Applying a highly specific and reproducible cDNA RDA method to clone garlic up-regulated genes in human gastric cancer cells

Abstract

AIM: To develop and optimize cDNA representational difference analysis (cDNA RDA) method and to identify and clone garlic up-regulated genes in human gastric cancer (HGC) cells.

METHODS: We performed cDNA RDA method by using abundant double-stranded cDNA messages provided by two self-constructed cDNA libraries (Allitridi-treated and paternal HGC cell line BGC823 cells cDNA libraries respectively). BamH I and Xho I restriction sites harbored in the library vector were used to select representations. Northern and Slot blots analyses were employed to identify the obtained difference products.

RESULTS: Fragments released from the cDNA library vector after restriction endonuclease digestion acted as good marker indicating the appropriate digestion degree for library DNA. Two novel expressed sequence tags (ESTs) and a recombinant gene were obtained. Slot blots result showed a 8-fold increase of glia-derived nexin/protease nexin 1 (GDN/PN1) gene expression level and 4-fold increase of hepatitis B virus x-interacting protein (XIP) mRNA level in BGC823 cells after Allitridi treatment for 72 h.

CONCLUSION: Elevated levels of GDN/PN1 and XIP mRNAs induced by Allitridi provide valuable molecular evidence for elucidating the garlic's efficacies against neurodegenerative and inflammatory diseases. Isolation of a recombinant gene and two novel ESTs further show cDNA RDA based on cDNA libraries to be a powerful method with high specificity and reproducibility in cloning differentially expressed genes.

Key Words: $[Keywords]

INTRODUCTION

cDNA representational difference analysis (cDNA RDA), with high specificity by eliminating those fragments present in two compared populations and leaving only the differences, has been employed with some degree of success[1-6]. While cDNA RDA, like other PCR-based difference screening methods, is prone to produce false positive results[7,8]. In our study, we used cDNA RDA method to isolate garlic inducible differentially expressed genes in human gastric cancer (HGC) cells. Allitridi is a critical constituent of garlic oil, mainly containing diallyl trisulfide (DATS) and diallyl disulfide (DADS), which is widely used in cancer chemoprevention and anti-cardiovascular disease research[9-16]. Differences between two double-stranded cDNAs populations derived from Allitridi-treated and paternal HGC cell line BGC823 cells cDNA libraries were identified by using cDNA RDA. BamH I and Xho I restriction sites harbored in the library vector were employed to select representations. We found another major source of false positives in cDNA RDA, which was inappropriate enzyme digestion of sample DNAs, and introduced improvements to minimize their production.

MATERIALS AND METHODS

Allitridi-treated and paternal HGC cell line BGC823 cells cDNA libraries

Allitridi is a critical constituent of garlic oil, containing 97.98% diallyl trisulfide (DATS) and 0.85% diallyl disulfide (DADS) in concentration. BGC823 cells were incubated in medium containing 25 mg•L⁻¹ Allitridi for 72 h. Total RNA isolated from paternal and Allitridi-treated BGC823 cells were extracted, followed by synthesis of double stranded cDNAs using λZAP-cDNA Gigapack III Gold Cloning Kit (Stratagene). cDNAs derived from Allitridi-treated and paternal BGC823 cells, with EcoR I cutting site at 5'-end and Xho I site at 3'-end, were unidirectionally cloned into λZAP II vector for cDNA library construction.

cDNA RDA based on cDNA libraries: generation of representations

Difference between Allitridi-treated BGC823 (Alli823) cDNA library DNA (used as tester) and paternal cell (BGC823) library DNA (used as driver) was identified by using cDNA RDA to isolate Allitridi up-regulated genes. Library DNAs of the two cDNA libraries were prepared and digested with BamH I respectively in a mixture containing 40 μg library DNAs, 8 μL BamH I (10kU•L⁻¹, Promega), and 10 μL 10 × buffer H in a final volume of 100 μL. The digestions were carried out at 37 °C for 2, 4, and 12 h respectively to obtain the most appropriately digested fragments. After phenol extraction and ethanol precipitation, digested DNAs (about 24 μg) were then ligated to 18 μL R-Bam-24 (10 μmol•L⁻¹) and 18 μL R-Bam-12 (10 μmol•L⁻¹) adapters. PCR reactions were set up to generate the initial representations by using R-Bam-24 as primer.

In order to expand the content of messages in the representations, we further employed BamH I and Xho I together to digest the library DNAs. Library DNAs (about 40 μg) of Alli823 and BGC823 cDNA libraries were digested with Xho I respectively. Digested DNAs were then ligated to Xho I Linker-15 (5'-TCGAGGATCCATTCA-3') and Xho I Linker-13 (5'-ACTGAATGGATCC-3'). Resulting ligations were digested with BamH I and then ligated to R-Bam-12/24 adapters (10 μmol•L⁻¹), followed by PCR amplification to generate amplicons.

