Oncogenic transformation induced by the Qin protein is correlated with transcriptional repression

Abstract

The retroviral oncogene qin codes for a protein that belongs to the family of the winged helix transcription factors. The viral Qin protein, v-Qin, differs from its cellular counterpart, c-Qin, by functioning as a stronger transcriptional repressor and a more efficient inducer of tumors. This observation suggests that repression may be important in tumorigenesis. To test this possibility, chimeric proteins were constructed in which the Qin DNA-binding domain was fused to either a strong repressor domain (derived from the Drosophila Engrailed protein) or a strong activator domain (from the herpes simplex virus VP16 protein). The chimeric transcriptional repressor, Qin–Engrailed, transformed chicken embryo fibroblasts in culture and induced sarcomas in young chickens. The chimeric activator, Qin–VP16, failed to transform cells in vitro or in vivo and caused cellular resistance to oncogenic transformation by Qin. These data support the conclusion that the Qin protein induces oncogenic transformation by repressing the transcription of genes which function as negative growth regulators or tumor suppressors.

The oncogene qin was isolated as a cell-derived insert in the genome of avian retrovirus ASV31, a virus that induces neoplastic transformation in cell culture and sarcomas in young chickens. The viral qin gene (v-qin) is the oncogenic determinant of ASV31. The v-qin gene differs from its cellular counterpart c-qin by an N-terminal leader sequence, several aa substitutions and a C-terminal deletion (1, 2). The Qin proteins (v-Qin and c-Qin) belong to the family of winged helix transcription factors (2, 3). More than forty members of this family have been described in recent years; they are found in yeast, Drosophila, Caenorhabditis elegans, Xenopus, rodents, and humans (4, 5). The functions of winged helix proteins are related to embryonal development and tissue differentiation; these proteins are expressed in a highly tissue-specific manner (6). A familiar example for winged helix-controlled differentiation is the nude mouse whose phenotype results from a defect in the whn gene (7).

The c-Qin protein is the avian homolog of the mammalian brain factor 1 (BF-1); expression of this protein is restricted to the telencephalon, the nasal half of the retina, and the optic stalk (1, 8). A null mutation of the mouse BF1 gene leads to death at birth accompanied by dramatic size reduction of the cerebral hemispheres (9). Some genes coding for winged helix proteins have oncogenic potential (3, 10, 11). Viral qin induces neoplastic transformation, presumably due to aberrant transcriptional regulation and ectopic expression. In transient transfection, the Qin proteins act as transcriptional repressors, suggesting that this activity may play a role in oncogenicity (2). To examine this possibility, we constructed chimeric transcription factors containing the Qin DNA-binding motif and heterologous regulatory domains of either the Drosophila Engrailed or herpes simplex virus VP16 proteins. The properties of these chimeric proteins support the conclusion that transcriptional repression is a requirement for neoplastic transformation induced by Qin.

MATERIALS AND METHODS

Plasmid Construction.

To create Qin–Engrailed (En) chimeras, the PCR-amplified fragments of engrailed (amino acids 2–298) (12) and v-qin (amino acids 1–251) (3) or chicken c-qin (amino acids 1–243) (1) were cloned consecutively into the HindIII–XbaI or BamHI–HindIII sites of the same pBK/RSV vector (Stratagene). Oligonucleotide primers designed for PCR included BamHI, HindIII, and XbaI restriction sites allowing in-frame cloning. All DNA fragments derived from PCR amplification were sequenced using primers internal to the qin or engrailed coding sequences or T3 and T7 primers. Qin–VP16 chimeras were constructed by exchanging an EspI–EcoRI fragment of either c-qin or v-qin with a BamHI–EcoRI fragment of plasmid 597.1 coding for the transactivation domain of the VP16 protein (13). The cohesive ends generated by BamHI or EspI digestion were filled in with Klenow polymerase. The resultant Qin–VP16 and Qin–En plasmids were cleaved with either BamHI–EcoRI or BamHI–ClaI for cloning into the retroviral expression vector RCAS (14) to yield RCAS–Qin–En and RCAS–Qin–VP16. Transfection of chicken embryo fibroblasts (CEF) with the RCAS vectors leads to the production of infectious RCAS retroviruses that replicate and express the cloned insert. The reciprocal recombinants between v-qin and c-qin, VC and CV, have been described (2).

Electrophoretic Mobility Shift Assays.

Circular pBK/RSV plasmids with inserts coding for c-Qin, v-Qin, c-Qin–En, v-Qin–En, c-Qin–VP16, or v-Qin–VP16 were used as templates for in vitro translation in a transcription and translation coupled system according to the manufacturer’s protocol (Promega). Wheat germ lysates containing translated proteins were used in electrophoretic mobility shift assays as described (2). The relative amounts of translated proteins were estimated by labeling the proteins with [³⁵S]methionine (40 μCi in 50 μl reaction volume; 1 Ci = 37 GBq) followed by SDS/PAGE and autoradiography.

