[page 53↓]

3.  Results

3.1. Identification of Proteins Interacting with H-REV107-1

3.1.1. Screening of a Human Kidney cDNA Library to Identify Potential Interacting Partners of the H-REV107-1 Protein

To understand the mechanism of H-REV107-1 – mediated growth suppression (Sers et al., 1997), a search for its potential interacting partners using a Lex-A Yeast Two-Hybrid system (Fields et al., 1989) was performed. In this approach an H-REV107-1 recombinant protein, pEG202-107 (bait), fused with a prokaryotic DNA-binding domain of the LexA protein was generated. The pEG202-107 protein and library cDNA inserts, fused with the E.coli B42 activation domain, were co-expressed in the yeast cell line EGY48 carrying the LexA and GAL4 reporter genes. Yeast transformants grew on a selection medium, in which activation of the reporter genes made an interaction phenotypically detectable. The approach is described in detail in chapter 2.2.1.

As a bait, a truncated form of the human H-REV107-1 protein, lacking the membrane binding domain that hampered the transport of the hybrid protein into the yeast nucleus where the interaction takes place, was used. To screen a human kidney cDNA library available from Clontech, three independent large scale co-transformations were performed. More than two hundred preliminary positive clones were identified (Fig. 6).

Although the yeast two-hybrid system is a convenient tool for detection of protein-protein interactions in vivo, one disadvantage of this approach is the abundance of false positives. To verify putative interacting partners of the H-REV107-1 protein, the following strategy has been chosen:

Positive colonies were restreaked and re-tested on a selective medium to verify stability of the interactions.

Inserts of positive colonies were amplified using PCR to find out whether some of the colonies contain more then one cDNA library plasmid.

The cDNA library plasmids were isolated from yeast and transformed into E. coli KC8, the plasmids isolated from the clones confirmed in previous steps were sequenced.

Common yeast two-hybrid false positives were excluded from further analysis. The mating assay was applied to confirm interactions with the candidates chosen in the previous steps.

The interacting partners confirmed in the mating step were tested using co-immunoprecipitation in mammalian cells and an in vitro binding assay. Confirmed interactions were analysed functionally.


[page 54↓]

Fig. 6 Yeast Two Hybrid screen for H-REV107-1 interacting proteins, and strategy for their validation


[page 55↓]

3.1.2.  Sequencing Analysis of Clones Encoding Putative Interaction Partners of the H-REV107-1 Protein

Screening of the cDNA library and exclusion of colonies containing more than one cDNA plasmid resulted in the selection of 168 clones appropriate for a sequencing analysis. cDNA library plasmids were isolated from the E.coli KC8 transformants, sequenced, compared to each other for elimination of duplicates, and analysed using NCBI Data Base http://www.ncbi.nlm.nih.gov/BLAST/ in order to determine which genes are present. The genes encoding potential interacting partners of the H-REV107-1 protein are listed in the Table 3. In addition, 37 yeast open reading frames, 33 mitochondrial genes, and 45 unknown genes were identified and excluded from further analysis.


[page 56↓]

Table 3 List of genes encoding putative interacting partners of the H-REV107-1 protein

cDNA

Human Homologue

Accession N.

1

23kD highly basic protein (HS23KD )

X56932

2

Activated RNA polymerase II transcription cofactor 4 (PC4)

NM_006713

3

Alpha-amylase

BAA14130

4

BAG-family molecular chaperone regulator-3 (BAG-3)

AF095193

5

Beta-2-microglobulin (B2M)

NM_004048

6

Calcyclin (PRA, S100A6)

NM_014624

7

Carboxypeptidase D (CPD)

NM_001304

8

Cathepsin D, lysosomal aspartyl protease (CTSD)

NM_001909

9

Cystatin C, amyloid angiopathy and cerebral hemorrhage (CST3)

NM_000099

10

Enoyl Coenzyme A hydratase 1, peroxisomal (ECH1)

NM_001398

11

Fasciculation and elongation protein zeta 2 (zyginII,FEZ2)

U69140

12

Ferritin L-chain

Y09188

13

Heat shock protein 90 (Hsp90)

NM_005348

14

L apoferritin (exons 3 and 4)

X03743

15

Leukocyte common antigen (CD45)

AJ006102

16

Na, K-ATPase beta subunit (ATP1B)

NM_001677

17

Pancreatic lipase (PNLIP)

NM_000936

18

Plasma membrane Na+/H+ exchanger isoform (NHE3)

U28043

19

Proteasome subunit HC2

D00759

20

Proteasome subunit, alpha type, 2 (PSMA1)

NM_148976

21

Protein phosphatase 2A regulatory subunit A (PR65/PPP2R1A)

NM_014225

22

Putative DNA-directed RNA polymerase III C11 subunit

AF126531

23

Putative permease-related protein (SLCA8)

Y18483

24

Retinoic acid receptor gamma (RARG)

NM_000966

25

Retinoid inducible gene 1 (RIG1/TIG3/H-REV107-2/RARRES3)

AF092922

26

Reversion-induced LIM protein (RIL)

P50479

27

Ribosomal protein S14

BC006784

28

RNA-binding protein regulatory subunit (DJ-1)

NM_007262

29

Small EDRK-rich factor 2 (SERF2)

NM_005770

30

Sterol carrier protein-2 (SCP-2)

M55421

31

Syndecan 1 (SDC1)

NM_002997

32

TATA element modulatory factor

L01042

33

Translation termination factor gene (ETF1)

NM_004730

34

Translational inhibitor protein (p14.5)

NM_005836

3.1.3. Verification of Specificity of Interactions Using the Mating Test

The mating test was applied to ensure that the proteins identified in the primary yeast two-hybrid screening activate the reporter genes only due to interaction with H-REV107-1 and not through unspecific interactions with one of the reporter genes (Russell et al., 1998). To choose candidates for a mating test, the results from the library screenings were compared with the list of common false positives, the result of 100 yeast two-hybrid library screenings done by E. Golemis, Fox Chase Cancer Centre, Philadelphia, PA, USA.(Estojak et al., 1995).

http://www.fccc.edu/research/labs/golemis/InteractionTrapInWork.html

The following genes were noticed there as common false positives:

Table 4 List of known false positives found in yeast two-hybrid screenings

16 heat shock proteins

2 cytoskeletal proteins

14 ribosomal proteins

ferritin

5 cytochrome oxydase

lamin

3 collagen-related proteins

tRNA synthase

3 mitochondrial proteins

ubiquitin

3 zink finger proteins

vimentin

After excluding these genes, the following candidates were selected for the mating assay:


[page 57↓]

Table 5 List of the genes encoding potential H-REV107-1 interacting partners chosen for the mating test

Activated RNA polymerase II transcription cofactor (PC4)

Beta-2-microglobulin (B2M)

Calcyclin (PRA, S100A6)

Cathepsin D (CTSD)

Cystatin C (CST3)

Protein phosphatase regulatory subunit A (PR65/PPP2R1A)

Retinoic acid receptor gamma (RARG)

Retinoid inducible gene 1 (RIG1/TIG3/H-REV107-2/RARRES3)

Syndecan 1 (SDC1)

Translation termination factor gene (ETF1)

Translational inhibitor protein P14.5 (P14.5)

In this test the specificity of interaction was examined by activation of both the Leu2- and LacZ- reporter genes in the diploid yeast strain, generated by fusion of the EGY48 and RFY206 haploid yeast strains. For every interaction three independent mating tests were performed. Interactions obtained as specific in at least two tests were chosen for further analysis. As a negative control, I analysed whether these proteins are able to activate the reporter genes without H-REV107-1. These negative controls are indicated with “-“ in Fig. 7. A PDZ domain, shown to activate both reporter genes, was kindly provided by E Cuppen, (Institute of Cellular Signaling, Nijmegen, The Netherlands), and used as a positive control.

Interaction of the H-REV107-1 protein with ETF1, PC4, and S100A6 was shown to be specific but relatively weak, because activation of the LacZ reporter gene resulted in a light blue colour of colonies (Fig., upper panel). Interaction with SDC1, CST3, p14.5, and RARG was strong, although p14.5 and SDC1 activated the Leu reporter gene in one of the negative control (Fig. 7). Interaction with B2M, CTSD, and H-REV107-2/RIG1 was demonstrated as unspecific. The B2M and CTSD were able to activate the Leu reporter gene in two negative control tests, and H-REV107-2/RIG1 activated the Leu reporter gene in three negative control tests, independent of H-REV107-1 presence (white colonies in the negative control tests indicated with “-“ in the Fig. 7).

[page 58↓]

Fig. 7 Interaction between H-REV107-1 and potential binding proteins found in the yeast two-hybrid screening is verified using the mating assay

Diploid yeast colonies were grown on Leu – deficient, X-Gal – containing medium. Stability of an interaction between H-REV107-1 and one of the tested proteins was monitored through the intensity of the blue colour of the colonies. The more stable interactions resulted in the development of a dark blue colour (CST3, SDC1, p14.5, RARG). Weak or transient interactions resulted in growth of lightly coloured colonies (ETF1, PC4, and S100A6). Colonies in the upper lanes signed with “+” expressed H-REV107-1 and one of the interacting proteins. Colonies in the lower lanes signed with “-“ do not contain H-REV107-1, but only the activation domain and the library cDNA insert. Three mating tests for every target are depicted.

