Mar 13, 1995 - Hanneke van der Gulden, John Allen and. Anton Berns3 .... A.Berns, unpublished results). Screening of a thymus cDNA library with the PCR.
The EMBO Journal vol.14 no.11 pp.2536-2544, 1995
Proviral tagging in Eg-myc transgenic mice lacking the Pim-1 proto-oncogene leads to compensatory activation of Pim-2 Nathalie M.T.van der Lugt1, Jos Domen2, Els Verhoeven, Koert Linders, Hanneke van der Gulden, John Allen and Anton Berns3 Department of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands 'Present address: Hubrecht Laboratory, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands 2Present address: Department of Pathology, B263 Beckman Center, Stanford University School of Medicine, Stanford CA 94305-5428, USA 3Corresponding author N.M.T.van der Lugt and J.Domen contributed equally to this work
The Pim-1 proto-oncogene is one of the most potent collaborators of the myc proto-oncogenes in inducing lymphomagenesis in mice. Contrary to the profound effects when overexpressed in vivo, Pim-l-deficient mice showed only subtle phenotypic alterations, which could indicate the presence of redundantly acting genes. In line with this, a PCR-based screen has led to the identification of a closely homologous gene, Pim-2. The X-linked Pim-2 gene is 53% identical to Pim-1 at the amino acid level and shares substrate preference and the usage of non-AUG initiation codons with Pim-1. We have used these data to test whether the strong synergistic interaction between Pim-1 and c-myc can be utilized to gain access to Pim-1 compensatory pathways. We reasoned that, upon proviral tagging in compound mutant mice (Eji-myc/Pim-L'-1 mice), the selective advantage of cells carrying provirally activated genes, that act downstream from or parallel to Pim-1, would increase. We show here that this is the case. A dramatic increase (from 15 to 80%) was found in the frequency of proviral activation of the Pim-2 gene. These data show that the described strategy of 'complementation tagging' represents a powerful new tool to identify components of pathways involved in processes as complex as multistep tumorigenesis. Key words: lymphomagenesis/Pim-l/Pim-2/protein serine/ proviral tagging/threonine kinase
Introduction The activation of the c-myc oncogene, by either chromosomal rearrangement, gene amplification or retroviral insertional mutagenesis, is one of the predominant steps in many neoplasias. Experimental model systems support the important role of myc in transformation. Ej-myc transgenic mice in which c-myc oncogene expression is targeted to the B-cell lineage, show an expansion of the pre-B-cell compartment from early on in life, and -50% of the mice develop pre-B-cell lymphomas before 5
months of age (Adams et al., 1985; Langdon et al., 1986). However, constitutive expression of c-myc is not sufficient for full B-cell transformation, as is evident from the variable latency period and the monoclonality of the tumors. Proviral tagging has been applied successfully to identify new oncogenes that cooperate with c-myc in lymphomagenesis. This technique takes advantage of slowtransforming retroviruses to activate genes via insertional mutagenesis, thereby facilitating their cloning and subsequent characterization (for reviews, see van Lohuizen and Bems, 1990; Adams and Cory, 1992). One of the oncogenes identified in this way is Pim-] (Cuypers et al., 1984; van Lohuizen et al., 1991). Proviral insertion near Pim-J leads to overexpression of Pim-] mRNA and protein. The Pim-J gene encodes two primary translation products, sized 34 and 44 kDa, by altemative initiation at AUG and CUG. Both proteins have protein serine/threonine kinase activity and show clear substrate preferences (Selten et al., 1986; Saris et al., 1991). Relatively high expression levels of the gene are found in hematopoietic organs such as thymus, spleen and fetal liver, but also in embryonic stem cells and testes. Expression is highly inducible by certain growth factors such as interleukin (IL)-2, IL-3, IL-7 and granulocyte-macrophage colonystimulating factor (GM-CSF) (Bems et al., 1987; Meijer et al., 1987; Dautry et al., 1988; Ihle et al., 1990; Domen et al., 1993a), but not by others, such as macrophage colony-stimulating factor (M-CSF) or steel factor (SF) (Lilly et al., 1992; Domen et al., 1993b). The oncogenic potential of the Pim-1 gene has been further assessed in Ep-Pim-J transgenic mice (van Lohuizen et al., 1989). T-cell lymphomas develop in 510% of these mice before the age of 7 months. However, when these mice are infected with Moloney murine leukemia virus (MoMuLV), T-cell lymphomagenesis is strongly accelerated. In almost 100% of the tumors either c-myc or N-myc were found to be activated by proviral insertion. In analogy, -35% of the MoMuLV-induced tumors in Ei-myc mice carried a proviral insertion near Pim-1. These data implicated a strong synergistic effect of Pim-l and myc in lymphomagenesis. This synergism was most dramatically shown by cross-breeding E,umyc and Ei-Pim-1 transgenic mice: bitransgenic mice succumbed in utero to lymphoblastic leukemia, the strongest synergism between myc and another oncogene recorded to date (Verbeek et al., 1991). Moreover, MoMuLV-induced tumors in Eg-myc mice showed that, besides proviral insertions near Pim-], frequent proviral integrations near Bmi-JlBla-J and Pal-i were observed (van Lohuizen et al., 1991; Haupt et al., 1991). More insight into the functioning of the Pim-1 proteins has come from the study of Pim- 1 -deficient mice. Analysis of hematopoietic cells in vitro revealed that Pim-1 is an important modulator of the IL-3 and IL-7 growth responses