Transfection and Focus Assays.

Assays for transcriptional regulation in CEF cultures were carried out using transient transfection with the calcium phosphate precipitation method (15). CEF were plated 20 h before transfection at a density of 5 × 10⁵ cells per 35-mm dish in DMEM. The use of the pBK/RSV expression vectors and the p6BCluc reporter plasmids in these transfections has been described (2). Cells were harvested for luciferase assays at 40 h posttransfection. Assays for oncogenic transformation were initiated by stable transfection with the RCAS constructs following the dimethyl sulfoxide-polybrene method (16). Transfected cultures were overlaid with nutrient agar and incubated at 37°C until foci of transformed cells developed.

Western Blot Analysis, Immunofluorescence, and Tumor Formation.

The protocols for Western blot analysis and immunofluorescent staining have been described (17, 18). Tumorigenicity was tested in 1-day-old Spafas chicks (Storrs, CT) that were injected subcutaneously in the wing web with 1 × 10⁶ transfected cells serving as source of infectious retrovirus. Chicks were examined weekly for 6–7 weeks.

RESULTS

v-Qin Is a Stronger Transcriptional Repressor Than c-Qin and Shows Enhanced Oncogenicity.

The v-Qin protein shows a reduced DNA-binding activity compared with c-Qin. This difference maps to an amino acid substitution (Gly → Asp) within the winged helix DNA-binding motif (2). Overexpression of v-Qin or of c-Qin in CEF induced similar transformed cell morphology and produced comparable numbers of neoplastic foci (Table 1). However, the RCAS–v-Qin retrovirus was significantly more oncogenic in chickens than the RCAS–c-Qin virus, although both viruses expressed their respective inserts at similar levels. v-Qin induced tumors in almost all injected birds within 2–3 weeks, whereas c-Qin seldom led to tumor formation and then only after a latent period of 6–7 weeks (Table 1). Even these rare and delayed tumors might not have been caused by wild-type c-Qin, but by a mutant that could have arisen during the extended course of RCAS–c-Qin replication. The difference in oncogenicity between c-Qin and v-Qin was correlated with other properties of the two proteins. Previous studies have shown that v-Qin acts as a stronger repressor than c-Qin on the p6BCluc reporter (2). Results with recombinants between v-qin and c-qin suggest that tumorigenicity and enhanced transcriptional repression are linked to each other and to the amino acid substitution in the winged helix domain of v-qin (Table 1) (2). These data support a role of transcriptional repression in oncogenic transformation by qin.

Table 1.

Transcriptional regulation and oncogenicity of Qin constructs

RCAS construct	Transcriptional regulation*	Focus formation in vitro†	Tumorigenicity
			Incidence‡	Latent period, weeks
v-Qin	1.8	34	10/11	2–6
c-Qin	11.2	35	3/12	6–7
CV§	3.2	25	3/3	3–5
VC§	17.3	30	1/4	6–7
RCAS + v-Qin¶	ND	5.0  ×  105	ND	ND
v-Qin-VP16 + v-Qin¶	ND	4.4  ×  104	ND	ND

^*Relative luciferase activity expressed as percent of RCAS control. Data are from a representative experiment. 

^†Number of foci per 50 ng DNA transfected (constructs v-Qin, c-Qin, CV, and VC), and transforming titer of RCAS-v-Qin virus expressed in focus forming units/ml (constructs RCAS + v-Qin and v-Qin-VP16 + v-Qin). Data are from a representative experiment. 

^‡Birds with tumors over birds injected. 

^§Recombinants between v-qin and c-qin (see Fig. 1A) (2). 

^¶CEF were first transfected with RCAS or vQ-VP16-RCAS coding for viral envelope protein of subgroup A, and after 1 week they were challenged with v-Qin-RCAS virus of envelope subgroup B. ND, not determined. 

Construction of Qin–En and Qin–VP16 Chimeras.

The relationship between transcriptional repression and oncogenic potential was explored with chimeras between qin and Drosophila engrailed or herpes simplex virus VP16. The Qin–En chimeric proteins consisted of N-terminal sequences derived from either c-Qin or v-Qin including the winged helix DNA-binding motif. These were fused to the repressor domain of the Drosophila Engrailed protein (19, 20) (Fig. 1A). The Qin–VP16 chimeras were generated by linking the same N-terminal Qin sequences to the acidic transactivation domain of the herpes simplex virus protein VP16 (21) (Fig. 1A). In both types of chimeras, the major transcriptional repressor region of the Qin protein was deleted. Because the specificity of DNA binding is determined by the winged helix motif (2, 8), these chimeric proteins are designed to be either negative or positive regulators operating through Qin-binding sequences.

Figure 1.

