Materials and Methods. The CAT reporter plasmids for these mutants were subsequently transfected into exponentially growing HeLa cells, along with the human growth hormone as an internal control for transfection efficiency. The cells were harvested 48 h posttransfection and assayed for CAT activity as described in the text.

FIG. 5.

The endogenous binding activities to the 4EBE and the myc CTE are indistinguishable in EMSAs. (A) We compared DNA binding to the eIF4E polypyrimidine element with binding to the myc CTE element first described as the DNA-binding site for hnRNP K (95). The sequences of the DNA oligonucleotides tested in this experiment are labeled as A to D. We compared two configurations of the 4EBE (sequences A and C) with identical configurations of the myc hnRNP K CTE binding site (sequences B and D). We compared a trimer of the eIF4E CT element to a complete myc CT element of the identical length (sequences A and B, respectively). We further compared dimers of each individual element separated by the myc CTE optimal distance (sequences C and D). The CT repeats in oligonucleotides A and B are highlighted by underlining and overlining. The CT elements in oligonucleotides C and D are highlighted by underlining and italics. (B) EMSA findings with HeLa whole-cell extracts bound to labeled oligonucleotide A. Binding was competed for by the indicated trimeric oligonucleotides A or B at the indicated fold excess concentrations. Finally, we compared competition with double-stranded oligonucleotides (ds) or single-stranded oligonucleotides (ss). (C) EMSA findings with HeLa whole-cell extracts bound to labeled oligonucleotide A. Binding was competed for by the indicated dimeric oligonucleotides C or D at the indicated fold excess concentrations. We further compared competition with double-stranded oligonucleotides (ds) or single-stranded oligonucleotides (ss).

FIG. 6.

Chromatin immunoprecipitation demonstrates binding of hnRNP K to the eIF4E promoter. Cell lysates from formaldehyde-treated TGR cells in the indicated growth conditions were immunoprecipitated with nonimmune rabbit serum (lane 2) or with anti-hnRNP K antibody that was coincubated with (lane 4) or without (lane 3) the peptide used to generate the antibody in rabbits (99). DNA was purified from the eluted complex and used as a template in PCRs with primers to a nontranscribed rat locus (AC120715), the rat c-myc promoter (c-myc), the eIF4E promoter (eIF4E), and the 5S rRNA region (5S). PCR products were separated by agarose gel electrophoresis and visualized with ethidium bromide. Input DNA was used as a positive control in lane 1.

FIG. 7.

Overexpression of hnRNP K and hnRNP K knockdowns regulate endogenous eIF4E mRNA. (A) Rat 1A cells were transfected with vector alone, myc, hnRNP K, or the combination of hnRNP K and myc. Pooled transfectants were selected in neomycin and hygromycin and grown to confluence. mRNA was harvested, and 5 μg was run in a Northern analysis. The same blot was probed for the indicated genes, including c-myc, hnRNP K, eIF4E, and actin. The probes used for the c-myc and hnRNP K blots were the human full-length cDNAs used to construct the expression vectors to increase the specificity of detection of the transfected sequences. The actin and eIF4E probes were mouse full-length cDNA probes. (B and C) TGR (myc+/+) (B) and HO15 (myc−/−) (C) cells were transiently transfected with an siRNA oligonucleotide for hnRNP K. Expression of hnRNP K protein and eIF4E mRNA were analyzed 48 h after transfection in confluent transfected cells. Loading controls include actin protein and the rRNA bands on an ethidium bromide-stained gel. The levels of hnRNP K protein, myc protein, and eIF4E mRNA were compared in cells transfected with an siRNA for hnRNP K (si) to those transfected with a scrambled siRNA (sc).

FIG. 8.

Overexpression of hnRNP K transforms cells and enhances passage through G1. (A) eIF4E protein levels were compared between Rat1A cells transfected with hnRNP K and cells transfected with an expression construct for eIF4E (55). Vector control transfected cells (C1) for the hnRNP K transfectants, the hnRNP K transfectants (HN), vector control transfected cells for the eIF4E transfectants (C2), and the eIF4E transfectants (4E) are shown. Immunoblots for eIF4E and actin are shown. (B) The cells presented in panel A were then evaluated for transformation in a standard soft agar assay. Plotted is the mean and standard error of the fold change in cells per well for six wells each in four separate repetitions of the experiment comparing transfected constructs to vector control cells for each repetition (n = 24 for each plot). The hnRNP K and eIF4E transfectants are indicated. (C) Rat1A cells transfected with the indicated constructs were analyzed by fluorescence-activated cell sorting for DNA content. Plotted is the percentage of cells in the G1, S, and G2 phases for three separate determinations for each of the two times the experiment was repeated, together with the standard error of the mean for these determinations. Cells studied were transfected with the indicated expression vectors, including the vector control (vec), myc, and hnRNP K. These determinations were made for actively proliferating subconfluent cells. (D) Myc−/− cells were infected with retroviruses expressing either GFP or hnRNP K. The infected cells were grown to confluence and were then growth arrested over 48 h by removing serum from the culture medium. DNA synthesis was measured as described in Materials and Methods in the subsequent absence (arrested) or presence (stimulated) of serum for 20 h. hnRNP K infection alone did not enhance DNA synthesis in the growth arrested cells but significantly increased DNA synthesis after cells were stimulated to initiate division.