2 6 Oxford University Press 2536
Compensatory activation of Pim-2 in Ei-myc mice
in vitro of bone marrow-derived mast cells and lymphoid cells, respectively (Domen et al., 1993a,b). Despite this, mice with two inactivated copies of the Pim-] gene show no obvious phenotypic alterations in their hematopoietic system, as discerned by flow cytometry (Laird et al., 1993). In view of the evolutionary conservation of the Pim-l gene, this would be most easily explained by functional redundancy between Pim-l and other genes. Functional redundancy could be a consequence of the existence of parallel pathways or of a Pim-1-like protein. To address this question, we have applied both a standard molecular biological strategy, i.e. cloning of Pim-1 family members by degenerate PCR, and a functional cloning strategy that would identify components of the Pim-1 signal transduction pathway. The latter approach was based upon the observation that Pim-J activation is an important step in lymphomagenesis. We reasoned that in the Pim-l-1- background, the selection pressure increases for proviral activation of genes that act downstream from or parallel to Pim-]. In addition, the Pim-l1 mice facilitated the search for Pim-] homologs by degenerate PCR. Here we describe the analysis of Pim-1-deficient mice and compound mutant mice (Pim-] null-mutant mice, carrying the Et-myc transgene) for their susceptibility to MoMuLV-induced lymphomagenesis and report on the frequent activation of a new Pim-] family member, Pim-2, by 'complementation tagging'.
Results PCR cloning of Pim-2 Previous attempts to find Pim-homologous genes in the murine genome using low-stringency hybridization have failed to reveal the presence of such genes. Therefore, we decided to employ a PCR-based approach using degenerate primers. Comparison of the mammalian Pim-1 sequences with Pim-like sequences that were recently deposited in the Caenorhabditis elegans data base (ACeDB), allowed us to delineate highly Pim-specific regions in the open reading frame (ORF). Amplification on wild-type DNA failed to yield anything but Pim- 1 clones. Therefore, amplifications were performed on DNA isolated from Pim-1 null-mutant mice, which lack the region to be amplified from the Pim-] gene. We obtained a Pim-like clone, distinct from Pim-], which we named Pim-2. A proviral insertion site which is preferentially occupied in later stages of lymphomagenesis, and which was previously named Pim-2 (Breuer et al., 1989), has been renamed Tic- 1. The gene affected by insertions in this latter locus has not yet been identified (J.Jonkers and A.Berns, unpublished results). Screening of a thymus cDNA library with the PCR probe and a brain cDNA library with a thymus Pim-2 cDNA probe yielded clones representing exons 4-6, and some of the introns of Pim-2. 5' End cDNA sequences were generated using a RACE protocol (Frohman et al., 1988) on RNAs from two different lymphomas with high Pim-2 expression levels due to proviral activation (see below). All PCRs indicated the same 5' end. The sequence of the Pim-2 cDNA (Figure 1) reveals a structure very similar to Pim-], with a long 3' UTR (1014 nucleotides), a 834 nucleotide long ORF starting with an AUG surrounded by a very weak Kozak consensus sequence
(Kozak, 1987) and a 190 nucleotide long, GC-rich, 5' UTR containing no in-frame stop codons, but several CUGs surrounded by a favorable Kozak sequence (see also below). Northern blot analysis showed Pim-2 expression at low levels in a variety of tissues and highest expression in brain and thymus. The pattern clearly differs from that of Pim-], which is hardly expressed in brain and shows prominent expression of a shortened transcript in testis (Meijer et al., 1987). Strikingly, high levels of both mRNAs are present in activated lymphocytes and bone marrow-derived mast cells grown in the presence of IL-3. Like Pim-1, Pim-2 is also expressed in embryonic stem cells and in various hematopoietic cell lines (not shown). Using a cDNA probe to screen a genomic library, clones representing two distinct loci were obtained. One locus contained sequences contiguous with the cDNA sequence, lacking the intron sequences between exons 4 and 5. In addition eight mutations, six of which were non-silent, were present in a sequenced region of 199 bp. This included changes in highly conserved residues, like Glu266 to Lys. We therefore concluded that this locus represented a pseudogene. The other locus contained a gene structurally very similar to Pim-]. It consists of six exons spread over 5.5 kb of genomic DNA (Figure 2A). Exon (and intron) sequences are identical to those of the cDNAs, showing that this locus contains the Pim-2 gene. Sequencing of 451 bp of the promoter region reveals that, while both Pim-1 and Pim-2 contain a GC-rich and TATA-less promoter (Figure 2B), there is little direct sequence homology. Using interspecies backcross hybrids, the Pim-2 gene was mapped proximal on the X chromosome, approximately halfway between the markers A-raf and Act-7, near Kv4. 1. The pseudogene maps to chromosome 8 (N.Copeland and N.Jenkins, personal communication).
