The primary structure of the putative oncogene pim-1 shows extensive homology with protein kinases

Cell,

Vol. 46,

603-611,

August

15, 1966, Copyright

0 1966 by Cell Press

The Primary Structure of the Putative Oncogene pim-1 Shows Extensive Homology with Protein Kinases Gerard Selten,‘§ H. Theo Cuypers:S Wilbert Els Robanus-Maandag,’ Joseph Verbeek,S Jos Domen,* Charles van Beveren,t and Anton Berns’* * Department of Biochemistry University of Nijmegen Geert Grootepiein N 21 6525 EZ Nijmegen, The Netherlands tThe Salk Institute I? 0. Box 85800 San Diego, California 92138 $Section of Molecular Genetics, The Netherlands Cancer Institute; and Department of Biochemistry, University of Amsterdam Plesmanlaan 121 1066 CX Amsterdam, The Netherlands

Boeiens,’

Summary We have shown previously that the putative oncogene @n-l is frequently activated by provirus insertion in murine leukemia virus-induced T cell lymphomas. Here we describe the structure of the pim-1 gene as determlned by sequencing genomic and cDNA clones. The gene has an open reading frame, encoding a protein of 313 amino acids, extending over six exons and preceded and followed by stop codons in all readlng frames. Proviruses always integrate outside the protein-encoding domain, showing a high preference for a small region in the Y-terminal exon; integration in the 3’ exon results in relatively high levels of pim-1 mRNA. Computer search reveals homology between @n-l and protein kinases: all the domains characteristic of protein kinases are conserved in the pim-1 amino acid sequence. The highest homologies were observed with the protein-serine klnases. Introduction Oncogenic tra&formation by the nonacute transforming retroviruses probably depends on the activation of cellular genes by integrated proviruses. Activation of the c-myc proto-oncogene by proviral DNA insertion has been described in avian leukosis virus (ALV)- and chicken syncytial virus (CSV)-induced B cell lymphomas in birds (Hayward et al., 1981; Neel et al., 1981; Payne et al., 1981; Westaway et al., 1984; Swift et al., 1985), in murine leukemiavirus (MuLV)-induced T cell lymphomas in mice (Corcoran et al., 1984; Selten et al., 1984; Li et al., 1984; Steffen, 1984; O’Donnell et al., 1985), and in feline leukemia virus-induced lymphoma in cats (Neil et al., 1984). A similar mechanism underlies the activation of SPresent address: The Netherlands.

Gist Brocades

N. V., f? 0. Box 1, 2600 MA Delft,

c-e& in ALV-induced erythroblastosis (Fung et al., 1983; Nilsen et al., 1985), the activation of c-myb in Abelson and Moloney murine leukemia virus-induced plasmacytoid lymphosarcomas (Sheng-Ong et al., 1984), the activation of in&l (Nusse and Varmus, 1982; Nusse et al., 1984) and int-2 (Peters et al., 1983) in mouse mammary tumor virus (MMTV)-induced mammacarcinomas, and the activation of pim-1 in MuLV-induced T cell lymphomas (Cuypers et al., 1984; Selten et al., 1985; O’Donnell et al., 1985). Other common integration regions include M/vi-l, Mlvi-2, Mlvi-3 (Tsichlis et al., 1985), mis-1 (Lemay and Jolicoeur, 1984), m-1 (Webb et al., 1984), and gin-l (Jolicoeur, personal communication). Integrated proviruses can activate cellular genes either by transcription from the viral promoter (Neel et al., 1981; Voronova and Sefton, 1986) or by virtue of the enhancer activity of the proviral long terminal repeat (LTR) (Payne et al., 1982; Nusse et al., 1984; Corcoran et al., 1984; Dickson et al., 1984; Selten et al., 1985). In addition, integrations within the transcription unit might modify the stability or translatability of the mRNA, or give rise to a truncated protein product with altered biological characteristics. Here we describe the structure of the putative oncogene @m-l, which is frequently activated in MuLVinduced T cell lymphomas. We report the nucleotide sequence of the pim-1 gene and the effects of the site of proviral integration on the level of mRNA expression. The deduced amino acid sequence indicates that the pim-1 protein is a member of the class of proteins with protein kinase activity. Results The @m-l Domain A schematic overview of the distribution of integrated proviruses within the pim-1 domain in 36 independent lymphomas is shown in Figure 1 (see also Cuypers et al., 1984; Selten et al., 1985). Proviral integrations are found within a region of 15 kb, showing a strong preference for the 3’ part of the pim-1 transcription unit. From a total of 36 lymphomas with a provirally activated pim-1 gene, 25 contained a provirus within the 3’ region of the transcription unit, in 8 lymphomas the provirus was inserted downstream of pim-1, and in 3 instances the provirus was integrated upstream of the pim-1 transcription unit. Proviral insertion is associated with increased pim-1 mRNA levels. Integration upstream of map position -6 or downstream of position +0.5 (see Figure 1) does not alter the size of the 2.8 kb pim-1 mRNA transcript on Northern blots (Selten et al., 1985). In lymphomas containing a provirus around position -0.5 of the pim-1 map (approximately 70% of the lymphomas), modified transcripts ranging in size from 2.0 to 2.6 kb, covalently linked to viral LTR sequences, were detected (Selten et al., 1985). The nucleotide sequences of the pim-1 genomic DNA and of cDNAs synthesized from pim-1 mFtNAs obtained from three independent MuLV-induced lymphomas were determined.

Cell 604

Figure

1. Proviral

Integration

Sites within the pim-I

Domain

Vertical large arrows indicate the regions in which proviral integrations were detected in the independently induced lymphomas. Numbers beneath the arrows give the number of lymphomas having an integrated provirus within these regions. The localization as well as transcriptional direction of the pim-1 gene is indicated by the horizontal arrow. Restriction sites: R-V, EcoFtV; RI, EcoRI; K, Kpnl; H, Hindlll; SC, Sacl; P, Pstl, and B, BamHI. Map coordinates (in kb) are shown above.

