Previously, analysis of cDNAs encoding the ltk tyrosine kinase suggested that the structure of this protein was unique among tyrosine kinases, containing a ...
The EMBO Journal vol.10 no.10 pp.2911 -2919, 1991
The Itk gene encodes tyrosine kinase
a
novel receptor-type protein
John J.Krolewski and Riccardo Dalla-Favera Department of Pathology and Comprehensive Cancer Center, Columbia University, College of Physicians and Surgeons, New York, NY 10032, USA Communicated by J.Schlessinger
Previously, analysis of cDNAs encoding the ltk tyrosine kinase suggested that the structure of this protein was unique among tyrosine kinases, containing a transmembrane domain but only a short, or virtually non-existent, extracellular domain. Further, it was suggested that translational initiation might occur predominantly at a CTG codon. We have now cloned and sequenced a putative full length human ltk cDNA which contains novel sequence information relative to previously identified cDNAs. This ltk cDNA encodes a protein product containing all of the features of typical receptor-type protein tyrosine kinase, including: an ATG translational initiation codon, a secretory signal sequence and a 347 amino acid extracellular domain as well as transmembrane and intracellular kinase domains. Ribonuclease protection analysis indicates that our cloned cDNA represents the most abundant species of mature Itk mRNA. In vitro transcription and translation of the ltk cDNA yields a 100 kDa protein, consistent with initiation at the putative ATG translational codon. In addition, transfection of the ltk cDNA into COS-1 cells produces a similar-sized, glycosylated protein possessing in vitro kinase activity. These data indicate that the Itk gene product likely functions as a cell surface receptor for an unidentified cellular growth factor. Key words: human ltk gene/protein tyrosine kinase/receptor
Introduction Protein tyrosine kinases (PTKs) were first identified as the protein products of the oncogenes carried by acutely transforming animal retroviruses. Subsequently, it was shown that the normal cellular homologues of these viral oncogenes encode a large family of proteins which share a conserved catalytic domain about 300 amino acids in length (Hanks et al., 1988). It is now known that most, if not all, of the tyrosine kinases normally function in the initial steps of intracellular signal transduction pathways (Cantley et al., 1991). Structurally, these proteins can be divided into two classes: receptor-type PTKs and non-receptor PTKs. Members of the non-receptor class are mostly found in the cytoplasm, associated with the plasma membrane via an N-terminal myristate (Resh, 1990). Based on the discovery that the Ick PTK acts as the signal transducer for the CD4 and CD8 cell surface molecules (Veillette et al., 1988; Rudd et al., 1988; Shaw et al., 1989), it has been proposed that members of this class of tyrosine kinases resemble 'broken Oxford University Press