Characteristics of the Pim-2 protein As indicated in Figure 1, the first AUG in the Pim-2 mRNA is surrounded by a very poor Kozak consensus sequence. This, combined with the fact that the upstream sequence is open and contains a number of CUG codons in a more favorable sequence context, suggests that translational initiation in Pim-2 may be analogous to Pim-I, with only a fraction of the initiations occurring at AUG. In vitro transcription and translation in rabbit reticulocyte lysate of the Pim-2 cDNA showed this to be indeed the case (Figure 3). Only a small proportion of the protein thus produced has the expected molecular weight for initiation at AUG (34 486 Da). The vast majority is present in two larger bands, migrating at 37 and 40 kDa. Initiation at CUG 14 and 90 would yield proteins of these sizes, 40 059 and 37 723 Da respectively. In order to demonstrate that the larger proteins indeed initiated from these CUG codons, both were deleted by site-directed mutagenesis. As shown in Figure 3, deletion of these CUGs prevents initiation of the larger proteins. These CUG initiation codons are likely to be used also in vivo, as the autophosphorylation patterns (see below) of Pim-2 immunoprecipitated from cells and reticulocyte lysate are identical (not shown). When the complete amino acid sequences of both proteins are aligned (Figure 4), the similarity is clearly the highest within the catalytic domain (Vib). Like Pim- 1, Pim-2 lacks an Asp residue at position 2537
N.M.T.van der Lugt et aL
1* L
101
A
R
T
A
N
L
N
A
A
S
P
G
A
G
S
A
P
P
S
D
L
S
P
A
L
T
P
P
29
agctgttcgcccgccggacagctcgccacgtcgggtattcgacctccAGtac S
G
P
S
P
A
L
A
P
R
S
A
T
C
P
G
G
S
L
S
F
G
I
N
L
R
S
S
T
S
M
L
T
K
63
201~~~~~~~~~~~~~~~P P
Q
L
G
H
S
P
P
V
P
T
P
T
Q
G
P
P
G
K
D
R
A
A
A
E
F
E
Y
R
L
P
G
L
96
301~~~~~~~~~~~~~~~~~P G
L
G
K
G
G
F
V
T
F
G
A
H
R
V
D
T
R
R
Q
V
I
A
K
V
S
I
N
R
R
V
G
L
129
401~~~P W
S
W
601
F
C
S
D
I
C
T
V
G
E
P
S
R
L
F
P
M
L
E
V
L
V
L
A
E
L
L
R
P
W
P
M
G
V
K
Q
A
G
E
D
L
D
F
N
G
N
I
Y
G
P
T
E
R
I
V
G
K
L
P
L
D
E
G
L
163 196
C
F
F
G
R
Q
T
S
I
V
K
V
L
A
I
F
D
Q
V
A
G
C
H
S
G
H
A
A
L
G
R
L
V
D
H
V
P
E
D
R
H
Y
T
I
K
F
D
G
D
N
E
D
T
L
I
V
R
229 S
Y
263
gctcgggacccaacgactccgcggcgcgtatgttcatttaagtttggctc E
P
P
F
E
P
1101
T
E
S
D
agttgccgtctacagctgacgcaccgcccctggttctggttagaggaact S
801
S
V
T
W
R
S
Q
D
S
C
I
R
E
L
I
S
P
Q
H
R
L
E
Y
E
E
H
A
I
A
E
L
D
A
F
H
L
L
P
L
P
T
P
W
A
M
S
P
G
L
S
V
H
Q
S
W
V
P
A
V
E
C
C
D
E
L
L
K
Y
L
A
P
D
I
I
N
R
S
C
V
M
C
R
S
K
D
G
A
L
G
I
S
P
P
P
296
K
329
T
P
363
cctccgtctctc__acacagtttgtgatcagcccaagaaatgctttgctgtt L
P
W
S
L
L
P
370
1201 1301 1401 1501
cgtgtaatcgatcagtagatggtaggcgaagtaacagcgcatcttcacga gaccctcaacggtctatggggggtcttttgtttaaagtatgtgatctctt gccagtctacgctgatctccgcctcaatcaaacctccctgtaagttatgg agggttctaatgcctccgacctagaagctnctgacctcgtcgcgcgctgc
2001
tgtggaaacctATTTAA.5aataatgaagtcacaaaaaagaaaaaaaaaaa
2051
Fig. 1. Mouse Pim-2 cDNA sequence. The translational initiation sites and stop codon are capitalized and underlined. In the 3' UTR, destabilizing motifs, as defined by Shaw and Kamen (1986), are capitalized and the poly(A) addition site consensus sequence is underlined. Exon boundaries are indicated by arrows.
A
GI'G CTFG ATG
TAG
1 kbp
240
.~~~~~~~~._______
~~~~~~~~~~360 ~~~~~~~~~~~~~~~480
Fig. 2. (A) Genomic structure of the murine Pim-2 gene. Exons are indicated by boxes. White boxes indicate the untranslated regions, black boxes the AUG-initiated open reading frame and hatched boxes the N-terminal extensions of the open reading frame that initiate with CUG codons. (B) Promoter sequence of the murine Pim-2 gene. Stippled boxes indicate the position of two octamer binding site consensus sequences (ATGCAAAT), the underlined regions two SP-1I binding site consensus sequences (CCGCCC). The transcribed region, as determined by the longest anchored PCR products, is indicated by the affow and lower case.