Construction and expression of Qin chimeric proteins. (A) Schematic structures of Qin (vQ and cQ), reciprocal v-Qin–c-Qin recombinants, and Qin chimeras. v-Qin has an 8-aa leader (black) at the N terminus and is truncated at the C terminus. The winged helix domain of v-Qin (dark gray) differs from that of c-Qin (light gray) by an amino acid substitution (G → D). Diagonal shading: repressor domain of Engrailed. Wave shading: activator domain of VP16. (B) Western blot analysis. CEF transfected with RCAS (lane 2), RCAS–c-Qin-En (lane 3), RCAS–v-Qin-En (lane 4), and RCAS–v-Qin–VP16 (lane 5) were lysed by boiling in 1× SDS containing sample buffer. Lysates were analyzed by SDS/PAGE followed by blotting onto nitrocellulose membranes, which were probed subsequently with a Qin-specific antibody at a dilution of 1:200 and alkaline phosphatase-conjugated anti-rabbit serum at a dilution of 1:2000. Mock-infected CEF (lane 1) were also included. The molecular weight markers are indicated on the left in kDa.

Expression levels of the RCAS–Qin–En and RCAS–Qin–VP16 vectors were measured by Western blot analysis. Nuclear extracts were prepared from transfected CEF; proteins were separated by PAGE, blotted, and probed with a polyclonal antibody against full-length c-Qin (Fig. 1B). The chimeras were stably expressed, although accumulating levels of Qin–VP16 were lower than those of Qin–En (Fig. 1B, lane 5, the 41-kDa band). Gel mobility shift assays were performed with the fusion proteins translated in vitro using a labeled B2 oligonucleotide (2) or the optimal Qin-binding sequence as determined by PCR selection (TGTAAACAAA) (J.L., unpublished data). The chimeric proteins c-Qin–En and c-Qin–VP16 bound DNA as did c-Qin (Fig. 2A). Binding of the proteins to the labeled oligonucleotide could be specifically competed with cold oligonucleotide. In contrast, protein–DNA complexes formed by v-Qin and v-Qin chimeras were barely detectable (data not shown) due to the weak affinity of v-Qin for DNA (2).

Figure 2.

DNA binding and transcriptional regulatory activities of Qin chimeras. (A) Electrophoretic mobility shift assay of fusion proteins translated in vitro in wheat germ lysates. Each of the translated proteins binds to the labeled B2 oligonucleotide probe containing the Qin-binding sequence (2). Lanes: 1, c-Qin; 2, c-Qin–En; 3, c-Qin–VP16, 4, vector control. (B) Luciferase assays from cotransfection experiments performed in CEF using reporter plasmid p6BCluc (2) and the individual fusion proteins expressed with the pBK/RSV vector. In each transfection mixture, the amount of expression plasmid was 2 μg per plate, and the reporter plasmid was kept at 1 μg per plate.

Qin Chimeras That Repress Transcription Are Oncogenic; Chimeras That Activate Transcription Inhibit Oncogenesis.

The transcriptional regulatory properties of the Qin chimeras were examined in transient transfection assays using the pBK/RSV expression vector and p6BCluc as a reporter. In addition to v-Qin, which repressed the background level of transcription seen with the pBK/RSV vector alone, repression was also observed with the Qin–En chimeras (Fig. 2B). Qin–VP16 functioned as an activator of transcription. If the p6BCluc was replaced by pCMX, a reporter without Qin-binding sites, only minimal changes in the background luciferase activity were observed with any of the chimeras, suggesting that repression and activation by these proteins is DNA sequence-specific. Oncogenic transformation by the Qin chimeras was tested in CEF using the RCAS expression vector constructs (13). Both RCAS–c-Qin–En and RCAS–v-Qin–En (Fig. 3A) induced transformed cell foci indistinguishable from those induced by v-Qin. However, the process of focus formation by the chimeras was slower than with RCAS–v-Qin (4 weeks versus 2 weeks), and focus titers were lower (20 foci/μg DNA of RCAS–v-Qin–En versus 60 foci/0.1 μg DNA of RCAS–v-Qin). As a negative control, a Qin–En chimera lacking the Qin DNA-binding domain was included in these assays but no foci were observed, suggesting that Qin–En caused oncogenesis specifically through Qin target genes (data not shown). CEF transfected with either RCAS–v-Qin–VP16, RCAS–c-Qin–VP16, or RCAS alone did not develop foci. Nontransformed RCAS and RCAS–v-Qin–VP16 transfected CEF were then transferred and were superinfected with the RCAS–v-Qin virus. This virus carried envelope proteins of subgroup B avian retroviruses and was therefore not excluded from the transfected cells which synthesized envelope protein of subgroup A. The RCAS-transfected cells developed foci of transformed cells in response to v-Qin. In contrast, the RCAS–v-Qin–VP16 transfectants showed an about 10-fold reduced susceptibility to v-Qin-induced transformation (Table 1). This result suggests that the Qin–VP16 chimera can function as a transdominant negative mutant of Qin.