FIG. 9.

Overexpression of hnRNP K alters translation rates. (A) Polysomal profiles of Rat 1A cells transfected with hnRNP K (dark black line) were compared those containing an empty vector control (gray line). Cells were grown to confluence for 48 h in medium lacking serum, and cytoplasmic extracts were run on standard sucrose gradients to separate mRNAs in monosomal versus polysomal fractions. The y axis identifies UV absorption in relative units, and the x axis identifies fractions of the gradient in relative time from least dense on the left to most dense on the right. (B) Polysomal profiles of myc−/− cells infected with pBABE-hnRNP K (dark black line) were compared to those infected with pBABE-GFP (lighter gray line). Cells were again evaluated at confluence in the absence of serum stimulation. Axes are as described in panel A. We further indicate the mean and standard error of the percentage of RNAs found in the subpolysomal versus polysomal fractions for two replicas for each analysis. (C) To confirm the global patterns of polysomal fractionation, we then analyzed the fractionation of reporter genes coupled to the translationally regulated leader sequence of ribosomal protein L32. We compared a reporter construct containing the 5′TOP from ribosomal protein L32 (rpL32) coupled to a GFP cDNA with a reporter construct containing the inactivated TOP sequence from rpL32 coupled to a growth hormone cDNA. These two constructs were transfected into Rat1A cells in the absence or presence of pcDNA-hnRNP K. The cells were grown to confluence over 72 h and serum was withdrawn for the last 48 h before harvest. Polysomal (P) and subpolysomal (S) pooled fractions were isolated by using sucrose density gradients. The mRNAs contained in each pool were isolated from the cells transfected either with (hnRNP K+) or without (pcDNA+) the hnRNP K expression vector. The pooled polysomal and subpolysomal RNAs were then blotted in standard RNA blots and probed for the GFP (rpl32) and growth hormone (μrpl32) reporters. The ratio of the intensity of the signal in the polysomal fraction (P) compared to the subpolysomal fraction (S) demonstrates increased translation initiation rates. (D) Subconfluent, proliferating Rat 1A cells from Fig. 7 transfected with eIF4E and hnRNP K were harvested, and the protein content per cell was measured for each construct.

FIG. 10.

hnRNP K transforms cells through effects on eIF4E expression. (A) Rat1A cells were transfected with a control RNA (CR) and two siRNAs for eIF4E (E1 and E2) as described in Materials and Methods. Protein extracts were made 48 h after transfection and evaluated for eIF4E and actin levels by immunoblotting. An untransfected control lane is also included (lane C). (B) Rat1A cells expressing hnRNP K from Fig. 7 above were transfected with the control siRNA and the two siRNAs from panel A. Additional hnRNP K-expressing cells were examined without transfecting them (control). At 48 h after transfection the cells were seeded for a soft agar assay. The mean and standard error of colonies per well, evaluated 8 days after the cells were put in soft agar for each condition, were plotted. (C) Rat1A transfectants were evaluated for transformation in a standard soft agar assay. Shown are cells transfected with a vector control, a dominant inhibitor of eIF4E function (4EBPμ), hnRNP K, eIF4E, the combination of the dominant inhibitor of eIF4E, together with hnRNP K, and the combination of eIF4E, together with the dominant inhibitor of eIF4E, as indicated in the figure. Plotted is the mean and standard deviation of the cells per well for four wells each in two separate repeats of the experiment.

FIG. 11.

Polypyrimidine elements in mammalian translation initiation factor promoters. The genomic sequences of thirty-eight human translation initiation factor genes (eIFs) were identified by using LocusLink (104). Five thousand nucleotides 5′ to and 3′ to the transcription initiation site of each gene were downloaded into Clone Manager Suite 7 for further manipulation (Scientific and Educational Software, Durham, N.C.). The immediate promoter regions of all of these sequences were inspected for sequences containing stretches of eight or more pyrimidines that were less than 20 nucleotides from the transcription initiation site or the translation initiation codon. Thirteen human eIFs revealed such sequences. Two thousand nucleotides surrounding the transcription initiation site of the 13 candidates were then used in a BLAST search of the mouse and rat genomes. Promoter regions were considered to be confirmed if this region showed >70% homology, and it fell within a calculated CpG island in all three species. The polypyrimidine stretches were conserved between mouse, rat, and human genomes in nine of the confirmed promoter candidates as shown. Listed are the 70 nucleotides containing the proximal promoter sequences from all three species for the nine candidates containing conserved polypyrimidine stretches, and the polypyrimidine stretches are identified by gray highlighting. The transcription initiation sites identified in the genome databases were confirmed or modified by a BLAST comparison of each genomic sequence against available expressed sequence tag sequences for each species. Indicated as white letters against a black background are either the site listed in LocusLink as the transcription initiation site, the site mapping of the 5′ end of the majority of all of the available ests for the indicated species, or the 5′-most expressed sequence tag sequence if no consensus 5′ end was obvious. The translation initiation codon (ATG) is identified for the eight candidates where it is positioned close to the transcription initiation site. Myc binding sites (

