Amplification of 3q25-q26 was one of the most frequent chromosomal alterations in human ovarian carcinoma. A chromosome microdissection-hybrid selection method was applied to isolate transcribed sequences from a primary ovarian cancer containing high-copy-number amplification of 3q26 using 3q26 band-specific DNAs generated by chromosome microdissection. Using this method, we have isolated a novel candidate oncogene eIF-5A2 (eukaryotic initiation factor 5A2). eIF-5A2 shares 82% identity of amino acid sequence with eIF-5A including the minimum domain needed for eIF-5A maturation by hypusine modification at lysine-50 residue. Amplification and overexpression of eIF-5A2 was frequently detected in primary ovarian cancers and ovarian cancer cell lines. The proliferation-related function of eIF-5A supports that eIF-5A2 is a candidate oncogene related to the development of ovarian cancer.

Ovarian cancer is the leading cause of death from female genitourinary cancers and will affect approximately 1% of American women during their lifetime (1). The prognosis of ovarian cancer is very poor, with a 5-year survival rate of <20% if not diagnosed early. Compared with other solid tumors, relatively little is known regarding the molecular pathogenesis and progression of ovarian cancer. Recently, a frequently amplified region at 3q25-qter was identified in primary ovarian carcinoma by CGH3(2, 3, 4). The minimum overlapping amplified region has been mapped to 3q26. In our CGH study, gain of 3q was detected in 15 of 30 (50%) primary ovarian cancers.4 High-copy-number amplification of 3q within a minimal overlapping region at 3q26 was observed in four cases (OV-4, OV-7, OV-13, and OV-27). Gain of 3q has also been frequently detected in other solid tumors including esophageal carcinoma (5), prostate cancer (6), and carcinoma of the fallopian tube (7). These studies strongly indicate that 3q26 may contain an oncogene(s) related to the pathogenesis and progression of various solid tumors, including ovarian cancer.

Identification of commonly amplified chromosomal regions and corresponding oncogenes within the region is imperative to understand the molecular mechanism of tumorigenesis. Recently, PIK3CA which encodes a catalytic subunit of phosphatidylinositol 3-kinase has been implicated as a candidate oncogene at 3q26 (8). Since only low-level amplification of PIK3CA was detected in OV-4, OV-27, and an ovarian cancer cell line UACC-1598 that contains high-level amplification at 3q26 in the form of double minutes (9), there is likelihood of another oncogene residing in this region. In this report, we used a hybrid selection method to rapidly isolate a novel expressed gene from the microdissected DNA of 3q26.

Tumor Samples and Cell Lines.

Primary ovarian cancer samples were collected at the time of surgical resections from the Cancer Institute, Sun Yat-sen University of Medical Sciences (Guangzhou, China). Ovarian cancer cell lines UACC-1598 and UACC-2727 were obtained from the Tissue Culture Core Service of the University of Arizona Comprehensive Cancer Center. Ovarian cancer cell line OVCAR3 was obtained from the American Type Culture Collection (Manassas, VA).

Chromosome Microdissection.

Chromosome microdissection and PCR amplification of microdissected DNA were performed as described previously (10). Briefly, five copies of the 3q26 band were dissected from G-banding normal metaphase chromosomes. The dissected DNA fragments were treated with topoisomerase I and then amplified by PCR with UN1 primer (5′-CCGACTCGAGNNNNNNATGTGG-3′).

cDNA Synthesis and Library Construction.

A random primed cDNA library was constructed in bacteriophage λGEM4 according to the manufacturer’s protocol (RiboClone cDNA Synthesis System; Promega, Madison, WI). For hybrid selection, random primed cDNA was prepared from OV-4 mRNA using a SMART PCR cDNA Library Construction kit (Clontech Laboratories, Palo Alto, CA).

Hybrid Selection.

