INTRODUCTION

The regulation of eukaryotic cell division, in response to external signals, requires numerous protein kinases. Cyclin-dependent kinases are a family of protein serine-threonine (Ser/Thr) kinases, and their activities play an essential role in normal cell cycle progression and neoplastic transformation (1, 2). Conversely, activation of chk-1-encoded protein Ser/Thr kinase is associated with cell arrest when the cells are exposed to DNA damage agents (3).

The polo (derived from Drosophila polo gene) family protein Ser/Thr kinases (such as mouse fnk and snk, human and mouse plk, Drosophila polo, and yeast cdc5) have been implicated in cell division (4, 5, 6, 7, 8, 9, 10, 11), although their precise mode of action remains unclear. snk encodes a serum-inducible enzyme involved in the early mitogenic response, and its activity is transcriptionally and post-transcriptionally regulated (4). The polo gene encodes a protein Ser/Thr kinase, mutations that result in abnormal mitotic and meiotic division in the fruit fly (5). A polo homolog encoded by cdc5 in budding yeast is required for nuclear division late in the mitotic cycle (6). Recently, several groups have cloned (7, 8, 9) both human and murine protein kinases that share a strong homology to Drosophila polo kinase and thus are termed plk (polo-like kinase). Microinjection of in vitro transcribed plk sense mRNA into serum-starved NIH3T3 cells induces DNA synthesis as evidenced by the uptake of [³H]thymidine (7). It has been also shown that plk binds to components of the mitotic spindle during mitosis (10). Donahue et al. (11) have recently reported cloning of a mouse Ser/Thr kinase that is activable by fibroblast growth factor and some other mitogens and thus is named fnk (fibroblast growth factor-inducible kinase). The deduced fnk amino acid sequence shares 49, 36, 33, and 32% overall identity to snk, plk, polo, and cdc5, respectively. Induction of fnk mRNA expression is first detectable in about 0.5 h after the addition of mitogens, and activation of fnk expression is a protein synthesis-independent process (11). These authors suggest that Fnk may be a transiently expressed protein kinase participating in the early events for growth factor-stimulated cell cycle progression.

In this report we describe the cloning and characterization of a cDNA coding for a putative protein Ser/Thr kinase termed prk (proliferation-related kinase) from a human megakaryocytic cell line Dami through a combination of PCR¹ and 5-RACE procedures. Deduced prk amino acid sequence shares a strong overall homology to mouse Fnk. It is also homologous to other polo family kinases, especially in the kinase domain. prk mRNA is expressed in limited human primary tissues and established cell lines. Its mRNA expression is activated rapidly by serum or cytokines in serum/cytokine-deprived cells with the peak induction around 2 h after treatment. Southern blotting analysis indicates that prk is a single copy gene in the genome. In addition, prk expression appears to be down-regulated in primary lung tumor samples.

EXPERIMENTAL PROCEDURES

Enzymes and chemicals were purchased from suppliers as follows. Restriction enzymes, 5-RACE kit, T4 DNA ligase, DNA labeling kit, various culture media (RPMI 1640, Iscove's modified Dulbecco's medium, minimal essential medium, Dulbecco's modified Eagle's medium, and McCoy's 5A medium) and antibiotics (penicillin/streptomycin) were purchased from Life Technologies, Inc. K562, HEL, HL-60, Dami, A549, Molt4, and Raji cell lines were purchased from the American Type Cell Collection. The GMOO637D cell line was kindly provided by Dr. Sohan Gupta at Hipple Cancer Research Center (Dayton, OH). SAM-1 and Ba/F3 cell lines were obtained from Dr. T. Kamesaki (Georgetown University), and Dr. A. D'Andrea (Children Hospital at Boston), respectively. The GeneClean system was from Bio 101 Inc. (Vista, CA). Multiple tissue Northern blots, high molecular weight genomic DNA, -actin cDNA, and a gt11 placental cDNA library were purchased from Clontech (Palo Alto, CA). The pT7Blue plasmid was obtained from Novagen Inc. (Madison, WI). [-³²P]dCTP (800 Ci/mmol) was from DuPont NEN. Recombinant interleukin-3 (IL-3), and transforming growth factor- (TGF-) were purchased from the R & D system (Minneapolis). Recombinant thrombopoietin was kindly provided by Genentech Inc. (South San Francisco). Fetal bovine serum (FBS) was purchased from HyClone Laboratories (Logan, UT).

