INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Marek's disease virus (MDV) is an oncogenic avian herpesvirus that induces T-cell lymphomas in its natural host, the chicken, 4 to 6 weeks after infection (28). Many MDV strains with differing oncogenicities have been isolated from field populations of fowl (26). Although candidates of oncogenic genes have been presented, oncogenic mechanisms have not been determined (16, 19, 25). A comparison of the whole genome structures of many strains of MDV is needed to understand the complicated oncogenic mechanisms by which this virus causes tumors. Unfortunately, purification of the MDV genome is complicated because its density is similar to that of chick cell DNA. Furthermore, because of the cell-associated nature of the virus, which limits the proportion of cells that can be infected, the yield of virus particles is usually poor (18). As for comparisons of nucleotide sequences, the size of the MDV genome (approximately 180 kbp) is too long for direct sequence analysis.

Lisitsyn et al. (15) described a method for the preparation of restriction enzyme (RE) fragments present in one population of genome DNA but not in others, using sequence-independent single-primer amplification, called representational difference analysis (RDA). The RDA method was successfully used to identify and obtain clones of Kaposi's sarcoma-associated virus (also called human herpesvirus 8 [reference 3]) DNA and TT virus DNA (17). Unfortunately, the sizes of isolated clones obtained by the original RDA method are too small for sequence analysis of the MDV genome.

In the present study, we modified the original RDA method. This was done by first omitting the initial amplification step and then using proofreading Taq polymerase for amplification of longer DNA fragments from total cellular DNA infected with a variety of oncogenic strains of MDV.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Virus strains, culture of cells, and infection of chickens. The strains of MDV used were very virulent Md5 (27) and RB-IB (23); virulent GA (5), JM (24), and MS2 (11); and nonvirulent CVI-988 clone C (4). Viruses were propagated in chicken embryo fibroblasts (CEF), which were prepared from specific pathogen-free embryos and cultivated in modified Eagle medium supplemented with 10% fetal calf serum, 2 mM glutamine, and 100 U of penicillin G/ml in a humidified atmosphere at 37°C.

RE fragments. Total cellular DNAs were extracted using proteinase K (Sigma Chemical Co., St. Louis, Mo.). Briefly, cells were lysed and incubated in 1% sodium dodecyl sulfate and proteinase K (100 µg/ml) at 65°C for 8 h (22). Proteins surrounding DNA were lysed and digested during incubation. The DNA solution was extracted with phenol and chloroform and precipitated using ethanol. The feather pulp DNA was also extracted from the root of the feather pulp, to which a small amount of feather follicle epithelium was attached. Glycogen (Roche Diagnostics GmbH, Mannheim, Germany) was used for precipitation of a small amount of DNA extracted from the feather.

RDA. RDA was performed according to the method of Lisitsyn et al. (15), except that RE fragments from MDV-infected cellular DNA were directly subtracted by those from uninfected cellular DNA. Namely, 2 to 5 ng of RE fragments from MDV-infected cells were ligated with specific linkers for each of the REs listed in Table 1 and mixed with 100 ng of the RE fragments extracted from uninfected CEF DNA. The mixture was precipitated with ethanol, dissolved in 4 µl of 3× EE buffer (30 mM EPPS [N-(2-hydroxyethyl)piperazine-N'-3-propanesulfonic acid] and 3 mM EDTA [pH 8.0]), and covered with mineral oil. After the mixture had been heated to 99°C for 4 min, 1 µl of 5 M NaCl was added, and the solution was incubated at 67°C for 21 h. During this incubation period, RE fragments of the chicken genome with a linker included in the DNA from MDV-infected cells were hybridized with those from uninfected cells without a linker. After hybridization, the reaction mixture was diluted to 100 µl with 1× reaction buffer of the Expand HiFi PCR system (Roche Diagnostics). One microliter of hybridized solution was transferred to 100 µl of preheated reaction mixture of the Expand HiFi PCR system (10 µl of 10× Expand HiFi buffer [including 15 mM MgCl₂] and a 0.2 mM concentration of each deoxynucleotide). The PCR mixture was heated to 72°C prior to the addition of 5 U of Taq polymerase mixture included in the Expand HiFi PCR system. The reaction mixture was maintained at 72°C for 5 min to fill the 5' overhanging ends of the linkers for RDA, providing a sequence complementary to the 24-mer linker at the 3' ends of the molecule that had hybridized to the RE fragments of the MDV-infected CEF. This was immediately followed by a denaturation step (94°C for 2 min), and 20 cycles of PCR (94°C for 1 min, and 72°C for 8 min). In the amplification step, RE fragments having linkers on both ends were selectively amplified in the PCR step. As a result, DNA fragments specific to MDV-infected cells, most of which were viral DNA, were amplified. The optimum ratio of DNA extracted from infected CEF to those from CEF was preliminarily determined by small-scale RDA. In most cases, the optimum ratio was in the range of 1:3 to 1:100. The amplified genome was analyzed on agarose gels.

