Identification of the SRC pyrimidine-binding protein (SPy) as hnRNP K: implications in the regulation of SRC1A transcription

Abstract

The human SRC gene encodes pp60(c-src), a non-receptor tyrosine kinase involved in numerous signaling pathways. Activation or overexpression of c-Src has also been linked to a number of important human cancers. Transcription of the SRC gene is complex and regulated by two closely linked but highly dissimilar promoters, each associated with its own distinct non-coding exon. In many tissues SRC expression is regulated by the housekeeping-like SRC1A promoter. In addition to other regulatory elements, three substantial polypurine:polypyrimidine (TC) tracts within this promoter are required for full transcriptional activity. Previously, we described an unusual factor called SRC pyrimidine-binding protein (SPy) that could bind to two of these TC tracts in their double-stranded form, but was also capable of interacting with higher affinity to all three pyrimidine tracts in their single-stranded form. Mutations in the TC tracts, which abolished the ability of SPy to interact with its double-stranded DNA target, significantly reduced SRC1A promoter activity, especially in concert with mutations in critical Sp1 binding sites. Here we expand upon our characterization of this interesting factor and describe the purification of SPy from human SW620 colon cancer cells using a DNA affinity-based approach. Subsequent in-gel tryptic digestion of purified SPy followed by MALDI-TOF mass spectrometric analysis identified SPy as heterogeneous nuclear ribonucleoprotein K (hnRNP K), a known nucleic-acid binding protein implicated in various aspects of gene expression including transcription. These data provide new insights into the double- and single-stranded DNA-binding specificity, as well as functional properties of hnRNP K, and suggest that hnRNP K is a critical component of SRC1A transcriptional processes.

Figure 1

Ultraviolet cross-linking analysis of a SPy-TC1 double-stranded oligonucleotide complex. (A) The SRC1A promoter region contains two critical Sp-family member DNA-binding sites, GC1 and GA2, as well as three TC tracts, TC1, TC2 and TC3. SPy double-stranded binding sites located within TC1 and TC2 are shaded gray. The major transcription start sites are located just downstream of TC3 and are marked by an arrow. (B) An EMSA was first performed by incubating wild-type TC1 double-stranded radiolabeled oligonucleotides with SW620 nuclear extract followed by exposure of the gel to UV light. The shifted species was isolated by excision, and the radiolabeled SPy-DNA complex either eluted out of the excised gel slice and subjected to analysis by SDS–PAGE (lane A), or electrophoresed out of the gel slice following insertion into the well of an SDS–polyacrylamide gel (lane B). The gels were then visualized by autoradiography.

Figure 2

EMSA analysis of ion-exchange and DNA-affinity chromatography fractions. (A) EMSA analysis of fractions from first round step-wise elution of ion-exchange column using a TC1 double-stranded radiolabeled oligonucleotide. (B) EMSA of fractions from second round linear elution ion-exchange column using a TC1 double-stranded radiolabeled oligonucleotide. Fraction numbers from each column are shown along the tops of each panel, including the salt concentration gradients used for each elution. (C) EMSA analysis of fractions eluted from the SPy DNA affinity column using a radiolabeled TC1 single-stranded pyrimidine oligonucleotide. DNA affinity fraction 17 was resolved by SDS–PAGE and fluorescently stained with SYPRO Ruby-Red.

Figure 3

MALDI-TOF mass spectrometric analysis of SPy tryptic peptides. The polypeptide present in affinity fraction 17 was resolved by SDS–PAGE, stained with Ruby-Red, excised from the gel and subject to tryptic digestion. MALDI-TOF mass spectrometric analysis of the subsequent tryptic peptides produced the spectrum shown in (A). The spectrum in (B) represents the masses of tryptic peptides present in a background piece from the SDS–PAGE.

Figure 4

ProteinProspector search results of SPy tryptic peptide masses. The peptide masses resulting from MALDI-TOF analysis of SPy tryptic peptides were searched using the ProteinProspector website against the masses of virtually derived tryptic peptides present in the NCBI protein database. Fourteen of 32 submitted peptide masses (A) were found to match precisely with virtually derived tryptic peptide masses of hnRNP K. The table shows submitted and matched masses of the corresponding peptides, including the difference in mass between the two (delta p.p.m.). Matches to within 50 p.p.m. are considered significant (32). (B) Amino acid sequence of hnRNP K showing the regions that matched to SPy tryptic peptides (bold and underlined). Numbered regions correspond to the locations within hnRNP K of identified tryptic peptides from (A). These peptides cover 32% of the matched protein sequence, providing unambiguous identification.