2538
Compensatory activation of Pim-2 in Ej.-myc mice
1
A
3
2
2
1
3
4
5
6
kD
4OkD 38kD
345k
-
-. .+0
34kD
-mwm .-'
-50
Fig. 3. In vitro transcription and translation of Pim-2 cDNA clones. Shown are 35S-labeled proteins translated from a wild-type clone (lane 1), a clone with CUG-15 deleted (lane 2) and a clone with deletions of both CUG-15 and -90 (lane 3).
1
37
GASGPPDSLPSTLAPPSPGSPAAL
MARATNLNAAPSA
-33
1 MGPAAPLALPPPALPDPAGEPARGQPRQRPQSSSDSPSALRASRSQSRNA 50
38
PRASTPCGLSGFSGLNIRSTSS
.............
MLTK
PLQGHP
69
. . . . . : . : . . : : : : . 51 TRSLSPGRRLSPSSLRRRCCSSRHRRRTDTLEVGMLLSKINSLAHLRARP 100
.
I
70 SPPVTPTQPPGGKDRAAFEAEYRLGPLLGKGGFGTVFAGHRVTDRRQVAI 119
101 CNDLHATKLAPGKEKEPLESQYQVGPLLGSGGFGSVYSGIRVADNLPVAI 150
1f. II . JI III . . IV JI 120 KVISRNRVLGWSTVSDSVTCPLEVALLWKVGEGNGHPGVIRLLDWFETPE 169
B
s .1
151 KHVEKDRISDWGELPNGTRVPMEVVLLKKV..SSDFSGVIRLLDWFERPD 198 . v VIa 170 GFMLVLERPMPAQDLFDYITEKGPLGESCSRSFFTQVVAAVQHCHARGVV 219
199 SFVLILERPEPVQDLFDFITERGALQEDLARGFFWQVLEAVRHCHNCGVL 248 VII . .VIII VIb 220 HRDIKDENILIDLCRGSIKLIDFGSGALLHDEPYTDFDGTRVYSPPEWIS 269
I1111111111111:11.111111111111.1..1111111111111111I. 249 HRDIKDENILIDLSRGEIKLIDFGSGALLKDTVYTDFDGTRVYSPPEWIR 298
14
ix
.
1.
x
270 RHQYHALPATVWSLGVLLYDMVCGDIPFERDQEILEAELHFPAHVSPDCC 319 299 YHRYHGRSAAVWSLGILLYDMVCGDIPFEHDEEIIKGQVFFRQTVSSECQ 348 .xi 320
ALIRRCLAPKPCSRPSLEEILLDPWMQS
PAEEKPINS.SKGSPTP
.:: . H .. : 1. 1111: 1: : 1. : 1: . 111: 111 349 HLIKWCLSLRPSDRPSFEEIRNHPWMQGDLLPQAASEIHLHSLSPGSSK
363
397
364 LPWSLLP 370
Fig. 4. Comparison of the mouse Pim-2 (top) and Pim-1 (bottom) amino acid sequences. Alignment was done by the GCG program GAP (Devereux et al., 1984). The major translation initiation sites indicated by double arrows. The subdomains of the protein kinase domain, as defined by Hanks and Quinn (1991), are indicated by Roman numerals and delineated by single arrows
are
308, which is highly conserved in other protein kinases. Furthermore, there is little direct sequence similarity outside of the catalytic domain. In order to determine a number of biochemical features of Pim-2, antibodies were raised against a GST fusion protein containing the C-terminal 148 amino acids of the Pim-2 protein (P2F serum). Furthermore, a Pim-2 expression plasmid containing a vesicular stomatitis virus (VSV) tag (de Wind et al., 1992) was made, since an antiserum directed against this tag was available. Incuba-
Fig. 5. Kinase activity of murine Pim-2 immunoprecipitated from COS cells. (A) SDS-gel electrophoresis of autophosphorylated Pim-2 proteins. Lanes I and 4, wild-type Pim-2; lanes 2 and 5, lysine to methionine mutant; lanes 3 and 6, untransfected COS cells. Lanes 1-3 precipitated with P2F serum; lanes 4-6 with antiserum directed against the VSV tag. (B) Phosphoamino acid analysis of autophosphorylated Pim-2 protein produced in reticulocyte lysate.
tion of antibody-bound protein immunoprecipitated from either a tumor cell line containing a MoMuLV-activated Pim-2 gene (not shown) or COS cells transfected with the Pim-2 gene (Figure 5A) with [y-32P]ATP resulted in the phosphorylation of the Pim-2 proteins. Immunoprecipitations from reticulocyte lysate and two other lymphomaderived cell lines yielded similar results (not shown). To check that these specific phosphorylations were the result of Pim-2 intrinsic protein kinase activity, and do not reflect the activity of a co-immunoprecipitated contaminating protein kinase, site-directed mutagenesis (Lys 120 into Met) was employed to generate an enzymatically inactive VSV-tagged protein. A similar mutation in many other protein kinases, including Pim-1 (Saris et al., 1991), has been shown to abolish enzyme activity. Indeed, no protein
2539
N.M.T.van der Lugt et al.