receptors' with sequences near the amino termini of these proteins specifically interacting with the intracellular portion of various transmembrane receptor molecules (reviewed by Sudol, 1991). The second class of these proteins, the receptor-type PTKs, are integral membrane proteins, spanning the plasma membrane by virtue of a hydrophobic domain of about 25 amino acids. The amino-terminal portion of the receptor PTK extends extracellularly and functions to bind ligand, inducing the receptor to dimerize within the membrane, activating the intracellular kinase domain and resulting in receptor autophosphorylation (Ullrich and Schlessinger, 1990). Ligands are known for about half of the receptor-type PTKs and all are polypeptide growth factors including, among others: the epidermal growth factor (Ullrich et al., 1984), insulin (Ullrich et al., 1985), macrophage colony stimulating factor (Sherr et al., 1985), the platelet-derived growth factors (Yarden et al., 1986; Matsui et al., 1989) and the fibroblast growth factors (Lee et al., 1989; Ruta et al., 1989). The recent application of both genetic and biochemical approaches has identified the cognate ligands of some of the remaining 'orphan' receptor PTKs, such as those encoded by the met (Bottaro et al., 1991), neu (Lupu et al., 1990) and kit (reviewed by Witte, 1990) genes, demonstrating the utility of these receptors as a route to novel growth factors. The ltk gene was initially identified in the mouse and was found to resemble the insulin receptor subfamily of tyrosine kinases, on the basis of amino acid homology within the kinase domain (Ben-Neriah and Bauskin, 1988). These authors detected two ltk mRNAs, of -2.3 and 2.9 kb, by Northern blot analysis. Subsequently, we and others independendy isolated human (Krolewski et al., 1990; Maru et al., 1990) and mouse (Bernards and de la Monte, 1990) ltk cDNA clones. The two cDNAs cloned from the mouse are nearly identical, measuring 2.2 kb in length and containing a single long open reading frame. Initially it was suggested that the first ATG within the open reading frame represented the translational initiation codon, producing a protein containing a hydrophobic transmembrane domain and a kinase domain, but only a tiny extracellular extension of eight amino acids (Ben-Neriah and Bauskin, 1988). It was proposed that the ltk protein was the prototype of a distinct class of PTKs which, while unable to function as a receptor in its own right, might be associated with another transmembrane molecule and serve as the signal transducing component of such a complex. Such a role would be functionally analogous to that subserved by the non-receptor tyrosine kinase Ick. Later, Bernards and de la Monte (1990) noted that the sequence surrounding the putative ATG start codon was a poor match to the consensus sequence for eukaryotic translational initiation (Kozak, 1987) and, further, that an upstream CTG codon provided a functional site for translational initiation both in vitro and in vivo. In this case, the predicted protein product would contain an extracellular domain of