Fig. 8 Expression of the S100A6, ETF1, PC4, and P14.5 proteins is confirmed in yeast

Protein extracts were prepared form the yeast colonies used in the mating test. 10 μ g of the protein lysate was subjected to SDS-PAGE and analysed by Western blotting. A: analysis with an anti-HA antibody. The S100A6, ETF1, and PC4 proteins were detected. B: analysis with an antibody against P14.5.


[page 59↓]

3.1.4.  Examination of Protein Expression in Yeast

To choose candidates for further analysis in mammalian cells, the expression of the protein encoded by the cDNA library inserts in yeast colonies, verified by mating assay, was examined. All cDNA inserts used in the interaction trap were fused to a sequence encoding an HA-epitope at their 5’-ends. To control the expression of the recombinant proteins, lysates from yeast expressing the genes of interest were analysed by Western blotting using antibodies recognising an HA-epitope or a recombinant protein. The S100A6, ETF1 and PC4 proteins were detected with an anti-HA antibody (Fig. 8 A), and the P14.5 protein was detected with an antibody against P14.5, kindly provided by C. Kerkhoff, Institute of Clinical Chemistry and Laboratory Medicine, Regensburg, Germany (Schmiedeknecht et al., 1996; Fig. 8 B). No signals were detected in the analysis of the yeast expressing SDC1, CST3, and RARG, possibly, because the conformation of these proteins hampered binding of the HA-antibody to N-terminal epitopes.

The PC4, ETF1, S100A6, and P14.5 proteins were potential interacting partners of H-REV107-1 complying the criteria of “true putative positives” of the yeast two-hybrid system. They were confirmed in the mating test, and expression of the recombinant proteins was demonstrated in yeast (Fig. 8).

To verify the interaction between proteins identified by yeast two-hybrid system, co-immunoprecipitation from cellular extracts is the standard procedure. Therefore I used recombinant mammalian expression vectors to co-express H-REV107-1 and the putative interacting partners in COS-7 cells.

In addition to the proteins validated by mating assay, co-immunoprecipitation was done with the RARG and PR65 proteins. Retinoic Acid Receptor Gamma (RARG) is an important mediator of the retinoic acid cellular response (Holmes et al., 2000). Retinoic acid regulates cell differentiation and proliferation. Furthermore, it has been demonstrated to inhibit growth in several tumor cell lines (Ogata et al., 1994). Therefore, the interaction between H-REV107-1 and RARG was tested by co-immunoprecipitation despite the fact that no RARG protein has been detected in Western Blot analysis (Fig. 8)

The second potential interaction partner of H-REV107-1 was the PR65 protein, the regulatory subunit of the protein phosphates 2A. The PR65 encoding cDNA clone identified in the original yeast two-hybrid screening was lost. However, PR65 is a known regulator of numerous cellular events appearing to play a privileged role in the regulation of cell growth (Janssens and Goris, 2001). The catalytic subunit of PP2A was demonstrated to interact with the ETF1 protein (Andjelkovic et al., 1996), a potential interacting partner of H-REV107-1 found in this yeast two-hybrid screening. It was possible that the interaction with PR65 mediates growth suppressive properties of the H-REV107-1


[page 60↓]

3.1.5.  Generation of the H-REV107-1V5 and ΔCH-REV107-1HA Expression Vectors.

For co-immunoprecipitation in COS-7 cells H-REV107-1 recombinant proteins fused to HA-, and V5-epitopes were designed. These two epitopes were used because we wished to test which antibody is better appropriate for the immunoprecipitation of protein complexes. Full length and truncated H-REV107-1 recombinant proteins were generated to estimate if the H-REV107-1 membrane binding domain might influence binding capacity of the H-REV107-1 protein. Thus, the ΔCH-REV107-1HA recombinant protein consisted of a truncated H-REV107-1 lacking the membrane binding domain, and fused at its C-terminus to an HA-epitope (Fig. 9 A). The H-REV107-1V5 protein consisted of full length H-REV107-1 fused at its C-terminus to an V5-epitope (Fig. 9 B).

Fig. 9 The H-REV107-1 HA and V5 fusion proteins

A: Δ CH-REV107-1HA, cDNA without membrane binding domain was amplified using PCR, and subcloned in frame with the HA-Tag into the pcDNA3 vector.
B: H-REV107-1V5, H-REV107-1 full length cDNA was amplified using PCR, and ligated into the pEF6/V5 vector in a way that the V5-encoded sequence was located in frame at the 3’-end behind the membrane-binding domain. Abbreviations: MBD – membrane binding domain, term – termination of transcription (stop codon).


[page 61↓]

3.2.  PC4

3.2.1. H-REV107-1 Interacts with PC4 in COS-7 Cells

The PC4 protein has been isolated as a RNA-polymerase II transcriptional coactivator (Ge et al., 1994). A mammalian expression vector containing full length PC4 cDNA fused to a V5-epitope (PC4-V5) was purchased from the GeneStorm collection of Invitrogen. To examine whether PC4 interacts with the H-REV107-1 protein, COS-7 cells were transiently transfected with the ΔCH-REV107-1HA and PC4-V5 expression vectors. The PC4-V5 protein was detected in the precipitated complex containing ΔCH-REV107-1HA, however the interaction was very weak (Fig. 10, lane 3, arrow). These results were reproduced several times.

Fig. 10 The PC4-V5 and ΔCH-REV107-1HA proteins interact with each other in COS-7 cells

COS-7 cells were transiently transfected with PC4-V5 and Δ CH-REV107-1 expression vectors. As a negative control epitope – containing plasmids without inserts were used. Cell lysates were harvested 48 hours after transfection and used for immunoprecipitation with a HA-conjugated Sepharose. The immunocomplexes were subjected to SDS-PAGE and analysed using Western Blot with anti-V5 and anti-HA antibodies.
Lane 1: both proteins are present in the whole protein extract used for the immunoprecipitation.
Lane 2: negative control. Lane 3: Δ CH-REV107-1HA precipitated with an HA-conjugated Sepharose, and co-precipitated PC4-V5 detected with an anti-V5 antibody.


[page 62↓]

3.2.2.  Examination of the Intracellular Localisation of the Ectopically Expressed H-REV107-1 and PC4 Proteins

The PC4 protein is a transcriptional coactivator localised in the nucleus (Ge et al., 1994), whereas H-REV107-1 was described as a non-nuclear protein (Sers et al., 1997). To identify potential intracellular compartments where the proteins might be colocalised, immunofluorescence staining of COS-7 cells transiently transfected with the PC4-V5 and H-REV107-1 expression plasmids was performed. The H-REV107-1 expression vector containing full length cDNA was used because the role of the membrane binding domain has been shown to be important for the cellular distribution of the protein (Sers et al., 1997). PC4-V5 was localised mainly in nucleus, and in some cells in the cytoplasm. The H-REV107-1 protein was detected only in the cytoplasm (LINKlink). Thus, the preferential localisation of PC4 in the nucleus prevented efficient interaction with the H-REV107-1 protein.

Fig. 11 The H-REV107-1 protein is distributed through the cytoplasm in COS-7 cells, whereas the PC4-V5 protein is localised preferably in the nucleus

COS-7 cells were transiently transfected with the PC4 and H-REV107-1 full length expression vectors. Immunofluorescence staining and subsequent confocal laser microscopy was done 48h post transfection. Cells were stained with an anti-H-REV107-1 and anti-V5 primary antibodies, and anti-mouse AlexaFluor 488 nm, and anti-rabbit AlexaFluor 594 nm secondary antibodies. DAPI was used for DNA – staining.

To define the intracellular localisation of H-REV107-1 more precisely, we performed a subcellular fractionation of the H-REV107-1 protein in COS-7 cells transiently transfected with the H-REV107-1 expression vector. H-REV107-1 was revealed in the cytoplasmic and in the nuclear fractions (Fig. 12). The presence of two bands was explained by possible post-translational modification of the protein, like it was described for the rat H-rev107 protein (Sers et al., 1997).

Thus, the intracellular localisation of H-REV107-1 and PC4 partially overlapped in the nucleus and in the cytoplasm. This result would explain the weak interaction demonstrated in the co-immunoprecipitation experiments (Fig. 10).


[page 63↓]

Fig. 12 The H-REV107-1 protein is distributed through the nuclear and cytoplasmic fractions of the transiently transfected COS-7 cells

COS-7 cells were transiently transfected with the H-REV107-1 expression plasmid. Proteins were harvested 48 hours after transfection and fractionated. For SDS-PAGE 15 μ g of protein extracts from every fraction were loaded. Immunoblotting was performed against a polyclonal anti-H-REV107-1 antibody. A specific H-REV107-1 band was detected in the cytoplasmic (Cytopl) and nuclear fractions (Nucl), but not in the membrane fraction (Mem).

3.2.3. H-REV107-1 Interacts with Endogenous PC4 in COS-7 Cells

A further proposition explaining the weakness of the PC4-V5 – ΔCH-REV107-1HA binding was that the C-terminated epitope of one of the proteins can handicap the interaction. An interaction of ΔCH-REV107-1HA with other proteins was demonstrated in previous experiments. The PC4 protein carried at its C-terminus a single-strand DNA-binding domain, that has been already shown to play a decisive role in a protein-protein binding (Weger et al., 1999). To prove whether the C-terminal epitope of PC4-V5 has an inhibitory effect, the interaction between ΔCH-REV107-1HA and endogenous PC4 was investigated in further experiments.