Various cell lines were grown normally in media supplemented with 10% FBS and antibiotics (100 µg/ml penicillin, 50 µg/ml streptomycin). K562, HEL, SAM-1, Dami, HIMeg-1, and HL-60 cell lines were cultured in RPMI 1640 medium; Ba/F3 cells in RPMI 1640 medium supplemented with recombinant murine IL-3 (5 ng/ml). Molt4 and Raji were using McCoy's 5A medium. A549 and GMOO637D cell lines were cultured in Dulbecco's modified Eagle's medium. In some experiments, MO7e, Dami, GMOO637D, and K562 cells were starved for 18 h by culturing cells in the medium supplemented with no FBS, and the serum-deprived cells were refed with the serum for various lengths of time and collected for RNA isolation. MO7e cells were supplemented with 5 ng/ml human IL-3. To determine prk induction by cytokines, MO7e cells were first cultured in the medium containing 10% FBS and 0.02 ng/ml human IL-3 for 18 h and then supplemented with IL-3 (10 or 100 ng/ml), thrombopoietin (500 ng/ml), and TGF- (100 ng/ml), respectively. These cells were cultured for an additional 2 h before harvesting for RNA isolation.

Total RNA was isolated from various cell lines using a kit obtained form Tel-Test Inc. (Friendwood, TX) according to the protocol provided by the supplier. Reverse transcriptase-mediated PCR was performed on Dami total RNA using a SuperScript^TM Preamplification system purchased from Life Technologies, Inc. A pair of oligonucleotide primers, corresponding to two stretches of conserved domains as described (12), was synthesized by Oligo Etc. Inc. The pair of primers has the following sequences: upstream, 5-aagatc/tgga/tagagga/c/ta/gct/c/g-3 with 72-fold degeneracy; downstream, 5-agaact/c/gtcagga/ggc/gt/g/ccg/tg/ata-3 with 144-fold degeneracy. The PCR products were analyzed on agarose gels, and DNA bands were excised and eluted from agarose gels using the GeneClean system. The eluted PCR products were cloned directly into pT7Blue vector for sequencing analysis.

The placental cDNA library was plated on Escherichia coli Y1090 (about 1 × 10⁴ plaque-forming units/150-mm plate), and the plaques were lifted onto nitrocellulose filters that were processed and screened by hybridization with a prk cDNA fragment (obtained by PCR) that was labeled with [³²P]dCTP. The putative positive clones were rescreened until they were plaque-purified. The phage DNA was purified using the plate lysate method (13). The purified DNA was digested with EcoRI and then analyzed on 1% agarose gels. Southern blotting was performed to confirm the positivity of the released insert. The cDNA insert was then cloned into the pBluescript II SK⁺ plasmid vector for restriction mapping and DNA sequencing analysis. DNA sequencing was performed at the University of Cincinnati's Core Facility.

The 5-RACE was carried out using a kit from Life Technologies, Inc. according to the protocol provided by the supplier. Two prk-specific primers (antisense) were synthesized and named prk-p1 (5-agatgttgtcagcgtcctca-3) and prk-p2 (5-tctgcgggatgactttgaca-3), respectively. The first strand cDNA was synthesized from Dami cell total RNA using the primer prk-p2. The synthesized cDNA was used for TdT-tailing reaction and then for PCR amplification using an anchor primer provided in the kit and prk-p2. The specificity of PCR amplification was confirmed by Southern blotting analysis of the amplified product using the prk probe. The amplified product with a size range of 200-500 base pairs was eluted from agarose gels and cloned into the pT7Blue vector for sequencing analysis.

Approximately equal amounts of total RNA (15 µg) were fractionated on 1% formaldehyde agarose gels, and the fractionated RNA was transferred to Nytran Plus membranes. The RNA blots were baked for 2 h, prehybridized for 2 h, and probed with a ³²P-labeled prk or -actin cDNA fragment overnight. After hybridization, the blots were washed and autoradiographed. High stringency hybridization and washing conditions were as described (14).