B % tumor free
A % tumor free
100
80
so
40
20
50
0
100
age
(days)
age (days)
Fig. 6. (A) Tumor incidence of 40 Pim-I+'+ mice (squares), 81 Pim-]+I- mice (circles) and 75 Pim-1-n- mice (triangles). (B) Tumor incidence of 14 Pim-J+'+ mice (squares), 45 Pim-J+k- mice (circles) and 30 Pim-r-L- mice (triangles), all carrying the Et-mvc transgene. Note that in (A) a significant delay in the latency period is observed in the absence of Pim-1. Even the presence of only one Pim-1 allele results in a significant delay.
kinase activity was detected in immunoprecipitates of mutant Pim-2 protein from transfected COS cells (Figure SA). The sequence of Pim-2 clearly places it in the group of protein serine/threonine kinases (Hanks and Quinn, 1991). While dual specificity for serine/threonine and tyrosine cannot be excluded on the basis of the primary amino acid sequence, no phosphotyrosine was detected in phosphoamino acid analysis of Pim-2 phosphorylated protein (Figure SB). When a mixture of the five histones was used as substrate, Pim-2, like Pim-1, only used histone H2B as a substrate for phosphorylation. Pim-2 exhibited similar reaction conditions for optimal phosphorylation as Pim-l (Saris et al., 1991), e.g. manganese is preferred over magnesium as divalent cation, while the presence of small quantities of zinc are inhibitory for catalytic activity (not shown).
Lymphoma induction Proviral tagging in Pim-i mutant mice was performed as an independent method to identify genes that are (functionally) homologous to Pim-i or act downstream of Pim-i. Therefore, 1 day-old wild-type (+/+) pups and littermates, heterozygous (+/-) or homozygous (-/-) for the inactivated Pim-i allele, were infected with MoMuLV. Half of these mice carried the Eg-myc transgene. Mice were sacrificed when moribund, and tumors were analyzed for lineage markers and proviral insertion pattems. In the absence of the E.-myc transgene, mice developed T-cell lymphomas. This was demonstrated by flow cytometric analysis, showing surface expression of the CD3 antigen (not shown). In the presence of the Ej-myc transgene, Bcell lymphomas develop, as demonstrated by the surface expression of the B220 antigen (not shown). No difference in tumor phenotype was observed between wild-type and Pim- 1-deficient mice. The latency period of T-lymphoma development in Pim-k-1- mice is significantly longer (-1 month) than in Pim-i+I+ mice (Pt-myc
Fig. 8. Coincidence of common proviral insertion sites in MoMuLV-induced tumors in (A) Pim-lI +'+IEt-myc, (B) Pim-I +'-'Et-myc and (C) Pim-l-'-fEg-myc mice. Respectively, 54, 29 and 13% of the tumors do not have proviral insertions in the sites that were tested. Tumors carrying proviral insertions in more than one common insertion site are indicated by overlapping bars. The data for Bmi-J and Pal-i are represented as one bar, since proviral integration in these two genes appeared to be mutually exclusive. The percentages of occupation of the individual sites are
denoted in Table
I.
in, for example, Drosophila to identify components of the sevenless signal transduction pathway (Simon et al., 1991). The term 'complementation tagging' relates to the suppression of defined mutations, e.g. loss-of-Pim-J, by proviral insertional mutagenesis. The identification of Pim-2 by 2542
PCR cloning and the observation that both Pim-] and Pim-2 could be activated by proviral insertion allowed us to test whether such a complementation tagging was feasible. The experiments presented here indicate that, in the presence of the Ej-myc transgene and only one functional
Compensatory activation of Pim-2 in E,t-myc mice
Pim-I allele (Pinm-1K/1EJ1p-mvc), the number of proviral insertions near Pimn-1 and Pimn-2 increases >2-fold as compared with insertions in tumors of Pim-I`'/+E.-mvc mice. If both alleles of Pim-1 are inactivated (Pimn-l-l Ep-ncX), >80% of the tumors carry a proviral insertion in Pim-2, a dramatic increase over the percentage seen in Pim-l+lE4tp-mvc mice. This dramatic rise in the selective pressure to activate Pim-2 is only observed in the presence of the Ep-mvc transgene and, therefore, might be a reflection of the synergistic interaction between Pim and Myc. Alternatively, it might reflect an increased requirement for the activation of the Pim pathway in B-cell lymphomagenesis (Ep-myc mice) as compared with T cell lymphomagenesis (wild-type and Pim-i mutant mice). As expected, proviral activations of Pim-! and Pim-2 were found to be mutually exclusive. This is reflected by the absence of insertions near Pim-2 in Ej-Pim-1 transgenic mice and by the lack of co-activation of Pim-1 and Pim-2 in MoMuLV-induced tumors in wild-type and E.trnvc transgenic mice. This shows that Pim-l and Pim-2 indeed act in a functionally redundant fashion in tumorigenesis. Had the Pimn-2 gene not been identified by a PCR-based method, we certainly would have identified Pim-2 as a new common insertion site in the MoMuLV-induced lymphomas in the Pim---l/Ep-mvc mice. Therefore, identification of proviral insertion sites in these compound mice permits identification of genes acting in a narrowly defined pathway. Together, the data unambigously show that 'complementation tagging' is a feasible approach in this in Oi'vo tumor progression system. Consequently, it might be expected that cloning of the proviral insertion sites present in the remaining 20% of the tumors that show no activation of Pim-2 will give access to other Pim family members or, more interestingly, to genes acting downstream of these kinases. This is currently being investigated. The methodology of 'complementation tagging' described here adds an intriguing gimmick to the repertoire of genetic tools to identify new components of signal transduction pathways. It can be applied in a number of different settings and should therefore have a wide applicability. In addition, it permits evaluation of the relative importance of a particular pathway to confer a specific biological effect.