PC4 expression in COS-7 cells was examined by Western Blot analysis using an anti-PC4 antibody, a generous gift of Professor R. Heilbronn, Free University, Berlin, Germany (Weger et al., 1999). PA1 cells already described as PC4-positive were used as a positive control (Kannan et al., 1999). Equal expression of PC4 protein in PA-1 and COS-7 cell lines was demonstrated (Fig. 13 A). Co-immunoprecipitation with a HA-conjugated Sepharose was done in COS-7 cells transiently transfected with the ΔCH-REV107-1HA and PC4-V5 expression vectors. More stringent washing of the precipitated protein complex were performed to check whether the endogenous PC4 or the recombinant PC4-V5 protein interacts more stably with ΔCH-REV107-1HA.


[page 64↓]

Fig. 13 H-REV107-1 interacts with the endogenous PC4 protein

A: COS-7 and PA1 cells were lysed in SDS-Sample buffer. 15 μ g of the lysate was subjected to SDS-PAGE. Upper panel - Western blotting against anti-PC4 antibody, bottom panel – actin was used as a loading control.
B: COS-7 cells were transiently co-transfected with PC4-V5 and
Δ CH-REV107-1HA expression vectors, or with an empty vector containing HA-epitope only as a negative control. Forty eight hours after transfection cells were lysed and immunoprecipitated with a HA-conjugated Sepharose. The whole protein extract (10 μ g), and the immunoprecipitated complex were resolved by SDS-PAGE, and analysed via Western blotting. Middle panel – Western blot analysis with an anti-PC4 antibody. Lane 2 - the immunoprecipitated endogenous and V5-tagged PC4 proteins were revealed (arrowheads). Upper panel – the membrane was stripped and incubated with an anti-V5 antibody. Bottom panel – after second stripping the membrane was incubated with an anti-HA antibody. Lane 1 – protein extracts, lane 2 – co-immunoprecipitation , lane 3 – negative control.


[page 65↓]

Western Blot analysis using the anti-PC4 antibody revealed two bands in the protein extract. The precipitated protein complex contained endogenous PC4 and very low amount of the ectopically expressed PC4-V5 protein (Fig. 13 B, middle panel, lane 2, arrowheads). The V5-tagged PC4 protein had a slightly lower mobility in electrophoresis gel due to the V5-epitope (Fig. 13 B, middle panel, lane 1, 2). Immunoblotting with the anti-V5 antibody revealed a single band in the protein extract (Fig. 13 B, upper panel, lane 1). Incubation with the anti-HA antibody exhibited ΔCH-REV107-1HA protein, expressed in the COS-7 cells, and precipitated with a HA-conjugated Sepharose (Fig. 13 B, bottom panel, lane 1 and 2, accordingly). This experiment demonstrated that H-REV107-1 interacts more stably with the endogenous PC4 protein.

Thus, interaction with the endogenous protein suggested that the H-REV107-1 – PC4 binding takes place in vivo. A potential role of this interaction in H-REV107 function was investigated in further experiments.

3.2.4. H-REV107-1, PC4, and STAT1 Form a Protein Complex related to IFNγ-signaling

The PC4 protein participates in the regulation of RNA-polymerase II activity, and interacts with a variety of other transcriptional regulators (Wu and Chiang, 2001). It has been demonstrated that PC4 enhances transcriptional activity of the BRCA1 protein (Haile and Parvin, 1999), which is involved in IFNγ - dependent growth control (Ouchi et al., 1999). A differential display from rat astrocytoma cells treated with IFNγ recovered the H-rev107 – sequence (Kuchinke et al., 1995), indicating that this gene is an IFNγ- target. This prompted us to suppose a potential role of the H-REV107-1 – PC4 complex in IFNγ - signaling.

Therefore, we investigated an effect of IFNγ on the H-REV107-1 expression in human cells. H-REV107-1 was demonstrated to be down-regulated in most tumor cell lines and in 50% of human ovarian carcinomas as compared to normal tissue. Increase of H-REV107-1 mRNA expression under IFNγ-treatment was shown in human ovarian carcinoma cell lines: A27/80 described as IFNγ-sensitive, and OVCAR-3 demonstrated to be resistant to the IFNγ-treatment. Further investigation revealed that H-REV107-1 is a direct target of interferon regulatory factor 1, IRF1 (Sers et al., 2002), a tumor suppressor and a mediator of the IFNγ-signalling (Tanaka and Taniguchi, 2000).

The increase of H-REV107-1 mRNA expression upon IFNγ treatment in OVCAR-3 and A27/80 cells suggested a potential involvement of the gene in IFNγ-mediated growth suppression (Sers et al., 2002). To prove this hypothesis, the restoration of H-REV107-1 expression upon IFNγ-induction was examined at the protein level.


[page 66↓]

OVCAR-3 and A27/80 cells were treated for 9, 12, 24, 36, and 48 hours with IFNγ, and analysed using Western blotting. A low but specific H-REV107-1 signal was detected in OVCAR-3 cells after 36 hours of treatment, which further increased during the following 12 hours (Fig. 14, upper panel). In A 27/80 cells the protein was below detectable level (data not shown), although up-regulation of mRNA expression was demonstrated (Sers et al., 2002).

To understand the mechanism of the IFNγ -dependent growth suppression, and to pursue a role of the H-REV107-1 protein in this process, the phenotype of A27/80 and OVCAR-3 cells treated with IFNγ was investigated using immunofluorescence analysis. In A27/80 cells no considerable phenotypic changes were detected after 48 hours of induction (data not shown). In a small fraction of OVCAR-3 cells (below 5%) the H-REV107-1 protein was detected. Most importantly, up-regulation of endogenous H-REV107-1 correlated with a nuclear morphology typical for apoptotic cells, and in a fraction of the cells, which underwent apoptosis, H-REV107-1 was localised in the nuclei (Fig. 15, red arrowheads). This suggested involvement of H-REV107-1 in the IFNγ-mediated cell death, and supposed that nuclear localisation of the protein might be important for its growth suppressive properties.

The small fraction of OVCAR-3 cells sensitive to IFNγ-treatment was probably not detected in growth analysis performed earlier. Therefore, OVCAR-3 cells were assumed to be resistant to IFNγ-treatment. In contrast, A27/80 cells showed considerable growth inhibition up to 50% (Sers et al., 2002), but no phenotypic changes were obtained in this cell lines, suggesting different down-stream mechanisms of the IFNγ-signaling in A27/80 and OVCAR-3 cell lines. To determine how IFNγ-responses diverse in OVCAR-3 and A27/80 cells, expression of the major down-stream IFNγ-effectors was investigated in these cell lines. Among them are the signal transducer and activator of transcription, STAT1 (Schindler and Darnel, 1995), and its down-stream targets involved in growth inhibitory effects, interferon regulatory factor 1 (IRF1) and cyclin-dependent kinase inhibitor, p21WAF1 (Naldini et al., 2001). The STAT1 and p21WAF1 proteins were shown to be induced after IFNγ-exposure in several ovarian cancer cell lines (Burke et al., 1999). The increase of IRF1 expression upon IFNγ-induction in OVCAR-3 and A27/80 cells has been described (Sers et al., 2002; Fig. 14).

The expression of the STAT1 and p21WAF1 proteins after incubation of OVCAR-3 and A27/80 cells with IFNγ was investigated using Western blot analysis. Up-regulation of STAT1 was detected after 24 hours and increased up to 48 hours in both cell lines (Fig. 16, upper panel). Immunoblotting with an anti- p21WAF1 antibody revealed a specific band in A27/80 cells after 24 hours of IFNγ-treatment. The amount of protein was stable during the following 72 hours. The p21WAF1 protein was not detected in OVCAR-3 cells (Fig. 16, middle panel).


[page 67↓]

Fig. 14 Up-regulation of H-REV107-1 expression in OVCAR-3 cells after IFNγ-induction

Western blot analysis of H-REV107-1 expression in OVCAR-3 cells was performed after incubation with 100 U/ml of IFN γ . H-REV107-1 was detected with a specific anti-H-REV107-1 antibody. Then the membrane was incubated with an antibody against IRF1, and after the next stripping with an anti-IRF2 antibody. Actin was used as a loading control.

Fig. 15 Induction of the H-REV107-1 expression upon IFNγ - treatment leads to cell death

Immunofluorescence analysis of OVCAR-3 cells treated for 48 hours with 100U/ μ l of IFN γ . Cells were incubated with a primary anti-H-REV107-1 antibody, and secondary anti-rabbit AlexaFluor546 antibody. Nuclei were stained with DAPI. Only cells expressing H-REV107-1 showed an altered morphology of the nucleus, indicative of apoptosis (white arrowheads).


[page 68↓]

Fig. 16 STAT1 and P21WAF1 expression after IFNγ - induction in OVCAR-3 and A27/80 cells.

OVCAR-3 and A27/80 cells were treated with 100 and 1000 Units of IFN γ , respectively for 24, 48, and 72 hours; 10 μ g of nuclear extracts were subjected to SDS-PAGE followed Western Blot analysis using antibodies against p21 WAF1 and STAT-1. Histone 3 was used as a loading control. C – p21 WAF1 control protein.

Thus, up-regulation of different STAT1-targets was demonstrated in OVCAR-3 and A27/80 cell lines in response to IFNγ-treatment. In A27/80 cells, the IFNγ-dependent growth suppression correlated with the up-regulation of p21WAF1 expression. In contrast, induction of apoptosis in OVCAR-3 cells was p21WAF1 independent, but correlated with up-regulation of the IRF-1 and H-REV107-1 proteins.