Genomic pim-1 Sequences The 11 kb EcoRl fragment, corresponding to the region between -9.6 and +1.3 on the pim-1 map (Figure l), was molecularly cloned from normal BALB/c DNA. The nucleotide sequence between position -6.6 (Sac1 site) and +1.3 (EcoRI site) was determined from both strands and with overlapping regions across every restriction site. A partial sequence of this region, containing 7993 nucleotides, is shown in Figure 2. pim-1 cDNA Clones To determine the pim-1 gene structure, pim-1 cDNAs were molecularly cloned and sequenced. mRNA from lymphomas with proviruses within the pim-1 transcription unit (tumors 10 and 54) and downstream of the pim-1 gene (lymphoma 9) were used as starting material. Proviral insertion as observed in lymphomas 10 and 54 gives rise to apim-1 mRNA of approximately 2.0 kb. pim-1 transcription in these lymphomas is terminated at the polyadenylation site of the Moloney MuLV LTR (Selten et al., 1985). Integration of proviruses downstream of the pim-1 gene, as in lymphoma 9, does not affect the size of the pim-1 mRNA transcript (Selten et al., 1985). cDNAs were cloned in kgtll using established protocols. The largest oligo(dT)primed pim-1 cDNA clone was obtained from lymphoma 10. The cDNA clone was 1410 bp long, including 508 bp of LTR sequences. This clone reached to position 2711 in the fourth exon. Using a synthetic primer on this mRNA preparation (position 4652 to 4667 in the fifth exon), we obtained two cDNA clones of 918 bp that reached to map position 1719 in the first exon. Oligo(dT)-primed cDNA synthesis using a 2.8 kbpim-1 mRNA preparation as template resulted in a series of relatively small cDNA clones. The largest clone in this series was 550 nucleotides long, reaching to map position 6444. These and several other pim-I cDNA clones were sequenced by the Ml3 dideoxyribonucleotide chain termination method (Sanger et al., 1980), and the sequences were compared with that of pim1 genomic DNA. This revealed both the structure of the pim-1 gene as well as proviral integration sites in lymphomas 10 and 54. Structure of the pim-1 Gene The pim-1 gene contains at least six exons, as indicated in Figure 2. These six exons are 419 (or slightly larger),

107, 51, 367, 177, and (probably) 1499 bp in size. They cover in total 2620 nucleotides of pim-1 mRNA, which was estimated to be 2.8 kb long by Northern blot analysis (Selten et al., 1985). We cannot completely exclude the possibility that the 3’exon of 1499 bp is interrupted around position 6500 (Figure 2) by a small intron of variable size. This is based upon the following observations. Fragments of approximately 1000,440,390, and 380 nucleotides, corresponding to the region between positions 5496 and 7000, were protected from Sl nuclease digestion after hybridization of the 2.8 kb pim-1 mRNA to an 11 kb genomic clone (-9.7 to +1.3). Hybridization with small Ml3 probes showed that the 1000 nucleotide fragment could be assigned to the region between positions 5446 and 6450, whereas the 440, 390, and 380 nucleotide fragments could be located between map positions 6550 and 7000. This was unexpected, since the cDNA clones starting at the second poly(A) addition site at position 6995 were colinear over their complete length with the genomic sequence; they thereby cover the region between map positions 6444 and 6994. This discrepancy might have resulted from instability in the Sl nuclease protection assay of the AT-rich stretches around map position 6500, or may be the result of differential splicing. The cloning of additional cDNAs is required to resolve this issue. At the 5’end the presence of one (small) exon located upstream of exon 1 is not yet completely excluded. Sl nuclease protection assays showed a heterogeneous pattern for exon 1, with protected fragments of sizes 450, 430, and 360 nucleotides. Since an acceptor sequence is found around position 1706 (a position that would correspond to an exon of 431 nucleotides), pim-1 mRNA transcription might start further upstream. However, pSP6 RNAase protection assays did not reveal a distinct exon between positions 0 and 1705. Furthermore, two independent cDNA clones reached to position 1719, suggesting that this position might be an initiation site for RNA transcription. Most likely, transcription can initiate in the region around map coordinate 1720 at several different positions, as has been observed for other genes. Primer extension studies have to be performed to determine the various start positions more accurately. Upstream of position 1720 the GC-rich motif CCGCCC is repeated eight times. In addition, the lymphoid-specific octamer motif ATGCAAAT is present at position 1438. RNA Processing Signals Comparison of the cDNA and genomic DNA nucleotide sequences showed consensus donor and acceptor site sequences (Breathnach and Chambon, 1981; Mount, 1982) for all exon-intron boundaries in the pim-1 gene (see Figure 2). At the 3’ end of the pim-1 gene, two sites with the consensus polyadenylation sequence AATAAA were found at positions 6931 and 6976, respectively. Although there was no evidence for utilization of the site at position 6931, the sequence of six independently derived pim-1 cDNA clones revealed poly(A) tracts starting exclusively at position 6995, 19 nucleotides downstream of the polyadenylation signal at position 6976. Since three A residues are present in the genomic sequence between

%;u.snce

of @n-l:

Homology

with Protein

Kinases

Figure 2. Sequence of thepim-1 Gene and the Predicted @m-l Protein Sequence The “plus” strand of the nucleotide sequence is shown. Numbering starts at the Sac1 restriction site at position -6.6 on the physical map (see Figure 1). Exon sequences are given in capital letters, starting with exon 1. Portions of large (sequenced) introns are replaced by abbreviations (e.g.-O.520 kb-) referring to the exact number of nucleotides in kb. Restriction sites for Sacl, SamHI, and EcoRl aswell as the polyadenylation signal AATAAA and the CG-rich motifs in the promoter region are underlined. The ATGCAAAT octanudeotide motif is double underlined. Six independent cDNA clones end within the AAA triplet at position 6995. Proviral integration sites of MuLV, which were derived by sequencing cloned host-virus junction DNA fragments (tumor 15) or cloned cDNAs (tumors 10 and 54), are marked by arrows. The proposed amino acid sequence for the @m-l protein is given above the nucleotide sequence. Numbering starts at the methionine within exon 1.

positions 6995 and 6997, the exact polyadenylation site could be situated between nucleotides 6995 and 6997. The Predicted @m-l Protein A single long open reading frame (ORF), extending over six exons, was found. In the 5’ region of exon 1, stop codons in all three reading frames are present. A single initiation codon is present in exon 1 at position 2056, downstream of these stop signals. The AUG initiator codon is followed by the ORF, which ends in the large exon at position 5650. The nucleotide sequence within this ORF corresponds to apim-1 polypeptide of 313 amino acid residues and a molecular mass of m35.5 kd. Proviruses Are integrated outside the Protein-Encoding Domain The three proviruses found at the 5’end of the pim-1 gene were located approximately 1 kb upstream of exon 1