Southern blot hybridization. One hundred nanograms to two micrograms of RDA-derived fragments or RE fragments of MDV were separated on agarose gels, blotted onto a Biodyne nylon membrane (Pall Co. Ltd., Port Washington, N.Y.) by capillary transfer in 20× SSC (20× SSC is 3 M sodium chloride plus 0.3 M sodium citrate) for 16 h, and fixed to the membrane by baking in an oven at 80°C for 30 min. Hybridization was carried out in a buffer containing 10.3% polyethylene glycol (8,000 molecular weight), 1.55× SSPE (20× SSPE is 3.6 M sodium chloride, 0.2 M sodium phosphate (pH 7.7), 20 mM EDTA), 7.24% sodium dodecyl sulfate, and ³²P-labeled DNA probes at 65°C for 16 h. As probes for the MDV genome, BamHI fragment clones of the MDV, a generous gift from M. Nonoyama (Tampa Bay Institute, St. Petersburg, Fla.) (7), were used as probes for the MDV genome. When RDA-derived fragments were used as probes, fragments were extracted from agarose gel using a Jet Sorb DNA extraction kit (GENOMED GmbH, Bad Oeynhausen, Germany) according to the manufacturer's protocol. Labeling was carried out using a random prime labeling kit (BcaBest Labeling kit; Takara Shuzo Co. Ltd, Kyoto, Japan) according to the manufacturer's protocol.

Cloning and sequence analysis. The RDA-derived fragments were cut with the RE that had been used for the RDA, separated on agarose gels, extracted as described in the preceding section, and cloned into pSPORT1 (Gibco-BRL Life Technologies, Co., Rockville, Md.). Plasmids that included RDA-derived fragments were selected by colony hybridization and extracted from a small culture using a Jet Prep plasmid extraction kit (GENOMED). When the fragment was sequenced, three clones of the plasmid were sequenced to detect PCR errors. Sequences of the RDA-derived fragments were determined using a Big Dye terminator kit (Perkin-Elmer Biosystems, Foster City, Calif.) and a Prism310 sequencer (Perkin-Elmer Biosystems) by the primer walking method. Construction of primers and DNA sequence analysis were carried out using a Vector NTI program (InforMax, North Bethesda, Md.). To determine the original sequence of the MDV genome, regions of the MDV genome were amplified and sequenced with the primers used for primer walking sequencing of the cloned RDA-derived fragments.

Nucleotide sequence accession numbers. Accession numbers for the two sequences determined in this study in the DDBJ database are AB041042 and AB041041 for BamHI-K₁ and -T, respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

When the BamHI-restricted fragments extracted from the MDV CVI strain-infected CEF were subtracted from those from uninfected CEF, 11 bands less than 3.5 kbp long were visible on the ethidium bromide-stained gel after electrophoresis (Fig. 1A). No amplified fragment was observed when BamHI-restricted fragments from mock-infected CEF were used as a template for amplification. Different patterns of amplified fragments were shown when BamHI-, BglII-, EcoRI-, HindIII-, and XhoI-restricted fragments were used as templates. Therefore, the sizes and numbers of the amplified fragments were dependent on the RE used in the RDA. All of the amplified fragments were hybridized with a ³²P-labeled MDV probe (a mixture of BamHI-digested MDV fragment clones (BamHI-A to BamHI-T) (Fig. 1B). These results indicated that RE fragments from MDV DNA, but not from cellular DNA, were selectively amplified. A comparison of the Southern blot patterns of the RDA-amplified fragments with those of RE-digested MDV DNA showed that almost all of the band patterns of the amplified fragments corresponded to those of RE-digested MDV DNA in the molecular size range of 0.3 to 3.5 kbp regardless of the RE (Fig. 1B and C). These results indicated that RE fragments less than 3.5 kbp long were efficiently amplified.