Two different STAT1-downstream pathways were described earlier (Ouchi et al., 2000). The first led to IRF1 up-regulation. The second, an IRF-1 independent pathway, led to the activation of a STAT1 – BRCA1 complex, and up-regulation of their target genes, including p21WAF1 (Ouchi et al., 2000). This pathway seemed to be activated in A27/80 cells. The tumor suppressor, breast cancer susceptibility gene 1 (BRCA1) was demonstrated to bind STAT1 in IFNγ-induced cells, thereby enhancing STAT1 activity (Ouchi et al., 2000). Furthermore, BRCA1 has been demonstrated to act as a transcription factor (Shuai et al., 1994).


[page 69↓]

Fig. 17 STAT1 and PC4 proteins interact with H-REV107-1

COS-7 cells were transiently transfected with the STAT1, PC4-V5, and Δ CH-REV107-1HA expression plasmids, or with the STAT1, PC4-V5, and pcDNA3-HA as a negative control. Protein extracts were harvested 42 hours after transfection. The clarified supernatant was incubated with the HA-Sepharose for 10 hours. The immunoprecipitated protein complex was subjected to SDS-PAGE and analysed by Western blotting.
Lane 1 – protein extracts, lane 2 – immunoprecipitation, lane 3 – negative control. Upper panel – incubation of the membrane with STAT1 antibody revealed the specific band in the primary protein extract used for the precipitation (lane 1), and a weak signal in the immunocomplex with the H-REV107-1 protein (lane 2). Middle panel – the PC4-V5 protein was detected with the epitope-specific anti-V5 antibody. Bottom panel –
Δ CH-REV107-1HA detected with an anti-HA antibody.

Experiments in vitro showed that the transcriptional activation by BRCA1 is maximal in the presence of the PC4 coactivator, although their direct interaction has not been confirmed (Haile and Parvin, 1999). Therefore, it was possible that the H-REV107-1 protein, interacting with PC4, could be a link between PC4 and the STAT1 – BRCA1 complex. This suggested the existence of a protein complex consisting of STAT1, PC4, H-REV107-1, and BRCA1.


[page 70↓]

The interaction between these proteins was tested using co-immunoprecipitation. To ensure their high expression level, COS-7 cells were simultaneously transfected with either ΔCH-REV107-1HA, PC4-V5, STAT-1, and BRCA1, or ΔCH-REV107-1HA, PC4-V5, and STAT-1 expressing plasmids. As a negative control, a transfection with PC4-V5, and STAT1 expression vectors, and a plasmid containing the HA-epitope only was done. Immunocomplexes were purified with a HA-conjugated Sepharose. The precipitated proteins were analysed using Western blotting. Incubation with an anti-STAT1 antibody revealed a specific band of approximately 90 kDa in the protein extract and in the immunocomplex precipitated with the ΔCH-REV107-1HA protein (Fig. 17, lane 1, 2 respectively). The PC4-V5 protein was detected with an anti-V5 antibody (Fig. 17, lane 1, 2). The BRCA1 protein was detected neither in the protein extract, nor in the precipitated protein complexes, suggesting that overexpression of this protein failed (data not shown).

Thus, the interaction between H-REV107-1, STAT-1 and PC4 proteins ectopically expressed in COS-7 cells was demonstrated. Participation of the BRCA1 protein in this complex is still a question to be answered. A potential role of this interaction in IFNγ-signalling has to be further investigated.

3.3. PR65

3.3.1. H-REV107-1 Interacts with PR65 in COS-7 Cells

PR65 is the regulatory subunit A of the protein phosphatase 2A (PP2A). PP2A is a highly conserved serine/threonine phosphatase essential for a number of cellular functions including signal transduction, translational control, and cell cycle regulation (Millward et al., 1999; Sontag, 2000; Janssens and Goris, 2001). PP2A was shown to act as a pro-apoptotic phosphatase (Klumpp and Krieglstein, 2002). The H-REV107-1 protein was demonstrated to possess growth suppressive properties (Sers et al., 1997), and to induce apoptosis in human ovarian carcinoma cells (Sers et al., 2002). Therefore, the possible interaction between these two proteins was analysed.

Mammalian expression vector containing full length PR65 cDNA fused to a V5-epitope (PR65-V5) was purchased from the GeneStorm collection of Invitrogen. To examine the interaction between H-REV107-1 and PR65, co-immunoprecipitation in COS-7 cells was performed. The cells were transiently transfected with the PR65-V5 and ΔCH-REV107-1HA expression vectors. Protein complexes were purified with an HA-conjugated Sepharose, and analysed with the anti-V5 antibody (Fig. 18 A). A PR65-V5 – specific band was present in the protein complex together with ΔCH-REV107-1HA (Fig. 18 A, upper panel, lane 2, arrow), but not in the control precipitation with the expression vector containing a HA-epitope only (Fig. 18 A, upper panel, lane 3). To ensure a presence of the ΔCH-REV107-1HA protein, Western blotting with an anti-HA antibody was done (Fig, 18 A, bottom panel).

[page 71↓]

Fig. 18 Co-immunoprecipitation of ΔCH-REV107-1HA and PR65-V5 in COS-7 cells

COS-7 cells were transiently transfected with the PR65-V5 and Δ CH-REV107-1HA expression plasmids. Protein extracts were harvested 48 hours after transfection, and co-immunoprecipitated with V5 and HA-antibodies. A: Protein complexes precipitated with the anti-HA antibody. Upper panel – co-precipitated PR65-V5 protein was detected with the anti-V5 antibody, bottom panel – precipitated Δ CH-REV107-1HA protein was detected with the anti-HA antibody. Lane 1 – expression of both Δ CH-REV107-1HA and PR65-V5 proteins in COS-7 cells used for the co-immunoprecipitation. Lane 2 – co-immunoprecipitation of PR65-V5 with Δ CH-REV107-1HA. Lane 3 and 4 – control immunoprecipitation where either PR65-V5 or Δ CH-REV107-1HA are present. B: Protein complex was precipitated with the anti-V5 antibody. Bottom panel – co-precipitated Δ CH-REV107-1HA protein detected with the anti-HA antibody; upper panel – precipitated PR65-V5 revealed with the anti-V5 antibody. Lane 2 - co-immunoprecipitation of the Δ CH-REV107-1HA protein with PR65-V5. Lanes 3, 4 – negative controls.

The reverse experiment, precipitation with the anti-V5 antibody, confirmed that the ΔCH-REV107-1HA – specific band precipitated together with the PR65-V5 protein (Fig. 18 B, bottom panel, lane 2, arrow).


[page 72↓]

3.3.2.  H-REV107-1 and PR65 are Co-Localised in COS-7 Cells

In addition to co-immunoprecipitation, the intracellular localisation of the H-REV107-1 and PR65 proteins was examined using immunofluorescence analysis of COS-7 cells transiently transfected with the PR65-V5 and H-REV107-1 expression vectors. Staining was analysed using a confocal laser microscopy (Fig. 19). Ectopically expressed H-REV107-1 and PR65-V5 had almost identical distribution patterns in COS-7 cells. Co-localisation of these proteins was obtained in the perinuclear region (Fig. 19, bottom panel – overlap of the H-REV107-1 and PR65 localisation is shown in yellow).

Fig. 19 The H-REV107-1 and PR65-V5 proteins ectopically expressed in COS-7 cells are co-localised in the cytoplasm

COS-7 cells were transiently transfected with PR65-V5 and H-REV107-1 expression vectors. Forty eight hours post transfection cells were fixed and incubated with a monoclonal anti-V5 antibody and polyclonal anti-H-REV107-1 antibody, and with the secondary anti-mouse AlexaFluor 488 and anti-rabbit AlexaFluor 594 antibodies, respectively. Nuclei were stained with DAPI.
Upper panel: PR65 – green, H-REV107-1 - red, Nuclei - blue. Bottom panel – overlay of all three staining. The perinuclear region where H-REV107-1 and PR65 proteins are co-localised had a yellow-green colour (arrow).


[page 73↓]

3.3.3.  H-REV107-1 Interacts with PR65 in a Cell-Free System

The GST pull-down assay is a well established method to verify the specificity of interaction between two proteins under cell-free conditions. This method allows excluding of an influence of other proteins on the interacting partners.

For this experiment the Glutathione S-transferase (GST) Gene Fusion System (Pharmacia Biotech) was used. GST-fusion proteins can be expressed in E.coli, and purified by affinity chromatography using Glutathione Sepharose. The ΔCH-REV107-1 protein fused at its N-terminus with Glutathione S-transferase (107-GST) was purified from E.coli, and incubated with the clarified cell lysate prepared from COS-7 cells over-expressing PR65-V5. As a control the COS-7 protein extract was incubated with the purified Glutathione S-transferase (GST) alone. After washing, proteins bound to the Glutathione Sepharose were separated by SDS-PAGE and subjected to Western blotting using the anti-V5 antibody. This analysis revealed PR65-V5 in a complex with the 107-GST fusion protein (Fig. 20, upper panel, lane 4). Control incubation with an anti-GST antibody revealed a specific 107-GST and GST bands of the predicted size in the E.coli cell lysate (Fig. 20, lane 2 and 1, respectively), and in the Sepharose-bound protein complexes (Fig. 20, lane 4 and 6, respectively).

