Decreased Virus Population Diversity in p53-Null M
http://www.100md.com
病菌学杂志 2005年第18期
Immunology Graduate Program
Departments of Pathology
Molecular Biology and Microbiology, Sackler School of Graduate Biomedical Sciences, Tufts University School of Medicine, Boston, Massachusetts 02111
ABSTRACT
The Abelson murine leukemia virus (Ab-MLV), like other retroviruses that contain v-onc genes, arose following a recombination event between a replicating retrovirus and a cellular oncogene. Although experimentally validated models have been presented to address the mechanism by which oncogene capture occurs, very little is known about the events that influence emerging viruses following the recombination event that incorporates the cellular sequences. One feature that may play a role is the genetic makeup of the host in which the virus arises; a number of host genes, including oncogenes and tumor suppressor genes, have been shown to affect the pathogenesis of many murine leukemia viruses. To examine how a host gene might affect an emerging v-onc gene-containing retrovirus, we studied the weakly oncogenic Ab-MLV-P90A strain, a mutant that generates highly oncogenic variants in vivo, and compared the viral populations in normal mice and mice lacking the p53 tumor suppressor gene. While variants arose in both p53+/+ and p53–/– tumors, the samples from the wild-type animals contained a more diverse virus population. Differences in virus population diversity were not observed when wild-type and null animals were infected with a highly oncogenic wild-type strain of Ab-MLV. These results indicate that p53, and presumably other host genes, affects the selective forces that operate on virus populations in vivo and likely influences the evolution of oncogenic retroviruses such as Ab-MLV.
INTRODUCTION
Abelson murine leukemia virus (Ab-MLV) is a replication-defective retrovirus that transforms NIH 3T3 cells and pre-B cells in vitro and induces pre-B-cell lymphoma in vivo by expressing the v-Abl protein tyrosine kinase (reviewed in reference 23). Kinase activity is required for tumor induction, and wild-type Ab-MLV can induce tumors in the absence of viral replication (9). However, a replicating helper virus is required for at least one weakly oncogenic mutant, the P90A strain (22), to induce tumors (14). In this instance, virus replication is accompanied by the emergence of viruses that have undergone additional mutation, rendering them more highly oncogenic (13, 14). Thus, viral determinants in addition to kinase activity play an important role in determining oncogenicity, and in vivo passaging can select for viruses with increased pathogenic potential.
Host determinants also influence tumor development. Laboratory mouse strains differ markedly in their susceptibilities to Ab-MLV, and at least two, as yet unknown, loci appear to be involved in mediating this response (19, 20). In addition, the use of genetically altered mice has revealed that both tumor suppressor genes and oncogenes can affect tumor induction and in vitro transformation by a variety of retroviruses, including Ab-MLV (2-4, 12, 28, 33, 34, 37). Such analyses have shown that the p53 tumor suppressor protein, a molecule activated following oncogenic insult (27), influences transformation by Ab-MLV in vivo and in vitro (33, 37). In vitro, p53 functions during the late stages of transformation and is required for the apoptotic crisis phase of the process, a response that is also influenced by the products of the Ink4a/Arf locus (18, 25, 33). Less is known about the way in which p53 affects tumor induction in vivo. However, analyses of p53-null animals infected with the weakly oncogenic P90A strain (37) revealed that the absence of p53 accelerated disease and relieved the selective pressure on the P90A strain for the generation of more highly oncogenic mutants.
The effects of p53 on the weakly oncogenic P90A strain suggest that host genes may play an important role in shaping the virus population present in infected animals. Analyses of newly arising avian erythroblastosis viruses demonstrated that several viruses can exist in a single animal (35), suggesting that v-onc gene-containing viruses probably arise from a pool of viruses generated during replication following the recombination event that joins viral and cellular sequences. To determine how a host gene such as p53 might influence the composition of emerging oncogenic viruses, we examined the viral populations recovered from p53+/+ and p53–/– mice infected with the P90A strain. These analyses revealed the presence of virus variants in both types of mice but demonstrated that viral populations are more heterogeneous in p53+/+ tumors than in p53–/– tumors. These results indicate that host genes can affect the composition of oncogenic retrovirus populations and likely contributed to the way in which v-onc gene-containing retroviruses evolved following oncogene capture.
MATERIALS AND METHODS
Viruses and animals. The Ab-MLV-P90A (13) virus stock was prepared from NIH 3T3 cells that were cotransfected with pBR-90, a plasmid containing the Ab-MLV-P90A genome (14), and pZAP, a plasmid containing Moloney MLV (Mo-MLV) sequences (29). Ab-MLV-P160 (22) was isolated from a clonally derived NIH 3T3 virus-transformed nonproducer cell line that had been superinfected with Mo-MLV-Cl2 (8). The Ab-MLV-2-3 and Ab-MLV-2-5 viruses were generated by recovering the SacI-FseI fragment (bp 2395 to 3016 of the Ab-MLV-P120 genome) from plasmids containing the C-terminal coding sequences of these variants and replacing the corresponding fragment in the pBR-90 vector. Viruses were recovered by cotransfection of NIH 3T3 cells as described above. All virus stocks were filtered through a 0.45-μm filter and titrated by using the NIH 3T3 transformation assay (26).
BALB/cJ and p53–/– mice were obtained from our breeding colony at Tufts University. The p53–/– mice were maintained by mating heterozygous animals originally obtained from a single breeding pair of p53+/– animals (Jackson Laboratory) that had been backcrossed to BALB/cJ mice five times and inbred for three generations. Neonatal mice were injected via the intraperitoneal route with approximately 1 x 104 focus-forming units of Ab-MLV stock containing 0.8 μg Polybrene (Sigma). Animals were monitored for a 90-day period and sacrificed once signs of tumor development (lymphadenopathy, cranial tumors, or hind limb paralysis) were evident. Tumor tissue was removed and frozen at –80°C. Tumor latency was compared by generating Kaplan-Meier survival plots and comparing the plots using Prism (GraphPad Software) and the Mantel-Haenszel log rank test, which calculates a P value comparing the two curves by considering the possibility that random chance would lead to curves that are as similar as those generated from the data. Comparisons of sequences recovered from tumor tissues were performed by using Fisher's exact test (GraphPad Software), which compares two unpaired groups of data, and the unpaired t test, which compares two groups of data.
DNA analysis. Genomic DNA was prepared by macerating tumor samples on dry ice with a mortar and pestle, and lysis buffer (10 mM Tris-HCl, pH 7.6, 10 mM EDTA, pH 8.0, 10 mM NaCl, 0.5% N-lauroyl sarcosine, 1 mg/ml proteinase K) was added. The lysate was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), and the DNA was recovered by precipitation and stored in TE buffer (10 mM Tris, pH 8.0, 1 mM EDTA). The C-terminal coding sequences of the viruses (bp 2380 to 3048 of the Ab-MLV-P120 genome) were amplified by using the primers 5'-AGAAGGTCTACGAGCTCATGC-3' and 5'-GCACAGGCTTTCTCAGTCCTT-3'. Amplification reaction mixtures contained 100 ng of genomic DNA, a 200 μM deoxynucleoside triphosphate mix (Pharmacia), a 0.4 μM concentration of each primer, 2.5 U Pfu polymerase (Stratagene), and Pfu buffer [20 mM Tris-HCl, pH 8.8, 2 mM MgSO4, 10 mM KCl, 10 mM (NH4)2SO4, 0.1% Triton X-100, 0.1 mg/ml bovine serum albumin]. After a 2-min incubation at 94°C, the reactions were run for 30 cycles of 1 min at 94°C, 1 min at 59°C, and 1 min at 72°C. After the final cycle, the reaction mixtures were incubated for 4 min at 72°C and cooled to 4°C. The reaction products were purified using a Qiaquik PCR purification kit (QIAGEN), and 3' A overhangs were added by diluting the entire purified product in a total volume of 50 μl containing 0.2 mM dATP, 0.5 U Taq polymerase (Applied Biosystems), and Taq buffer (10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, and 0.001% [wt/vol] gelatin) and incubating the mixture at 72°C for 15 min. The PCR products were cloned into the pCR4-TOPO vector (Invitrogen) following the manufacturer's protocol, and the inserts were sequenced on an ABI 3100 DNA sequencer (Perkin-Elmer) at the DNA Facility, Department of Physiology, Tufts University School of Medicine. As an additional control, the viral sequence was amplified from a vector containing the P90A COOH-terminal coding region; no changes were detected in the 6,500 bases analyzed.
The numbers of proviruses in the tumors were determined using quantitative real-time PCR and an Opticon 1 thermocycler (MJ Research); the data were analyzed using Opticon Monitor v1.08 software. All amplification reactions were performed in triplicate with reaction mixtures containing 10 ng template, 1x SYBR green dye (Applied Biosystems), and a 0.2 μM concentration of each primer. The number of cells in each reaction was evaluated by amplifying a portion of the Rag1 locus and calculating the number of copies based on amplification of a plasmid standard curve with the assumption that each cell contained two copies of the Rag1 locus. A similar plasmid standard curve was used to assess the number of v-abl copies, and this value was divided by the number of cells in the sample to calculate the number of proviruses per cell. A cell line containing a single copy of v-abl was used as an additional control. v-abl proviral sequences were amplified with the primers 5'-GATCCATCTCGCTGCGGTAT-3' and 5'-ACTAACTCAGCCAGAGTGTTGAAGC-3', and sequences from the Rag1 locus were amplified with the primers 5'-ATCATCTGTGGTTAGCCGTCTGT-3' and 5'-ATTATGTATCAGCTCTCACGCCC-3'. All amplification reaction mixtures were incubated at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and either 59°C for 1 min and 81°C for 1 s (for v-abl amplification) or 58.5°C for 1 min and 79°C for 1 s (for Rag1 amplification).