High molecular weight genomic DNA (9 µg/digestion) was digested with a panel of restriction enzymes (10 units/µg of DNA) to completion using appropriate buffers. The digested DNA was fractionated on agarose gels (0.8%), and the fractionated DNA was transferred to Nytran Plus membranes. The DNA blots were baked for 2 h, prehybridized for 2 h, and probed overnight with a full-length prk cDNA labeled with ³²P under either high stringency (50% formamide) or low stringency (40% formamide) conditions as described (14). After hybridization, high stringency-hybridized membranes were washed with 2 × SSC for 20 min at room temperature and with 0.1 × SSC, 0.1% sodium dodecyl sulfate for 1 h at 60 °C. Low stringency-hybridized membranes were washed with 2 × SSC for 20 min at room temperature and then with 2 × SSC, 0.1% sodium dodecyl sulfate for 1 h at 60 °C. Washed membranes were autoradiographed at 70 °C for 3 days.

RESULTS

One salient feature of megakaryocyte terminal differentiation and maturation is continued DNA synthesis uncoupled from cytokinesis. Thus, the nucleus becomes polyploid after a series of endomitoses. We have been interested in the molecular mechanism regulating differentiation of megakaryocytes (15, 16, 17). Since cells undergoing DNA synthesis and division recruit sets of regulatory molecules such as cyclin-dependent kinases and cyclins (1, 2), we have proposed that megakaryocyte precursors may express some unique sets of cyclin-dependent kinases or other protein kinases necessary for the endomitosis. To identify protein kinases that may play a role in megakaryocytic cell differentiation and maturation, we performed reverse transcriptase-mediated PCR to amplify total RNA isolated from the megakaryocytic cell line Dami using degenerate oligodeoxynucleotides corresponding to conserved domains in cyclin-dependent kinases (12). Amplified products with a molecular mass around 400-600 base pairs were subcloned into pT7Blue vector and sequenced by the dideoxy termination method. Out of 35 clones whose DNA inserts were sequenced, a clone with a novel sequence was identified and subsequently named prk. Since placenta is known to express many cellular genes, a gt11 cDNA library made from human placental mRNA was screened for a full-length cDNA clone corresponding to the identified novel Ser/Thr kinase fragment. More than 30 independent positive clones were identified and plaque purified. Restriction digestion analysis of purified gt11 DNA revealed that there were four different sizes of overlapping cDNA inserts. The longest insert was about 2.0 kilobase pairs which was fully mapped and sequenced. The nucleotide sequence contained a single long open reading frame. However, the initiating codon ATG for translation was not present. Partial DNA sequencing analysis of other cDNA inserts isolated during library screening indicated that they did not contain additional 5-coding sequences.

We then employed the 5-RACE method using a kit purchased from Life Technologies, Inc. Amplified cDNA fragments were analyzed by agarose gel electrophoresis, eluted from agarose, and cloned into pT7Blue plasmid vector. Plasmid DNA isolated from selected transformants was sequenced, revealing that an insert extended the 5-end by about 200 base pairs. Deduced amino acid analysis revealed an in-frame ATG codon coding for a putative translation initiation amino acid methionine. The complete nucleotide sequence as well as the deduced amino acid sequence are shown in Fig. 1. The long open reading frame encodes a protein of 607 amino acids. The ATP binding site (GKGGFARC) in the kinase domain is shown in the box (Fig. 1), and it differs slightly from the consensus site (GXGXXG(S/A)V) as described (18). There is a putative nuclear targeting signal (highlighted with a broken underline in Fig. 1) in the middle of the predicted amino acid sequence. Three copies of the ATTTA sequence element (Fig. 1, underlined), a feature for short half-life mRNA (19), were present in the 3-untranslated regions. In addition, there exists one copy of a consensus polyadenylation signal (AATAAA, double underlined) nine nucleotides 5 to the poly(A) tail.

Recently, Donohue et al. (11) reported the cloning and characterization of an immediate-early gene encoding the putative murine Ser/Thr kinase fnk, which belongs to the polo family of protein kinases. A search of sequence data banks revealed that prk shares the strongest homology with murine fnk. An alignment of the deduced Prk amino acid sequence with Fnk and the rest of polo family kinases is shown in Fig. 2. It appears that there are two domains in the polo family kinases: namely, the amino-terminal kinase domain (between arrows) and the carboxyl-terminal regulatory (presumably) domain. Amino acid homology between Prk and mouse Fnk shows a 91% residue identity. However, Prk differs from Fnk in the amino terminus and the joint region between two putative domains. Prk is 40 amino acids shorter than Fnk at the amino terminus. The amino-terminal 9 amino acid residues of Prk share little homology with Fnk. In addition, Prk has a 17-amino acid long insert in the domain hinge region, and this insert is absent from Fnk. Prk is homologous to other polo family members such as mouse Snk (50% residue identity), human Plk (50% residue identity), Drosophila polo (48% residue identity), and Saccharomyces cdc5 (38% residue identity). In the presumed regulatory domain, there exists a very conserved stretch of sequence KWVDYS (Fig. 1, shaded region, and Fig. 2) in polo family kinases.