(Selten et al., 1985). The integration site in tumor 15 was determined precisely by sequencing the provirus-host DNA junction fragment from a genomic clone. The provirus in this tumor was integrated 900 bp upstream of exon 1 (see Figure 4), and it showed the same transcriptional orientation as pim-1. At the 3’end of the gene, the majority of the proviruses were inserted in a small region within the 3’exon (Selten et al., 1985). To verify that truncation of the pim-1 protein occurred by proviral insertion in this exon, the proviral integration sites in six lymphomas were investigated in more detail (tumors 1, 10, 17, 32, 34, and 54). Proviral positions within the pim-1 gene in these lymphomas were determined by sequencing of genomic (lymphoma 1) and cDNA clones (lymphomas 10 and 54) or by an RNAase protection assay (lymphomas 1732, and 34) in which labeled antisense RNA was protected from RNAase digestion by hybridization topim-1 mRNA (Figure 3). The precise integration sites of the proviruses in these

Cell 606

lymphomas the proviral integration event occurred within the @m-l transcription unit. This suggests strongly that the intact pin+1 protein is required to confer a selective advantage to the tumor cell. Remarkably, the steady-state level of pim-1 mRNA appeared to be dependent on the proviral integration site (Figure 5). Lymphomas bearing a provirus within the 3’ region of the pim-1 transcription unit exhibited significantly higherpim-1 mRNA levels than lymphomas having a proviral insertion upstream or downstream of the pim-1 gene. In the lymphomas with high pim-1 expression levels, the pim-1 mRNA sequences are covalently linked to viral LTR sequences (Selten et al., 1985). We have not yet determined whether these differences in pim-1 mRNA level are caused by a difference in transcription rate or result from an altered mRNA stability, as observed for myc in murine plasmacytomas (Piechaczyk et al., 1985).

TUI VI01? No .

94-c 89--

Figure 3. Determination Protection Assay

of Proviral

Integration

Sites

by an RNAase

pim-1 Shows Homology with Protein Kinases Computer analysis shows that the predicted amino acid sequence of the pim-1 protein is not identical to that corresponding to any of the (onco)genes described to date. However, the pim-1 protein sequence reveals all the domains characteristic of protein kinases. Figure 6 shows a comparison between the predicted amino acid sequence of pim-1 and the amino acid sequences of protein kinases encoded by V-SK (Czernilofsky et al., 1980), v-abl (Reddy et al., 1983) v-mos (Van Beveren et al., 1981; Donoghue, 1982) and tck(Voronova and Sefton, 1988), as well as the y subunit of phosphorylase kinase (Reimann et al., 1984). All protein kinases show a 30 kd catalytic domain (Levinson et al., 1981; Brugge and Darrow, 1984; Weinmaster et al., 1983; Hunter and Cooper, 1985; Van Beveren and Verma, 1985), and this domain is also found within the pim-1 protein (e.g., compare the pim-1 sequence with residues 261-512 in pp60v-s’c, Figure 8). The consensus sequences Gly-X-Gly-X-Phe-Gly-X-Val-X-X-Gly (residues 274-284 for pp60v-“‘c) and Val-Ala-X-Lys (residues 292295 in pp60v-s’c) are completely conserved in the pim-1 protein. In this region the amino acid sequence homology is closest between pim-1 and v-mos. Other highly conserved sequences present in protein kinases as well as in pim-1 are His-ArgAsp-X-X-X-X-Asn-X (residues 384-392) and Lys-X-X-Asp-Phe-Gly (401-406 in pp60v-). The tyrosine at position 416 in p~60~-~” that is a target for auto-

A 32P-labeled antisense RNA fragment corresponding to the region between positions -0.95 and +0.3 on the pimI map (positions 5639 to 6660 in Figure 2) was annealed with poly(A)+ RNA and treated with RNAase as described in Experimental Procedures. The size of the protected fragment is a measure of the distance between the start of the BamHl recognition sequence at position 5639 and the proviral integration site. Proviral integration sites for lymphomas 17, 32, and 34 were determined by this assay. mRNA from lymphomas 10 and 54, for which the proviral integration sites were determined by cDNA sequencing, sewed as reference. RNA from lymphoma 9, which harbors a provirus downstream of the pim-1 gene at map position +9, was used as a control. The 92 base fragment present in all lanes is caused by nonspecific hybridization.

lymphomas are shown in Figure 4. The data show that proviruses in the 3’ exon were clustered in a small region. The provirus in lymphoma 32 was located 74 nucleotides downstream of the @m-l termination codon, and that in lymphoma 54 was located 5 bases further downstream (Figures 2, 3, and 4). Integration in lymphomas 10 and 17 occurred about 140 nucleotides downstream of the @m-l coding region at the same nucleotide position (Figures 3 and 4) whereas the provirus in lymphoma 34 was integrated 3 nucleotides further downstream from this latter site (Figures 3 and 4). This result shows that within this region integration is nonrandom. All proviruses were integrated outside the ORF, although in at least 25 out of 36

-e

5

4

P

PB

-3

P

.1

2

8

-

*zw

0

H B

‘WV

Figure

E

4. Structure

of the pim-1 Gene

The intron/exon structure as determined by comparison of cDNA and genomic DNA sequences is shown. Exons are indicated by blocks coding sequences are filled in. Horizontal arrows show’the direction as well as the position of integrated proviruses in the numbered lymphomas. Proviral integration sites that were determined by sequencing of cloned virus-host DNA junctions (lymphomas 1 and 15) by sequencing cDNA clones (lymphomas 10 and 64) or by RNAase mapping (lymphomas 17, 32, and 34) are shown here in detail. Abbreviations are the same as in Figure 1. except that E represents EcoRl restriction sites.