Southern analyses were performed as previously described (36). Briefly, digested DNAs were fractionated overnight through 0.8% agarose-1x Tris-borate-EDTA gels, and after denaturation and neutralization, the DNAs were transferred to nitrocellulose membranes (Schleicher & Schuell). The membranes were baked for 2 h at 80°C in a vacuum oven and incubated at 42°C for 2 h in prehybridization buffer containing 6x SSC (0.9 M NaCl, 90 mM sodium citrate, pH 7.0), 10x Denhardt's solution (0.2% Ficoll 400 [Pharmacia], 0.2% polyvinylpyrrolidone, and 0.2% bovine serum albumin), and 50 μg/μl of single-stranded salmon sperm DNA. Membranes were probed with the SalI-BspEI fragment of Ab-MLV-P120 (bp 3073 to 3556 of the genome) labeled by the random primer method with 50 μCi of [-32P]dCTP (3,000 Ci/mmol) using a Prime-It RmT random primer labeling kit (Stratagene) and purified using a G-50 Sephadex column (Roche). The probe was mixed with hybridization buffer (6x SSC, 5% dextran sulfate, 1% sodium dodecyl sulfate [SDS], 10 mM EDTA) and applied to the membranes, which were then incubated at 65°C overnight. The following day, the membranes were washed twice with 5x SSC-0.5% SDS and twice with 0.5x SSC-0.15% SDS and exposed to Kodak X-AR-5 film at –80°C with an intensifier screen.
For unblot experiments, the genomic DNAs were digested and fractionated as described above, and the gels were dried on a slab gel dryer using a house vacuum (31) and probed with an oligonucleotide probe (5'-CTGTACACTTTCTGTGTGTGAGCTATGT-3'; bp 3360 to 3384 of the Ab-MLV-P120 genome). The oligonucleotide was labeled with 750 μCi of [-32P]ATP (6,000 Ci/mmol) using T4 polynucleotide kinase (NEBiolabs) and purified using a G-25 Sephadex column (Roche). The probe was mixed with buffer containing 5x SSPE (0.9 M NaCl, 0.05 M NaH2PO4, 0.005 M EDTA, pH 7.4), 0.1% SDS, and 0.01 μg/ml of single-stranded salmon sperm DNA and applied to the dried gels, which were then incubated at 55°C overnight. Following hybridization, the gels were washed four times in wash buffer (2x SSC, 0.2% SDS), twice at room temperature and then twice at 55°C. The gels were air dried and exposed to Kodak X-AR-5 film at –80°C with an intensifier screen. Proviral integrations were excised by aligning the dried gel with the autoradiogram. The excised agarose was melted in 10 mM Tris-HCl, pH 8.0, and a sample was amplified using the PCR protocol described above, cloned, and sequenced.
Protein analysis. Tumor cell lysates were prepared by macerating tumor samples on dry ice with a mortar and pestle, and protein lysis buffer (10 mM Tris-HCl, 1% SDS, pH 7.5, 0.1 mM sodium vanadate, and 100 μM phenylmethylsulfonyl fluoride) was added. The lysates were boiled for 5 min and sheared by passing the lysates through a 25-gauge needle. The amount of protein present in the lysates was determined by using a bicinchoninic acid protein assay kit (Pierce), and 50 μg of protein was fractionated through an 8% SDS-polyacrylamide gel. The proteins were electrotransferred to polyvinylidene difluoride membranes (U.S. Biochemicals) that were blocked with phosphate-buffered saline containing 0.2% I-Block (Tropix) and 0.1% Tween 20 for at least 1 h. The blots were probed with an anti-Gag/v-Abl antibody (H548) (6) according to the Western Light kit protocol (Tropix), utilizing an alkaline phosphatase-conjugated secondary antibody with a CSPD substrate (Tropix). The blots were exposed to Kodak X-AR-5 film.
RESULTS
Ab-MLV-P90A generates variants in both p53+/+ and p53–/– mice. The P90A strain was derived from the wild-type P120 strain and contains a 19-base deletion that shifts the reading frame, resulting in the expression of a truncated v-Abl protein that resembles the wild-type protein up to amino acid 751 and contains an additional 33 amino acids appended as a result of the reading frame change (Fig. 1A). In preparation for experiments to assess the effects of p53 on Ab-MLV populations in vivo, we infected p53+/+ and p53–/– mice with P90A and monitored their survival. Consistent with the results of other work (37), P90A accelerated tumor formation when disease induction was compared in p53+/+ and p53–/– mice in the BALB/cJ background used in our laboratory (Fig. 1B). All of the null animals, compared to 60% of the wild-type animals, developed disease, and the mean survival period was somewhat shorter. A comparison of the survival curves by using a log rank test revealed that the null animals were somewhat more susceptible to disease (P = 0.0351). Despite this difference, the gross and microscopic appearances of the tumors were similar (data not shown). In addition, analyses of five null and five wild-type tumor samples tested revealed the presence of rearranged immunoglobulin heavy chain genes (data not shown) typically associated with Ab-MLV-induced pre-B-cell lymphoma. Consistent with these data, analyses of B, T and myeloid surface markers in a different series of tumors from p53-null and wild-type mice indicated that all of the tumors were of the pre-B-cell phenotype (37).
Analyses of the v-Abl proteins present in tumor lysates revealed the presence of variants in both wild-type and null mice (Fig. 1C). Some samples from both types of mice (M21, M22, M100, and M103) expressed v-Abl proteins that were larger than the P90A protein, and M102 expressed a v-Abl protein that was smaller than the P90A protein. Other samples expressed v-Abl proteins that were similar in size to the P90A protein. These data indicate that the loss of p53 is not fully sufficient to complement the defect in oncogenicity characteristic of the P90A mutant in the BALB/cJ background.
Different types of mutations are recovered from P90A- and P160-infected mice. Western analysis indicated that at least one variant was present in many of the infected mice. To analyze the virus population more fully, the sequences encoding the COOH terminus of the v-Abl protein were amplified directly from tumors arising in P90A-infected p53+/+ and p53–/– mice. For comparison, sequences from wild-type and null animals infected with Ab-MLV-P160, a wild-type strain that shows accelerated disease induction in p53–/– mice (33, 37), were also analyzed. The products were cloned, sequenced, and compared to the virus injected into the animal.
Analyses of sequences recovered from P90A-infected mice revealed that 12 of 21 variants recovered from either wild-type or null mice contained mutations that altered the reading frame and changed the structure of the v-Abl protein (Table 1). Six of these (M21 5-1, M21 6-2, M22 7-1, M22 8-1, M100 2-6, and M103 1-9) contained changes that allowed translation to proceed to the normal P120 terminator. Six others contained mutations that generated a termination codon prior to the P90A deletion. These variants resembled the P120, P118, and P85 series of variants isolated from P90A-infected mice and shown previously to be highly oncogenic (13, 14). No variants with this type were recovered from P160-infected mice (Table 1).
Although nearly 60% of the viruses recovered from P90A-infected mice had acquired changes similar to those shown to enhance the oncogenic potential of the virus, others contained base-pair substitutions that did not alter the reading frame of the viral RNA. This type of variant was also recovered from P160-infected mice (Table 1). When the frequencies of base substitutions were compared among the different groups, these changes were recovered from P90A-infected wild-type mice at the highest frequency (Table 2). In addition, a comparison of the ratios of silent to missense mutations revealed that more missense mutations were recovered from P90A-infected wild-type mice than from P160-infected wild-type mice (P = 0.0192; Fisher's exact test). However, this difference was not observed when the data for p53–/– animals infected with the two viruses were compared (P = 0.6148) or when the sequences from P90A-infected wild-type and null mice were compared (P = 0.5147). Together, these data suggest that missense mutations may be selected in P90A-infected mice but that the presence of p53 does not strongly influence selection for this type of change.
Not all P90A variants are more highly oncogenic. Variants that express v-Abl proteins smaller or larger than the P90A protein have previously been shown to be more oncogenic than P90A (13). To determine if viruses with substitutions that alter the amino acid sequence without changing the reading frame confer enhanced oncogenic potential, the C-terminal coding regions of two variants containing shared mutations, M23 2-3 and M23 2-5 (Fig. 2A), were cloned into a viral vector containing the full-length sequence of Ab-MLV. Viruses were isolated from the supernatant of transformed cells, titrated, and used to infect neonatal mice. Mice that were infected with either the 2-3 or 2-5 virus did not succumb to disease more rapidly than animals infected with P90A (Fig. 2B). In addition, protein analysis (Fig. 2C) and an analysis of virus sequences recovered from these tumors (data not shown) revealed that these tumors contained variant proteins. Thus, even though the increased frequency with which missense mutations are recovered following P90A infection of wild-type mice indicates that they may be selected, not all changes of this type enhance the oncogenic potential of the virus. The 2-5 virus shares two mutations with 2-3 and with a third isolate from M23, suggesting that further mutation of 2-5 gave rise to the other two isolates. Because the 2-5 isolate appears to be less oncogenic than both P90A and 2-3, it is possible that the additional changes in 2-3 partially correct for the changes that are present in the 2-5 isolate. Additional passaging of these isolates might amplify these apparent differences in oncogenicity.