2911
J.J.Krolewski and R.DalIa-Favera
end relative to the previously reported mouse and human Itk cDNAs. Our results demonstrate that this ltk cDNA encodes a typical receptor PTK of 100 kDa apparent
108 amino acids, probably still too small to bind a ligand by itself. We now report the cloning of a 3.0 kb human ltk cDNA which contains novel sequence information at its 5'
TGA
ATG
A
ATG
CTG
An
_
PTK
TM
SS
-
Rl xi
Rl
B 1
75
150
225
AGGTGGGCG TGGCCAGACCAGCTT TAGGTCCGGGGAGTG TC TCGCCAGCGGCAGCACACCCC TG TAAGTGGT.GGCCAGGGCTGC CGTGGCAAAA TGAGC TGTCAACTTTAGG TTGACAGGGGTG TGGCC GCGAC CGCAA GGGCT T TTGT
TGCCGGGTGGACCCAACAGGGATGGGC TGC TGGGC,ACAGC TGCTGGTGTGGTTCGGAGCCGCGGGCGCCA TTCTCTGCTC TAGCC CGGGG TCCCAGGAGAC TTTTCTGCGGTCCTCGCCCCTGCC GCTGGCAAGT CCCAGCCCCCGGGAC MDG C W G Q L L V W F G A A G A] I L C S S P G S Q E T F L R S S P L P L A S P S P R D 375 300 CCGAAAGTCAGCGCCCCGCC TAGTA TCTTGGAGCCAGCCTCCCCGCTGAA TTCTCCGGGCACCGAGGGGTCTTGGCTGTT TTCTACCTGCGGGGCCAGCGGCCGGCATGGGCCCACACAGACACAATGTGACGGGGCGTACGCGGGGACC P X V S A P P S I L E P A S P L N S P G T E G S W L F S T C G A S G R H G P T Q T Q C D G A Y A G T 525 4 50 AGCGTGGTGG TGACCGTGGGGGCCGCCGGGCAGCTGAGAGGCGTGCAGCTGTGG;CGCGTGCCGGGCCCTGGCCAG TATCTGATCTCAGCC TACGGAGCCGCGGGCGGCAAAGGCGCCAAGAACCACCTGTCGCGGGCGCA TGGCG TCTTC S V V V T V G A A G Q L R G V Q L W R V P G P G Q Y L I S A Y G A A G G K G A K N H L S R A H G V F 675 600 GTCTCAGCAA TCTTC TCCCTCGGTC TCGGGGAGTCGCTGT ACATCCTGGTGGGGCAGCAGGGAGAGGACGCCTGTCCCGGAGGTAGCCCGGAGAGCCAGC TCGTC TGCCTCGGGGAGTCTCGAGC CGTTGAAGAGCACGC GGCGA TGGAT V S A I F S L G L G E S L Y I L V G Q Q G E D A C P G G S P E S Q L V C L G E S R A V E E H A A M D
S T P S W V TCGACCCCCAGCTGGGTA
825 750 GGGAGCGAAGGGGTCCCGGGGTCGC GGCGCTGGGC GGGAGGTGGCGGGGG TGGCGGGGGCGCCACCTACG TTTTC CGGTTGCGCGCTGGCGAGCTGGAACCGTTGCTGGTGGCGGCCGGAGGCGGCGGTC GGGCC TACCTGAGGC CGCGG G G GG G G A T Y V F R L R A G E L E P L L V A A G G G G R A Y L R P R G S E G V P G S R R W A G S R P R .. A A P D R P L A P * Q D P A Q A P L S L W V P R
GCCCAGGCTCCGCTCTCCCTCTGGGTTCCCCGGGCTCAGGACCCGGGTGCGGCCCCTGACCGGCCACTGGCCCCATAG 975 900 GACCGAGGCC GGAC TCAGGC CTCCCCCGAGAAAC TGGAGAACCGC TCGGAGGCGC CCGGGAGCGGCGGGAGAGGCGGGGCGGCAGGGGGCGACGC TTCAGAGACTGACAACCTC TGGGC TGATGGGGAAGA TGGAGTATC C TTCA TACAC D R G R T Q A
S
P
E
K
L
E
N
R
T
R
V.
S
A
E A
A
P
G
S
G G
R
G G A
A
G
G
.
D A T
S
E
T
.S
D
N
L
L
W A
D G
E D
G V
.T
S
F
I
H
V
1125
1050
CCCAGCAGCGAGC TC TTCC TGCAGC CTCTGGCAGTCACCGAGAAC CACGGAGAGG TAGAGATCCGAAGGCACCTCAACTGCAGTCACTGCCCTTTGAGAGAC TGC CAATGGCAGGCAGAGC TCCAGC TGGC TGAA TGCC TGTGCC CAGAA P S S E L F L Q P L A V T E N H G E V E I R R H L N C S H C P L R D C Q W Q A E L Q L A E C L C P E T W T F K P K . . . G Y G 1275 1200 A TGA GCCTCCTTA T ;GTGT(.TCEG(,GTCCTGATTCT(GGGAGCAGAAGAAGTGG GGCATGGAGC TAGCTGTGGA TAACG TCACC TGCATGGACC TGCACAAGCCCCCAGGCCCT(;TGGrT TrTr. TrGlTGrT.TGGCCACTA K W G M E L A V D N V T C M D L H K P P G P IL V L M V A V V A T S T L S L L M V C G V LI L V K QG C A A L A .M P T T A S. .I. G . T N. .