Hybrid selection was performed as described previously (11). The method was to apply random primed cDNA from ovarian cancer OV-4 containing 3q26 amplification to select coding region sequences from microdissected DNA fragments of 3q26. Briefly, 5 μg of random primed cDNAs prepared from the OV-4 cells were immobilized on a 7 mm in diameter nylon membrane and then hybridized with PCR-amplified microdissected DNA fragments from the 3q26 region in 200 μl of hybridization solution (5× SSC, 5× Denhardt’s solution, 0.1% SDS, 100 μg/ml salmon sperm DNA, and 100 μg/ml human Cot-1 DNA) at 65°C with gentle rotation overnight. After a stringent wash, the hybridized microdissected DNA was eluted and recovered by PCR.

cDNA Library Screen.

PCR-recovered, selected DNA fragments (100 ng) were labeled with 32P by random priming (Life Technologies, Inc., Rockville, MD) and then purified over Sephadex G50. The probe was prehybridized with 100 μg of human Cot-1 DNA at 65°C for 1 h and then hybridized to cDNA library plaque lifts by the standard method.

Chromosome Mapping.

A BAC clone selected through BLAST database searching was labeled with Spectrum Orange-dUTP by nick translation (Life Technologies, Inc.). The labeled probe was then hybridized to prebanded normal lymphocyte metaphase chromosomes by FISH according to the method described previously (11).

Southern and Northern Blot Analyses.

Genomic DNA from human placenta, primary ovarian cancer specimens, and ovarian cancer cell lines was isolated using the SDS/phenol/chloroform method. DNA was digested with EcoRI, fractionated on 1% agarose gel, transferred to a nylon membrane (Zeta-probe; Bio-Rad, Hercules, CA), and hybridized overnight at 42°C with 32P-labeled probes for the eIF-5A2 and PIK3CA genes. Total cellular RNA was prepared using the TRIzol/chloroform method, size fractionated on 1% agarose/2.2 m formaldehyde gel, transferred to a nylon membrane, and hybridized with 32P-labeled probes for the eIF-5A2 and PIK3CA genes.

Chromosome Microdissection.

To identify the candidate oncogene(s) within the 3q26-amplified region, chromosome microdissection of 3q26 was performed from normal metaphase chromosomes. These microdissected DNA sequences were demonstrated specifically hybridizing to 3q26 by FISH (Fig. 1 B).

Hybrid Selection.

3q26 region-specific expressed sequences were selected from microdissected DNA by hybrid selection. Thirty-one positive cDNA clones were isolated and sequenced by screening 5 × 104 plaques. Sequencing results revealed that 4 of 31 cDNA clones (sized from 0.5 to 2.1 kb) represent a novel gene with an open reading frame encoding 153 amino acids (Fig. 2). Database searching with BLAST indicated that the cDNA shared a significant sequence homology (126 of 153, 82% amino acid identity) to eIF-5A (Fig. 2). The sequence identity includes the domain needed for hypusine modification and the lysine-50 residue where the hypusine residue can be formed by posttranlational modification (12). eIF-5A is the only known cellular protein that contains a hypusine residue, which is formed by posttranslational modification of the lysine-50 residue (13). Hypusination of eIF-5A has been shown to be essential for the growth of some tumorigenic cell lines. The novel gene was named eIF-5A2. The GenBank accession number for eIF-5A2 is AF262027. To determine the expression pattern of eIF-5A2 in normal tissues, RT-PCR using primers designed based on the eIF-5A2 sequence was performed on normal ovary tissue and tissues from liver, placenta, breast, skin, brain, and colon. The results showed that eIF-5A2 is expressed in all of these tissues (Fig. 3 C). Expressed sequence tag database search through BLAST revealed that eIF-5A2 is also expressed in human testis, uterus, lymphocyte, and lung tissues.

Map of eIF-5A2 to 3q26.2.

To confirm the chromosome localization, a BAC clone containing the eIF-5A2 sequence was isolated through BLAST searching. The BAC clone was confirmed containing eIF-5A2 by Southern blot analysis (Fig. 3,A) and PCR using designed primers within an exon (204 bp; Fig. 3,B). This BAC clone was then mapped to 3q26.2 by FISH (Fig. 1, C and D).