Sequence 607

of pim-1 : Homology

Site of Integration

with Protein

I

3’

5’ v

/g

wly-

RNA b

TUMOR

NO)

1

Kinases

v

02

1

DOWNSTREAM v

02

1 02

servation of amino acids in the pim-1 protein at positions that are required for the enzymatic activity of protein kinases strongly suggests that pim-1 belongs to the class of proto-oncogenes encoding protein kinases. Experiments are in progress to determine whether thepim-1 protein has serine/threonine kinase or tyrosine kinase activity.

PIM

Discussion

ACTIN

Figure

5. pim-1 mRNA

Levels

Relate

to Proviral

Integration

Sites

Poly(A)+ RNA (1 and 0.2 pg) from 12 lymphomas was bound to nitrocellulose filters. The integration regions denoted as 5’, 3’, and downstream are as indicated in Figure 1 by vertical arrows. The filter was hybridized first to apim-l-specific probe (upper), then washed and subsequently hybridized to an actin probe as a control for the quantity of the mRNA in the samples (lower).

phosphorylation and that is conserved in the catalytic domain of all tyrosine kinases (Patschinsky et al., 1982; Snyder and Bishop, 1984) is also present in thepim-1 protein at amino acid residue 198 (Figure 8). The consensus sequence Asp-Val-Tp-Ser-X-GIy-lle-Leu-Leu-X-X-X-X-X-XGly-X-X-Pro-Tyr seen in tyrosine kinases (residues 444483 in p~80~-~~) is almost completely retained in pim-1 (residues 224-242). At the carboxy terminus the homology is greatest between pim-1 (residues 275-284) and v-abl. The overall homology of pim-1 with the various protein kinases is reflected by the alignment score shown in Table 1. From this alignment it appears that pim-1 has the highest homology with protein-serine kinases. The con-

The Structure of the piml Gene The results reported here show that the pim-1 gene contains six coding exons. The total length of these six exons is approximately 2800 nucleotides, close to the mRNA size of 2.8 kb, which includes the poly(A) tail (Selten et al., 1985). The splice sites conform to the typical canonical sequences (Breathnach and Chambon, 1981; Mount, 1982). The 3’ exon is much larger than the others, and 1344 nucleotides of this exon are noncoding. Although two poly(A) addition signals are found at the 3’ end, the sequences of a number of independently derived pim-1 cDNAs revealed that only the second poly(A) addition signal, at position 8978, was used. The 5’ end of the pim-1 gene remains less well-defined. Two oligonucleotideprimed cDNAs ended at position 1719; thus we predict a 5’ exon of 419 nucleotides. Exons of slightly larger size as well as somewhat smaller size were found by Sl nuclease protection analysis, suggesting the presence of multiple initiation sites for mRNA transcription. No TATA consensus sequence was found upstream of this region. However, eight CCGCCC motifs were present in the 250 nucleotides preceding one of the putative transcription initiation sites. This sequence, which is found in multiple copies in a number of promoters, such as the human and mouse hypoxanthine ribosyltransferase genes (Pate1 et al., 1988; Melton et al., 1988) the human Harveyms protooncogene (Ishii et al., 1985a), the epidermal growth factor receptor gene (Ishii et al., 1985b), and the adenosine deaminase gene (Valerio et al., 1985), has been shown to bind specifically the transcription factor SP-1 (reviewed by Dynan and Tjian, 1985). Transcription from these pro-

Figure 6. Homologies between the Amino Acid Sequences of pim-1 and Protein Kinases residues 21-291, The pim-I protein sequence, is aligned with sequences from they subunit of phosphorylase kinase (PhK) and from the protein kinases encoded by v-mos, v-ah/, tck. and v-sm. The single-letter amino acid code is used. The numbering below indicates V-SIC sequences from residue 250 to the carboxy terminus (residue 519). Homology of @m-l with any of the other sequences is indicated below; underlined residues are highly conserved in protein kinases (van Beveren and Verma, 1985).

Cell 606

Table 1. Comparison of the Amino pim-I and Protein Kinases Alignment Sequence

Versus

Acid Sequences

of

Scores

@n-l

Versus

src

5.5

-

Yes 8 LSTRA p56 abl fps fes (ST)

6.9 5.4 6.1 7.3 7.2 9.0

103.1 69.1 72.6 46.0 33.0 35.0

ros fms erb0 trk

3.9 5.3 3.7 2.4

35.2 25.7 24.4 23.6

mil raf mos

5.4 4.4 6.1

21.1 19.1 11.1

10.3 11.1 6.3 11.6

4.9 6.9 10.9 4.9

cAPK cGPK MLCK PhK

src

Comparisons were made using the ALIGN program (Dayhoff, 1979) with a unitary matrix, with the exception that cysteine was scored as 2. The gap penalty was 3, and 100 random runs were carried out for each comparison. The average standard deviation for the group of random comparisons was 2.01 f 0.33. The alignment score is the real score (the number of identities minus penalties for gaps) minus the mean random score divided by the average standard deviation. Scores greater than two plus the average standard deviation are considered significant. Sequences were compared in the region between amino acid residues 250-519 (for src) (see also Figure 6). Sequences for comparison were obtained from the following sources: src, yes, fgr, abl, fps, fes (ST), ros, fms, erbS, mil, raf, and mos (Weiss et al., 1965); LSTRA p56 (tck) (Voronova and Sefton, 1966); trk(Martin-Zancs et al., 1966); cAPK (Shoji et al., 1961); cGPK (Takio et al., 1964); MLCK (Takio et al., 1965); PhK (Reimann et al., 1964). Abbreviations: cAPK, CAMP-dependent protein kinase; cGPK, cGMPdependent protein kinase; MLCK, myosin light chain kinase; PhK, phosphorylase kinase.

moters lacking TATA box sequences has been shown to start at multiple sites. The octamer motif ATGCAAAT, which is present upstream of the promoter of the immunoglobulin genes (Falkner and Zachau, 1984; Parslow et al., 1984) and which interacts with lymphoid-specific tram+acting factors (Singh et al., 1988), is also found approximately 280 bp upstream of pim-1 exon 1 at position 1438. This box is normally found approximately 70 bp upstream of the RNA cap site. However, in some immunoglobulin genes this motif is present at a position similar to the one it has in the pim-1 gene (Neuberger, 1983). In addition, cruciform structures with significant stability can be formed in this region. Although these characteristics are indicative of a promoter region, we still have to prove that this region has promoter activity and is used for pint-1 transcription. Furthermore, it is important to recall that the cDNAs used to identify the 5’ region was obtained from mRNAs present in tumors in which pim-1 was transcriptionally activated by integrated proviruses. Therefore, it is possible that in normal cells transcription is initiated at different sites.