The virus population in P90A-infected wild-type mice is more diverse than that in p53-null mice. Although many of the P90A-infected mice contained variants, all of the P160-infected mice and many of the P90A-infected mice contained sequences that were identical to those of the starting virus for the region analyzed (Table 1). To assess the possible effects of the virus strain and host genotype on the viral populations present in the tumors, the number of variants with altered coding sequences that were recovered from each animal was examined (Fig. 3). As described above, variants (open boxes) were recovered from most P90A-infected mice, irrespective of their p53 status. P90A (black boxes) was also recovered from the majority of the animals, including 4/5 p53+/+ mice and 3/5 p53–/– mice. A comparison of the frequencies of P90A and the variants indicated that variants were recovered only somewhat more frequently than P90A from p53+/+ animals (P = 0.0446) and were not recovered more frequently from null animals (P = 0.0900). A comparison of the total viral sequences present in each type of mouse revealed that the virus population in the p53+/+ tumors was more diverse than that in p53–/– mice (P = 0.0076). Because the particular animals studied in depth succumbed to disease within similar latent periods, these differences are not likely to reflect the time that virus replicated in the host.
When data for P160-infected mice were compared in the same way (Fig. 3), sequences identical to the starting virus were recovered more frequently than variants from both the wild-type mice (P < 0.0001) and the p53-null mice (P = 0.0004), indicating that fewer sequence variants are present in mice infected with wild-type virus than in those infected with P90A. In addition, the absence of p53 had no effect on the diversity of the virus population following P160 infection (P = 0.8196). Together, these data indicate that p53 affects the diversity of the virus population when mice are infected with a weakly oncogenic virus, but not with a highly oncogenic strain, suggesting that p53 increases selective pressure for virus variation.
Tumors arising in p53+/+ animals contain more viral integrations. The virus population in P90A-infected wild-type animals is more diverse than that in null mice. These data could indicate that more proviruses are required for tumor development following P90A infection or that the tumors contain a more complex pattern of clones than that observed in the null animals. To investigate these possibilities, the frequency of integrated proviruses in the tumor tissue was determined by using a quantitative real-time PCR assay (Fig. 4A). Although the numbers of proviral copies varied among the different tumors, samples taken from different sites within an animal displayed similar numbers of proviruses. In addition, the average numbers of proviruses detected in samples from p53–/– and p53+/+ mice were not statistically different (P = 0.7321). Because Ab-MLV-induced tumors contain relatively small numbers of healthy cells (30) and since no differences in pathology, tumor size, or composition were evident among samples from the different mice, these data indicate that differences in the increased diversity in the viral population do not reflect an increased number of proviruses in the wild-type mice.
To test the possibility that the tumors in the two types of animals differ in clonality, Southern blotting using an abl probe was used to examine the integration sites (Fig. 4B). These analyses revealed that the p53+/+ tumors contain more proviral integrations than the tumors from p53–/– animals (P = 0.0134). Because the tumors in both groups of mice have similar numbers of proviruses (Fig. 3), the p53–/– animals likely have a larger number of proviral integrations that are present in very small numbers of cells and not resolved by Southern analysis. Because wild-type tumors display integration sites that are not all of equivalent intensities, at least some of the tumors in wild-type mice are made up of multiple clones of cells that are present in sufficient numbers to be detectable by Southern analysis. Such a pattern is consistent with the idea that more than one of the variants contributes to the tumor burden. These data could also indicate that insertional mutagenesis involving either Mo-MLV or Ab-MLV proviruses may play a role in tumor induction. One common Ab-MLV integration site has been identified (11), but it is used at a low frequency and might not be detected in the relatively small sample of tumors examined here.
P90A can initiate the tumorigenic process. Although sequence variants were recovered from the majority of the tumors and many of these had mutations consistent with enhanced oncogenic potential (14), P90A was also recovered from the wild-type animals almost as frequently as the variants. To assess which of the viruses founded the clone(s) characteristic of the wild-type tumors, unblotting (31) was used to resolve the integration site(s) of the provirus(es); bands corresponding to the most prominent integration sites were recovered and the sequences encoding the COOH terminus were amplified. For two of the tumors, M22 and M76, variant sequences were recovered, and in each instance, the sequence was the same as that most frequently recovered from the bulk DNA (Table 3). However, for two other samples (M23 and M75), the sequence recovered was identical to that of P90A. While these two tumors contained diverse viral populations, P90A was commonly recovered using conventional cloning methods. The P90A sequence was also recovered from an integration in M21, an animal from which variant sequences were recovered from bulk tumor DNA much more frequently. Because this sample contained two proviral integrations but only one yielded sequence despite repeated PCR attempts, it is possible that variant sequences are present in the second integrated provirus. In addition, because the band that yielded the P90A sequence was less intense than the second integration band, it is likely that only a portion of the cells in the tumor contained P90A. Taken together, these data demonstrate that both P90A and some variants are capable of initiating a tumor clone. Therefore, P90A is capable of providing the growth signals necessary to initiate the tumorigenic process. These data also suggest that in some animals, variants give rise to the clone that initiates the tumor even though P90A remains present in the animal. Additional variants present in tumor tissue may result from replication as the tumor grows and arise after the initial oncogenic event; some of these may be represented by less intensive bands observed in the Southern analysis, which were difficult to resolve by the less-sensitive unblot approach.
DISCUSSION
Our analysis of the weakly oncogenic Ab-MLV-P90A strain has revealed that the genetic make-up of the host affects the composition of v-onc-containing retrovirus populations in a host animal. Virus populations in P90A-infected mice are likely to mimic those present when newly captured v-onc gene-containing viruses are emerging in vivo. Very few experiments have directly assessed the events that can occur following the recombination events that incorporate cellular sequences into retroviruses. However, limited information (35) and the inherent ability of retroviruses to undergo further recombination and mutation (1, 10, 32) strongly suggest that the structure of newly captured v-onc genes changes following v-onc gene capture. Our experiments used animals lacking the p53 tumor suppressor gene, a genetic background documented to affect Ab-MLV transformation (33, 37) and therefore a situation that might reveal the effects of host background on the virus population. Nonetheless, our results indicate that the genetic constitution of the host(s) infected with viruses such as Ab-MLV helped to shape the v-onc gene-containing retroviruses.
Host genes that alter the outcome of retrovirus infection are well known. The majority of these genes restrict replication and spread of the virus or affect the immune response to the virus (21). Much less is known about how host genes affect the population of viruses in an individual host. Because the numbers of proviruses in the tumors from p53-null and wild-type animals are similar, the presence of p53 does not appear to affect virus replication. Although at least one gene linked to the Mhc locus has been implicated in Ab-MLV pathogenesis, these effects are revealed only when adult mice are infected (19). Other possible effects of the immune response were not addressed in our experiments. However, we used neonatal animals, and p53-null animals are not immunocompromised (7). Thus, immune mechanisms are not likely to be responsible for the differences observed. The tumors in p53-null animals contain fewer clonally integrated proviruses, consistent with the idea that the p53 protein influences the ability of cells infected with Ab-MLV to expand and contribute to the tumor mass (33, 37).
Our experiments were stimulated by work from the Calame group (37) that failed to detect variants by Western analysis in p53-null animals and suggested that the presence of p53 affected the virus population. In our series, v-Abl protein variants were detected by Western analysis of tissues from the null animals, and the frequencies with which P90A and variants were recovered were not statistically different between wild-type and null mice. For one null animal, P90A was the only virus recovered. The reasons for the difference in the recovery of variants require further study, but differences in the strains of mice used may contribute to the results obtained. The Calame group used mixed C57BL6/129 animals and the mice used in our study were backcrossed for a limited number of generations, raising the possibility that host genes in addition to p53 may have affected the diversity of Ab-MLV populations in both studies. Even though analyses of simple sequence length polymorphisms on chromosome 11 near the targeted p53 allele and on several other chromosomes using DNAs from our mice revealed the presence of BALB/c-derived sequences (our unpublished data), sufficient backcrossing to ensure a homogeneous background was not carried out. Thus, differences in the results obtained from the two studies could reflect the particular constellation of genes present in the specific mice that were analyzed. By extension, the products of such genes might cooperate with p53 and influence tumor latency or viral diversity. Future experiments using larger groups of mice and different strain combinations will be useful in determining the contribution that such genes might play.
The generation of variants in P90A-infected mice is strongly dependent on the presence of helper virus (14), indicating that they almost certainly arise during viral replication. Consistent with this idea, features known to affect the fidelity of reverse transcription can be found at the sites of many of the mutations. For example, all of the single base insertions (4/4) and single base deletions (1/1) occurred at sites where the same nucleotide was repeated four to six times. Regions similar to these have been implicated as hotspots for mutation in studies that investigated retroviral mutation (5, 16). A second process involving template switching and recombination (15, 17) likely generated the larger deletions observed in many of the variants. Regions of sequence homology, even as small as three nucleotides (15), can influence this process, and in one variant (6-2) the deletion recovered occurred across a stretch of eight identical nucleotides.
Our earlier work, conducted prior to readily adaptable PCR approaches (13), suggested that most tumors arising in a P90A-infected mouse contained a variant with increased oncogenic potential and led to the hypothesis that this variant was responsible for induction of the tumor. Similar to these data, variants of this type were recovered in the present study. However, the newer results indicate that the earlier interpretation was overly simple. All of the P90A-infected animals contained a diverse population of viruses, some of which did not display enhanced oncogenic potential compared to P90A. In addition, in several instances P90A initiated the tumorigenic process, since these sequences were recovered from the proviral integration that marks the tumor clone. These data may suggest that additional evolution within the P90A population would occur if tumor extracts were passaged in additional mice, a hypothesis that is under investigation. Because most v-onc gene-containing retroviruses were isolated before molecular characterization was possible and passaged multiple times, often through different species, examining the effects of passaging should help us to understand the dynamics of host-virus interactions that helped to shape viruses like Ab-MLV.