Amplification of eIF-5A2 in Ovarian Cancer.

DNA sequence amplification of eIF-5A2 in 30 primary ovarian cancers was studied by Southern blot analysis. Amplification of eIF-5A2 was detected in 15 of 30 cases and high-copy-number amplification was observed in 4 cases (OV-4, OV-7, OV-13, and OV-27; Fig. 4,A). High-copy-number amplification of eIF-5A2 was also detected in the ovarian cancer cell line UACC-1598, which contains 3q26 amplification in the form of double minutes (Fig. 4,A). Expression level of eIF-5A2 in four primary ovarian cancers (OV-4, OV-7, OV-24, and OV-27) and ovarian cancer cell line UACC-1598 was analyzed by Northern blot and overexpression of eIF-5A2 was observed in all cases (Fig. 4,B). To confirm whether the double minutes contain eIF-5A2, metaphase prepared from UACC-1598 was hybridized with the BAC clone containing eIF-5A2 by FISH. The result demonstrated that the BAC clone hybridized to all double minutes (Fig. 5,A). Amplification of eIF-5A2 was also observed in another ovarian cancer cell line (OVCAR3) by FISH (Fig. 5,B). According to our CGH results, a minimal overlapping amplicon at 3q26.1-26.2 was defined and eIF-5A2 is mapped exactly within the region (Fig. 5 C).

We describe in this report the application of a chromosome microdissection-hybrid selection technique to isolate a novel member of the eIF-5A family, eIF-5A2, from a frequent amplified region at 3q26 in ovarian cancer. The identity of the amino acid sequence between eIF-5A2 and eIF-5A is 82%, including the minimum domain needed for hypusine modification and the lysine-50 residue where the hypusine residue can be formed by posttranlational modification. This indicates that eIF-5A2 is an eIF-5A family member and may have similar function.

It has been suggested that eIF-5A plays a role in the translation initiation; however, the initiation of protein synthesis in eIF-5A-depleted yeast cells was not significantly affected (14). Although the precise function of eIF-5A is unclear, the necessity for the hypusine-containing eIF-5A for cell proliferation is well studied (14, 15). It has been demonstrated that complete intracellular depletion of eIF-5A by gene deletion resulted in inhibition of cell growth (13). Other studies indicated that the inhibition of deoxyhypusine synthase, the enzyme involved in the hypusination reaction of eIF-5A, inhibited Chinese hamster ovary cell proliferation (16), suppressed the growth of HeLa cells and v-src-transformed NIH3T3 cells (17). Hanauske-Abel et al.(18) suggested that hypusine containing eIF-5A may directly affect the expression of a selective set of genes involved in the G1 to S transition of the cell cycle in the eukaryotic cells since the cell cycle could be arrested at the G1-S boundary by inhibition of eIF-5A hypusine modification. Recently, Tome et al.(19) reported that suppression of eIF-5A hypusine modification by excess putrescine accumulation or treated cells with diaminoheptane is one mechanism by which cells may be induced to undergo apoptosis. A recent study showed that the hypusine formation activity is serum responsive and significantly increased by >30-fold in Ras oncogene-transformed NIH3T3 cells (20). All of these studies strongly suggested the association between eIF-5A and cancer development.

In the present study, amplification of eIF-5A2 was detected in 15 of 30 primary ovarian cancers and several ovarian cancer cell lines, including UACC-1598 which contains high-copy-number amplification of the 3q26 region. Overexpression of eIF-5A2 was observed in all four tested primary ovarian cancers with 3q26 amplification. In addition, a BAC clone containing eIF-5A2 was hybridized to double minutes in the ovarian cancer cell line UACC-1598. Therefore, based on its chromosomal location, its amplification status in ovarian cancer, and its possible proliferation-related function, eIF-5A2 is considered as the putative oncogene within the minimal overlapping amplicon at 3q26.2 in ovarian cancer.