The first exon contains a single AUG codon, which is followed by a 313 amino acid long ORF extending over six exons. The AUG codon is preceded by stop codons in all three reading frames. The noncoding portion of exon 1 is at least 337 nucleotides long (based on cDNA clones). Thus, in spite of the uncertainties concerning the 5’ end of the pim-1 gene, there is little doubt about the deduced amino acid sequence of the corresponding protein. Transcriptional Activation The primary effect of proviral insertion appears to be enhanced pim-1 expression. Although the majority of proviral integrations are clustered within the 3’ exon, the pim-1 coding sequence itself is spared in all integrations examined to date. All proviruses integrated in this 3’ exon are oriented away from the body of the pim-1 transcription unit, in accordance with the enhancement model of gene activation. This situation is reminiscent of the activation of in&l, in which integration in the noncoding region of the 3’ exon is also observed (Van Ooyen and Nusse, 1984). Tumors in which proviruses were integrated within the 3’ exon showed the highest pim-1 mRNA levels. Proviral integration in a transcription unit is likely to affect the transcription rate, and it may also affect the stability of the mRNA. The contribution of each of these factors to the steady-state level of pim-1 mRNA has not been determined. The high pim-1 expression level seen in tumors with proviruses in the 3’exon might have provided the driving force for the selective outgrowth of these tumors. Proviruses upstream of exon 1, which were found in both transcriptional orientations, gave rise to only a moderately enhanced expression level. The pim-1 mRNA in these tumors was indistinguishable from the normal pim-1 mRNA, and no hybrid transcripts were detected (Selten et al., 1985). The @n-l Protein The deduced amino acid sequence of the pim-1 gene shows that the pim-1 polypeptide is 313 amino acids long. A hydrophobicity plot does not reveal a long hydrophobic amino terminus or any other signal peptide that might direct the gene product to a membrane within the cell. pim-1 shows a remarkable amino acid sequence homology with protein kinases. The highest homologies were observed with they-subunit of phosphorylase kinase (Reimann et al., 1984) and with cGMP-dependent protein kinase (Takio et al., 1984) (see Figure 8 and Table 1). In the amino-terminal region of the catalytic domain, the homology with mos was remarkably high, whereas at the carboxy terminus close resemblance with abl was observed. All the domains conserved in protein kinases are found in pim-1. At present we cannot predict whether pim-1 is a serine/threonine kinase or a tyrosine kinase. The tyrosine present at position 416 in v-src, which can serve as an autophosphorylation site in all tyrosine kinases, is also conserved in pim-1. pim-1 probably belongs to a class of lymphoid-specific protein kinases that fulfill a role in the intracellular transduction of signals involved in the proliferative stimulation of lymphocytes. The recently identified T-cell-specific tyrosine-protein kinase from LSTRA cells

Seuence

of pim-1:

Homology

with Protein

Kinases

(tck) might serve a similar function (Voronova et al., 1984; Marth et al., 1985). Both pint-1 and rck are activated by proviral insertions that spare the protein-encoding domains (Voronova and Sefton, 1986; this work); neither gene has undergone mutations affecting the amino acid sequence. Apparently, the increased expression of pim-1 and tck is sufficient to confer a selective (growth) advantage to the cell. Experiments are in progress to identify the enzymatic activity and cellular localization of the pim-1 gene product. Experimental

Pmcedures

MuLWnduced

Lymphomas

The origin of MuLV-induced (Cuypers et al., 1984; Selten

Molecular

Cloning

lymphomas was described et al., 1984, 1985).

of Genomic

previously

DNA Fragments

An 11 kb EcoRl pim-1 DNA fragment, located between positions -9.6 and +1.3 on thepim-1 map (see Figure l), was molecularly cloned from normal BALB/c DNA. From the cloned pim-1 DNA fragment, subclones were derived in pBR322 and/or pSP64/65 plasmids. The cloning strategy for virus-host DNA junctions of lymphomas 1 and 15 has been described previously (Cuypers et al., 1984; Selten et al., 1965).

Hybridization

Probes

pim-l-specific probes were described previously (Cuypers et al., 1984; Selten et al., 1985). The aotin cDNA clone was described by Dodemont et al. (1982).

DNA Sequence

Analysis

From thepim-1 region, 8000 bp was sequenced from both strands, with overlapping sequences across every restriction site. pim-1 DNA fragments, subcloned in pBR322 covering the Sacl-EcoRI DNA fragment between positions -6.6 and +I.3 on thepiml map, were digested with SauBA, Taql, or Hpall, and were then subcloned in phage M13. DNA sequencing was performed by the dideoxynucleotide chain termination method (Sanger et al., 1980).

Synthesis

and Molecular

Cloning

of cDNAs

DNAs complementary to poly(A)+ RNA from lymphomas 9, 10, and 54 were synthesized as follows. Ten micrograms of poly(A)+ RNA, treated with methylmercuric hydroxide, was neutralized with f+mercaptoethanol. This mixture was adjusted to 50 mM Tris-HCI (pH 8.3) 40 mM KCI, 6 mM MgClz, 10 mM DTT, 3000 U RNAasin per milliliter, 4 mM Na.,PzOr, 50 9g actinomycin D per milliliter, 0.1 mg oligo(dT) (12-18) per milliliter or 0.1 mg syntheticpim-l-specific oligomer per milliliter, 0.5 mM dGTR 0.5 mM dATR 0.5 mM dTTR 0.25 mM dCTR and 400 l&i [a-32P]dCTP per milliliter, all in a final volume of 100 ~1. The mixture was put on ice for 10 min, heated for 2 min at 42OC, and the synthesis was started by addition of 150 U AMV reverse transcriptase; incubation was performed for 1 hr at 42OC. Second-strand synthesis was performed by adding DNA polymerase and RNAase H according to Gubler and Hoffman (1983). ARer treatment of the dsDNA with T4 DNA polymerase to obtain blunt ends, decameric EcoRl linkers were ligated to the dsDNA fragment. After digestion with EooRI, doublestranded cDNA fragments were separated from EcoRl linkers by Biogel A15m chromatography. Approximately 200 ng linker-containing double-stranded cDNA was ligated with 10 vg of EcoRIdigested and calf intestine phosphatase-treated lgtll vector DNA by T4 DNA ligase (Huynh et al., 1985). An average of ml.5 x lOa plaque-forming (units of phage per 200 ng cDNA was obtained after in vitro packaging of the ligation mixture. Libraries were screened with 32P-labeled pim1 probes. Clones containing pim-1 EcoRl inserts were subcloned in pSP64.