Analyses of the infected tissues revealed that the P90A virus was retained in 4/5 wild-type tumors and 3/5 null tumors. In addition, P90A was clonally integrated in three of five wild-type mice, demonstrating that this virus played an important role in initiating the tumorigenic process. These data are consistent with the ability of P90A to transform pre-B cells in vitro (24) and with the observation that a small number of tumors arise in animals infected with P90A in the absence of helper virus (14). While the possibility exists that these viruses have acquired mutations outside the region sequenced, these results predict that viruses that have incorporated cellular sequences must have the ability to stimulate cell growth in order to emerge as v-onc gene-containing retroviruses. Consistent with this idea, P120/D484N, a kinase-inactive strain that cannot stimulate cell growth, fails to establish an infection in mice (our unpublished data). The need to incorporate cellular sequences in a way that allows the product to stimulate growth, albeit poorly, places additional constraints on the generation of v-onc gene-containing retroviruses. Thus, for viruses such as Ab-MLV to have arisen originally, the recombination events that occurred initially needed to encode a protein that was capable of stimulating cell replication. Additional studies using the P90A model should help to elucidate the mechanisms that have shaped the v-onc gene-containing retroviruses that are studied today.
ACKNOWLEDGMENTS
We are grateful to Erika de Leon Vazquez, Jyoti Mathur, Todd Ashworth, Warangkhana Songsungthang, Rebekah Stackpole, and Laila Sheikh for assistance with analyses of variant viruses, to Chris Schmid for assistance with statistical analysis, and to John Coffin and Caleb Lee for helpful discussions.
This work was supported by grant CA 24220 from the National Cancer Institute.
REFERENCES
Anderson, J. A., R. J. Teufel II, P. D. Yin, and W. S. Hu. 1998. Correlated template-switching events during minus-strand DNA synthesis: a mechanism for high negative interference during retroviral recombination. J. Virol. 72:1186-1194.
Baxter, E. W., K. Blyth, E. R. Cameron, and J. C. Neil. 2001. Selection for loss of p53 function in T-cell lymphomagenesis is alleviated by Moloney murine leukemia virus infection in myc transgenic mice. J. Virol. 75:9790-9798.
Baxter, E. W., K. Blyth, L. A. Donehower, E. R. Cameron, D. E. Onions, and J. C. Neil. 1996. Moloney murine leukemia virus-induced lymphomas in p53-deficient mice: overlapping pathways in tumor development? J. Virol. 70:2095-2100.
Blyth, K., A. Terry, N. Mackay, F. Vaillant, M. Bell, E. R. Cameron, J. C. Neil, and M. Stewart. 2001. Runx2: a novel oncogenic effector revealed by in vivo complementation and retroviral tagging. Oncogene 20:295-302.
Burns, D. P., and H. M. Temin. 1994. High rates of frameshift mutations within homo-oligomeric runs during a single cycle of retroviral replication. J. Virol. 68:4196-4203.
Chesebro, B., K. Wehrly, M. Cloyd, W. Britt, J. Portis, J. Collins, and J. Nishio. 1981. Characterization of mouse monoclonal antibodies specific for Friend murine leukemia virus-induced erythroleukemia cells: Friend-specific and FMR-specific antigens. Virology 112:131-144.
Donehower, L. A., M. Harvey, B. L. Slagle, M. J. McArthur, C. A. Montgomery, Jr., J. S. Butel, and A. Bradley. 1992. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature (London) 356:215-221.
Fan, H., and M. Paskind. 1974. Measurement of the sequence complexity of cloned Moloney murine leukemia virus 60 to 70S RNA: evidence for a haploid genome. J. Virol. 14:421-429.
Green, P. L., D. A. Kaehler, and R. Risser. 1987. Cell transformation and tumor induction by Abelson murine leukemia virus in the absence of helper virus. Proc. Natl. Acad. Sci. USA 84:5932-5936.
Hu, W. S., and H. M. Temin. 1990. Genetic consequences of packaging two RNA genomes in one retroviral particle: pseudodiploidy and high rate of genetic recombination. Proc. Natl. Acad. Sci. USA 87:1556-1560.
Jiang, X., Z. Hanna, M. Kaouass, L. Girard, and P. Jolicoeur. 2002. Ahi-1, a novel gene encoding a modular protein with WD40-repeat and SH3 domains, is targeted by the Ahi-1 and Mis-2 provirus integrations. J. Virol. 76:9046-9059.
Lund, A. H., G. Turner, A. Trubetskoy, E. Verhoeven, E. Wientjens, D. Hulsman, R. Russell, R. A. DePinho, J. Lenz, and M. van Lohuizen. 2002. Genome-wide retroviral insertional tagging of genes involved in cancer in Cdkn2a-deficient mice. Nat. Genet. 32:160-165.
Murtagh, K., G. Skladany, J. Hoag, and N. Rosenberg. 1986. Abelson murine leukemia virus variants with increased oncogenic potential. J. Virol. 60:599-606.
Parmar, K., R. C. Huebner, and N. Rosenberg. 1991. Carboxyl-terminal determinants of Abelson protein important for lymphoma induction. J. Virol. 65:6478-6485.
Pathak, V. K., and H. M. Temin. 1990. Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: deletions and deletions with insertions. Proc. Natl. Acad. Sci. USA 87:6024-6028.
Pathak, V. K., and H. M. Temin. 1990. Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: substitutions, frameshifts, and hypermutations. Proc. Natl. Acad. Sci. USA 87:6019-6023.
Pulsinelli, G. A., and H. M. Temin. 1991. Characterization of large deletions occurring during a single round of retrovirus vector replication: novel deletion mechanism involving errors in strand transfer. J. Virol. 65:4786-4797.
Radfar, A., I. Unnikrishnan, H.-W. Lee, R. A. DePinho, and N. Rosenberg. 1998. p19Arf induces p53-dependent apoptosis during Abelson virus-mediated pre-B cell transformation. Proc. Natl. Acad. Sci. USA 95:13194-13199.
Risser, R., D. Kaehler, and W. W. Lamph. 1985. Different genes control the susceptibility of mice to Moloney or Abelson murine leukemia viruses. J. Virol. 55:547-553.
Risser, R., M. Potter, and W. P. Rowe. 1978. Abelson virus-induced lymphomagenesis in mice. J. Exp. Med. 148:714-726.
Rosenberg, N., and P. Jolicoeur. 1997. Retrovirus pathogenesis, p. 475-586. In J. M. Coffin, S. E. Hughes, and H. E. Varmus (ed.), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Rosenberg, N., and O. N. Witte. 1980. Abelson murine leukemia virus mutants with alterations in the virus-specific P120 molecule. J. Virol. 33:340-348.
Rosenberg, N., and O. N. Witte. 1988. The viral and cellular forms of the Abelson (abl) oncogene. Adv. Virus Res. 35:39-81.
Rosenberg, N. E., D. R. Clark, and O. N. Witte. 1980. Abelson murine leukemia virus mutants deficient in kinase activity and lymphoid cell transformation. J. Virol. 36:766-774.
Sachs, Z., N. Sharpless, R. DePinho, and N. Rosenberg. p16Ink4a stimulates apoptosis during Abelson virus-mediated pre-B cell transformation. Submitted for publication.
Scher, C. D., and R. Siegler. 1975. Direct transformation of 3T3 cells by Abelson murine leukemia virus. Nature (London) 253:729-731.
Sherr, C. J., and F. McCormick. 2002. The RB and p53 pathways in cancer. Cancer Cell 2:103-112.
Shinto, Y., M. Morimoto, M. Katsumata, A. Uchida, K. Aozasa, M. Okamoto, T. Kurosawa, T. Ochi, M. I. Greene, and Y. Tsujimoto. 1995. Moloney murine leukemia virus infection accelerates lymphomagenesis in E mu-bcl-2 transgenic mice. Oncogene 11:1729-1736.
Shoemaker, C., J. Hoffman, S. P. Goff, and D. Baltimore. 1981. Intramolecular integration within Moloney murine leukemia virus DNA. J. Virol. 40:164-172.
Siegler, R., and S. Zajdel. 1972. Pathogenesis of Abelson-virus-induced murine leukemia. J. Natl. Cancer Inst. 48:189-218.
Stoye, J. P., W. N. Frankel, and J. M. Coffin. 1991. DNA hybridization in dried gels with fragmented probes: an improvement over blotting techniques. Technique 3:123-128.
Temin, H. M. 1993. Retrovirus variation and reverse transcription: abnormal strand transfers result in retrovirus genetic variation. Proc. Natl. Acad. Sci. USA 90:6900-6903.
Unnikrishnan, I., A. Radfar, J. Jenab-Wolcott, and N. Rosenberg. 1999. p53 mediates apoptotic crisis in primary Ab-MLV-transformed pre-B cells. Mol. Cell. Biol. 19:4825-4831.
van Lohuizen, M., S. Verbeek, P. Krimpenfort, J. Domen, C. Saris, T. Radaszkiewicz, and A. Berns. 1989. Predisposition to lymphomagenesis in pim-1 transgenic mice: cooperation with c-myc and N-myc in murine leukemia virus-induced tumors. Cell 56:673-682.
Vennstrom, B., C. Raynoscheck, L. Jansson, G. Doederlein, V. Lhotak, A. Johnsson, and H. Beug. 1994. Retroviral capture of c-erbB proto-oncogene sequences: rapid evolution of distinct viral genomes carrying mutant v-erbB genes with different transforming capacities. Oncogene 9:1307-1320.
Wang, L. C., Y.-Y. Chen, and N. Rosenberg. 1995. Pre-B cells transformed by ts Abelson virus rearrange k and l light chain genes in early G1. Curr. Top. Microbiol. Immunol. 194:355-361.