The amplified region of 3q26 may contain more than one important gene, including PIK3CA and eIF-5A2 which are the biological targets of amplification events in ovarian cancer. eIF-5A2 is telomeric to PIK3CA and the genomic distance between them is about 7cM. Further studies of the amplification and expression of eIF-5A2 in a large set of ovarian cancers using tissue microarray as well as determination of its function will be necessary to address the relationship between eIF-5A2 and ovarian cancer.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

      
1

Supported in part by the Leung Kwok Tze Foundation.

            
3

The abbreviations used are: CGH, comparative genomic hybridization; BAC, bacterial artificial chromosome; FISH, fluorescence in situ hybridization.

      
4

X-Y. Guan, T. C-M. Tang, K-F. Huo, J. S. T. Sham, Y. Fang, and J-M. Yang, unpublished data.

Fig. 1.

A, amplification of 3q25-q26 in two primary ovarian cancers detected by CGH. Tumor DNA and normal reference DNA was labeled with green and red color, respectively. B, microdissected DNA from 3q26 (red) was hybridized to normal metaphase chromosomes (partial metaphase was shown). C, G-banded partial metaphase with arrow to region of hybridization at 3q26.2. D, identical partial metaphase as c after FISH with BAC clone (red) containing eIF-5A2, which was localized at 3q26.2.

Fig. 1.

A, amplification of 3q25-q26 in two primary ovarian cancers detected by CGH. Tumor DNA and normal reference DNA was labeled with green and red color, respectively. B, microdissected DNA from 3q26 (red) was hybridized to normal metaphase chromosomes (partial metaphase was shown). C, G-banded partial metaphase with arrow to region of hybridization at 3q26.2. D, identical partial metaphase as c after FISH with BAC clone (red) containing eIF-5A2, which was localized at 3q26.2.

Close modal
Fig. 2.

Sequence alignment of the human eIF-5A2-deduced amino acid sequence with human eIF-5A and chicken eIF-5A; those identical sequences are highlighted in black. The minimum domain of eIF-5A needed for hypusine modification is boxed. ∗, lysine residue for posttranslationally modified to hypusine.

Fig. 2.

Sequence alignment of the human eIF-5A2-deduced amino acid sequence with human eIF-5A and chicken eIF-5A; those identical sequences are highlighted in black. The minimum domain of eIF-5A needed for hypusine modification is boxed. ∗, lysine residue for posttranslationally modified to hypusine.

Close modal
Fig. 3.

The BAC clone containing eIF-5A2 gene was confirmed both by Southern blot analysis (A) and by PCR using designed primers within an exon (204 bp; B). The genomic DNAs prepared from normal placenta and the BAC clone were digested with EcoRI and hybridized with eIF-5A2 cDNA. C, expression of eIF-5A2 in six normal tissues was detected using RT-PCR.

Fig. 3.

The BAC clone containing eIF-5A2 gene was confirmed both by Southern blot analysis (A) and by PCR using designed primers within an exon (204 bp; B). The genomic DNAs prepared from normal placenta and the BAC clone were digested with EcoRI and hybridized with eIF-5A2 cDNA. C, expression of eIF-5A2 in six normal tissues was detected using RT-PCR.

Close modal
Fig. 4.

A, Southern blot analysis demonstrates amplification of eIF-5A2 in five primary ovarian cancers and the ovarian cancer cell line UACC-1598. Fifteen micrograms of EcoRI-digested DNAs were separated by electrophoresis in 0.8% agarose gel and hybridized with a 1.2-kb eIF-5A2 cDNA probe. A probe for β-actin was used for loading control. B, Northern blot analysis of eIF-5A2 in four primary ovarian cancers and UACC-1598. Ten micrograms of total RNA were loaded in each lane. The Northern blot was hybridized with a 1.2-kb eIF-5A2 cDNA probe. The RNA separating gel stained with ethidium bromide for the Northern blot was used as loading control.