Materials and methods PCR amplifications PCR amplification of the Pimti-2-fragment was performed on DNA from homozygous Pirtn-I null-mutant mice (te Riele et al., 1990) with the primers A: 5'-GCGGATCCGGIGTIATIAG/AICTICTG/CGAC/TTGGT-3' and B: S'-GCGAATTCGCICCIGAICCGAAG/ATCGATIAGC/TTT-3'. The PCR mixture was heated to 75°C, followed by addition of Taq polymerase and 40) cycles of 45 s 94°C, 2 min 45°C and 45 s 72°C. PCR amplification of RNA was performed by a first strand reverse transcriptase reaction (Boehringer Mannheim) followed by 5' end cloning with the primers RI: 5'-ATGGACAACTCCACGGGC-3', R2: 5'-CCCTTCTCTGTGATATAGTCG-3' and R3: 5'-GCGGAATTCAAGAAGGCGTATCACACCCG-3'. RACE conditions were essentially as described (Frohman et al., 1988). To change Lys 120 into a Met, cDNA was amplified with the primers 4: 5'-GGTGGCCATCATGGTAATCTCC-3' and 5: 5'-GCGGAATTCAAGAAGGCGTATCACACCCG-3'. The resulting fragment was cleaved with Mscl and used to replace the endogenous MscI fragment. CTG deletions in the cDNA were made by amplifying cDNA with primers
1: 5'-GCATCGATAGTAGCGTTGGGGGCGCGCGCG-3' and 2: 5'-
TCTATCCGTGACGCGGTG-3' (CTG-15) or primers I and 3: 5'TGCCCGGGGCAGAGCGGCCGGAGAGCCAGGGCTGGGCGCGGTGGAA-3' (CTG-15 and 90). PCR conditions were 30 cycles of I min 94°C, I min 55°C and I min 72°C. The PCR product was digested with SinaI and Clal and cloned in the corresponding cDNA sites.
Cloning Cloning and subcloning of cDNAs and genomic DNAs were done according to established procedures (Sambrook et al., 1989). Clones were obtained from the following libraries: oligo-d(T)-primed thymus cDNA (Stratagene), oligo-d(T)-primed brain cDNA (Meijer et al., 1990) and 129/Ola genomic phage library. The Pim-2 expression plasmids containing a VSV tag (de Wind et al., 1992) were made by changing the Pimn-2 stop codon by site-directed mutagenesis into a Notd site. The tag was added as a NotI-XbaI fragment and the resulting cDNA was cloned (SalI-Sacl) into the pJ3W expression plasmid. A kinase-inactive form of Pim-2 (see above) was cloned into the same plasmid. In vitro transcription and translation In vitro transcription and translation of plasmids containing Pim-2 inserts (from position I to the Sacl site at position 1302) were done in rabbit reticulocyte lysates using the TnT system (Promega Biotec) with T7 RNA polymerase according to the manufacturer's instructions.
Immunological methods An antiserum (P2F) was raised against a GST-Pim-2 fusion protein. To produce the fusion protein, the EcoRV-HindII portion of the Pim-2 cDNA was coupled to GST sequences. Inclusion bodies were purified from bacteria by spinning the bacterial lysates made by sonication through a 40%7 sucrose cushion and used for immunizing rabbits. For booster injections, ureum-solubilized protein was used. Immunoprecipitations were essentially as described (Saris et al., 1991; Domen et al., 1993a). Briefly, cells were lysed by freeze-thawing in lysis buffer (20 mM PIPES pH 7.0, 30 mM NaCI, 5 mM MgCl,, 14 mM f-mercaptoethanol, 1% aprotinin, 1 mM PMSF, 1 mM leupeptin and I ,ug/tl soybean trypsin inhibitor. Immunoprecipitations were performed on the cleared lysates (10 min microfuge) using protein A-Sepharosebound antibodies in 10 mM sodium phosphate pH 8.0, 150 mM NaCl, 5 mM CHAPS and 14 mM P-mercaptoethanol.
Kinase assays Kinase assays were done essentially as described by Saris et al. ( 1991) and Domen et al. (I 993a). Briefly, immunoprecipitates were taken up in 4 ,ul kinase buffer (20 mM PIPES pH 7.0, 15 mM MnCI2, 7 mM fI-mercaptoethanol, 0.25 mM 3-glycerophosphate, 0.4 mM spermine). After pre-incubation on ice for at least 5 min, 4 gl of kinase buffer containing 10 gCi [y-32P]ATP (3000 Ci/mmol) were added, and reactions were incubated at 30°C for 30 min. The reactions were terminated by adding 8 gl 2x sample buffer or, in the absence of exogenous substrate, by washing the beads once with ice-cold wash buffer (20 mM PIPES pH 7.0, 14 mM f3-mercaptoethanol) followed by resuspension in 16 pl l x sample buffer. Samples were boiled (5 min) and analyzed on 15% SDS-polyacrylamide gel. Exogenous substrates were added at 1 .g per 8 ,l and included a mixture of all five bovine histones and purified histone H2B (Boehringer).