Zou, X., F. Cong, M. Coutts, G. Cattoretti, S. P. Goff, and K. Calame. 2000. p53 deficiency increases transformation by v-Abl and rescues the ability of a C-terminally truncated v-Abl mutant to induce pre-B lymphoma in vivo. Mol. Cell. Biol. 20:628-633.(Erica Marchlik, Richard K)
Departments of Pathology
Molecular Biology and Microbiology, Sackler School of Graduate Biomedical Sciences, Tufts University School of Medicine, Boston, Massachusetts 02111
ABSTRACT
The Abelson murine leukemia virus (Ab-MLV), like other retroviruses that contain v-onc genes, arose following a recombination event between a replicating retrovirus and a cellular oncogene. Although experimentally validated models have been presented to address the mechanism by which oncogene capture occurs, very little is known about the events that influence emerging viruses following the recombination event that incorporates the cellular sequences. One feature that may play a role is the genetic makeup of the host in which the virus arises; a number of host genes, including oncogenes and tumor suppressor genes, have been shown to affect the pathogenesis of many murine leukemia viruses. To examine how a host gene might affect an emerging v-onc gene-containing retrovirus, we studied the weakly oncogenic Ab-MLV-P90A strain, a mutant that generates highly oncogenic variants in vivo, and compared the viral populations in normal mice and mice lacking the p53 tumor suppressor gene. While variants arose in both p53+/+ and p53–/– tumors, the samples from the wild-type animals contained a more diverse virus population. Differences in virus population diversity were not observed when wild-type and null animals were infected with a highly oncogenic wild-type strain of Ab-MLV. These results indicate that p53, and presumably other host genes, affects the selective forces that operate on virus populations in vivo and likely influences the evolution of oncogenic retroviruses such as Ab-MLV.
INTRODUCTION
Abelson murine leukemia virus (Ab-MLV) is a replication-defective retrovirus that transforms NIH 3T3 cells and pre-B cells in vitro and induces pre-B-cell lymphoma in vivo by expressing the v-Abl protein tyrosine kinase (reviewed in reference 23). Kinase activity is required for tumor induction, and wild-type Ab-MLV can induce tumors in the absence of viral replication (9). However, a replicating helper virus is required for at least one weakly oncogenic mutant, the P90A strain (22), to induce tumors (14). In this instance, virus replication is accompanied by the emergence of viruses that have undergone additional mutation, rendering them more highly oncogenic (13, 14). Thus, viral determinants in addition to kinase activity play an important role in determining oncogenicity, and in vivo passaging can select for viruses with increased pathogenic potential.
Host determinants also influence tumor development. Laboratory mouse strains differ markedly in their susceptibilities to Ab-MLV, and at least two, as yet unknown, loci appear to be involved in mediating this response (19, 20). In addition, the use of genetically altered mice has revealed that both tumor suppressor genes and oncogenes can affect tumor induction and in vitro transformation by a variety of retroviruses, including Ab-MLV (2-4, 12, 28, 33, 34, 37). Such analyses have shown that the p53 tumor suppressor protein, a molecule activated following oncogenic insult (27), influences transformation by Ab-MLV in vivo and in vitro (33, 37). In vitro, p53 functions during the late stages of transformation and is required for the apoptotic crisis phase of the process, a response that is also influenced by the products of the Ink4a/Arf locus (18, 25, 33). Less is known about the way in which p53 affects tumor induction in vivo. However, analyses of p53-null animals infected with the weakly oncogenic P90A strain (37) revealed that the absence of p53 accelerated disease and relieved the selective pressure on the P90A strain for the generation of more highly oncogenic mutants.
The effects of p53 on the weakly oncogenic P90A strain suggest that host genes may play an important role in shaping the virus population present in infected animals. Analyses of newly arising avian erythroblastosis viruses demonstrated that several viruses can exist in a single animal (35), suggesting that v-onc gene-containing viruses probably arise from a pool of viruses generated during replication following the recombination event that joins viral and cellular sequences. To determine how a host gene such as p53 might influence the composition of emerging oncogenic viruses, we examined the viral populations recovered from p53+/+ and p53–/– mice infected with the P90A strain. These analyses revealed the presence of virus variants in both types of mice but demonstrated that viral populations are more heterogeneous in p53+/+ tumors than in p53–/– tumors. These results indicate that host genes can affect the composition of oncogenic retrovirus populations and likely contributed to the way in which v-onc gene-containing retroviruses evolved following oncogene capture.
MATERIALS AND METHODS
Viruses and animals. The Ab-MLV-P90A (13) virus stock was prepared from NIH 3T3 cells that were cotransfected with pBR-90, a plasmid containing the Ab-MLV-P90A genome (14), and pZAP, a plasmid containing Moloney MLV (Mo-MLV) sequences (29). Ab-MLV-P160 (22) was isolated from a clonally derived NIH 3T3 virus-transformed nonproducer cell line that had been superinfected with Mo-MLV-Cl2 (8). The Ab-MLV-2-3 and Ab-MLV-2-5 viruses were generated by recovering the SacI-FseI fragment (bp 2395 to 3016 of the Ab-MLV-P120 genome) from plasmids containing the C-terminal coding sequences of these variants and replacing the corresponding fragment in the pBR-90 vector. Viruses were recovered by cotransfection of NIH 3T3 cells as described above. All virus stocks were filtered through a 0.45-μm filter and titrated by using the NIH 3T3 transformation assay (26).
BALB/cJ and p53–/– mice were obtained from our breeding colony at Tufts University. The p53–/– mice were maintained by mating heterozygous animals originally obtained from a single breeding pair of p53+/– animals (Jackson Laboratory) that had been backcrossed to BALB/cJ mice five times and inbred for three generations. Neonatal mice were injected via the intraperitoneal route with approximately 1 x 104 focus-forming units of Ab-MLV stock containing 0.8 μg Polybrene (Sigma). Animals were monitored for a 90-day period and sacrificed once signs of tumor development (lymphadenopathy, cranial tumors, or hind limb paralysis) were evident. Tumor tissue was removed and frozen at –80°C. Tumor latency was compared by generating Kaplan-Meier survival plots and comparing the plots using Prism (GraphPad Software) and the Mantel-Haenszel log rank test, which calculates a P value comparing the two curves by considering the possibility that random chance would lead to curves that are as similar as those generated from the data. Comparisons of sequences recovered from tumor tissues were performed by using Fisher's exact test (GraphPad Software), which compares two unpaired groups of data, and the unpaired t test, which compares two groups of data.
DNA analysis. Genomic DNA was prepared by macerating tumor samples on dry ice with a mortar and pestle, and lysis buffer (10 mM Tris-HCl, pH 7.6, 10 mM EDTA, pH 8.0, 10 mM NaCl, 0.5% N-lauroyl sarcosine, 1 mg/ml proteinase K) was added. The lysate was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), and the DNA was recovered by precipitation and stored in TE buffer (10 mM Tris, pH 8.0, 1 mM EDTA). The C-terminal coding sequences of the viruses (bp 2380 to 3048 of the Ab-MLV-P120 genome) were amplified by using the primers 5'-AGAAGGTCTACGAGCTCATGC-3' and 5'-GCACAGGCTTTCTCAGTCCTT-3'. Amplification reaction mixtures contained 100 ng of genomic DNA, a 200 μM deoxynucleoside triphosphate mix (Pharmacia), a 0.4 μM concentration of each primer, 2.5 U Pfu polymerase (Stratagene), and Pfu buffer [20 mM Tris-HCl, pH 8.8, 2 mM MgSO4, 10 mM KCl, 10 mM (NH4)2SO4, 0.1% Triton X-100, 0.1 mg/ml bovine serum albumin]. After a 2-min incubation at 94°C, the reactions were run for 30 cycles of 1 min at 94°C, 1 min at 59°C, and 1 min at 72°C. After the final cycle, the reaction mixtures were incubated for 4 min at 72°C and cooled to 4°C. The reaction products were purified using a Qiaquik PCR purification kit (QIAGEN), and 3' A overhangs were added by diluting the entire purified product in a total volume of 50 μl containing 0.2 mM dATP, 0.5 U Taq polymerase (Applied Biosystems), and Taq buffer (10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, and 0.001% [wt/vol] gelatin) and incubating the mixture at 72°C for 15 min. The PCR products were cloned into the pCR4-TOPO vector (Invitrogen) following the manufacturer's protocol, and the inserts were sequenced on an ABI 3100 DNA sequencer (Perkin-Elmer) at the DNA Facility, Department of Physiology, Tufts University School of Medicine. As an additional control, the viral sequence was amplified from a vector containing the P90A COOH-terminal coding region; no changes were detected in the 6,500 bases analyzed.
The numbers of proviruses in the tumors were determined using quantitative real-time PCR and an Opticon 1 thermocycler (MJ Research); the data were analyzed using Opticon Monitor v1.08 software. All amplification reactions were performed in triplicate with reaction mixtures containing 10 ng template, 1x SYBR green dye (Applied Biosystems), and a 0.2 μM concentration of each primer. The number of cells in each reaction was evaluated by amplifying a portion of the Rag1 locus and calculating the number of copies based on amplification of a plasmid standard curve with the assumption that each cell contained two copies of the Rag1 locus. A similar plasmid standard curve was used to assess the number of v-abl copies, and this value was divided by the number of cells in the sample to calculate the number of proviruses per cell. A cell line containing a single copy of v-abl was used as an additional control. v-abl proviral sequences were amplified with the primers 5'-GATCCATCTCGCTGCGGTAT-3' and 5'-ACTAACTCAGCCAGAGTGTTGAAGC-3', and sequences from the Rag1 locus were amplified with the primers 5'-ATCATCTGTGGTTAGCCGTCTGT-3' and 5'-ATTATGTATCAGCTCTCACGCCC-3'. All amplification reaction mixtures were incubated at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and either 59°C for 1 min and 81°C for 1 s (for v-abl amplification) or 58.5°C for 1 min and 79°C for 1 s (for Rag1 amplification).