Figure 5

Antibodies specific for hnRNP K recognize SPy. EMSA analysis of a SPy-TC1 double-stranded DNA complex in the presence of increasing amounts of anti-hnRNP K (A) polyclonal or (B) monoclonal 3C2 antibody. Addition of Ets antibody to the EMSA reactions in both (A) and (B) is shown as a control. (C) Western analysis of DNA affinity fractions 16–18 (of Fig. 2C) using anti-hnRNP K monoclonal antibody 3C2.

Figure 6

Effect of SPy/hnRNP K DNA-binding mutations and titration of SPy/hnRNP K protein on SRC1A promoter activity. (A) Mutations that abolish SPy single-stranded binding (i.e. four incorporated purines within either the pyrimidine strand of TC1, TC2 or both together) were introduced into 0.38SRC-CAT and reporter levels assayed by transient transfection in SW480 cells. Solid black bars represent SPy double-stranded mutations; gray bars represent SPy single-stranded mutations. CAT levels are expressed relative to wild-type 0.38SRC-CAT (open bar). Sequences of the mutations are shown below. (B) Single-stranded oligonucleotides capable of SPy/hnRNP K binding (TC1 CT) as well as (C), the mutant form, incapable of SPy/hnRNP K binding (TC1ss mut CT) were co-transfected with 0.38SRC-CAT into SW480 cells. 0.38SRC-CAT harboring TC1+TC2ss mutations was cotransfected with both TC1 CT (D) and TC1ss mut CT (E), single-stranded oligonucleotides into SW480 cells. The black triangle represents increasing concentrations (see Materials and Methods) of the competitor oligonucleotide in the presence of consistent quantities of reporter plasmid. CAT levels were standardized to total protein and β-galactosidase expression, and are the average of three duplicate experiments.

Figure 6

Effect of SPy/hnRNP K DNA-binding mutations and titration of SPy/hnRNP K protein on SRC1A promoter activity. (A) Mutations that abolish SPy single-stranded binding (i.e. four incorporated purines within either the pyrimidine strand of TC1, TC2 or both together) were introduced into 0.38SRC-CAT and reporter levels assayed by transient transfection in SW480 cells. Solid black bars represent SPy double-stranded mutations; gray bars represent SPy single-stranded mutations. CAT levels are expressed relative to wild-type 0.38SRC-CAT (open bar). Sequences of the mutations are shown below. (B) Single-stranded oligonucleotides capable of SPy/hnRNP K binding (TC1 CT) as well as (C), the mutant form, incapable of SPy/hnRNP K binding (TC1ss mut CT) were co-transfected with 0.38SRC-CAT into SW480 cells. 0.38SRC-CAT harboring TC1+TC2ss mutations was cotransfected with both TC1 CT (D) and TC1ss mut CT (E), single-stranded oligonucleotides into SW480 cells. The black triangle represents increasing concentrations (see Materials and Methods) of the competitor oligonucleotide in the presence of consistent quantities of reporter plasmid. CAT levels were standardized to total protein and β-galactosidase expression, and are the average of three duplicate experiments.

Figure 7

Possible model for SRC1A transcriptional regulation by Sp1 and hnRNP K. Our data suggest a model by which hnRNP K could recognize and bind specifically to double-stranded sequences within TC1 and TC2 followed by strand separation that would be facilitated by the increased affinity of hnRNP K for single-stranded DNA. The resulting single-stranded ‘bubble’ could encompass the entire TC-tract region or more discrete regions therein. For example, the ability of hnRNP K to bind TBP could recruit TFIID to the TC3 region and aid in the assembly of a pre-initiation complex. Transactivation would be accomplished by Sp1, most notably in this model, from binding to the GC1 site. Such a structure would be in agreement with the location of the preferred transcription start sites, which are located ∼20 bp downstream of TC3 and would also account for other minor sites, which are located throughout the TC-tract region.