To shed light on the potential biological role of prk, we first examined the pattern of prk gene expression in primary human tissues. Northern blots containing poly(A)⁺ RNA (2 µg/lane) isolated from a variety of human tissues were probed with a ³²P-labeled cDNA fragment under high stringency conditions as described (14). Fig. 3A shows that there is a single species of prk-specific mRNA which is about 2.5 kilobase pairs (lane 3). prk expression was very tissue-restricted. Among 16 human organs/tissues examined, a moderate level of prk mRNA was detected only in the placenta (Fig. 3A, lane 11). A low level of prk expression was also present in ovary (lane 5), peripheral blood leukocyte (lane 8), and lung (lane 12) samples. Little or no prk mRNA was present in spleen, thymus, prostate, testis, small intestine, colon, heart, brain, liver, skeletal muscle, kidney, and pancreas. As a loading control, the same blot was stripped and reprobed with ³²P-labeled -actin probe, and the result is shown in Fig. 3B. We also examined prk expression in a variety of cell lines. The summarized results are shown in Table I. Low, but detectable, levels of prk expression were present in Dami, MO7e, and three glioma cell lines. Little or no prk expression was found in other cell lines that were examined.

Table I. Summary of prk expression in various cell lines Plus (+) symbol indicates the degree of hybridization signal on Northern blots. Minus () symbol indicates no hybridization signals detected. The amount of RNA used for analysis and integrity of the RNA were confirmed by ethidium bromide staining of the gels. Cell line Type Expression by Northern blotting K562 Erythrocytic   HEL Erythrocytic   Dami Megakaryocytic + MO7e Megakaryoblastic + SAM-1 Megakaryoblastic   HIMeg-1 Myeloid progenitor   HL-60 Promyelocytic   Molt4 Lymphocytic   Raji Lymphocytic   GMOO637D Lung fibroblastic   A549 Lung fibroblastic   Ba/F3 Pre-B cell   U373 Brain glioma + U118 Brain glioma + U138 Brain glioma +

Our initial objective was to identify novel protein kinases that might be involved in regulating the cell cycle of megakaryocytic precursors. Since Prk shares sequence homology to Fnk and other polo family protein kinases, which all appear to be involved in various stages of the cell cycle regulation (4, 5, 6, 7, 8, 9, 10, 11), we next examined the inducibility of prk expression by mitogens such as FBS. MO7e, an IL-3-dependent megakaryoblastic cell line, was deprived of FBS for 18 h and then refed with the serum. Northern blotting analysis revealed (Fig. 4A) that little or no prk mRNA expression was detected in serum-deprived cells (lane 1). prk mRNA expression was induced about 30 min after serum addition (lane 2). Its expression peaked at about 2 h after serum addition (lane 4) and then decreased rapidly. To examine whether the prk induction requires new protein synthesis, the protein synthesis inhibitor cycloheximide (50 µg/ml) was added at the same time as when the serum was supplemented to the cultures. It was shown that cycloheximide did not inhibit but rather superinduced prk mRNA expression (lane 8). Cycloheximide alone had no effect on prk mRNA induction (lane 9).

To determine whether or not the activation of prk expression by the serum is unique to MO7e, we next examined prk expression and its inducibility by the serum in Dami (megakaryocytic), K562 (erythroleukemic), and GMOO637D (lung fibroblastic) cell lines. These cell lines were first deprived of FBS for 18 h and then refed with the serum for various lengths of time. Total RNA isolated from the treated cells was analyzed for prk expression by Northern blotting. It was observed that there were low levels of prk mRNA in serum-deprived Dami cells (Fig. 5A, lane 1), which were induced by the serum with the peak induction around 2 h after serum addition (lane 4). No prk mRNA expression was detected in serum-deprived GMOO637D (lane 6) or K562 (lane 10) cells, and serum addition activated prk mRNA expression in both of these cell lines (lanes 7-9 and lanes 12 and 13). Serum also induced the steady-state level of prk mRNA in U118 glioma cell line (data not shown). Fig. 5B shows the rRNA bands as loading controls.