Figure 3.

(A) Focus assay of Qin chimeric constructs. CEF were transfected with (a) RCAS, (b) RCAS- v-Qin- En, and (c) RCAS- v-Qin -VP16 and cultured under nutrient agar for 4 weeks. (B) Immunofluorescent staining of CEF transfected with (a) RCAS vector alone, (b) RCAS–v-Qin-En, and (c) RCAS–v-Qin–VP16 using a Qin specific polyclonal antiserum raised against full length c-Qin and fluorescein isothiocyanate conjugated goat anti-rabbit IgG serum (×400).

CEF transfected with the fusion constructs were studied by immunofluorescence with a polyclonal antibody against intact c-Qin (Fig. 3B). CEF transformed by RCAS–v-Qin–En showed an elongated spindle shape and relatively small cell body. In contrast, CEF producing RCAS–v-Qin–VP16 were polygonal and flat, with an enlarged cell body. RCAS–c-Qin–En and RCAS–c-Qin–VP16 induced the corresponding changes in CEF (data not shown), and all fusion proteins were localized in the nucleus.

To test for oncogenicity in vivo, CEF transfected with the chimeric RCAS constructs and releasing the corresponding infectious RCAS virus were injected subcutaneously into the wing web of newly hatched chickens. All chickens that received the control construct RCAS–v-Qin came down with sarcomas at the site of injection within two weeks. Four out of six chickens injected with the Qin–En chimeras also developed tumors but only after a 4- to 5-week latent period. In contrast, none of six animals injected with Qin–VP16 showed a tumor even after eight weeks. These results demonstrate that Qin–En constructs are oncogenic, albeit less so than v-Qin, and Qin–VP16 chimeras lack detectable oncogenic potential.

DISCUSSION

Transcriptional repression by the Qin oncoprotein is correlated with oncogenic potential. This fact suggests that repression plays a role in neoplastic transformation and that Qin-induced oncogenesis may result from aberrant repression of specific genes. One would expect that these genes are negative growth regulators; they may normally function as tumor suppressors, and their down-regulation may cause uncontrolled growth. Identification and isolation of direct Qin target genes that are differentially regulated in Qin-transformed cells are now important tasks. This search may lead to the discovery of new tumor suppressor genes. Target identification has become the central problem for the understanding of all oncogenic transcription factors. If the number of relevant targets is small, it may be possible to induce the oncogenic cellular phenotype using other means of target gene regulation; in the case of the transcriptional repressor Qin, these may be transdominant negatives or antisense technology applied to the targets.

There are other oncogenes that transform by down-regulating the transcription of specific targets. A particularly clear example is erbA. The highly oncogenic retroviral version of this gene acts as a transdominant negative of the retinoic acid receptor and the thyroid hormone receptor. The result is a down-regulation of genes controlled by these two receptors. For oncogenic transformation by erbA, the effect on the retinoic acid receptor appears to be essential and sufficient (22).

The molecular mechanism of transcriptional repression by the Qin protein is not known. The requirement for sequence-specific DNA binding rules out titration of an activator (“squelching”). Qin may compete with an activator for the same or an overlapping DNA target sequence; it may bind next to an activator and thereby prevent the interaction between activator and initiation complex (“quenching”) or Qin may block the initiation complex directly (“silencing”). The repressor Engrailed has been suggested to act by quenching (20), and the functional equivalence of its repressor domain to that of Qin could imply a similar repression mechanism for Qin. However, other possibilities, including direct repression as in Drosophila Even-skipped (23) have not been ruled out. These considerations on mechanisms define a second important task in the study of Qin oncogenesis, the identification and cloning of proteins that interact with the oncogenic transcription factor. Among these would be proteins that transmit the regulatory signal from the Qin–DNA complex to the basal transcriptional machinery of the cell. Such proteins may interact with any of the functional domains of Qin. Particularly interesting would be proteins that interact with the repressor domain; they may also be binding to the repressor domain of Engrailed as this domain is able to transmit some of the repression signals that are correlated with Qin-induced transformation.

It has been suggested that the striking effect of null mutations in the mammalian Qin homolog BF-1 on brain development may result from a reduction of cortical progenitor cells (24). BF-1/Qin may play a role in regulating the proliferation of these progenitors. A recent study showed that retinal overexpression of Qin can distort the visual projection map (25). Whether BF-1/Qin acts as a transcriptional repressor in carrying out its normal mission in brain development is not known. The identification of specific target genes that are differentially expressed in BF-1⁺ as compared with BF-1⁻ telencephalic neuroepithelium could help answer this question.

ABBREVIATIONS

BF-1

brain factor 1

En

Engrailed

CEF

chicken embryo fibroblasts