Preparation

of RNA

Total cellular RNA was obtained as described previously (Selten et al., ‘1965). To obtain poly(A)+ RNA, total cellular RNA samples were heated at 65% for 5 min prior to poly(A)+ selection by oligo(dT) chromatography.

Dot-Blot

RNA Analysis

For dot-blot analysis, poly(A)+ RNA was dissolved in distilled water and then applied to nitrocellulose filters as described by Mtiller et al. (1982). Before exposure to X-ray film, filters were washed as previously described (Selten et al., 1985).

RNAase

Mapping

First, 0.2 vg of a pim-1 BamHl DNA fragment (probe B, between positions 5642 and 6858 in Figure 2) subcloned in pSP64 and linearized at the EcoRl site (in the polylinker), was transcribed to 32P-labeled piml RNA by RNA polymerase using [a-32P]CTP as labeled ribonucleotide under conditions indicated by the supplier (Promega Biotee). The reaction mixture was treated with DNAase, extracted with a phenol/chloroform mixture, passed over a Sephadex G-50 column, and precipitated with ethanol. Next, l-2 ng =P-labeled pim-1 RNA was mixed with 3 ug lymphoma poly(A)+ RNA in 30 pl of a hybridization mix containing 80% formamide, 40 mM Pipes (pH 6.4), 1 mM EDTA, and 400 mM NaCI. After heating to 8S’C for 10 min, samples were incubated for 3-6 hr at 3% and for IO-12 hr at 30%. Samples were treated with RNAase A and RNAase Tl, deproteinized, and precipitated twice with ethanol in the presence of 10 ug tRNA and I M ammonium acetate. Protected RNA fragments were separated on a 6% polyacrylamide gel containing 8 M urea and were analyzed by exposure to X-ray film.

Acknowledgments This work was performed in the laboratory of Dr. H. Bloemendal, to whom we are particularly indebted. We appreciate the assistance of Jos Rewinkel in the sequencing of the pim-1 region, We thank Dr. C. Klaassen for the preparation of pim-l-specific oligonucleotides and Mrs. M. H. Sonne-Gooren and Mrs. l? E. T van Nuland-van Wijngaarde for typing the manuscript. This work was supported by grants from the Koningin Wilhelmina Fonds (G. S.) and by the Netherlands Organization for the Advancement of Pure Research (ZWC) through the Foundation for Fundamental Medical Research (FUNGO). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked hdwfisement” in accordance with 18 USC. Section 1734 solely to indicate this fact. Received

March

17, 1986; revised

May 12. 1986.

References Breathnach, of eucaryotic 349-383.

R., and Chambon, l? (1981). Organization and expression split genes coding for proteins. Ann. Rev. Biochem. 50,

Brugge, J. S., and Darrow, D. (1984). Analysis of the catalytic domain of phosphotransferase activity of two Avian Sarcoma Virus transforming proteins. J. Biol. Chem. 259, 4550-4557. Corcoran, L. M., Adams, J. M., Dunn, A. R., and Cory, S. (1984). Murine T lymphomas in which the cellular myc oncogene has been activated by retroviral insertion. Cell 37 113-122. Cuypers, H. T, Selten, G., Quint, W., Zijlstra, M., Robanus-Maandag. E., Boelens, W., van Wezenbeek, P, Melief, C., and Berns, A. (1984). Murine leukemiavirus-inducedToe lymphomagenesis: integration of proviruses in a distinct chromosomal region. Cell 37, 141-150. Czernilofsky. A. F!, Levinson, A. D., Varmus, H. E.. and Bishop, J. M. (1980). Nucleotide sequence of an Avian sarcoma virus oncogene (sro) and proposed amino acid sequence for gene product. Nature 287; 196-203. Dayhoff, M. O., ed. (1979). Atlas of Protein Sequence Volume 5, Supplement 3. (Washington, D. C.: National search Foundation).

and Structure, Biomedical Re-

Dickson, C., Smith, R., Brookes, S., and Peters, G. (1964). Tumorigenesis by mouse mammary tumor virus: proviral activation of a cellular gene in the common integration region irk-2. Cell 37 529-536. Dodemont, H. J., Soriano, P, Quax, W. J., Ramaekers, P., Lenstra, J. A., Groenen, M. A., Bernardi, G., and Bloemendal, H. (1962). The genes coding for the cytoskeletal proteins actin and vimentin in warmblooded vertebrates. EMBO J. 7, 167-171.

Cell 610

Donoghue, D. I. (1962). Demonstration of biological activity and nucleotide sequence of an in vitro synthesized clone of the Moloney murine sarcoma virus mos gene. J. Virol. 42, 538-546. Dynan, W. S., and Tjian, R. (1985). Control of eukaryotic messenger RNA synthesis by sequence-specific DNA-binding proteins. Nature 316, 774-770. Falkner, F. G., and Zachau, H. G. (1984). Correct transcription of an immunoglobulin K gene requires an upstream fragment containing conSeNSd sequence elements. Nature 370, n-74. Fung, Y.-K.T., Lewis, W. G., Hung, H.J., and Crittenden, L. B. (1963). Activation of the cellular oncogene c-erb8 by LTR insertion: molecular basis for induction of erythroblastosis by avian leukosis virus. Cell 33, 357-360. Gubler, method

L., and Hoffman, B. J. (1983). A simple and very for generating cDNA libraries. Gene 25, 263-269.

efficient

Hayward, W. S., Neel, B. G., and Astrin, S. M. (1981). Activation cellular one gene by promoter insertion in ALV-induced lymphoid kosis. Nature 290, 475-480.