Mice and lymphoma induction
Outbred (129 Ola\Balb/c) Pim-l-1- mice were generated as described (Laird et al., 1993). Further breeding of these mice with FVB or B/CBA mice resulted in Pimn-I mutant and wild-type mice (FVB\129 Ola\Balb/ c or B/CBA\129 Ola\Balb/c), which were used for the induction of lymphomas. The genotype of the mice was monitored as described (Laird et al., 1991). Mice of this same genetic background were used to cross in the EI-t-lvc transgene (Verbeek et al., 1991; founder-line 186), resulting in Pim-l mutant and wild-type El-mvc transgenic mice with a mixed genetic background (FVB\B/CBA\B6\129 Ola\Balb/c). These mice were used for the induction of lymphomas. We did not observe significant differences in the survival of mice with the same genotype, but obtained from litters of a different genetic background. To induce lymphomas, I day-old mice were injected with 104_-10 p.f.u. of MoMuLV clone IA as described (Jaenisch et al., 1975). Mice were sacrificed, and moribund and tumor tissues were frozen at -80°C. Statistical analysis was performed with the Mantel-Cox (log-rank) method (BMDP statistical software).
2543
N.M.T.van der Lugt et aL
DNA and RNA isolation and analysis High molecular weight DNA from lymphoma tissues was prepared as described (van der Putten et al., 1979). For Southern analysis, 10 .g of total genomic DNA of each tumor were digested with restriction enzymes as recommended by the supplier, separated on a 0.7% agarose gel, transferred to nitrocellulose and hybridized to 32P-labeled probes. For Northern analysis, 20 tg of total RNA, prepared by the LiCl-urea method, were separated on 1 % agarose formaldehyde gels (Sambrook et al., 1989) and transferred to nylon membranes. The following probes were used: c-myc, 3 kb XbaI-HindIII fragment (Shen-Ong et al., 1982); N-myc, 3.5 kb PstI fragment (Taya et al., 1986); Pim-J probe A, 0.9 kb BamHI fragment (Cuypers et al., 1984); MoMuLV U3 LTR, 180 bp fragment from U3 LTR region (Cuypers et al., 1984); Pal-/, 1 kb BglIIEcoRI fragment (van Lohuizen et al., 1991); Bla-l, 2.2 kb Bglll-EcoRI fragment (van Lohuizen et al., 1991); Bmi-J probe B (van der Lugt et al., 1994); Pim-2 3' probe, 0.7 kb Sacl fragment and Pim-2 5' probe, 0.5 kb BamHI fragment.
Flow cytometric analysis 1 X 106 cells were incubated in 96-well plates for 30-45 min at 4°C in 20 ,ul PBS+ + (phosphate-buffered saline with 1% BSA and 0.1 % sodium azide) and saturating amounts of monoclonal antibody. Cells were washed twice with PBS + + and incubated with streptavidin-phycoerythrin for biotinylated antibodies or PBS++. The cells were analyzed on a FACSscan (Becton Dickinson). The following antibodies were used: CD45R/B220 (6B2), CD3 (145-2C I ), both from Pharmingen (San Diego, CA).
Database accession numbers The GenBank accession numbers for the sequences reported in this paper are: L41495 and L41496.
Acknowledgements We would like to thank Drs R.Plasterk for drawing our attention to the C.elegans Pim sequences, N.Copeland and N.Jenkins for the chromosomal mapping of the Pim-2 gene and pseudogene, B.Scheijen for providing the Pal-I probe, D.Acton, M.Vlaar and G.Habets for providing libraries, J.Jonkers for help in setting up the RACE-PCR, G.Gil-Gomez for providing the cell lines 537 and 556, N.de Wind for providing the VSV tag plasmids, E.de Vries for help with the phosphoamino acid analysis, T.Hamersfeld for immunization of rabbits and T.Maidment, L.Rijswijk and N.Bosnie for taking care of the mice. We thank M.Alkema for critically reading the manuscript. This work was supported by the Dutch Cancer Society (KWF) (K.L.; E.V.; N.v.d.L.), the Netherlands Organization for Scientific Research (NWO) (N.v.d.L.) and by a fellowship of the Leukemia Society of America (J.A.).