View larger version (81K): [in this window] [in a new window]   FIG. 1.   RDA-derived RE fragments of MDV (strain CVI). The RE fragments were prepared by subtraction of BamHI-, BglII-, EcoRI-, HindIII-, or XhoI-digested CEF DNA from those of MDV-infected CEF (see Materials and Methods). (A) Agarose gel electrophoresis of RDA-derived fragments amplified from BglII-, EcoRI-, HindIII-, or XhoI-digested MDV DNAs. CEF indicates the negative control of RDA. (B) Southern blot analysis of the gel shown in panel A. A sequence corresponding to that of the MDV genome was detected with 32P-labeled MDV genomic clones. (C) Southern blot analysis of RE fragments of MDV. The probe was the same as above. Numbers between panels indicate sizes of markers in kilobase pairs.

To confirm that the RE fragments were amplified exactly, a 3.5-kbp RDA-derived fragment amplified from BamHI-digested MDV was hybridized to a subset of BamHI clone DNAs (Fig. 2). The 3.5-kbp fragment only hybridized with the same-sized MDV BamHI-K fragments (BamHI-K₁, K₂, and K₃ [Fig. 2]). In the same manner, 10 residual amplified fragments corresponded to each MDV BamHI fragment of the same size (data not shown). Furthermore, to estimate the fidelity of amplified fragments, sequences of RDA-derived fragments were compared with their counterparts of the original genome. Two RDA-derived fragments (3.5 and 0.7 kbp) amplified from BamHI-digested DNA extracted from MDV GA strain-infected CEF were cloned and sequenced. These clones were hybridized with BamHI-K₁ and -T of published BamHI fragments, respectively (data not shown). The sequences of three randomly selected clones of the 0.7-kbp fragment were exactly the same, and two randomly selected clones of the 3.5-kbp fragment were also exactly the same. Only one base on one randomly selected clone was not the same as that of the other two clones. Thus, sequencing of multiple clones of RDA-derived fragments covered the PCR errors in relatively long sequences. As a result, sequences determined from the three clones coincided throughout a total of 4,327 bases (data not shown but accessible in DDBJ with accession no. AB041041 and AB041042). No deletion or insertions were detected. This result suggests that nucleotide sequences determined from the RDA-derived fragments are reliable enough for gene analysis of a viral genome.

View larger version (56K): [in this window] [in a new window]   FIG. 2.   Hybridization of RDA-derived fragments from BamHI-digested MDV with MDV BamHI fragment clones (7). The lower subset of published BamHI fragments (fragment K to T clones) were separated in agarose gel, stained (upper panel), and hybridized with a 32P-labeled 3.4-kbp RDA-derived fragment. Numbers on the left indicate sizes of markers in kilobase pairs.

As the RDA-derived fragments could be observed directly on ethidium bromide-stained gel, interstrain differences in RE fragments can be detected from comparisons of the RDA-derived fragments between strains. In this study, we detected an interstrain difference in SalI-digested MDV fragment size (Fig. 3A). Probing the hybridization product with a 1.8-kbp SalI fragment indicated that the size of a SalI-digested fragment (1.8 kbp in GA, MS2, and RB-IB strains) was changed to 2 kbp in CVI and JM strains (Fig. 3B). This result indicates that interstrain difference in sizes of RE fragments can be detected from direct comparisons of the RDA-derived RE fragments on agarose gels. The sequences of the 1.8- and 2.0-kbp fragments were determined. A 200-bp insertion was seen in the Meq gene of CVI and JM strains (Fig. 3C). The sequence and the location of the insertion were the same as those in a recent report on MDV (14).

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Interstrain comparisons of RDA-derived RE fragments of MDV. (A) Agarose gel electrophoresis of RDA-derived fragments prepared from SalI-digested MDV DNAs. Strains of MDV are indicated above the gel. Numbers on the left indicate sizes of markers in kilobase pairs. (B) Southern blot analysis of SalI-digested MDV. The 1.8-kbp RDA-derived fragment which appeared is indicated by an arrow in the agarose gel in panel A and was used as a probe. (C) Diagram of the 1.8-kbp SalI fragment of GA strain MDV. The shaded box indicates a 0.2-kbp insert detected in the genome of CVI and JM strains of MDV. The open arrow indicates the Meq gene. Restriction sites are also indicated.