of a leu-

den, L. B., Raines, M. A., and Kung, H. (1985). c-erhB activation in ALVinduced erythroblastosis: novel RNA processing and promoter insertion result in expression of an amino-truncated EGF receptor. Cell 41, i’l9-726. Nusse, R., and Varmus, H. E. (1982). Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Cell 37, 99-109. Nusse, R., van Ooyen, A., Cox, D., Kai, T., Fung, Y., and Varmus, H. (1984). Mode of proviral activation of a putative mammary oncogene (inf-1) on mouse chromosome 15. Nature 307, 131-136. O’Donnell, P V., Fleissner, E., Lonial, H., Koehne, D. F., and Reicin, A. (1985). Early clonality and high-frequency proviral integration into the c-myc locus in AKR leukemias. J. Virol. 55, 500-503. Parslow, T. G., Blair, D. L., Murphy, W. J., and Granner, D. K. (1984). Structure of the 5’ ends of immunoglobulin genes. A novel COnSeNSd sequence. Proc. Natl. Acad. Sci. USA 87, 2850-2654.

Ann. Rev.

Patel, P I., Franson, P E., C&key, C. T, and Chinault, A. C. (1986). Fine structure of the human hypoxanthine phosphoribosyltransferase gene. Mol. Cell. Biol. 6, 393-403.

Huynh, T. V., Young, R. A., and Davis, R. W. (1985). Constructing and screening cDNA libraries in 1gtll. In DNA Cloning Techniques: A Practical Approach, D. M. Glover, ed. (Oxford: IRL Press), pp. 49-78.

Patschinsky, T, Hunter, T., Esch, F. G., Cooper, J. A., and Sefton, B. M. (1982). Analysis of the sequence of amino acids surrounding sites of tyrosine phosphorylation. Proc. Natl. Acad. Sci. USA 79, 973-977.

Ishii, S., Merlino, G. T., and Pastan, I. (1985a). Promoter region of the human Harvey ras proto-oncogene: similarity to the EGF receptor proto-oncogene promoter. Science 230, 1378-1381.

Payne, G. S., Bishop, J. M., and Varmus, H. E. (1982). Multiple arrangements of viral DNA and an activated host oncogene in bursal lymphomas. Nature 295, 209-213.

Ishii, S., Xu, Y., Stratton, R. H., Roe, B.A., Merlino, G. T, and Pastan, I. (1965b). Characterization and sequence of the promoter region of the human epidermal growth factor receptor gene. Proc. Natl. Acad. Sci. USA 82, 4920-4924.

Payne, G. S., Courtneidge, S. A., Crittenden, L. B., Fadly, A. M., Bishop, J. M., andvarmus, H. E. (1981). Analysisof avian leukosisvirus DNA and RNA in bursal tumors: viral gene expression is not required for maintenance of the tumor state. Cell 23, 311-322.

Lemay, G., and Jolicoeur, P (1984). Rearrangement of a DNA sequence homologous to a cell-virus junction fragment in several Moloney murine leukemia virus-induced rat thymomas. Proc. Natl. Acad. Sci. USA 81, 39-42.

Peters, G., Brookes, S., Smith, R., and Dickson, C. (1983). Tumorigenesis by mouse mammary tumor virus: evidence for a common region for provirus integration in mammary tumors. Cell 33, 369-377.

Hunter, T., and Cooper, Biochem. 54, 897-930.

J. A. (1985). Protein-tyrosine

kinases.

Levinson, A. D., Courtneidge, S. A., and Bishop, J. M. (1981). Structural and functional domains of the Rous sarcoma virus transforming protein (pp60m). Proc. Natl. Acad. Sci. USA 78, 1624-1628. Li, Y., Holland, C. A., Hartley, J. W., and Hopkins, N. (1984). Viral integration near c-myc in 10-200~ of MCF 247-induced AKR lymphomas. Proc. Natl. Acad. Sci. USA 81, 6808-6811. Marth, J. D., Peet, R., Krebs, E. G., and Perlmutter, R. M. (1985). A lymphocyte-specific protein-tyrosine kinase gene is rearranged and overexpressed in the murine T cell lymphoma LSTRA. Cell 43, 393404. Martin-Zanca, D., Hughes, S. H., and Barbacid, M. (1986). A human oncogene formed by the fusion of truncated tropomyosin and proteintyrosine kinase sequences. Nature 319, 743-748. Melton, D. W., McEwan, C., McKie, A. B., and Reid, A. M. (1986). Expression of the mouse HPRT gene: deletional analysis of the promoter region of an X-chromosome linked housekeeping gene. Cell 44, 319-328. Mount, S. (1982). A catalogue Acids Res. 10, 459-472.

of splice

junction

sequences.

Nucl.

MOller, R., Slamon, D. J., Tremblay, J. M., Cline, M. J., and Verma, I. M. (1982). Differential expression of cellular oncogenes during preand postnatal development of the mouse. Nature 299, 840-644. Neel, B. G., Hayward, W. S., Robinson, H. L., Fang, J.. and Astrin. S. M. (1981). Avian leukosis virus-induced tumors have common proviral integration sites and synthesize discrete new RNAs: oncogenesis by promoter insertion. Cell 23, 323-334. Neil, J. C., Hughes, D., McFarlane, R., Wilkie, N. hf., Onions, Lees, G., and Jarett, 0. (1984). Transduction and rearrangement myc gene by feline leukemia virus in naturally occurring T-cell mias. Nature 308, 814-820.

D. E.. of the leuke-

Neuberger, M. S. (1983). Expression and regulation of immunoglobulin heavy chain gene transfected into lymphoid cells. EMBO J. 2, 13731376. Nilsen,

T. W., Maroney,

P A., Goodwin,

Ft. G., Rottman,

F. M., Critten-

Piechaczyk, M., Yang, J. Cl., Blanchard, J.-M., Jeanteur, P., and Marcu, K. 8. (1985). Posttranscriptional mechanisms are responsible for accumulation of truncated c-myc RNAs in murine plasma cell tumors. Cell 42, 589-597. Reddy, E. P., Smith, M. J., and Srinivasan, A. (1983). Nucleotide sequence of Abelson murine leukemia virus genome: structural similarities of its transforming gene product to other one gene products with tyrosine-specific kinase activity. Proc. Natl. Acad. Sci. USA 80, 36233627. Reimann, E. M., Titani, K., Ericsson, L. H., Wade, R. D., Fischer, E. H., and Walsh, K. A. (1984). Homology of the y subunit of phoshorylase b kinase with c-AMP dependent protein kinase. Biochemistry 23, 4185-4192. Sanger, F., Coulson, A. R., Barrell, B. G., Smith, A. J. H., and Roe, B. A. (1980). Cloning in single-stranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 743, 161-178. Schwartz, D. E., Tizard, quence of Rous sarcoma

R., and Gilbert, W. (1983). virus. Cell 32, 853-869.