1425 1350 CAGGGCC TGCAGGAGA TGAG GCTGCCGAGC CCTGAGC TTGAGC TGAGCAAGC TTCGAACC TCTGC CATCAGGACAGCCCCCAATCCCTA TTATTGCCAGG TGGGGC TTGGCCCGG CCCAG TCCTG GCC TC TGCCA CCAGG TGTCACCGAG Q G L Q E M R L P S P E L E L S K L R T S A I R T A P N P Y Y C Q V G L G P A Q S W P L P P G V T E P L S . S G W G T
1575 1500 t GTTTC CCCAGCCAATGTTAC TCTGC TCAGAGCCC TGGGCCATGGTGCCTT TGGGGAGGTGTATGAGGGAC TGGTAAT TGGCCT TCCTGGGGACTC CAGTC CCCTGCAGGTAGCTA TCAAGACCC TGCCAGAAC TC TGC TC GCC TCAGGA T V S P A N V T L L R A L G H G A F G E V Y E G L V I G L P G D S S P L Q V A I K T L P E L C S P Q D H
P
T
1725
1650
GAGC TGGATT TCCTCATGGAGGCCC TCA TCA TCAGCAAGT TTCGC CATCAGAACA TTGTGCGGTG TGTGGGGCTCAGCCTCAGGGCCACC CCTCGCCTCA TTCTGC TGGAAC TGA TGTC TGGAGG GGACA TGAAGAGT TTCC TGAGGCAC E L D F L M E A L I I S K F R H Q N I V R C V G L S L R A T P R L I L L E L M S G G D M K S F L R H F S A S 1875 1800 AGTCGGCCACACC TGGGCCA GCCATCACC TCTGGTCATGC GGGACCTGCTGCAAC TGGCCCAGGACATAGCCCAGGGC TGCCACTACC TGGAGGAAAATCAC TTCATCCACAGGGATATTGCCGC CCGGAAC TGC CTGCT GAGC TGCGCT S R P H L G Q P S P L V M R D L L Q L A Q D I A Q G C H Y L E E N H F I H R D I A A R N C L L S C A S . . A . . T P . . Q 2025 1950 GGACC CAGCC GAGTGGCCAAGATTGGGGAC TTTGGGATGGCACGAGATATCTACCGGGCCAGTTA TTACCGCAGGGGGGACCGGGCCTTGCTCCCAGTCAAGTGGATGCCCCCAGAGGCC TTCC TGGAGGGCATC TTCACATCCAAGACA G P S R V A K I G D F G M A R D I Y R A S Y Y R R G D R A L L P V E W M P P E A F L E G I F T S K T L L T . G Q .K A
2175
2100
GATTCCTGGTCTTTTGGGGTGCTGC TCTGGGAGATCTTCTCACTGGGCTACATGCCCTATCCTGGGCGCACCAACCAGGAGGTGC TGGAC TTCGTCGTTGGAGGAGGCCGGATGGACCCTCCTAGGGGCT GCCCAGGGCC TGTGTACCGC D S W S F G V L L W E I F S L G Y M P Y P G R T N Q E V L D F V V G G G R M D P P R G C P G P V Y R N I A T N . . H . 4 2325 2250 ATCATGACCCAGTGT TGGCAGCACGAGCCTGAGC TCCGCCC TAGC TTTGC CAGCA TCTTGGAGCG TCTGCAGTAC TGCAC TCAGGACCCGGA TGTGCTGAATTCACTCCTGCCAA TGGAGC TGGGGCCCA CCCCAGAGGA GGAAGGGAC T I M T Q C W Q H E P E L R P S F A S I L E R L Q Y C T Q D P D V L N S L L P M E L G P T P E E E G T E A P I L V . P G. . D Q
2475
2400
TC TGGGCTGGGGAAC AGA TC T TTGGAGTGC CTAAGACCCC CACAGCCCCAGGAAC TGAGTCCAGAGAAGT TGAAAAGCTGGGGAGGTAGC CCTCTTGGCC CCTGGC TGTCCTCTGGCC TCAAGCC CCTCAAATCCAGGGGCC TCC AACC T P Q E L S P E K L K S W G G S P L G P W L S G L K P L K S R G L Q P S G L G N R S L E C L R P P S K S Q N G L P G L S . . R T P . C 2550
2625
CAGAACC TT TGGAAT CCCAC TTATC GC TCC TGAGC CCCAAGGGGC CCTGAGGGTAAGGAC TGAGGCAC TGAGGGTCCCTC CCTATACTCC TCAGGCTCCTGGGTGGCC TG TTATGCCAGC GGCCT CTGT TCCCTGCAGTC TGTGC TGTGT Q
N
2700
L
I
W
N
P
T
Y
R
G
S
I
W
T
P
R G
P Q G E D
T G
I
E H C N
G
S
S
S
S
S
I
P
G I
Q
2775
G TCTGGGCC T GTC TCGGGGC TGGCC TGGCAGCGC TGCAC TTGCCA TGCTGGAAACCAGCCCAGGCCTCCC AGGGAAGGGCCCAGCCAC TT CCAGC TTTTGATCTTGGGGC CAGAGGCCGC CTTACACACACCCCAGGTG TCCATGGGGAG 25 2850 CAC2TGGATTGCTTTCCCATATGAGCGTCCTTCATCTGGGCAGA9CCCCCACCCTGCAGATGCTTCTAATAAAAGGCTCTTCTCATCCTCCAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
291 2
The Itk encodes receptor tyrosine kinase
molecular weight containing an extracellular domain of 347 amino acids, which appears to be sufficiently large to bind a ligand.