References Adams,J.M. and Cory,S. (1992) Cancer Surv., 15, 119-141. Adams,J.M., Harris,A.W., Pinkert,C.A., Corcoran,L.M., Alexander,W.S., Cory,S., Palmiter,R.D. and Brinster,R.L. (1985) Nature, 318, 533-538. Berns,A., Cuypers,H.T., Selten,G. and Domen,J. (1987) In Kjeldgaard, N. and Forchhammer,J. (eds), Pim-1 Activation in T-cell Lymphomas. Munksgaard, Copenhagen, pp. 211-224. Breuer,M.L., Cuypers,H.T. and Berns,A. (1989) EMBO J., 8, 743-747. Cuypers,H.T., Selten,G., Quint,W., Zijlstra,M., Robanus-Maandag,E., Boelens,W., van Wezenbeek,P. and Berns,A. (1984) Cell, 37, 141-150. Dautry,F., Weil,D., Yu,J. and Dautry-Varsat,A. (1988) J. Biol. Chem., 263, 17615-17620. Devereux,J., Haeberli,P. and Smithies,O. (1984) Nucleic Acids Res., 12, 387-395. de Wind,N.N., Domen,J. and Berns,A. (1992) J. Virol., 66, 5200-5209. Domen,J., van der Lugt,N.M.T., Laird,P.W., Saris,C.J., Clarke,A.R., Hooper,M.L. and Berns,A. (1993a) Blood, 82, 1445-1452. Domen,J., van der Lugt,N.M.T., Acton,D., Laird,P.W., Linders,K. and Bems,A. (1993b) J. Exp. Med., 178, 1665-1673. Domen,J., van der Lugt,N.M.T., Laird,P.W., Saris,C.J.M. and Berns,A. (1993c) Leukemia, 7, S 108-S 112. Frohman,M.A., Dush,M.K. and Martin,G.R. (1988) Proc. Natl Acad. Sci. USA, 85, 8998-9002. Hanks,S.K. and Quinn,A.M. (1991) Methods Enz-ymol., 200, 38-62. Haupt,Y, Alexander,W.S., Barri,G., Klinken,S.P. and Adams,J.M. (1991) Cell, 65, 753-763.
2544
Ihle,J.N., Morishita,K., Bartholomew,C., Matsugi,T. and Askew,D. (1990) Int. J. Cell Cloning, 8, 130-146. Jaenisch,R., Fan,H. and Croker,B. (1975) Proc. Natl Acad. Sci. USA, 72, 4008-4012. Knighton,D.R., Zheng,J., Ten Eyck,L.F., Ashford,V.A., Xuong,N.-H., Taylor,S.S. and Sowadski,J.M. (1991) Science, 253, 407-414. Kozak,M. (1987) J. Mol. Biol., 196, 947-950. Laird,P.W., Zijderveld,A., Linders,K., Rudnicki,M.A., Jaenisch,R. and Berns,A. (1991) Nucleic Acids Res., 19, 4293. Laird,P.W., Van der Lugt,N.M.T., Clarke,A., Domen,J., Linders,K., Mcwhir,J., Berns,A. and Hooper,M. (1993) Nucleic Acids Res., 21, 4750-4755.
Langdon,W.Y., Harris,A.W., Cory,S. and Adams,J.M. (1986) Cell, 47, 11-18. Lilly,M., Le,T., Holland,P. and Hendrickson,S.L. (1992) Oncogene, 7, 727-732. Meijer,D., Hermans,A., von Lindern,M., van Agthoven,T., de Klein,A., Mackenbach,P., Grootegoed,A., Talarico,D., Della Valle,G. and Grosveld,G. (1 987) EMBO J., 6, 4041-4048. Meijer,D., Graus,A., Kraay,R., Langeveld,A., Mulder,M.P. and Grosveld,G. (1990) Nucleic Acids Res., 14, 7357-7365. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Saris,C.J.M., Domen,J. and Berns,A. (1991) EMBO J., 10, 655-664. Selten,G., Cuypers,H.T., Boelens,W., Robanus-Maandag,E., Verbeek,J., Domen,J., van Beveren,C. and Berns,A. (1986) Cell, 46, 603-611. Shaw,G. and Kamen,R. (1986) Cell, 46, 659-667. Shen-Ong,G.L.C., Keath,E.J., Piccoli,S.P. and Cole,M.D. (1982) Cell, 31, 443-452. Simon,M.A., Bowtell,D., Dodson,G.S., Laverty,T.R. and Rubin,G.M. (1991) Cell, 67, 701-716 Taya,Y., Mizusawa,S. and Nishimura,S. (1986) EMBO J., 5, 1215-1219. te Riele,H., Robanus-Maandag,E., Clarke,A., Hooper,M. and Berns,A. (1990) Nature, 348, 649-651. van der Lugt,N.M.T. et al. (1994) Genes Dev., 8, 757-769. van der Putten,H., Terwindt,E., Berns,A. and Jaenisch,R. (1979) Cell, 18, 109-116. van Lohuizen,M., Verbeek,S., Krimpenfort,P., Domen,J., Saris,C., Radaszkiewicz,T. and Berns,A. (1989) Cell, 56, 673-682. van Lohuizen,M. and Berns,A. (1990) Biochim. Biophys. Acta, 1032, 213-235. van Lohuizen,M., Verbeek,S., Scheijen,B., Wientjens,E., van der Gulden,H. and Berns,A. (1991) Cell, 65, 737-752. Verbeek,S., van Lohuizen,M., Van der Valk,M., Domen,J., Kraal,G. and Berns,A. (1991) Mol. Cell Biol., 11, 1176-1179. Received on February 8, 1995; revised on March 13,1995