Using BamH I alone or BamH I and Xho I together to digest library DNAs and to prepare representations is showed schematically in Figure 1.

Figure 1 Schematic diagram of preparing representations for cDNA RDA based on cDNA libraries. Using BamH I alone (A) or BamH I and Xho I together (B) to digest library DNAs. Black boxes represent insert cDNAs of cDNA libraries. White boxes represent two arms of cDNA library vector (λZAP II vector). “B” epresents BamH I and “X”Xho I.

The followed three rounds of hybridizations and selective amplifications were performed according to the protocol supplied by typical cDNA RDA. Sequences of adapters used here are as follows: R-Bam-24 5'-AGCACTCTCCAGCCTCTCACCGAG-3'; R-Bam-12 5'-GATCCTCGGTGA-3'; J-Bam-24 5'-ACCGACGTCGACTATCC ATGAACG-3'; J-Bam-12 5'-GATCCGTTCATG-3'; N-Bam-24 5'-AGGCAACTGTGCTATCCGAGGGAG-3'; N-Bam-12 5'-GATCCTCCCTCG-3'.

Slot and Northern blots analyses

Reamplified difference products (100 ng each) were mixed with 2.5 μL 3 mmol•L⁻¹ NaOH respectively and incubated for 1 h at 65 °C, and then transferred onto Nitrocellulose filters (S&S Com) by using Slot Minifold^® II (Schleicher&Schull). Two same filters were prepared and then cross-linked using a UV Stratalinker. Reverse transcription products (the first stranded cDNAs) of 5 μg total RNA of Allitridi-treated and paternal BGC823 cells were used as probes respectively. Probes were labeled with ³²P using a Random-Primer labeling kit (Promega) and hybridization carried out in 1 mmol•L⁻¹ EDTA, 0.25 mmol•L⁻¹ Na₂HPO₄ and 70 g•L⁻¹SDS solution for 16 h at 60 °C. Following hybridization, the filters were washed twice in 1 mmol•L⁻¹ EDTA, 40 mmol•L⁻¹ Na₂HPO 4 and 50 g•L⁻¹ SDS for 25 min at 60 °C and twice in 1 mmol•L⁻¹ EDTA, 40 mmol•L⁻¹ Na₂HPO₄ and 10 g•L⁻¹ SDS for 25 min at 60 °Cagain. Filters were exposed to a phosphor screen for 48 h and analyzed.

Total RNA (10 μg each) isolated from HGC cell lines BGC823, MGC803, PAMC82, SGC7901 and MKN45 were transferred onto Nitrocellulose filters. Probe (the third difference product, DP3) was labeled by random primer extension and Northern blots hybridization carried out in the same manner described above.

Sequencing and Database Searching

Difference products were amplified and cloned into pGEM-T Easy Vector (Promega). Double stranded plasmid DNAs were prepared using miniprep columns (Promega) and sequenced with Ultra VGI 1280 (applying User Manual version 3.0). Resulting sequences were compared to the GenBank database by using the BLAST program.

RESULTS

Generation of representations for cDNA RDA based on cDNA libraries

Double stranded cDNAs, with EcoR I cutting site at 5'-end and Xho I site at 3'-end, were unidirectionally cloned into λZAP II vector for cDNA library construction. Library vector (λZAP II vector) has three BamH I restriction sites with one site located very near to 5'-end of insert cDNAs. Thus BamH I adapters can be introduced into the 5'-end of insert cDNAs. With the same copies three distinct fragments appeared in the products of library DNA digested with BamH I at 37 °C for 4 h (about 22.1, 11.3 and 6.4 kb in size respectively), which indicated appropriate digestion degree. However, library DNAs were not fully digested at 37 °C for 2 h, and the smears of library DNAs after digestion for 12 h showed excessive reactions. When BamH I and Xho I were used together to digest library DNAs (1 cut/2 kb insert cDNAs in average), another vector-derived fragment (850 bp in size) was released from 3'-end of insert cDNAs. We further designed Xho I Linker-15/13 to convert Xho I cutting site to BamH I site. Thus the content of cDNA messages available for difference analysis was expanded. The results for library DNAs after digestion with different enzymes or different conditions are showed in Figure 2. Subsequently, PCR reactions were set up to generate the initial representations (amplicons) (Figure 3).

Figure 2 Agarose gel electrophoresis of digestion products of Alli823 (lanes 1, 3, 5) and BGC823 (lanes 2, 4, 6) cDNA library DNAs digested with BamH I at 37 °C for 2 (lanes 1, 2), 4 (lanes 3, 4), and 12 h (lanes 5, 6) respectively. The digestion products of Alli823 (lane 7) and BGC823 (lane 8) library DNAs digested with BamH I and Xho I together (A fragment, 850 bp in size, appeared in the digestion products). λPhage/Hind III size marker (lane M).