These experiments demonstrated that the PR65 and H-REV107-1 proteins interacted in COS-7 cells as well as in cell-free conditions. Further experiments aimed at identifying the H-REV107-1 protein domains responsible for the interaction, and understanding the functional role of the H-REV107-1 – PR65 interaction.

3.3.4. Homodimer Formation of H-REV107-1

Western blot analysis of the COS-7 cells over-expressing the ΔCH-REV107-1HA truncated protein revealed two bands of different sizes: one band of 16 kDa, corresponding to the ΔCH-REV107-1HA protein, and the second band of approximately 32 kDa (Fig. 21, middle panel, lane 2). Co-transfection of COS-7 cells with the ΔCH-REV107-1HA and H-REV107-1V5 expression vectors, and subsequent Western blot analysis with an anti-HA antibody resulted also in the observation of two bands of different sizes. This suggested that the H-REV107-1 protein formed homodimers in COS-7 cells.

To prove this hypothesis, COS-7 cells were transiently transfected with the ΔCH-REV107-1HA and H-REV107-1V5 expression vectors, and with the ΔCH-REV107-1HA and PR65-V5 expression vectors as a positive control. Immunoprecipitation was performed with an anti-V5 antibody. Immunoprecipitated protein complexes were analysed by Western blotting with an anti-HA antibody. The ΔCH-REV107-1HA protein was detected in a complex with H-REV107-1V5 or PR65-V5 (Fig. 21, bottom panel, lanes 3 and 4, respectively), confirming the ability of H-REV107-1 to form homodimers in COS-7 cells.


[page 74↓]

Fig. 20 H-REV107-1 interacts with PR65 in a cell-free conditions

The H-REV107-1 protein fused at its N-terminus with a GST-epitope (107-GST), and the GST-epitope alone were over-expressed in E.coli B21, and purified with GST-Sepharose. Purified proteins were incubated with a lysate from COS-7 cells over-expressing PR65-V5. 10 μ g of total protein extract, and protein complexes bound to the Glutathione Sepharose were subjected to SDS-PAGE and analysed using Western blotting.
Upper panel: Western Blot analysis was done with an anti-V5 antibody. Bottom panel: Western Blot analysis using an anti-GST antibody. Lane 1 – E.coli B21/GST protein extract after sonication; lane 2 – E.coli B21/107-GST sonicate; lane 3 - 10
μ g of the protein extract from COS-7 cells over-expressing PR65-V5; lane 4 – GST pull-down of PR65-V5 with the recombinant 107-GST protein; lane 5 – GST pull-down of PR65-V5 with the Glutathione-Sepharose; lane 6 – GST pull-down of PR65-V5 with the GST-protein.


[page 75↓]

Fig. 21 The H-REV107-1 protein forms a homodimer in COS-7 cells

COS-7 cells were transiently transfected with the Δ CH-REV107-1HA and H-REV107-1V5 expression plasmids, with Δ CH-REV107-1HA and PR65-V5, and with every plasmid alone as a negative control. Forty eight hours after transfection immunoprecipitation with an anti-V5 antibody and Protein G agarose was performed. The co-immunoprecipitated Δ CH-REV107-1HA protein was detected by Western blotting against anti-HA antibody depicted on the bottom panel.
Lane 1 – cell lysate of COS-7 cells transfected with the H-REV107-1V5 and
Δ CH-REV107-1HA expression plasmids; lane 2 – cell lysate of COS-7 cells transfected with the PR65-V5 and Δ CH-REV107-1HA expression plasmids; lane 3 – co-immunoprecipitation of Δ CH-REV107-1HA with H-REV107-1V5; lane 4 – co-immunoprecipitation of Δ CH-REV107-1HA with PR65-V5; lane 5, 6 – negative controls.


[page 76↓]

3.3.5.  Determination of the H-REV107-1 Domains Responsible for Interaction with PR65 and Homodimer Formation

3.3.5.1. Generation of the H-REV107-1 Mutant Proteins

To map the region of H-REV107-1 responsible for protein binding, three mutants were generated and tested by co-immunoprecipitation. Selection of the domains to be mutated was based on the analysis of the protein sequence described in the Introduction. In brief, members of the H-REV107 protein family share a highly conserved domain (H-box), the function of which is unknown (Hughes and Stanway, 2000). The ΔC107-HWAY mutant was generated with a point mutation in the H-box leading to the substituting His by Ala at the position 23. The NCE domain was shown to play an important role for the activity of the lecithin retinol acyltransferase (LRAT), a protein which belongs also to the H-REV107 family (Mondal et al., 2000). The ΔC107-NCE mutant comprised the amino acid exchange Cys-112 to Ser-112 in this region. The third ΔC107-ΔN mutant carried a deletion of 20 amino acids at the proline-rich N-terminus (Fig. 22). Such proline-rich domains were shown to be important for protein binding (Kay et al., 2000). All mutants were fused at the C-terminus with the HA-epitope.

3.3.5.2. The N-terminal Domain of the H-REV107-1 Protein is Required for the Interaction with PR65 and for Homodimer Formation

The ability of the ΔC107- NCE, ΔC107-HWAY and ΔC107-ΔN mutants to associate with PR65-V5 was tested using co-immunoprecipitation in COS-7 cells. Proteins were precipitated with the HA-conjugated Sepharose, and examined using Western blot analysis with an anti-V5 antibody. The PR65-V5 protein was detected in a complex with the ΔCH-REV107-1HA protein used as a positive control (Fig. 24 A, upper panel, lane 1), and with the ΔC107-NCE, and ΔC107-HWAY mutant proteins (Fig. 24 A, upper panel, lanes 2, 3 accordingly). The ΔC107-ΔN mutant was unable to bind PR65-V5 (Fig. 24 A, upper panel, lane 4). Upon co-immunoprecipitation with an anti-V5 antibody similar results were obtained. The ΔC107-NCE, and ΔC107-HWAY mutants were shown to associate with PR65-V5 (Fig. 24 B, bottom panel, lane 2, 3). The ΔN107 mutant was not detected in a complex with PR65-V5 (Fig. 24 B, bottom panel, lane 4). This demonstrated that the N-terminal proline-rich domain of the H-REV107-1 protein is required for the interaction with PR65.

To define the protein domain responsible for a homodimer formation, COS-7 cells were transiently transfected with the H-REV107-1V5 expression vector, and a vector expressing each of the mutants, respectively. Immunoprecipitation was done with the anti-V5 antibody. The ΔNH-REV107-1 mutant was also deficient in homodimer formation (Fig. 23, bottom panel, lane 5).


[page 77↓]

Fig. 22 H-REV107-1 mutants generated for search of the domains responsible for interaction

Fig. 23 The ΔC107-ΔN mutant does not form homodimers

COS-7 cells were simultaneously transfected with expression vectors: H-REV107-1V5 and one of the mutants. Protein complexes were precipitated with an anti-V5 antibody. Upper panel – Western blot analysis of the precipitated proteins with an anti-V5 antibody. Bottom panel - Western blotting against anti-HA antibody. Lane 1 – 10 μ g of protein extract from COS-7 cells transfected with the Δ CH-REV107-1HA expression vector. Lane 2 – immunoprecipitation of Δ CH-REV107HA with H-REV107-1V5 used as a positive control. Lane 3 - negative control. Lane 4 – immunoprecipitation of Δ C107-NCE with H-REV107-1V5. Lane 5 – immunoprecipitation of the Δ C107- Δ N mutant with H-REV107-1V5. Lane 6 – negative control.

[page 78↓]
Fig. 24 The ΔN107 mutant fails to interact with PR65

COS-7 cells were co-transfected with the following expression vectors: PR65-V5 and Δ CH-REV107-1 (WT), PR65-V5 and Δ C107-NCE mutant (NC), PR65-V5 and Δ C107-HWAY mutant (HW), and PR65-V5 and Δ C107- Δ N mutant ( Δ N). For negative controls plasmids containing epitopes only were transfected ( - ).
A: immunoprecipitation with HA-conjugated Sepharose. Upper panel – precipitated protein complexes were analysed with an anti-V5 antibody for the PR65-V5 detection. Lane 1 – PR65-V5 was obtained in a complex with
Δ CH-REV107-1HA, lane 2 – with the Δ C107-NCE mutant, lane 3 – with the Δ C107-HWAY mutant. Lane 4 – the PR65-V5 protein does not associated with the Δ C107 Δ N mutant. Bottom panel – control Western blotting against anti-HA antibody.
B: reciprocal co-immunoprecipitation with an anti-V5 antibody. H-REV107-1 associating with PR65-V5 was revealed by Western blotting against anti-HA antibody. Lane 1 – H-REV107-1 was obtained in the complex with PR65-V5, lane 2 –
Δ C107-NCE mutant, lane 3 – Δ C107-HWAY mutant. Lane 4 – the Δ C107- Δ N mutant did not associate with PR65-V5. Upper panel – incubation with an anti-V5 antibody revealed precipitated PR65-V5.


[page 79↓]

Thus, the N-terminal domain of the H-REV107-1 protein was shown to be responsible for both association with the PR65 protein, and for the homodimer formation. Preliminary experiments were performed to test a potential competition between H-REV107-1 homodimer formation and PR65 binding. In this experiment the PR65 binding eliminated homodimer formation (data not shown).