Results Sequence of the 3.0 kb ftk cDNA Previously, we isolated a 1.2 kb partial ltk cDNA from a human B-cell cDNA library by low stringency hybridization with an fins tyrosine kinase probe (Krolewski et al., 1990). Sequencing revealed that this clone (Xl in panel A of Figure 1) contained only two thirds of a typical tyrosine kinase domain. Further, Northern blot analysis of human lymphoid cell lines had indicated that the ltk mRNA was 3.1 kb long (Krolewski et al., 1990). Attempts to isolate longer, full length ltk cDNAs from the original B-cell library failed, so we surveyed a panel of human cell lines for ltk expression by a ribonuclease protection assay (data not shown), and used the highest expressing cell line (the human neuroblastoma SK-N-MC) as a source of RNA for the construction of ltk gene specific cDNA libraries. Panel A of Figure 1 shows the clones (X2, X3 and X4) isolated from the successive construction and screening of two such libraries. Two clones (X2 and X3) contain a 180 bp insert relative to the third (X4). This insert apparently represents an unspliced intron in the cDNA (see below) and therefore is not included in the sequence data of Figure 1. The sequence of these overlapping cDNA clones (Figure 1, panel B) contains a large open reading frame encoding an 803 amino acid protein beginning with the ATG at nucleotide (nt) position 171. The sequence surrounding this ATG fits the definition of a 'strong' translational start site (Kozak, 1987). This reading frame is closed 19 codons upstream; in addition, the 5' untranslated region contains an upstream ATG, in another reading frame, which encodes a putative five amino acid mini-protein. Upstream ATGs are found frequently in proto-oncogene loci, and in the case of the mouse Ick gene, have been shown to serve a regulatory role (Marth et al., 1988). As this upstream ATG is in a relatively strong context it may function to modulate the translation of this gene (Kozak, 1987). The most important feature of the Itk cDNA sequence in Figure 1 is that it extends -750 bp farther at its 5' end than the longest of the previously identified Itk cDNA clones. In addition, the 5'-most 96 bp of the murine cDNA diverge completely from our ltk sequence (positions 732-828 in Figure IB), as seen by comparing the predicted amino acid sequences in this region. Interestingly, in the murine
sequence the 3' end of this divergent stretch encodes a stop codon and contains the nucleotide sequences found near a typical 3' splice site (Lewin, 1990): an AG dinucleotide and pyrimidine-rich stretch and a lariat branch site 18 bp 3' to the AG (Figure 1). These features suggest that the murine cDNA may actually represent a partially processed mRNA containing an intron at the 5' end. As a result of this novel 5' sequence, our cDNA clone encodes a protein 254 amino acids longer than, and in the same reading frame as, the protein product predicted from the murine cDNA by Bernards and de la Monte (1990). Importantly, examination of a hydrophobicity plot (Figure 2) and the computer-assisted application of the method of von Heijne (1986), predicts that the amino terminus of our deduced protein contains a secretory signal sequence which is presumed to be necessary for insertion of nascently synthesized receptor-type PTKs into the endoplasmic reticulum. Thus, we predict that the Itk protein has the features of a typical receptor-type PTK in the form of a 347 amino acid extracellular domain and a 16 residue signal sequence. In contrast, the predicted protein sequence beginning at the CTG start codon of Bernards and de la Monte (1990) is neither hydrophobic (Figure 2) nor does it contain an identifiable signal sequence. In other respects our sequence