Figure 3 Agarose gel electrophoresis of amplicons derived from different enzyme digestions. A: The amplicons obtained by using BamH I to digest Alli823 (lane 1) and BGC823 (lane 2) cDNA library DNAs respectively; B: The amplicons obtained by using BamH I and Xho I together to digest Alli823 (lane 1) and BGC823 (lane 2) library DNAs respectively. 1 kb size marker (lane M).

Identification of differentially expressed genes in Allitridi-treated BGC823 cells

Results of Slot blots analysis showed that expression of 4 cDNAs was up-regulated by Allitridi treatment and the degree to which each cDNA was up-regulated ranged from 3- to 8-fold (Figure 4). Sequencing and GenBank database searching results showed that two isolated difference fragments (SH2 and SH3) had remarkable homology over 97% with glia-derived nexin/protease nexin 1 (GDN/PN1) mRNA and hepatitis B virus x-interacting protein (XIP) mRNA respectively. No homology was found in fragment SH1, which indicated SH1 to be a novel EST. SH4 showed remarkable homology as high as 98% with a piece of human DNA sequence on chromosome 20 and was also showed to be a novel EST. The sequences of SH1 and SH4 were showed in Figure 5.

Figure 4 A: Agarose gel electrophoresis of reamplified difference products SH1-4 (lanes 1-4 respectively). PCR product of GAPDH (400 bp in size) used as quantity control (lane G); B: Slot blots analysis showing differentially expressed cDNAs. Reverse transcription products (first stranded cDNA) of total RNA of BGC823 and Alli823 cells used as probe respectively; C: The degree to which each cDNA was up-regulated.

Figure 5 Sequences of two novel ESTs (SH1 and SH4) isolated by cDNA RDA based on cDNA libraries.

Northern blots analysis was employed to detect the expression level of another obtained difference product (derived from calcyclin mRNA) and a novel transcript about 2 kb in size appeared only in the hybridization result of BGC823 total RNA (Figure 6). Further cDNA library screening and database searching showed this novel transcript to be a recombinant gene merged from calcyclin gene and homo sapiens heterogeneous nuclear ribonucleoprotein A0 (HNRPA0) gene.

Figure 6 Northern blotting result showing the expression level of calcyclin gene in human gastric cancer cell lines BGC823, MGC803, PAMC82, SGC7901 and MKN45 (lanes 1-5) respectively. The arrow showing the novel transcript derived from the recombinant gene merged from calcyclin gene and HNRPA0 gene.

DISCUSSION

When cDNA RDA[1-6] or RDA[17-19] is performed, sample cDNAs or genome DNAs are digested into fragmented populations to prepare representations, and the digestion products showed as smears by agarose gel electrophoresis. Therefore, it is difficult to determine the most appropriate digestion degree. However, deficient or excessive enzyme digestion of sample DNAs can undoubtedly result in false positive results. In our study, cDNA RDA was performed based on cDNA libraries. Our data showed that time for enzyme digestion affected the digestion degree obviously. With the same copies, three or four fragments released from library vector after enzyme digestion acted as good marks indicating appropriate digestion, which increased the specificity and reproducibility of cDNA RDA. Moreover, when BamH I and Xho I were used together to digest library DNAs, the content of cDNA messages available for difference analysis was expanded and more specific products were obtained.

Thrombin, as the principal component of the blood coagulation cascade, also prevents neurite outgrowth and modulates morphologic changes in both neurons and astrocytes[20-22]. Recent studies have showed that thrombin mediates polyneuronal synapse elimination, both in vivo and in vitro[23]. Glia-derived nexin/protease nexin 1 (GDN/PN1) is the most potent vertebrate inhibitor for thrombin, which, acting as neurite-promoting factor, plays an important role in neurotrophic and neuroprotective properties[24-29]. Garlic extract contains antioxidant phytochemicals that prevent the oxidant-mediated brain cell damage and produce neurotrophic effects[30-33]. Our results showed the potential effect of garlic on reducing the risk of neurodegenerative disease through up-regulation of GDN/PN1 expression.

The hepatitis B virus (HBV) has been reported to be a risk factor in the development of hepatocellular carcinoma[34-42]. The HBV X protein (HBx) is a small transcriptional activator and has been showed capable of transactivating many different viral and cellular promoters. It is essential for virus infection and is implicated in the development of hepatocellular carcinoma[43-49]. An HBx-interacting protein that specifically binds to the carboxy terminus of wild-type HBx was identified in 1998 and designated as HBx interacting-protein (XIP), which could inhibit HBx activity and thus decrease HBV replication[50]. Elevated level of XIP mRNA induced by garlic extract provided valuable molecular evidence for elucidating the garlic's efficacies against inflammatory diseases.

Identification of recombinant gene in our study showed the feasibility of employing cDNA RDA method to the discovery of probes for anonymous loci that suffered genetic rearrangements. Cloning of two novel ESTs further showed cDNA RDA based on cDNA libraries to be a powerful method in isolating novel genes.