Fig. 4.

A, Southern blot analysis demonstrates amplification of eIF-5A2 in five primary ovarian cancers and the ovarian cancer cell line UACC-1598. Fifteen micrograms of EcoRI-digested DNAs were separated by electrophoresis in 0.8% agarose gel and hybridized with a 1.2-kb eIF-5A2 cDNA probe. A probe for β-actin was used for loading control. B, Northern blot analysis of eIF-5A2 in four primary ovarian cancers and UACC-1598. Ten micrograms of total RNA were loaded in each lane. The Northern blot was hybridized with a 1.2-kb eIF-5A2 cDNA probe. The RNA separating gel stained with ethidium bromide for the Northern blot was used as loading control.

Close modal
Fig. 5.

Amplification of eIF-5A2 in two ovarian cancer cell lines detected by FISH. A, hybridization of the BAC clone (red) containing eIF-5A2 to the double minutes in the ovarian cancer cell line UACC-1598. B, six copies of eIF-5A2 were observed in the ovarian cancer cell line OVCAR3. C, eIF-5A2 is mapped to the minimal overlapping amplification region at 3q26.2. Blackvertical lines, gains of 3q; thick lines, high-copy-number amplifications; redvertical line, amplification region in UACC-1598; grayhorizontal line, the minimal overlapping amplification region.

Fig. 5.

Amplification of eIF-5A2 in two ovarian cancer cell lines detected by FISH. A, hybridization of the BAC clone (red) containing eIF-5A2 to the double minutes in the ovarian cancer cell line UACC-1598. B, six copies of eIF-5A2 were observed in the ovarian cancer cell line OVCAR3. C, eIF-5A2 is mapped to the minimal overlapping amplification region at 3q26.2. Blackvertical lines, gains of 3q; thick lines, high-copy-number amplifications; redvertical line, amplification region in UACC-1598; grayhorizontal line, the minimal overlapping amplification region.

Close modal

We thank Dr. Fu-Yuan Liu for providing primary ovarian cancer samples and assistance in DNA and RNA preparation by Qiwan Liang.