3.4. Investigation of a Role of the H-REV107-1 - PR65 Interaction in Apoptosis

3.4.1. H-REV107-1 Does not Induces Apoptosis in Rat Fibroblasts FE-8

The aim of further investigations was to explore the mechanism of H-REV107-1 mediated cell death and to define a role of the PR65 – H-REV107-1 interaction in this process. Examination of the role of H-REV107-1 in cell cycle progression by flow cytometric analysis required establishment of stable transfectants over-expressing H-REV107-1. All attempts to generate stable clones in human ovarian carcinoma cell lines OVCAR-3 and A27/80 were unsuccessful, and resulted in the loss of cells over-expressing H-REV107-1. However, an H-rev107 – tetracycline-inducible system in rat fibroblasts FE-8 was available (Sers et al., 1997). Therefore, a potential implication of the H-REV107-1 protein in cell cycle control was tested in this system. Two independent FE-8 clones stably transfected with tet-inducible H-rev107, and one clone transfected with the empty vector as a control (pUHD) were tested in this experiment. Cells were cultured for 24, 48, 72, 96, and 120 hours with or without doxycycline then stained with propidium-iodide (PI) followed flow cytometric analysis. These experiments were performed by Cornelia Giseler, Institute of Pathology, Charité, Berlin. H-rev107 expression was controlled by Western blotting against an anti-H-rev107 antibody (Fig. 25).

Fig. 25 H-rev107 expression after doxycycline induction in FE-8 cells

Cells were cultured for 48 hours with (+) our without (-) doxycycline, then lysed and subjected to SDS-PAGE. Upper panel – western blot analysis was performed with a polyclonal anti-Hrev107 antibody, bottom panel – anti-actin antibody was used as a loading control. Lane 5 - H-rev107 expression in the FE-8 cells harbouring H-rev107 expression plasmid after induction with doxycycline


[page 80↓]

Table 6 Cell cycle analysis of FE-8 cells harbouring tet-inducible H-rev107

Cell lines

 

G1(%)

G2(%)

S(%)

FE-8 pUHD

 

38,8

27,8

33,4

FE-8 pUHD + Tet

48h

40,2

21,4

38,4

 

72h

38,3

25

36,7

 

96h

35,6

25,3

39,1

 

120h

37,1

30,3

32,6

FE-8 H-rev107

 

42,3

22,9

34,8

FE-8 H-rev107 + Tet

48h

51,4

22,9

25,7

 

72h

55,9

23,1

21

 

96h

59,7

22,6

17,7

 

120h

60

18,8

21,2

FE-8 cells stably transfected with the empty pUHD vector, and cells harbouring H-rev107 under the control of a tetracycline-inducible promoter in pUHD were cultured in Tet-free medium with or without doxycycline. 48, 72, 96 and 120 hours after the induction, cells were stained with propidium iodide and subjected to flow cytometry. Cell cycle progression analysis revealed a tendency of the cells over-expressing H-rev107 to arrest in G1 phase.

Flow cytometry analysis demonstrated that the FE-8 cells over-expressing H-rev107 were arrested in G1, reaching 60% G1 phase after 120 hours. In contrast, only 37% of the pUHD-transfected cells were in the G1 phase at the same time (Table 6). Thus, the H-rev107 protein appeared to induce G1 arrest in rat fibroblasts. At the same time nuclear morphology was analysed using laser confocal microscopy. However no changes were observed after 48, 96, and 120 hours of H-rev107 induction excluding the possibility of apoptosis.

3.4.2. The ΔC107-ΔN Interaction Deficient Mutant Fails to Induce Apoptosis in Human Ovarian Carcinoma Cell Lines A27/80 and OVCAR-3

In contrast to FE-8 cells, forced expression of H-REV107-1 in OVCAR-3 and A27/80 cell lines was already demonstrated to induce apoptosis (Fig. 15). Therefore, we suggested that in different cell types and (or) different species H-REV107-1 exerts different effects on cellular growth. Further investigation of a potential role of the PR65 – H-REV107-1 interaction in the H-REV107-1 – mediated cell death was performed in these cells.


[page 81↓]

The growth suppressing properties of H-REV107-1 full length, ΔCH-REV107-1, and the interaction deficient mutant (ΔC107-ΔN) were compared using a colony assay. These experiments were kindly performed by Jaqueline Hellwig (Institute of Pathology, Charité, Berlin, Germany). The expression vectors were introduced in A27/80 and OVCAR-3 cells. The number of G418 – resistant colonies was calculated after 2-3 weeks of selection. In OVCAR-3 cells the full length H-REV107-1 protein reduced the number of colonies to 45% of the control. The truncated form (ΔCH-REV107-1) reduced colonies only to 65%, and the mutant (ΔC107-ΔN) was unable to reduce colony formation in OVCAR-3 cells. In A27/80 cells, the colony numbers with full length H-REV107-1 were 78%, the truncated form of the protein inhibited growth up to 55%, and the (ΔC107-ΔN) mutant again failed to reduce number of colonies (Fig. 26).

For further analysis of the H-REV107-1 role in apoptosis induction, the induction of apoptosis by the H-REV107-1 mutants was analysed in OVCAR-3 cells. The cells were transiently transfected with the H-REV107-1, ΔCH-REV107-1, ΔC107-NCE, ΔC107-HWAY, and ΔC107-ΔN expression vectors. Nuclei with apoptotic phenotype were calculated 42 hours after transfection. Expression of the full length H-REV107-1 cDNA resulted in 18,5% apoptotic nuclei, the truncated form lacking the membrane binding domain with or without point mutations led to induction of apoptosis in around 11% of the cells. After transfection of the ΔC107-ΔN deletion mutant only 4,3% of the cells underwent apoptosis. The background was calculated as the number of apoptotic nuclei after transfection with the empty vector, and was found to be 2,6% (Table 7).

These experiments demonstrated that two H-REV107-1 protein domains were important for the growth suppression and induction of apoptosis in OVCAR-3 cells: the C-terminal domain is responsible for the intracellular localisation of H-REV107-1 (Sers et al., 1997), the N-terminal domain mediates the binding of the PR65 protein, and is responsible for homodimer formation.

OVCAR-3 cells were transiently transfected with 6 different expression plasmids. Immunofluorescence analysis was performed 42 hours after transfection. Cells were fixed and stained with DAPI. Apoptotic nuclei were visually determined and calculated using Laser Confocal microscopy. For every transfection more than 300 hundred cells were counted. +


[page 82↓]

Fig. 26 Reduction of colony formation in OVCAR-3 and A27/80 cells

A27/80 and OVCAR-3 cells were transfected with four different expression plasmids contained: H-REV107-1 full length cDNA, Δ CH-REV107-1, and H-REV107-1 lacking N-terminal and C-terminal domains ( Δ C107- Δ N), or pcDNA3 used as a control. G418-resistant colonies were calculated. Number of colonies transfected with pcDNA3 was equated to 100%. The results of four independent experiments were evaluated

3.4.3. Cellular Re-Distribution of the Endogenous PR65 Protein Correlates with the H-REV107-1 Induced Apoptosis in OVCAR-3 Cells

The N-terminal domain of the H-REV107-1 protein was shown to be required for apoptosis in OVCAR-3 and A27/80 cells. This suggested that the interaction of H-REV107-1 with PR65 can be important for this process. Therefore the phenotype of OVCAR-3 cells transiently transfected with the H-REV107-1 and ΔC107-ΔN expression vectors were investigated at a single cell level.

In untransfected cells, PR65 was localised mainly around the nucleoli throughout the interphase, and in the cytoplasm during mitosis (Fig. 27 C, arrow). H-REV107-1 transfection induced a dramatic change in the intracellular distribution of PR65 and in the appearance of the nuclei. In the cells expressing H-REV107-1, PR65 immunofluorescence was seen only in the cytoplasm, that accompanied apoptotic nuclear morphology (Fig. 27 A, yellow arrows). The ΔNΔC107-1 mutant was distributed through the cytoplasm and the nucleus (Fig. 27 B, upper right panel).


[page 83↓]

The protein was unable either to induce apoptosis, or to change PR65 localisation (Fig. 27 B, yellow arrows).

Colocalisation analysis of H-REV107-1 and PR65 immunostaining revealed that the change in the distribution of the PR65 protein induced by H-REV107-1 correlated with a co-distribution of these two proteins. This can be clearly observed in the overlay images (Fig. 27 A, B, icons “Overlay”) in which yellow colour (a result of the addition of green and red spectra) indicate colocalisation. Yellow coloured areas were seen only for H-REV107-1 and PR65 staining, demonstrating that these proteins are colocalised (Fig. 27 A, icons “Overlay”). For the ΔNΔC107-1 and PR65 proteins colocalisation regions were not obtained. (Fig. 27 B, icon “Overlay”). In order to quantify these observations, the yellow, red and green pixels were counted and depicted as a diagram (Fig. 27 A, B, icons “Colocalisation”; these data were kindly provided by Kerstin Lehmann, MetaGen, Berlin, Germany).

These results demonstrated a direct correlation between interaction of H-REV107-1 with the PR65 protein, and the ability of H-REV107-1 to induce apoptosis in OVCAR-3 cells accompanied by the transport of PR65 from the nucleus to the cytoplasm.

The PP2A protein appears in cells mostly as a dimeric core complex consisting of the 36-kDa catalytic C subunit (PR36, α or β isoforms) and the 65-kDa “scaffolding” regulatory A subunit (PR65, α or β isoforms), or as a trimeric complex containing variable regulatory B subunits bound to the AC dimer (Janssens and Goris, 2001; Millward et al., 1999). The PR65α was demonstrated to interact with H-REV107-1. We asked whether H-REV107-1 – PR65α protein complex might interact with the PP2A catalytic subunits, and if over-expression of the H-REV107-1 protein in OVCAR-3 cells can alter the intracellular distribution of other PP2A subunits. The localisation of the catalytic subunit PR36 was determined in OVCAR-3 cells using immunofluorescence analysis.