Southern analyses were performed as previously described (36). Briefly, digested DNAs were fractionated overnight through 0.8% agarose-1x Tris-borate-EDTA gels, and after denaturation and neutralization, the DNAs were transferred to nitrocellulose membranes (Schleicher & Schuell). The membranes were baked for 2 h at 80°C in a vacuum oven and incubated at 42°C for 2 h in prehybridization buffer containing 6x SSC (0.9 M NaCl, 90 mM sodium citrate, pH 7.0), 10x Denhardt's solution (0.2% Ficoll 400 [Pharmacia], 0.2% polyvinylpyrrolidone, and 0.2% bovine serum albumin), and 50 μg/μl of single-stranded salmon sperm DNA. Membranes were probed with the SalI-BspEI fragment of Ab-MLV-P120 (bp 3073 to 3556 of the genome) labeled by the random primer method with 50 μCi of [-32P]dCTP (3,000 Ci/mmol) using a Prime-It RmT random primer labeling kit (Stratagene) and purified using a G-50 Sephadex column (Roche). The probe was mixed with hybridization buffer (6x SSC, 5% dextran sulfate, 1% sodium dodecyl sulfate [SDS], 10 mM EDTA) and applied to the membranes, which were then incubated at 65°C overnight. The following day, the membranes were washed twice with 5x SSC-0.5% SDS and twice with 0.5x SSC-0.15% SDS and exposed to Kodak X-AR-5 film at –80°C with an intensifier screen.
For unblot experiments, the genomic DNAs were digested and fractionated as described above, and the gels were dried on a slab gel dryer using a house vacuum (31) and probed with an oligonucleotide probe (5'-CTGTACACTTTCTGTGTGTGAGCTATGT-3'; bp 3360 to 3384 of the Ab-MLV-P120 genome). The oligonucleotide was labeled with 750 μCi of [-32P]ATP (6,000 Ci/mmol) using T4 polynucleotide kinase (NEBiolabs) and purified using a G-25 Sephadex column (Roche). The probe was mixed with buffer containing 5x SSPE (0.9 M NaCl, 0.05 M NaH2PO4, 0.005 M EDTA, pH 7.4), 0.1% SDS, and 0.01 μg/ml of single-stranded salmon sperm DNA and applied to the dried gels, which were then incubated at 55°C overnight. Following hybridization, the gels were washed four times in wash buffer (2x SSC, 0.2% SDS), twice at room temperature and then twice at 55°C. The gels were air dried and exposed to Kodak X-AR-5 film at –80°C with an intensifier screen. Proviral integrations were excised by aligning the dried gel with the autoradiogram. The excised agarose was melted in 10 mM Tris-HCl, pH 8.0, and a sample was amplified using the PCR protocol described above, cloned, and sequenced.
Protein analysis. Tumor cell lysates were prepared by macerating tumor samples on dry ice with a mortar and pestle, and protein lysis buffer (10 mM Tris-HCl, 1% SDS, pH 7.5, 0.1 mM sodium vanadate, and 100 μM phenylmethylsulfonyl fluoride) was added. The lysates were boiled for 5 min and sheared by passing the lysates through a 25-gauge needle. The amount of protein present in the lysates was determined by using a bicinchoninic acid protein assay kit (Pierce), and 50 μg of protein was fractionated through an 8% SDS-polyacrylamide gel. The proteins were electrotransferred to polyvinylidene difluoride membranes (U.S. Biochemicals) that were blocked with phosphate-buffered saline containing 0.2% I-Block (Tropix) and 0.1% Tween 20 for at least 1 h. The blots were probed with an anti-Gag/v-Abl antibody (H548) (6) according to the Western Light kit protocol (Tropix), utilizing an alkaline phosphatase-conjugated secondary antibody with a CSPD substrate (Tropix). The blots were exposed to Kodak X-AR-5 film.
RESULTS
Ab-MLV-P90A generates variants in both p53+/+ and p53–/– mice. The P90A strain was derived from the wild-type P120 strain and contains a 19-base deletion that shifts the reading frame, resulting in the expression of a truncated v-Abl protein that resembles the wild-type protein up to amino acid 751 and contains an additional 33 amino acids appended as a result of the reading frame change (Fig. 1A). In preparation for experiments to assess the effects of p53 on Ab-MLV populations in vivo, we infected p53+/+ and p53–/– mice with P90A and monitored their survival. Consistent with the results of other work (37), P90A accelerated tumor formation when disease induction was compared in p53+/+ and p53–/– mice in the BALB/cJ background used in our laboratory (Fig. 1B). All of the null animals, compared to 60% of the wild-type animals, developed disease, and the mean survival period was somewhat shorter. A comparison of the survival curves by using a log rank test revealed that the null animals were somewhat more susceptible to disease (P = 0.0351). Despite this difference, the gross and microscopic appearances of the tumors were similar (data not shown). In addition, analyses of five null and five wild-type tumor samples tested revealed the presence of rearranged immunoglobulin heavy chain genes (data not shown) typically associated with Ab-MLV-induced pre-B-cell lymphoma. Consistent with these data, analyses of B, T and myeloid surface markers in a different series of tumors from p53-null and wild-type mice indicated that all of the tumors were of the pre-B-cell phenotype (37).
Analyses of the v-Abl proteins present in tumor lysates revealed the presence of variants in both wild-type and null mice (Fig. 1C). Some samples from both types of mice (M21, M22, M100, and M103) expressed v-Abl proteins that were larger than the P90A protein, and M102 expressed a v-Abl protein that was smaller than the P90A protein. Other samples expressed v-Abl proteins that were similar in size to the P90A protein. These data indicate that the loss of p53 is not fully sufficient to complement the defect in oncogenicity characteristic of the P90A mutant in the BALB/cJ background.
Different types of mutations are recovered from P90A- and P160-infected mice. Western analysis indicated that at least one variant was present in many of the infected mice. To analyze the virus population more fully, the sequences encoding the COOH terminus of the v-Abl protein were amplified directly from tumors arising in P90A-infected p53+/+ and p53–/– mice. For comparison, sequences from wild-type and null animals infected with Ab-MLV-P160, a wild-type strain that shows accelerated disease induction in p53–/– mice (33, 37), were also analyzed. The products were cloned, sequenced, and compared to the virus injected into the animal.
Analyses of sequences recovered from P90A-infected mice revealed that 12 of 21 variants recovered from either wild-type or null mice contained mutations that altered the reading frame and changed the structure of the v-Abl protein (Table 1). Six of these (M21 5-1, M21 6-2, M22 7-1, M22 8-1, M100 2-6, and M103 1-9) contained changes that allowed translation to proceed to the normal P120 terminator. Six others contained mutations that generated a termination codon prior to the P90A deletion. These variants resembled the P120, P118, and P85 series of variants isolated from P90A-infected mice and shown previously to be highly oncogenic (13, 14). No variants with this type were recovered from P160-infected mice (Table 1).
Although nearly 60% of the viruses recovered from P90A-infected mice had acquired changes similar to those shown to enhance the oncogenic potential of the virus, others contained base-pair substitutions that did not alter the reading frame of the viral RNA. This type of variant was also recovered from P160-infected mice (Table 1). When the frequencies of base substitutions were compared among the different groups, these changes were recovered from P90A-infected wild-type mice at the highest frequency (Table 2). In addition, a comparison of the ratios of silent to missense mutations revealed that more missense mutations were recovered from P90A-infected wild-type mice than from P160-infected wild-type mice (P = 0.0192; Fisher's exact test). However, this difference was not observed when the data for p53–/– animals infected with the two viruses were compared (P = 0.6148) or when the sequences from P90A-infected wild-type and null mice were compared (P = 0.5147). Together, these data suggest that missense mutations may be selected in P90A-infected mice but that the presence of p53 does not strongly influence selection for this type of change.
Not all P90A variants are more highly oncogenic. Variants that express v-Abl proteins smaller or larger than the P90A protein have previously been shown to be more oncogenic than P90A (13). To determine if viruses with substitutions that alter the amino acid sequence without changing the reading frame confer enhanced oncogenic potential, the C-terminal coding regions of two variants containing shared mutations, M23 2-3 and M23 2-5 (Fig. 2A), were cloned into a viral vector containing the full-length sequence of Ab-MLV. Viruses were isolated from the supernatant of transformed cells, titrated, and used to infect neonatal mice. Mice that were infected with either the 2-3 or 2-5 virus did not succumb to disease more rapidly than animals infected with P90A (Fig. 2B). In addition, protein analysis (Fig. 2C) and an analysis of virus sequences recovered from these tumors (data not shown) revealed that these tumors contained variant proteins. Thus, even though the increased frequency with which missense mutations are recovered following P90A infection of wild-type mice indicates that they may be selected, not all changes of this type enhance the oncogenic potential of the virus. The 2-5 virus shares two mutations with 2-3 and with a third isolate from M23, suggesting that further mutation of 2-5 gave rise to the other two isolates. Because the 2-5 isolate appears to be less oncogenic than both P90A and 2-3, it is possible that the additional changes in 2-3 partially correct for the changes that are present in the 2-5 isolate. Additional passaging of these isolates might amplify these apparent differences in oncogenicity.