FBS contains many growth factors and cytokines. To determine whether defined cytokines were capable of activating prk expression, we deprived of MO7e cells of IL-3 for 18 h and then supplemented the cells with IL-3, thrombopoietin, or TGF- for 2 h. Northern blotting analysis showed that there were no detectable levels of prk mRNA in MO7e cells deprived of IL-3 (Fig. 6A, lane 1). Addition of IL-3 (lanes 2 and 3) but not of TGF (lane 5) activated a low level of prk mRNA expression. In addition, thrombopoietin activated a moderate level of prk mRNA expression (lane 4). Interestingly, a prk-specific transcript with a substantially large size (about 5 kilobase pairs) was also activated by thrombopoietin (arrow prk-hi). Fig. 6B shows rRNA bands as loading controls.

Aberrant expression of protein kinases has been correlated with neoplastic transformation (20, 21). To determine whether or not deregulated prk expression was correlated with tumorigenesis, RNA samples isolated from the tumors as well as the uninvolved tissues of 18 lung cancer patients with either adenocarcinoma or squamous cell carcinoma were analyzed for prk expression via Northern blotting (examples in Fig. 7). It was shown that although the steady-state level of prk mRNA varied greatly among the individual patients, the tumor (Fig. 7A, odd lanes) consistently expressed less (or no detectable level) prk mRNA compared with the uninvolved ``normal'' lung tissues from the same patients (even lanes). Specimens of 13 of these patients were available for histopathologic study. Significant tissue heterogeneity was present in all samples, but tumor cells were found to make up the majority of 12 of the 13 specimens (data not shown). The remaining tumor (patient 872) in which only a small number of malignant cells were present showed no significant decrease in prk expression by Northern blot (lanes 7 and 8). Fig. 7B shows rRNA bands as loading controls.

To determine whether there were genes whose structures were closely related to prk in the genome, we analyzed genomic DNA for the complexity of its restriction fragments hybridizable to prk cDNA. High stringency Southern blotting (Fig. 8) showed that KpnI, BamHI, SstI, and SalI digestions all yielded two major bands, and HindIII digestion gave a single band. As the prk cDNA used for the probe contains one site for KpnI and one site for SstI, these two enzyme digestion patterns are consistent with our predictions. BamHI and SalI have no site in prk cDNA used for the probe but yielded two bands, suggesting the presence of an intron(s) with these restriction sites. Low stringency hybridization was also performed on a duplicate Southern blot, and no difference in the hybridizing band pattern (except for the intensity) was observed (data not shown).

The results presented in Fig. 7 demonstrate that prk expression was down-regulated in lung carcinoma samples. We wondered whether there were some gross abnormalities (e.g. deletion or DNA rearrangement) in the prk genomic locus. Since several pairs of DNA samples from the same lung cancer patients whose RNA had been analyzed for prk expression were available, we also performed high stringency Southern blotting analysis on those DNA samples after restriction enzyme digestions. No apparent difference in the band patterns and their sizes was observed between tumor DNA samples and those of uninvolved normal samples (data not shown).

DISCUSSION

In the present study, we have reported the cloning of a full-length cDNA coding for a putative protein Ser/Thr kinase that we have named Prk. Analysis of the deduced amino acid sequence suggests that Prk may have a kinase domain and a regulatory domain. Except for the amino terminus and the putative hinge region of two domains, Prk is highly homologous to murine Fnk. Prk has a 9-amino acid long amino-terminal sequence that shares little homology with Fnk (Fig. 2). In addition, in the hinge region of the two domains Prk has a 17-amino acid long insert that is absent from Fnk. The kinase domain for the polo family proteins is very conserved across wide spectrum of species (Fig. 2). It is interesting to note that the putative ATP binding domain (GKGGFARC, Fig. 1) for Prk and other polo family kinases is different from the consensus (GXGXXG(S/A)XV; Ref. 18) derived from other protein kinases, and the valine residue in the consensus is repalced with cysteine in the polo family kinases. Considering the rather distinct and conserved kinase structure of the polo family kinases, it is reasonable to predict that these proteins are essential for cellular survival or functions and that there may be a rather strict substrate specificity. Identification of the cellular substrate(s) for these enzymes will certainly open a new avenue of research for elucidation of their biological role. The regulatory domain is less conserved among polo family members. It is likely that different members may have similar functions, but their expression may be developmentally controlled. On the other hand, it is equally possible that each polo member may have rather distinct functions, for example, being involved in different stages of the cell cycle progression since the cdc5 product in budding yeast is required for nuclear division late in the mitotic cycle (22), whereas fnk is implicated in regulating the early signaling events required for cell cycle progression stimulated by growth factors (11).