The commercially available antibody used recognised only the PR36 α isoform. Immunofluorescence analysis revealed that the catalytic subunit is distributed invariably through the cytoplasm (Fig. 27 C), and that over-expression of H-REV107-1 did not alter the protein localisation (data not shown). The expression and the intracellular distribution of other PP2A subunits is currently investigated.

3.4.4. H-REV107-1 Inhibits PP2A Activity in vitro

Finally, we asked if the interaction between H-REV107-1 and PR65α might influence the PP2A catalytic activity in OVCAR-3 cells. Therefore, the ability of PP2A, precipitated from OVCAR-3 cells, to dephosphorylate the synthetic substrate p-Nitrophenyl phosphate (pNPP) was measured in vitro. OVCAR-3 cells were transfected either with the pcDNA3 vector, or with the H-REV107-1, ΔCH-REV107-1, and ΔC107-ΔN expression vectors. As a control, OVCAR-3 cells were treated during 1 hour with the PP2A inhibitor okadaic acid at 2 nM.


[page 84↓]

Fig. 27 Induction of apoptosis in OVCAR-3 cells after H-REV107-1 over-expression is correlated with a re-distribution of the PR65 protein.

[page 85↓]
Fig. 27 Induction of apoptosis in OVCAR-3 cells after H-REV107-1 over-expression is correlated with a re-distribution of the PR65 protein.

A, B: OVCAR-3 cells were transiently transfected with the H-RE107-1 (A) and Δ C107- Δ N (B) expression vectors. Forty six hours after transfection immunofluorescence was performed. Cells were stained with monoclonal anti-PR65, and a polyclonal anti-H-REV107-1 primary antibodies, and anti-mouse AlexaFluor 488 and anti-rabbit AlexaFluor 594 secondary antibodies. Nuclei were stained with DAPI. Immunofluorescence was analysed with the help of laser confocal microscopy. Overlay was performed using LCS Multicolour-Software.
C: Immunofluorescence of untransfected OVCAR-3 cells was performed to investigate the intracellular localisation of the endogenous PP2A catalytic subunit, PR36. Cells were stained with monoclonal anti-PR36 or anti PR65 antibodies, and anti-mouse Alexa-Fluor 488 secondary antibody


[page 87↓]

Fig. 28 The H-REV107-1 protein inhibits the catalytic activity of PP2A in vitro

OVCAR-3 cells were transiently transfected with the H-REV107-1, Δ CH-REV107-1, and the Δ C107- Δ N mutant expression vectors. Cells transfected with the pcDNA3 empty vector were used as a positive control. For the negative control untransfected cells were treated for 1 hour with 2 nM okadaic acid. Forty six hours after transfection cells were lysed in phosphatase lysis buffer and immunoprecipitated with an anti-PP2A/A antibody. Protein phosphatase activity of the precipitated protein complexes was measured by pNPP hydrolysis at 405 nm. Absorbance units reflected the phosphatase activity. PP2A activity of cells transfected with a pcDNA3 vector was set 100%.

The amount of precipitated phosphatase, and H-REV107-1 expression were controlled by Western Blot analysis of the precipitated complexes and protein extracts, respectively (Fig. 29). The PP2A activity measured in the pcDNA3 transfected OVCAR-3 cells was set 100%. Treatment of OVCAR-3 cells with okadaic acid inhibited the phosphatase activity up to 94.5%. Surprisingly, the H-REV107-1 protein also inhibited PP2A activity by almost 80%. The ΔC107-1ΔN mutant was unable to inhibit PP2A (Fig. 28).

Thus, PP2A is active in untransfected OVCAR-3 cells, although regulatory and catalytic subunits are preferentially localised in different cellular compartments. Overexpression of the H-REV107-1 protein leads to the inhibition of the PP2A activity, and the N-terminus of H-REV107-1 appeared to be important for the inhibition, suggesting that the interaction between H-REV107-1 and PR65α is required for this process.

[page 88↓]
Fig. 29 Control Western blot analysis of the precipitated protein complexes

Upper panel – precipitated proteins were analysed by western blotting with an anti-PR65 antibody. Bottom panel – control western blot analysis of the protein extracts with an anti-H-REV107-1 antibody.
Lane 1 – OVCAR-3 cells transfected with an pcDNA3 plasmid, lane 2 – OVCAR-3 treated for 2 hours with okadaic acid 2 nM (OA), lane 3 – OVCAR-3 cells transfected with the H-REV107-1 expression vector, lane 4 –
Δ CH-REV107-1, lane 5 – Δ C107- Δ N protein.

3.4.5. Okadaic Acid Induces Apoptosis in OVCAR-3 Cells

Since the induction of apoptosis correlated with the H-REV107-1 ability to bind PR65 and to inhibit PP2A activity, the natural PP2A inhibitor, okadaic acid was used to study the potential role of the phosphatase in the cell survival. OVCAR-3 cells were treated with 10 nM okadaic acid (OA) for 48 hours and 72 hours, as a control DMSO was added to the medium. The percentages of cells undergoing apoptosis were quantified by dual-coloured flow cytometric analysis after staining with Propidium iodide and Annexin-V. Treatment of OVCAR-3 cells with 10 nM OA for 48 hours resulted in 12,1 % apoptotic cells, after 72 hours 54,8% of cells underwent apoptosis (Fig. 30).

Although in other cell lines PP2A was demonstrated to play a pro-apoptotic role (Klumpp and Krieglstein, 2002; Moon and Learner, 2003), the protein phosphatase 2 A was likely to be important for the survival in OVCAR-3 cells.


[page 89↓]

Fig. 30 Induction of apoptosis by OA treatment in OVCAR-3 cells

Subconfluent OVCAR-3 cells were treated either with 10 nM OA for 48 and 72 hours or with its vehicle DMSO for 72 hours. After harvesting the cells were subsequently stained with Propidium iodide (P) and fluorescein isothiocyanate (FITC)-conjugated anti-Annexin-V antibody (A). The cytograms demonstrate Annexin – binding (ordinate) versus Propidium iodide – binding (abscissa) in DMSO and OA – treated cells. Vital cells (A-/P-), pre-apoptotic cells (A+/P-), apoptotic cells (A+/P+), and residual damaged cells (A-/P+) are shown in respective quadrants.

3.4.6. PP2A Inhibition in OVCAR-3 Cells Leads to the Activation of Procaspase-9

Okadaic acid, when applied in a concentration of 10 nM, inhibits the activity of two protein phosphatases: PP2A and PP1 (Ishihara et al., 1989). The PP2A – specific concentration of 0.5 nM was used in further experiment to test if the induction of apoptosis in OVCAR-3 cells required only PP2A inhibition. Two known apoptotic pathways result in the activation of caspases. The first one is receptor-autonomous, depends on the participation of mitochondria and results in the activation of procaspase-9. The second pathway involves the interaction of death receptors with their ligands, and results in procaspase-8 activation (Zimmerman et al., 2001). Induction of these two pathways was investigated in OVCAR-3 cells.


[page 90↓]

Fig. 31 H-REV107-1 overexpression and OA treatment of OVCAR-3 cells activate procaspases –9 and –3

OVCAR-3 cells were transiently transfected with the H-REV107-1, Δ CH-REV107-1, Δ C107-NCE, Δ C107-HWAY, and Δ C107- Δ N expression vectors, or treated with 0.5 nm and 10 nm OA. Protein extracts were harvested 44 hours after transfection, and subjected to SDS-PAGE followed by Western Blot analysis.
A: Western blotting against cleaved caspase-9 antibody. Activated caspase-9 was revealed in OVCAR-3 cells after OA, or LY294002 treatment, or after H-REV107-1 expression. B: Activated downstream caspase-3 was detected after treatment with OA 0,5 nM and after H-REV107-1 ectopic expression. DMSO and pcDNA3 are negative controls. C: Induction of caspase-9 was tested in OVCAR-3 cells expressing H-REV107-1 wild type or the mutants.
Δ C107- Δ N failed to induce cleavage of the caspase-9.


[page 91↓]

Ectopic expression of the H-REV107-1 protein and okadaic acid treatment resulted in activation of the procaspase-9. Cleaved caspase-8 was not detected (data not shown). Cleaved caspase-3 was also tested as one of the down-stream caspases. The cleaved product was revealed after H-REV107-1 transfection and OA exposition (Fig 31 B). This demonstrated that the H-REV107-1 protein, similar to the PP2A inhibitor okadaic acid, induces apoptosis through the mitochondrial-dependent pathway.

High level of cleaved caspase-9 and -3 were only found after H-REV107-1 overexpression, and after treatment with OA of 10 nM, whereas incubation with 0.5 nM OA resulted in considerably weaker activation of both caspases (Fig. 31 A, B). This suggests that in OVCAR-3 cells 0.5 nM okadaic acid is insufficient for a complete inactivation of the PP2A activity.