The virus population in P90A-infected wild-type mice is more diverse than that in p53-null mice. Although many of the P90A-infected mice contained variants, all of the P160-infected mice and many of the P90A-infected mice contained sequences that were identical to those of the starting virus for the region analyzed (Table 1). To assess the possible effects of the virus strain and host genotype on the viral populations present in the tumors, the number of variants with altered coding sequences that were recovered from each animal was examined (Fig. 3). As described above, variants (open boxes) were recovered from most P90A-infected mice, irrespective of their p53 status. P90A (black boxes) was also recovered from the majority of the animals, including 4/5 p53+/+ mice and 3/5 p53–/– mice. A comparison of the frequencies of P90A and the variants indicated that variants were recovered only somewhat more frequently than P90A from p53+/+ animals (P = 0.0446) and were not recovered more frequently from null animals (P = 0.0900). A comparison of the total viral sequences present in each type of mouse revealed that the virus population in the p53+/+ tumors was more diverse than that in p53–/– mice (P = 0.0076). Because the particular animals studied in depth succumbed to disease within similar latent periods, these differences are not likely to reflect the time that virus replicated in the host.
When data for P160-infected mice were compared in the same way (Fig. 3), sequences identical to the starting virus were recovered more frequently than variants from both the wild-type mice (P < 0.0001) and the p53-null mice (P = 0.0004), indicating that fewer sequence variants are present in mice infected with wild-type virus than in those infected with P90A. In addition, the absence of p53 had no effect on the diversity of the virus population following P160 infection (P = 0.8196). Together, these data indicate that p53 affects the diversity of the virus population when mice are infected with a weakly oncogenic virus, but not with a highly oncogenic strain, suggesting that p53 increases selective pressure for virus variation.
Tumors arising in p53+/+ animals contain more viral integrations. The virus population in P90A-infected wild-type animals is more diverse than that in null mice. These data could indicate that more proviruses are required for tumor development following P90A infection or that the tumors contain a more complex pattern of clones than that observed in the null animals. To investigate these possibilities, the frequency of integrated proviruses in the tumor tissue was determined by using a quantitative real-time PCR assay (Fig. 4A). Although the numbers of proviral copies varied among the different tumors, samples taken from different sites within an animal displayed similar numbers of proviruses. In addition, the average numbers of proviruses detected in samples from p53–/– and p53+/+ mice were not statistically different (P = 0.7321). Because Ab-MLV-induced tumors contain relatively small numbers of healthy cells (30) and since no differences in pathology, tumor size, or composition were evident among samples from the different mice, these data indicate that differences in the increased diversity in the viral population do not reflect an increased number of proviruses in the wild-type mice.
To test the possibility that the tumors in the two types of animals differ in clonality, Southern blotting using an abl probe was used to examine the integration sites (Fig. 4B). These analyses revealed that the p53+/+ tumors contain more proviral integrations than the tumors from p53–/– animals (P = 0.0134). Because the tumors in both groups of mice have similar numbers of proviruses (Fig. 3), the p53–/– animals likely have a larger number of proviral integrations that are present in very small numbers of cells and not resolved by Southern analysis. Because wild-type tumors display integration sites that are not all of equivalent intensities, at least some of the tumors in wild-type mice are made up of multiple clones of cells that are present in sufficient numbers to be detectable by Southern analysis. Such a pattern is consistent with the idea that more than one of the variants contributes to the tumor burden. These data could also indicate that insertional mutagenesis involving either Mo-MLV or Ab-MLV proviruses may play a role in tumor induction. One common Ab-MLV integration site has been identified (11), but it is used at a low frequency and might not be detected in the relatively small sample of tumors examined here.
P90A can initiate the tumorigenic process. Although sequence variants were recovered from the majority of the tumors and many of these had mutations consistent with enhanced oncogenic potential (14), P90A was also recovered from the wild-type animals almost as frequently as the variants. To assess which of the viruses founded the clone(s) characteristic of the wild-type tumors, unblotting (31) was used to resolve the integration site(s) of the provirus(es); bands corresponding to the most prominent integration sites were recovered and the sequences encoding the COOH terminus were amplified. For two of the tumors, M22 and M76, variant sequences were recovered, and in each instance, the sequence was the same as that most frequently recovered from the bulk DNA (Table 3). However, for two other samples (M23 and M75), the sequence recovered was identical to that of P90A. While these two tumors contained diverse viral populations, P90A was commonly recovered using conventional cloning methods. The P90A sequence was also recovered from an integration in M21, an animal from which variant sequences were recovered from bulk tumor DNA much more frequently. Because this sample contained two proviral integrations but only one yielded sequence despite repeated PCR attempts, it is possible that variant sequences are present in the second integrated provirus. In addition, because the band that yielded the P90A sequence was less intense than the second integration band, it is likely that only a portion of the cells in the tumor contained P90A. Taken together, these data demonstrate that both P90A and some variants are capable of initiating a tumor clone. Therefore, P90A is capable of providing the growth signals necessary to initiate the tumorigenic process. These data also suggest that in some animals, variants give rise to the clone that initiates the tumor even though P90A remains present in the animal. Additional variants present in tumor tissue may result from replication as the tumor grows and arise after the initial oncogenic event; some of these may be represented by less intensive bands observed in the Southern analysis, which were difficult to resolve by the less-sensitive unblot approach.
DISCUSSION
Our analysis of the weakly oncogenic Ab-MLV-P90A strain has revealed that the genetic make-up of the host affects the composition of v-onc-containing retrovirus populations in a host animal. Virus populations in P90A-infected mice are likely to mimic those present when newly captured v-onc gene-containing viruses are emerging in vivo. Very few experiments have directly assessed the events that can occur following the recombination events that incorporate cellular sequences into retroviruses. However, limited information (35) and the inherent ability of retroviruses to undergo further recombination and mutation (1, 10, 32) strongly suggest that the structure of newly captured v-onc genes changes following v-onc gene capture. Our experiments used animals lacking the p53 tumor suppressor gene, a genetic background documented to affect Ab-MLV transformation (33, 37) and therefore a situation that might reveal the effects of host background on the virus population. Nonetheless, our results indicate that the genetic constitution of the host(s) infected with viruses such as Ab-MLV helped to shape the v-onc gene-containing retroviruses.
Host genes that alter the outcome of retrovirus infection are well known. The majority of these genes restrict replication and spread of the virus or affect the immune response to the virus (21). Much less is known about how host genes affect the population of viruses in an individual host. Because the numbers of proviruses in the tumors from p53-null and wild-type animals are similar, the presence of p53 does not appear to affect virus replication. Although at least one gene linked to the Mhc locus has been implicated in Ab-MLV pathogenesis, these effects are revealed only when adult mice are infected (19). Other possible effects of the immune response were not addressed in our experiments. However, we used neonatal animals, and p53-null animals are not immunocompromised (7). Thus, immune mechanisms are not likely to be responsible for the differences observed. The tumors in p53-null animals contain fewer clonally integrated proviruses, consistent with the idea that the p53 protein influences the ability of cells infected with Ab-MLV to expand and contribute to the tumor mass (33, 37).
Our experiments were stimulated by work from the Calame group (37) that failed to detect variants by Western analysis in p53-null animals and suggested that the presence of p53 affected the virus population. In our series, v-Abl protein variants were detected by Western analysis of tissues from the null animals, and the frequencies with which P90A and variants were recovered were not statistically different between wild-type and null mice. For one null animal, P90A was the only virus recovered. The reasons for the difference in the recovery of variants require further study, but differences in the strains of mice used may contribute to the results obtained. The Calame group used mixed C57BL6/129 animals and the mice used in our study were backcrossed for a limited number of generations, raising the possibility that host genes in addition to p53 may have affected the diversity of Ab-MLV populations in both studies. Even though analyses of simple sequence length polymorphisms on chromosome 11 near the targeted p53 allele and on several other chromosomes using DNAs from our mice revealed the presence of BALB/c-derived sequences (our unpublished data), sufficient backcrossing to ensure a homogeneous background was not carried out. Thus, differences in the results obtained from the two studies could reflect the particular constellation of genes present in the specific mice that were analyzed. By extension, the products of such genes might cooperate with p53 and influence tumor latency or viral diversity. Future experiments using larger groups of mice and different strain combinations will be useful in determining the contribution that such genes might play.
The generation of variants in P90A-infected mice is strongly dependent on the presence of helper virus (14), indicating that they almost certainly arise during viral replication. Consistent with this idea, features known to affect the fidelity of reverse transcription can be found at the sites of many of the mutations. For example, all of the single base insertions (4/4) and single base deletions (1/1) occurred at sites where the same nucleotide was repeated four to six times. Regions similar to these have been implicated as hotspots for mutation in studies that investigated retroviral mutation (5, 16). A second process involving template switching and recombination (15, 17) likely generated the larger deletions observed in many of the variants. Regions of sequence homology, even as small as three nucleotides (15), can influence this process, and in one variant (6-2) the deletion recovered occurred across a stretch of eight identical nucleotides.
Our earlier work, conducted prior to readily adaptable PCR approaches (13), suggested that most tumors arising in a P90A-infected mouse contained a variant with increased oncogenic potential and led to the hypothesis that this variant was responsible for induction of the tumor. Similar to these data, variants of this type were recovered in the present study. However, the newer results indicate that the earlier interpretation was overly simple. All of the P90A-infected animals contained a diverse population of viruses, some of which did not display enhanced oncogenic potential compared to P90A. In addition, in several instances P90A initiated the tumorigenic process, since these sequences were recovered from the proviral integration that marks the tumor clone. These data may suggest that additional evolution within the P90A population would occur if tumor extracts were passaged in additional mice, a hypothesis that is under investigation. Because most v-onc gene-containing retroviruses were isolated before molecular characterization was possible and passaged multiple times, often through different species, examining the effects of passaging should help us to understand the dynamics of host-virus interactions that helped to shape viruses like Ab-MLV.