To further show the advantage of this method based on increased sensitivity, we tried to prepare RDA-derived MDV fragments from single feather pulp. Fragments were amplified from some single-feather pulp (Fig. 4). This amplification was not observed with the feather pulp of a mock-infected chicken. When amplified products were digested with BamHI and ligated with the plasmid, 0.3-, 0.6-, 1.0-, 1.3-, 1.5-, 1.9-, 2.2-, 2.6-, 3.1-, and 3.5-kbp fragments were cloned (data not shown).

View larger version (71K): [in this window] [in a new window]   FIG. 4.   Agarose gel electrophoresis of RDA-derived BamHI-digested MDV from feathers of MDV-infected chickens. Five feathers from each infected chicken were subjected to RDA. Numbers beside the gel indicate sizes of markers in kilobase pairs. Mock indicates RDA-derived DNA from a feather of a mock-infected chicken.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The results of this study showed that most of the RE fragments of MDV of less than 3.5 kbp were simultaneously amplified to amounts sufficient for analysis on ethidium bromide-stained agarose gel from a small amount of infected cells. The yield of fragments per infected cell obtained in the present study was much higher than the yields obtained by other methods, such as Hirt's method or pulsed-field gel electrophoresis (9, 12). These fragments could be easily extracted from the gel and cloned into plasmid. As for nucleotide sequences, the viral RE fragments are expected to have relatively high fidelity. The inherent 3'-5' exonuclease proofreading activity of Pwo DNA polymerase results in a fourfold-increased fidelity of DNA synthesis (8.5 × 10⁶ errors per nucleotide incorporation) compared to Taq DNA polymerase (2.6 × 10⁵ errors per nucleotide incorporation) (1). The manufacturer estimated that approximately 90% of 3-kbp amplified fragments had no error made by misreading of Taq DNA polymerase after 20 cycles of amplification (Roche Diagnostics). The sequence data of the cloned RDA fragments, which correspond to BamHI-K₁ and -T fragments of the MDV genome, indicated high fidelity throughout the fragments when the sequence was compared to that determined from direct sequencing of the genome. A comparison of the sequence with the published GA sequence (13) showed that the entire sequence coincided within the BamHI-T fragment, while 18 bases did not coincide within the BamHI-K₁ fragment. Since the sequence was compared with that from the genome, inconsistencies of the nucleotides have been caused by the differences in viral clones maintained in our laboratory. It has been suggested that the MDV genome is susceptible to mutation (26). Thus, our sequence data suggest that the modified method of RDA can be successfully adapted to isolation of virus DNA fragments, which could be used for determination of sequences and other investigations.

RDA has also been used to compare interstrain differences in RE fragments of MDV (6). In this study, we detected a recently identified interstrain difference in RE fragments from comparisons of patterns of amplified fragments. Compared with the restricted number of known interstrain polymorphisms of RE fragments (8, 20), comparisons of RDA-derived fragments seemed to be an effective approach for detecting interstrain differences in RE fragments. We detected a newly identified insert within the Meq gene. The nucleotide sequence of the insert and the predicted change of the Meq gene were determined and identified to be the same as those of a recently published polymorphism of the Meq gene (14). These data suggested that the modified RDA method could be applied to interstrain comparisons.

The modified RDA can be started from a small quantity of DNA. MDV DNA may be isolated from small biopsy samples. It is well known that MDV-infected feather follicle epithelium is attached to the bottom of feather pulp (2). MDV fragments isolated from a small number of feathers would be a useful sample for analysis of field isolates of MDV. Although amplification of all feather pulps was not successful, our data suggest that five or six feathers are sufficient for the amplification of MDV fragments. The cause of the failure of amplification in some feather pulps is currently being investigated.

Since the modified method of RDA is not dependent on viral sequences, it may be used for isolation of other DNA viruses. Since RNA can also be subtracted by RDA (10), it is possible that the modified method of RDA could be further adapted to isolation of the genome sequence of RNA virus. Sallie (21) presented a theoretical PCR-based strategy for identification of infectious pathogens using a cDNA and DNA mixture to detect pathogens. We are now further modifying the method for isolation of both DNA and RNA viruses.