data confirm the presence of many of the features identified in the previously cloned mouse and human ltk cDNAs, including: a 25 amino acid hydrophobic transmembrane domain (Figure 2); a protein tyrosine kinase domain which is most closely related to the kinase domain of the c-ros proto-oncogene (Ben-Neriah and Bauskin, 1988) and a carboxy terminus which is 28 amino acids shorter than the murine equivalent (Maru et al., 1990). The previously reported human Itk cDNA sequence (Maru et al., 1990) is identical with our sequence in the region of overlap, which begins at position 985 in Figure 1, with two exceptions. First, the cDNA of Maru et al. (1990) contains a 51 bp (17 amino acid) insert just 3' to the sequence encoding the transmembrane domain. Alignment of these two human cDNAs indicates that this insertion probably represents an aberrantly spliced mRNA, as its 3' end is coincident with a splice junction these authors have identified. Second, three nucleotides are deleted from the Maru et al. (1990) cDNA relative to ours, within positions 1942 -1949. The 3.0 Itk cDNA represents the most abundant species of mature ftk mRNA In order to confirm that the 3.0 kb Itk cDNA represents a mature mRNA, we carried out ribonuclease protection analysis using a series of probes spanning the cDNA
Fig. 1. Sequence of the human Itk cDNA and its predicted protein product. Panel A. The top line is a schematic drawing illustrating the human Itk cDNA and some of the salient features of its protein product. The predicted translational start site (ATG at the far left) and stop codon (TGA) are indicated, along with two putative translational start sites (lower case CTG and ATG) suggested by others based on the analysis of murine cDNA clones. Regions corresponding to the secretory signal sequence (SS), transmembrane domain (TM) and the catalytic protein tyrosine kinase domain (PTK) are shown as rectangles filled with different patterns. The dashed line corresponds to an apparent intron sequence contained in the phage cDNA clones X2 and X3 (the absence of this sequence in X4 is indicated by a V-shaped line). The bottom part of the panel shows the extent of the Itk cDNA phages. EcoRI adaptors are indicated as small open rectangles at the end of the phage clones, while internal EcoRI sites are designated 'RI'. Panel B. The complete nucleotide sequence of the 3.0 kb Itk cDNA, derived from the overlapping phage clones shown in panel A, and the deduced Itk amino acid sequence are shown. In addition, the predicted mouse Itk amino acid sequence of Ben-Neriah and Bauskin (1988) and the divergent 5' end of the mouse cDNA are shown, where applicable, on the two lines below the human amino acid sequence. Identity between the two amino acid sequences is indicated by dots within the murine amino acid sequence, termination codons are indicated by asterisks and a dash is used to indicate a gap in the murine sequence relative to the human sequence (positions 1346-1349). Other key features of the sequence are indicated as follows: a putative upstream mini-protein (heavy underline); putative hydrophobic signal sequence and transmembrane domains (boxed amino acid sequences); nucleotide sequences characteristic of the 3' end of an intron, including (in a 5' to 3' direction): a potential lariat branch point, and the pyrimidine-rich stretch and AG dinucleotide characteristic of the 3' splice site (light underlines); CTG and ATG codons proposed as translational start sites in the mouse cDNA clones (heavy overlines), PTK domain (borders identified by downward arrowheads) and the putative polyadenylation signal (light overline). These sequence data are available from EMBL/GenBank/DDBJ under accession number X60702.