We have shown that prk mRNA expression is tissue-restricted. Human placenta and, to a lesser degree, ovaries, peripheral blood leukocytes, and lung exhibit measurable prk expression. In cultured cell lines, prk expression is generally low or unmeasurable but can be activated rapidly by mitogens such as serum and cytokines (Figs. 4, 5, 6). Considering that other polo family kinases are required or implicated in regulating cell cycle progression (4, 5, 6, 7, 8, 9, 10, 11, 22, 23), it is reasonable to speculate that Prk may be also involved in cell cycle regulation. The mammalian polo homologs such as murine snk and fnk belong to the immediate-early genes activated by a variety of mitogens (4, 11). Like snk and fnk, prk expression is a protein synthesis-independent process, and its activation is transient and correlated with cell proliferation. The short half-life of prk mRNA is evident in the rapid decrease of its steady-state level. This is consistent with several structural features (AUUUAs), characteristic of RNAs with a very short half-life (19), in the 3-untranslated region (Fig. 1). Since most protein kinases are regulated at the post-transcription level via reversible phosphorylation or by second messengers (24), polo family kinases represent a new group of kinases that are regulated at least partially at the transcription level. The kinase activity of polo family proteins may therefore need to be controlled tightly at the transcriptional level, and when expressed inappropriately, the activity may be disadvantageous to the cell survival or integrity.

Megakaryocyte maturation is characterized by polyploidization and increase in cell size and cytoplasmic mass. It has been long speculated that there are unique mechanisms that regulate the endoduplication of DNA in megakaryocytes. Wilhide et al. (25) have shown recently that overexpression of cyclin D1 in Dami cells causes growth arrest. This led these authors to suggest that cyclin D1 may participate in megakaryocyte differentiation by promoting endomitosis and/or inhibiting cell division (25). However, overexpression of cyclin D1 does not cause an increase in DNA polyploidy. We have observed that there is a higher basal level of prk mRNA in megakaryocytic cell lines Dami and MO7e than in other cell lineages. In addition, we have shown that thrombopoietin induces relatively high levels of prk gene expression in the MO7e cell line. Considering that thrombopoietin is a cytokine regulating megakaryocytopoiesis, it is conceivable that prk may play a role in proliferation and differentiation of megakaryocyte progenitors.

Many protein kinases are proto-oncogenes, and their activation or elevated expression is often correlated with cell cycle progression or cell proliferation (20). Thus, deregulated expression of these kinases often leads to oncogenic transformation (21). Although prk expression is correlated with cell proliferation, it is interesting to note that its mRNA expression is consistently down-regulated in more than a dozen lung tumor samples when compared with the uninvolved tissues from the same patients (Fig. 7). The subsequent Southern blotting studies revealed no gross abnormalities in tumor DNA samples compared with the uninvolved tissue DNA samples (data not shown) from the same lung cancer patients. This suggests that down-regulation of prk mRNA expression in primary lung tumors is likely due to alterations in a mechanism other than a major DNA deletion or chromosomal rearrangement. It is possible that prk is a ``safeguard'' gene for DNA replication or cell cycle progression. Diminished expression or activity of prk product may compromise the cell's ability to divide correctly or to pass correct genetic information to daughter cells. However, a caveat to the interpretation of these results is the inevitable tissue heterogeneity of the specimens analyzed. Although histopathologic analysis was able to confirm that tumor cells comprised the majority of all but one of the specimens examined, analysis of prk expression on a cell by cell basis is needed to assign down-regulation to the tumor cells definitely. Furthermore, the use of uninvolved lung from the same patient as a frame of reference entails potential artifacts. Normal lung is comprised a variety of differentiated pulmonary, vascular, and stromal elements which may not be directly comparable to the cell of origin of these specific tumors. Interestingly, it has been observed that other immediate-early genes are expressed at lower levels in tumor than normal tissue in lung cancer and other tumors (26).

Southern blotting analysis of human normal genomic DNA digested with a panel of restriction enzymes has revealed rather simple band patterns hybridizable to prk cDNA under either high or low stringency conditions (Fig. 8), suggesting the presence of a single copy of prk gene in human genome and the absence of genes whose structures are closely related to prk at the nucleotide level.