1
Lynch H. T., Casey M. J., Lynch J., White T. E., Godwin A. K. Genetics and ovarian carcinoma.
Semin. Oncol.
,
25
:
265
-280,  
1998
.
2
Iwabuchi H., Sakamoto M., Sakunaga H., Ma Y. Y., Carcangiu M. L., Pinkel D., Yang-Feng T. L., Gray J. W. Genetic analysis of benign, low-grade, and high-grade ovarian tumors.
Cancer Res.
,
55
:
6172
-6180,  
1995
.
3
Arnold N., Hagele L., Walz L., Schempp W., Pfisterer J., Bauknecht T., Kiechle M. Overrepresentation of 3q and 8q material and loss of 18q material are recurrent findings in advanced human ovarian cancer.
Genes Chromosomes Cancer
,
16
:
46
-54,  
1996
.
4
Sonoda G., Palazzo J., du Manoir S., Godwin A. K., Feder M., Yakushiji M., Testa J. R. Comparative genomic hybridization detects frequent overrepresentation of chromosomal material from 3q26, 8q24, and 20q13 in human ovarian carcinomas.
Genes Chromosomes Cancer
,
20
:
320
-328,  
1997
.
5
Pack S. D., Karkera J. D., Zhuang Z., Pak E. D., Balan K. V., Hwu P., Park W. S., Pham T., Ault D. O., Glaser M., Liotta L., Detera-Wadleigh S. D., Wadleigh R. G. Molecular cytogenetic fingerprinting of esophageal squamous cell carcinoma by comparative genomic hybridization reveals a consistent pattern of chromosomal alterations.
Genes Chromosomes Cancer
,
25
:
160
-168,  
1999
.
6
Sattler H. P., Rohde V., Bonkhoff H., Zwergel T., Wullich B. Comparative genomic hybridization reveals DNA copy number gains to frequently occur in human prostate cancer.
Prostate
,
39
:
79
-86,  
1999
.
7
Heselmeyer K., Hellstrom A. C., Blegen H., Schrock E., Silfversward C., Shah K., Auer G., Ried T. Primary carcinoma of the fallopian tube: comparative genomic hybridization reveals high genetic instability and a specific, recurring pattern of chromosomal aberrations.
Int. J. Gynecol. Pathol.
,
17
:
245
-254,  
1998
.
8
Shayesteh L., Lu Y., Kuo W. L., Baldocchi R., Godfrey T., Collins C., Pinkel D., Powell B., Mills G. B., Gray J. W. PIK3CA is implicated as an oncogene in ovarian cancer.
Nat. Genet.
,
21
:
99
-102,  
1999
.
9
Brooks D. J., Woodward S., Thompson F. H., Dos Santos B., Russell M., Yang J. M., Guan X-Y., Trent J., Alberts D. S., Taetle R. Expression of the zinc finger gene EVI-1 in ovarian and other cancers.
Br. J. Cancer
,
74
:
1518
-1525,  
1996
.
10
Guan X-Y., Trent J. M., Meltzer P. S. Generation of band-specific painting probes from a single microdissected chromosome.
Hum. Mol. Genet.
,
2
:
1117
-1121,  
1993
.
11
Guan X-Y., Xu J., Anzick S. L., Zhang H., Trent J. M., Meltzer P. S. Hybrid selection of transcribed sequences from microdissected DNA: isolation of genes within an amplified region at 20q11–q13.2 in breast cancer.
Cancer Res.
,
56
:
3446
-3450,  
1996
.
12
Kyrpides N. C., Woese C. R. Universally conserved translation initiation factors.
Proc. Natl. Acad. Sci. USA
,
95
:
224
-228,  
1998
.
13
Park M. H., Wolff E. C., Folk J. E. Is hypusine essential for eukaryotic cell proliferation?.
Trends Biochem. Sci.
,
18
:
475
-479,  
1993
.
14
Kang H. A., Hershey J. W. B. Effect of initiation factor eIF-5A depletion on protein synthesis and proliferation of Saccharomyces cerevisiae.
J. Biol. Chem.
,
269
:
3934
-3940,  
1994
.
15
Park M. H., Lee Y. B., Joe Y. A. Hypusine is essential for eukaryotic cell proliferation.
Biol. Signals
,
6
:
115
-123,  
1997
.
16
Park M. H., Wolff E. C., Lee Y. B., Folk J. E. Antiproliferative effects of inhibitors of deoxyhypusine synthase. Inhibition of growth of Chinese hamster ovary cells by guanyl diamines.
J. Biol. Chem.
,
269
:
27827
-27832,  
1994
.
17
Shi X-P., Yin K. C., Ahern J., Davis L. J., Stern A. M., Waxman L. Effects of N1-guanyl-1, 7-diaminoheptaane, an inhibitor of deoxyhypusine synthase, on the growth of tumorigenic cell lines in culture.
Biochim. Biophys. Acta
,
1310
:
119
-126,  
1996
.
18
Hanauske-Abel H. M., Park M. H., Hanauske A. R., Popowicz A. M., Lalande M., Folk J. E. Inhibition of the G1-S transition of the cell cycle by inhibitors of deoxyhypusine hydroxylation.
Biochim. Biophys. Acta
,
1221
:
115
-124,  
1994
.
19
Tome M. E., Fiser S. M., Payne C. M., Gerner E. W. Excess putrescine accumulation inhibits the formation of modified eukaryotic initiation factor 5A (eIF-5A) and induces apoptosis.
Biochem. J.
,
328
:
847
-854,  
1997
.
20
Chen Z. P., Chen K. Y. Marked elevation of hypusine formation activity on eukaryotic initiation factor 5A in v-HA-RAS transformed mouse NIH3T3 cells.
Cancer Lett.
,
115
:
235
-241,  
1997
.