2913
J.J.Krolewski and R.Dalia-Favera TGA
ATG A.C
ks5 f :m:--Ss
IIIIll11 4321-
1
.".. ..". . i
A
PTK
1IIIII1 II
A IA
IIIlI
l 11111111 1111111111111 1111111111I III
i 1it
Li
-1 -2 -3
.E
TM
L
r 1W11 ji1
-
4k1
A
s
1 11 llIII I1 I III
100
200
III I I I'III I II IIi I I Iiii I
300
400
500
600
700
800
Fig. 2. Kyte-Doolittle hydrophobicity plot of the deduced protein sequence of the human ltk cDNA. Across the top of the figure is a schematic diagram, drawn to scale, of the Itk cDNA and its predicted protein structure, with features as described in the legend to Figure 1. Beneath this is a hydrophobicity plot performed according to the method of Kyte and Doolittle (1982). The numbering of the abscissa corresponds to the amino acid position of the deduced Itk protein (Figure 1) and is drawn to the same scale as the schematic diagram, thus indicating the relative hydrophobicity of various predicted structural domains. The ordinate is a relative scale of hydrophobicity according to Kyte and Doolittle (1982). We are unable to ascribe any functional significance to the two hydrophobic regions located between the signal sequence and the transmembrane domain.
(Figure 3). In each case, an antisense RNA probe consisting of cloned cDNA and flanking plasmid sequences was hybridized with RNA from each of two cell lines: the neuroblastoma cell line SK-N-MC and CLL, created by Epstein-Barr virus immortalization of B cells from a patient with chronic lymphocytic leukemia. If the cloned cDNA represents mature mRNA, the expected protected band will be equal in size to the cDNA sequences contained in the probe. Figure 3 shows that this pattern is observed for probes B, C and D in both cell lines. In all cases ltk expression is considerably higher in SK-N-MC cells. Since probe C spans the region which contains a 180 nt insert in two of the original phage clones (Figure 1, panel A), the data also indicate that this insert is not a part of the mature cDNA and probably represents an unspliced intron. However, since it contains an open reading frame in register with the Itk reading frame, it remains a possibility that this insert is an infrequently used alternative exon. In the case of probe A, a number of minor bands and two major fragments of 160 nt and 140 nt are protected. The major bands are about 30 and 50 nt shorter, respectively, than the cDNA sequence in the probe. While this might represent a complicated pattern of alternative splicing events within the 5' end of our cDNA, the most likely explanation for the observed results is that the cDNA overlaps the two major transcriptional initiation sites for the ltk gene. In this case, the minor bands would likely represent additional, infrequent sites of initiation. Finally, although the protection analysis was not performed in a rigorously quantitative manner, the fragments protected by probes B -D are of similar intensity, suggesting that the 3.0 kb ltk cDNA is the predominant form of this message. Other Itk mRNAs which do not contain the 5' sequences detected by probe B, but do contain the 3' sequences detected by probe D, -
2914
apparently constitute only a minor portion of the ltk mRNA population. This conclusion is further strengthened by our identification of only a single, 3.1 kb ltk mRNA species in Northern blots of SK-N-MC cells (data not shown) which is identical in size to the mRNA previously observed in lymphoid cell lines (Krolewski et al., 1990). In vitro translation of the Itk cDNA produces a 100 kDa protein product Next, we sought to demonstrate that the ATG start codon and the accompanying open reading frame produced an Itk protein of the expected size, by constructing a plasmid template containing the 3.0 kb ltk cDNA (pB-LTK-3.0), transcribing the template in vitro, and then translating the resulting RNA in vitro. Lane A on the left side of Figure 4 shows that a single protein product is produced with an apparent molecular weight of 100 kDa, consistent with the use of the ATG and open reading frame identified in Figure 1. As a control, we constructed a 5' truncated version of the full length cDNA (pB-LTK-AX) which represents, approximately, the human analogue of the 2.2 kb murine cDNA isolated by Ben-Neriah and Bauskin (1988) and Bernards and de la Monte (1990) (Figure 4). This truncated template directed the synthesis of a 61 kDa product (lane A on the right side of Figure 4), essentially confirming the results of Bernards and de la Monte (1990) that the CTG is the site of translational initiation in this putative mRNA. Further, since the 61 kDa product is absent when the 3.0 kb template is used, these results indicate that the ATG start codon in the putative full length cDNA effectively suppresses translational initiation at this downstream CTG. This observation fits the prediction of Kozak (1987) that the 40S ribosomal subunit binds at the 5' end of an mRNA, scans
The Itk encodes receptor tyrosine kinase
PROBE:
A
RNA:
P
S
C C
T
P
P
S
C T
- C T
310 -_281 -271I
234
-_
t14
_
s16
_
_---
31C
-a-
281 27,
_-w
2 34
a.