Analyses of the infected tissues revealed that the P90A virus was retained in 4/5 wild-type tumors and 3/5 null tumors. In addition, P90A was clonally integrated in three of five wild-type mice, demonstrating that this virus played an important role in initiating the tumorigenic process. These data are consistent with the ability of P90A to transform pre-B cells in vitro (24) and with the observation that a small number of tumors arise in animals infected with P90A in the absence of helper virus (14). While the possibility exists that these viruses have acquired mutations outside the region sequenced, these results predict that viruses that have incorporated cellular sequences must have the ability to stimulate cell growth in order to emerge as v-onc gene-containing retroviruses. Consistent with this idea, P120/D484N, a kinase-inactive strain that cannot stimulate cell growth, fails to establish an infection in mice (our unpublished data). The need to incorporate cellular sequences in a way that allows the product to stimulate growth, albeit poorly, places additional constraints on the generation of v-onc gene-containing retroviruses. Thus, for viruses such as Ab-MLV to have arisen originally, the recombination events that occurred initially needed to encode a protein that was capable of stimulating cell replication. Additional studies using the P90A model should help to elucidate the mechanisms that have shaped the v-onc gene-containing retroviruses that are studied today.
ACKNOWLEDGMENTS
We are grateful to Erika de Leon Vazquez, Jyoti Mathur, Todd Ashworth, Warangkhana Songsungthang, Rebekah Stackpole, and Laila Sheikh for assistance with analyses of variant viruses, to Chris Schmid for assistance with statistical analysis, and to John Coffin and Caleb Lee for helpful discussions.
This work was supported by grant CA 24220 from the National Cancer Institute.
REFERENCES
Anderson, J. A., R. J. Teufel II, P. D. Yin, and W. S. Hu. 1998. Correlated template-switching events during minus-strand DNA synthesis: a mechanism for high negative interference during retroviral recombination. J. Virol. 72:1186-1194.
Baxter, E. W., K. Blyth, E. R. Cameron, and J. C. Neil. 2001. Selection for loss of p53 function in T-cell lymphomagenesis is alleviated by Moloney murine leukemia virus infection in myc transgenic mice. J. Virol. 75:9790-9798.
Baxter, E. W., K. Blyth, L. A. Donehower, E. R. Cameron, D. E. Onions, and J. C. Neil. 1996. Moloney murine leukemia virus-induced lymphomas in p53-deficient mice: overlapping pathways in tumor development? J. Virol. 70:2095-2100.
Blyth, K., A. Terry, N. Mackay, F. Vaillant, M. Bell, E. R. Cameron, J. C. Neil, and M. Stewart. 2001. Runx2: a novel oncogenic effector revealed by in vivo complementation and retroviral tagging. Oncogene 20:295-302.
Burns, D. P., and H. M. Temin. 1994. High rates of frameshift mutations within homo-oligomeric runs during a single cycle of retroviral replication. J. Virol. 68:4196-4203.
Chesebro, B., K. Wehrly, M. Cloyd, W. Britt, J. Portis, J. Collins, and J. Nishio. 1981. Characterization of mouse monoclonal antibodies specific for Friend murine leukemia virus-induced erythroleukemia cells: Friend-specific and FMR-specific antigens. Virology 112:131-144.
Donehower, L. A., M. Harvey, B. L. Slagle, M. J. McArthur, C. A. Montgomery, Jr., J. S. Butel, and A. Bradley. 1992. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature (London) 356:215-221.
Fan, H., and M. Paskind. 1974. Measurement of the sequence complexity of cloned Moloney murine leukemia virus 60 to 70S RNA: evidence for a haploid genome. J. Virol. 14:421-429.
Green, P. L., D. A. Kaehler, and R. Risser. 1987. Cell transformation and tumor induction by Abelson murine leukemia virus in the absence of helper virus. Proc. Natl. Acad. Sci. USA 84:5932-5936.
Hu, W. S., and H. M. Temin. 1990. Genetic consequences of packaging two RNA genomes in one retroviral particle: pseudodiploidy and high rate of genetic recombination. Proc. Natl. Acad. Sci. USA 87:1556-1560.
Jiang, X., Z. Hanna, M. Kaouass, L. Girard, and P. Jolicoeur. 2002. Ahi-1, a novel gene encoding a modular protein with WD40-repeat and SH3 domains, is targeted by the Ahi-1 and Mis-2 provirus integrations. J. Virol. 76:9046-9059.
Lund, A. H., G. Turner, A. Trubetskoy, E. Verhoeven, E. Wientjens, D. Hulsman, R. Russell, R. A. DePinho, J. Lenz, and M. van Lohuizen. 2002. Genome-wide retroviral insertional tagging of genes involved in cancer in Cdkn2a-deficient mice. Nat. Genet. 32:160-165.
Murtagh, K., G. Skladany, J. Hoag, and N. Rosenberg. 1986. Abelson murine leukemia virus variants with increased oncogenic potential. J. Virol. 60:599-606.
Parmar, K., R. C. Huebner, and N. Rosenberg. 1991. Carboxyl-terminal determinants of Abelson protein important for lymphoma induction. J. Virol. 65:6478-6485.
Pathak, V. K., and H. M. Temin. 1990. Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: deletions and deletions with insertions. Proc. Natl. Acad. Sci. USA 87:6024-6028.
Pathak, V. K., and H. M. Temin. 1990. Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: substitutions, frameshifts, and hypermutations. Proc. Natl. Acad. Sci. USA 87:6019-6023.
Pulsinelli, G. A., and H. M. Temin. 1991. Characterization of large deletions occurring during a single round of retrovirus vector replication: novel deletion mechanism involving errors in strand transfer. J. Virol. 65:4786-4797.
Radfar, A., I. Unnikrishnan, H.-W. Lee, R. A. DePinho, and N. Rosenberg. 1998. p19Arf induces p53-dependent apoptosis during Abelson virus-mediated pre-B cell transformation. Proc. Natl. Acad. Sci. USA 95:13194-13199.
Risser, R., D. Kaehler, and W. W. Lamph. 1985. Different genes control the susceptibility of mice to Moloney or Abelson murine leukemia viruses. J. Virol. 55:547-553.
Risser, R., M. Potter, and W. P. Rowe. 1978. Abelson virus-induced lymphomagenesis in mice. J. Exp. Med. 148:714-726.
Rosenberg, N., and P. Jolicoeur. 1997. Retrovirus pathogenesis, p. 475-586. In J. M. Coffin, S. E. Hughes, and H. E. Varmus (ed.), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Rosenberg, N., and O. N. Witte. 1980. Abelson murine leukemia virus mutants with alterations in the virus-specific P120 molecule. J. Virol. 33:340-348.
Rosenberg, N., and O. N. Witte. 1988. The viral and cellular forms of the Abelson (abl) oncogene. Adv. Virus Res. 35:39-81.
Rosenberg, N. E., D. R. Clark, and O. N. Witte. 1980. Abelson murine leukemia virus mutants deficient in kinase activity and lymphoid cell transformation. J. Virol. 36:766-774.
Sachs, Z., N. Sharpless, R. DePinho, and N. Rosenberg. p16Ink4a stimulates apoptosis during Abelson virus-mediated pre-B cell transformation. Submitted for publication.
Scher, C. D., and R. Siegler. 1975. Direct transformation of 3T3 cells by Abelson murine leukemia virus. Nature (London) 253:729-731.
Sherr, C. J., and F. McCormick. 2002. The RB and p53 pathways in cancer. Cancer Cell 2:103-112.
Shinto, Y., M. Morimoto, M. Katsumata, A. Uchida, K. Aozasa, M. Okamoto, T. Kurosawa, T. Ochi, M. I. Greene, and Y. Tsujimoto. 1995. Moloney murine leukemia virus infection accelerates lymphomagenesis in E mu-bcl-2 transgenic mice. Oncogene 11:1729-1736.
Shoemaker, C., J. Hoffman, S. P. Goff, and D. Baltimore. 1981. Intramolecular integration within Moloney murine leukemia virus DNA. J. Virol. 40:164-172.
Siegler, R., and S. Zajdel. 1972. Pathogenesis of Abelson-virus-induced murine leukemia. J. Natl. Cancer Inst. 48:189-218.
Stoye, J. P., W. N. Frankel, and J. M. Coffin. 1991. DNA hybridization in dried gels with fragmented probes: an improvement over blotting techniques. Technique 3:123-128.
Temin, H. M. 1993. Retrovirus variation and reverse transcription: abnormal strand transfers result in retrovirus genetic variation. Proc. Natl. Acad. Sci. USA 90:6900-6903.
Unnikrishnan, I., A. Radfar, J. Jenab-Wolcott, and N. Rosenberg. 1999. p53 mediates apoptotic crisis in primary Ab-MLV-transformed pre-B cells. Mol. Cell. Biol. 19:4825-4831.
van Lohuizen, M., S. Verbeek, P. Krimpenfort, J. Domen, C. Saris, T. Radaszkiewicz, and A. Berns. 1989. Predisposition to lymphomagenesis in pim-1 transgenic mice: cooperation with c-myc and N-myc in murine leukemia virus-induced tumors. Cell 56:673-682.
Vennstrom, B., C. Raynoscheck, L. Jansson, G. Doederlein, V. Lhotak, A. Johnsson, and H. Beug. 1994. Retroviral capture of c-erbB proto-oncogene sequences: rapid evolution of distinct viral genomes carrying mutant v-erbB genes with different transforming capacities. Oncogene 9:1307-1320.
Wang, L. C., Y.-Y. Chen, and N. Rosenberg. 1995. Pre-B cells transformed by ts Abelson virus rearrange k and l light chain genes in early G1. Curr. Top. Microbiol. Immunol. 194:355-361.
Zou, X., F. Cong, M. Coutts, G. Cattoretti, S. P. Goff, and K. Calame. 2000. p53 deficiency increases transformation by v-Abl and rescues the ability of a C-terminally truncated v-Abl mutant to induce pre-B lymphoma in vivo. Mol. Cell. Biol. 20:628-633.(Erica Marchlik, Richard K)