Nucleotide

se-

Selten, G., Cuypers, H. T., Zijlstra, M., Melief, C., and Barns, A. (1984). Involvement of c-myc in MuLV-induced T-cell lymphomas in mice: frequency and mechanisms of activation. EMBO J. 3. 3215-3222. Selten. G., Cuypers, H. T., and Berns, A. (1985). the putative oncogene pirn-l in MuLV-induced EMBO J. 4, 1793-1798.

Proviral T-cell

activation of lymphomas.

Shen-Ong, G. L. C., Potter, M., Mushinski, J. F., Lavu. S., and Reddy, E. P. (1984). Activation of the c-m@ locus by viral insertional mutagenesis in plasmacytoid lymphosarcomas. Science 226, 1077-1060. Shoji, S., Parmelee, E. C., Wade, R. D., Kumar, S., Ericsson, L. H., Walsh, K. A., Neurath, H., Long, G. L., Demaille, J. G.. Fisher, E. H.. and Titani, K. (1981). Complete amino acid sequence of the catalytic subunit of bovine cardiac muscle cyclic AMP-dependent protein kinase. Proc. Natl. Acad. Sci. USA 78, 848-851. Singh, H., Sen, R., Baltimore, D., and Sharp, P. A. (1986). A nuclear factor that binds to a COnSSNed sequence motif in transcriptional control elements of immunoglobulin genes. Nature 379, 154-158.

Sequence 611

of @m-l:

Homology

with Protein

Kinases

Snyder, M. A., and Bishop, J. M. (1984). A mutation at the major phosphotyrosine in pp60v-src alters oncogenic potential. Virology 736, 375-366. Steffen, D. (1984). Proviruses are adjacent leukemia virus-induced lymphomas. Proc. 2097-2103.

to c-myc in some murine Natl. Acad. Sci. USA 81,

Swift, Ft. A., Shaller, E., Witter, B. L., and Kung, H. J. (1965). Insertional activation of c-myc by reticuloendotheliosis virus in chicken B lymphoma: non-random distribution and orientation of the proviruses. J. Virol. 54, 869-672. Takio, K., Wade, Ft. D., Smith, S. B., Krebs, E. G., Walsh, K. A., and Titani, K. (1984). Guanosine cyclic 3’,5’-phosphate dependent protein kinase, a chimeric protein homologous with two separate protein families. Biochemistry 23, 4207-4216. Takio, K., Blumenthal, D. K., Edelman, A. M., Walsh, K. A., Krebs, E. G., and Titani, K. (1965). Amino acid sequence of an active fragment of rabbit skeletal muscle myosin light chain kinase. Biochemistry 24, 6028-6037. Tsichlis, P M., Lohse, M. A., Szpirer, C., Szpirer, J., and Levan, G., (1985). Cellular DNA regions involved in the induction of rat thymic lymphomas (Mlvi-1, Mlvi-2, Mlvi-3, and c-mu) represent independent loci as determined by their chromosomal map locations in the rat. J. Virol. 50, 936-942. Valerio, D., Duyvensteyn, M. G. C., Dekker, 8. M. M., Weeda, G., Berkvens, Th. M., van der Voorn, L., van Ormondt, H., and van der Eb. A. J. (1985). Adenosine deaminase: characterization and expression of a gene with a remarkable promoter, EMBO J. 4, 437-443. van Beveren, C., Galleshaw, J. A., Jonas, V., Berns, A. J. M., Doolittle, R. F., Donoghue, D. J., and Verma, I, M. (1981). Nucleotide sequence and formation of the transforming gene of a mouse sarcoma virus. Nature 289, 258-262. van Beveren, C., and Verma, genes. Curr. Topics Microbial.

1. M. (1985). Homology among and Immunol. 123, 73-98.

onco-

Van Doyen, A., and Nusse, Ft. (1984). Structure and nucleotide sequence of the putative mammary oncogene inr-1; proviral insertions leave the protein-encoding domain intact. Cell 39, 233-240. Voronova, A. F., and Sefton, protein kinase is stimulated 379, 682-685.

B. M. (1986). Expression by retrovirus promoter

Voronova, A. F, Buss, J. E., Patschinsky, T, Hunter, B. M. (1984). Characterization of the protein apparently the elevated tyrosine protein kinase activity of LSTRA Biol. 4, 2705-2713.

of a new tyrosine insertion. Nature T., and Sefton, responsible for cells. Mol. Cell.

Webb, E., Adams, J. M., and Cory, S. (1984). Variant (6;15) translocation in a murine plasmacytoma occurs near an immunoglobulin K gene but far from the myc oncogene. Nature 372, 777-779. Weinmaster, G., Hinze, E., and Pawson, T. (1983). Mapping of multiple phosphorylation sites within the structural and catalytic domains of the Fujinami avian sarcoma virus transforming protein. J. Virol. 46,29-41. Weiss, R., Teich, N., Varmus, H., and Coffin, J., eds. (1985). RNA Tumor Viruses. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). Westaway, D., Payne, G., and Varmus, H. E. (1984). Proviral deletions and oncogene base-substitutions in insertionally mutagenized c-myc alleles may contribute to the progression of avian bursal tumors. Proc. Natl. Acad. Sci. USA 87. 843-847. Yamamoto, Toyoshima, a member

T, Nishida, T., Miyajima, N., Kawai, S., Ooi, K. (1983). The e&B gene of avian erythroblastosis of the src family. Cell 35, 71-78.

T., and virus is

Yanisch-Perron, C., Vieira, J., and Messing, J. (1985). Improved Ml3 phage cloning vectors and host strains: nucleotide sequence of the Ml3 mp18 and pUCl9 vectors. Gene 33, 103-119.