CACGTG

) were identified by locating Pm1I sites in each promoter and are shown in boldface, underlined italics with gray highlighting. Remarkably, Myc binding sites were located at extremely short distances from the initiation site of seven of the nine candidate promoters; four were within 50 nucleotides, two were at about 150 nucleotides, and one was about 500 nucleotides distant. A nuclear respiratory factor 1α palindromic sequence (NRF-1=α PAL) binding site that also binds max and may be an alternative Myc binding site is identified in the eIF2S2 promoters. LocusLink identifications include eIF2S2 (human 8894, mouse 67204, and rat 296302), eIF2S3 (1968, 26905, and 299027), eIF3S1 (8669, 78655, and 311371), eIF3S3 (8667, 68135, and 299899), eIF3S5 (8665, 66085, and 293427), eIF3S6 (3646, 16341, and 299872), eIF3S10 (8661, 75705, and 300253), eIF4B (1975, 75705, and 300253), and eIF4E (1977, 13684, and 117045). The eIF3S1 mouse promoter sequences are assumed based on their homology to the rat and human sequences. We found no mouse cDNA or EST sequence that matched this sequence, and it was not annotated as the promoter sequence in the genome database.

Cited by

Targeting the eIF4E/β-catenin axis sensitizes cervical carcinoma squamous cells to chemotherapy.
Xu H, Wang Z, Xu L, Mo G, Duan G, Wang Y, Sun Z, Chen H. Xu H, et al. Am J Transl Res. 2017 Mar 15;9(3):1203-1212. eCollection 2017. Am J Transl Res. 2017. PMID: 28386346 Free PMC article.
The Eµ-hnRNP K Murine Model of Lymphoma: Novel Insights into the Role of hnRNP K in B-Cell Malignancies.
Malaney P, Velasco-Estevez M, Aguilar-Garrido P, Aitken MJL, Chan LE, Zhang X, Post SM, Gallardo M. Malaney P, et al. Front Immunol. 2021 Apr 12;12:634584. doi: 10.3389/fimmu.2021.634584. eCollection 2021. Front Immunol. 2021. PMID: 33912162 Free PMC article. Review.
Ezrin Binds to DEAD-Box RNA Helicase DDX3 and Regulates Its Function and Protein Level.
Çelik H, Sajwan KP, Selvanathan SP, Marsh BJ, Pai AV, Kont YS, Han J, Minas TZ, Rahim S, Erkizan HV, Toretsky JA, Üren A. Çelik H, et al. Mol Cell Biol. 2015 Sep;35(18):3145-62. doi: 10.1128/MCB.00332-15. Epub 2015 Jul 6. Mol Cell Biol. 2015. PMID: 26149384 Free PMC article.
c-myc Repression of TSC2 contributes to control of translation initiation and Myc-induced transformation.
Ravitz MJ, Chen L, Lynch M, Schmidt EV. Ravitz MJ, et al. Cancer Res. 2007 Dec 1;67(23):11209-17. doi: 10.1158/0008-5472.CAN-06-4351. Cancer Res. 2007. PMID: 18056446 Free PMC article.

References

1. Atchison, M. L., O. Meyuhas, and R. P. Perry. 1989. Localization of transcriptional regulatory elements and nuclear factor binding sites in mouse ribosomal protein gene rpL32. Mol. Cell. Biol. 9:2067-2074. - PMC - PubMed
1. Azizkhan, J. C., D. E. Jensen, A. J. Pierce, and M. Wade. 1993. Transcription from TATA-less promoters: dihydrofolate reductase as a model. Crit. Rev. Eukaryot. Gene Expr. 3:229-254. - PubMed
1. Bergemann, A. D., Z. W. Ma, and E. M. Johnson. 1992. Sequence of cDNA comprising the human pur gene and sequence-specific single-stranded-DNA-binding properties of the encoded protein. Mol. Cell. Biol. 12:5673-5682. - PMC - PubMed
1. Blackwell, T. K., J. Huang, A. Ma, L. Kretzner, F. W. Alt, R. N. Eisenman, and H. Weintraub. 1993. Binding of myc proteins to canonical and noncanonical DNA sequences. Mol. Cell. Biol. 13:5216-5224. - PMC - PubMed
1. Bomsztyk, K., O. Denisenko, and J. Ostrowski. 2004. hnRNP K: one protein multiple processes. Bioessays 26:629-638. - PubMed

hnRNP K binds a core polypyrimidine element in the eukaryotic translation initiation factor 4E (eIF4E) promoter, and its regulation of eIF4E contributes to neoplastic transformation - PubMed (original) (raw)