The ability of the H-REV107-1 mutants to activate cleavage of procaspase-9 was examined in further experiments. The result confirmed the previous observations. Overexpression of the ΔCH-REV107-1, ΔC107-NCE and ΔC107-HWAY mutants resulted in a weaker activation of procaspase-9 in comparison with the wild type H-REV107-1. The ΔC107-ΔN truncated protein failed to induce procaspase-9 cleavage in OVCAR-3 cells (Fig. 31 C).

3.5. Confirmation of Interaction between H-REV107-1 and RARG, S100A6, ETF1, and P14.5

3.5.1. RARG

The retinoic acid receptor gamma (RARG) belongs to the steroid/thyroid hormone nuclear receptor superfamily (Evan, 1988). This family contains two distinct sub-classes of proteins, the all-trans retinoic acid (ATRA) and the 9-cis retinoic acid receptors RAR and RXR. Each sub-class consists of α, β, and γ isoforms (Gudas, 1992). The RAR and RXR proteins form homodimers or heterodimers, and act as nuclear transcription factors (Zhang et al., 1992). In view of the growth suppressive properties of ATRA, and the participation of RARG in this signaling, we suggested a possible functional co-operation of the H-REV107-1 and RARG proteins in growth control. Therefore, we analysed the interaction between H-REV107-1 and RARG in COS-7 cells.

The mammalian expression vector containing full length of the RARG cDNA fused to a V5-epitope (RARG-V5) was purchased from GeneStorm Collection of Invitrogen. For co-immunoprecipitation of ΔCH-REV107-1HA and RARG-V5, COS-7 cells were transiently transfected with the corresponding expression plasmids, however no interaction was detected (data not shown). From the literature it was known that binding with ATRA or its antagonists can alter the binding capacity of RARG (Zhang et al., 1992).


[page 92↓]

Potential candidates that could improve the interaction between RARG and H-REV107-1 were all-trans retinoic acid (ATRA), an antagonist of ATRA, (E)-4-(2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthelenyl)-propen-1yl benzoic acid) TTNPB, and 27 bp double-strand oligonucleotides containing the retinoic acid responsive element (RARE), DR5. The mutated RARE, DR5M was used as a negative control in co-immunoprecipitation (Fig. 32).

H-REV107-1HA did not interact with RARGV5 without supplementary ligands (Fig. 32 A, lane 4). Addition of either TTNPB, or DR5 weakly enhanced the interaction (Fig. 32 A, lane 5; Fig. 32 B, lanes 3, 5). A combination of TTNPB and DR5 had an additive effect (Fig. 32A, lane 6; Fig. 32 B, lane 2), ATRA did not affect the interaction.

3.5.2. S100A6, ETF1, and P14.5

To verify the interaction between H-REV107-1 and the S100A6, ETF1, and P14.5 proteins by co-immunoprecipitation, we generated several mammalian expression vectors. The open reading frames (ORF) encoding S100A6, and ETF1 proteins fused with an HA-epitope, and P14.5 protein without epitope, were amplified by PCR, and cloned into the pcDNA3.1 mammalian expression vector resulting in the S100A6-HA, ETF1-HA, and P14.5 expression plasmids. Sequencing analysis of the cloned inserts confirmed that S100A6, and p14.5 contained full length cDNAs. The ETF1 insert included 711 bp encoding the C-terminal domain of 237 amino acids of the ETF1 protein, which has a full length of 437 aminoacids.

3.5.2.1. S100A6

The S100A6/Calcyclin protein belongs to the S100 subfamily of Ca(2+)-binding proteins (Ferrari et al., 1987). It was shown to be over-expressed in a variety of tumor cell lines and human tumours (Maelandsmo et al., 2000), to be associated with invasion in colorectal adenocarcinoma and metastasis in liver cancer (Komatsu et al., 2002). The first co-immunoprecipitation experiment under standard conditions revealed a negative result (data not shown). However, the protein-protein interaction between S100A6 and S100A2 proteins required an activation of Calcyclin by Ca2+ and Zn2+ ions (Filipek et al., 1999). In the following experiments a lysis buffer containing different concentration of Ca2+ and Zn2+- ions ranging from 1 nM to 100 nM was tested. A concentration of 2 nM of both ions was found optimal. For the co-immunoprecipitation, COS-7 cells were transiently transfected with the H-REV107-1V5 and S100A6HA expression vectors. The cells were lysed in the lysis buffer containing 2 nM CaCl2 and ZnCl2, and precipitated using the anti-V5 antibody (Fig. 33). Co-immunoprecipitation of the S100A6 protein with H-REV107-1 has been reproduced two times.


[page 93↓]

Fig. 32 TTNPB and DR5 are essential for the interaction between RARG and H-REV107-

COS-7 cells were transiently transfected with the H-REV107-1HA and RARGV5 expression vectors, and lysed 48 hours post transfection. As a negative control COS-7 cells were transfected with the epitope – containing plasmids without insert. The whole cell extract was used for immunoprecipitation with an anti-V5 antibody. H-REV107-1 was identified in the immunocomplexes using anti-HA antibody. RARGV5 was identified using an antibody against V5 epitope. A: lane 2, 3 – negative controls, lane 4 – immunoprecipitation in the absence of additional ligands; lane 5 – in presence of TTNPB; lane 6 – with both TTNPB ligand and DR5 RARE. B: Examination of a role of different ligands on the interaction. Lane 2 – DR5 and TTNPB, lane 3 – DR5, lane 4 – DR5 and ATRA, lane 5 – TTNPB.


[page 94↓]

3.5.2.2. ETF1

Eukaryotic translation termination factor 1 (ETF1) belongs to a family of tightly related proteins catalysing termination of protein synthesis (Guenet et al., 1999). Immunoprecipitation with an anti-V5 antibody was performed in COS-7 cells transiently transfected with the H-REV107-1V5 and ETF1-HA expression vectors. Immunoprecipitated protein complexes were analysed by Western blotting with anti-HA, and anti-V5 antibodies. The interaction between these two proteins was very weak (Fig. 34). This correlated with the results obtained in the mating test where the interaction between H-REV107-1 and ETF1 was also shown to be weak (Fig. 7). A co-immunoprecipitation experiment vice versa with an anti-HA antibody revealed a negative result (data not shown). These experiments demonstrated that the interaction between ETF1 and H-REV107-1 in COS-7 cells is very weak, and it was not investigated further.

Fig. 33 Calcyclin (S100A6) interacts with H-REV107-1 in COS-7 cells

COS-7 cells were transiently transfected with the H-REV107-1V5 and S100A6HA expression vectors and lysed in the Ca 2+ and Zn 2+ containing lysis buffer 48 hours after transfection. As a negative control cells were transfected with epitopes- containing plasmids without inserts. One mg of the protein extract was used for the co-immunoprecipitation with an anti-V5 antibody. Protein complexes were subjected to SDS-PAGE and immunoblotted against anti-V5 and -HA antibodies.
Upper panel – H-REV107-1V5 was immunoprecipitated and detected with an anti-V5 antibody, Lower panel – co-precipitated S100A6-HA was detected with an anti-HA antibody. Lane 1 – 15
μ g of protein extract used for the co-immunoprecipitation. Lane 2 – negative control. Lane 3 – co-immunoprecipitation of the H-REV107-1V5 and S100A6HA proteins.


[page 95↓]

Fig. 34 H-REV107-1 interacts weakly with ETF1 in COS-7 cells

COS-7 cells were transiently transfected with the H-REV107-1V5 and ETF1-HA expression vectors. Co-immunoprecipitation was performed 48 hours after transfection with an anti-V5 antibody.
Upper panel – protein complexes were analysed by Western blotting against an anti-HA antibody, lower panel – against an anti-V5 antibody. Lane “PE” – 10
μ g of protein extract used for the immunoprecipitation. Lane “IP“ – co-immunoprecipitation of the H-REV107-1V5 and ETF1HA proteins.

Fig. 35 H-REV107-1 fails to bind P14.5 in COS-7 cells

H-REV107-1-V5 and p14.5 – expression vectors were simultaneously transfected into COS-7 cells. As a negative COS-7 cells were transfected with epitope- containing plasmids without insert. Immunoprecipitation with an anti-V5 antibody was performed 48 hours after transfection.
Upper panel: Western blotting with an anti-p14.5 antibody; lower panel: Western blotting with an anti-V5 antibody.
Lane 1 - protein extracts used for the immunoprecipitation; lane 2, 3 – negative controls; lane 4 – co-immunoprecipitation of the H-REV107-1 and P14.5 proteins. Abbreviations: PE – protein extract, IP – co-immunoprecipitation.


[page 96↓]

3.5.2.3.  P14.5

The human protein P14.5 was identified in 1996 as the homologue of the rat perchloric acid-soluble protein. Its weak expression in a variety of liver and kidney tumor cells and high expression in fully differentiated cells suggested that the protein is involved in differentiation. The protein was described as an inhibitor of protein synthesis in vitro (Schmiedeknecht et al., 1996). To prove whether H-REV107-1 interacts with P14.5, COS-7 cells were transiently transfected with P14.5 and H-REV107-1V5 expression vectors. The cells were lysed 48 hours post transfection, and H-REV107-1V5 was immunoprecipitated using the anti-V5 antibody. Western Blot analysis of the immunoprecipitated protein complexes revealed H-REV107-1V5 (Fig. 35, lane 4) but not the P14.5 protein, although P14.5 was presented in the protein extract used for co-immunoprecipitation (Fig. 35, lane 4 and 1, respectively). It was concluded that under the conditions used, H-REV107-1 does not interact with the P14.5 protein in COS-7 cells.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
15.03.2004