U a
1 '8
ATG
TGA
A
. As
A
E
-
TM
A
PTK
B
D
Fig. 3. Ribonuclease protection analysis of the ltk cDNA. Uniformly labeled anti-sense RNA probes (panels A-D) derived from four subclones of the Itk cDNA were hybridized with 50 ,g of total RNA from the cell line SK-N-MC (lanes marked S) or the cell line CLL (lanes marked C), or with 50 sg of yeast tRNA (lanes marked T). Following digestion with a mixture of ribonucleases TI and A, the products were resolved on a 6% urea-polyacrylamide gel and subsequently autoradiographed at -70°C for 5 days with an intensifying screen. In each panel lane P contains a small amount of the undigested probe. The size of 32P-labeled HaeIlH fragments of 4X-174 bacteriophage DNA are indicated, in nt, along the sides of panels A and D. The bottom part of the figure contains a reproduction of the schematic diagram from Figure 1 illustrating the Itk cDNA and its predicted protein product, along with the location of the cDNA sequences in probes A-D. Note that probe C spans the point where a 180 bp insertion is found in some of the phage clones (see panel A, Figure 1). The sizes of the cDNA sequences in each of the probes (along with their location within the sequence of Figure 1) are as follows: probe A, 189 nt (positions 1- 189 of the Itk cDNA); probe B, 162 nt (positions 190-352); probe C, 190 nt (positions 953-1143) and probe D, 221 nt (positions 2132-2354). The major protected bands produced in the analysis using probe A measure 160 and 140 nt. -
until a 'strong' start codon is encountered, and then initiates protein synthesis. To confirm that the proteins synthesized are bonafide ltk proteins, we prepared an anti-ltk antiserum directed against a synthetic peptide corresponding to the carboxy terminus of the predicted protein product, and used this to immunoprecipitate the products of the in vitro translation reactions. In both cases, the translation product was immunoprecipitated by the antiserum (Figure 4, lanes B) and this interaction was efficiently competed by the addition of an excess of immunizing peptide (lanes C). Our attempts to prepare an antiserum against sequences near the amino terminus of the predicted protein, which would discriminate between the two different sized ltk proteins, were unsuccessful. In vivo expression of the 100 kDa Itk protein kinase in COS- 1 cells Next, we transfected COS-1 cells with derivatives of an
SV40-based eukaryotic expression vector (Wong et al., 1985) containing ltk cDNAs, in order to determine the size
of the ltk protein expressed in vivo and to demonstrate its kinase activity. Figure 5 shows the results of an immune complex kinase assay performed on anti-ltk immunoprecipitates of lysates recovered from these transiently transfected cells. Typically, the formation of such immune complexes activates the kinase activity of the protein, which phosphorylates itself as well as other proteins present in the immunoprecipitate. Cells transfected with a vector expressing the sense orientation of the 3.0 kb cDNA produce a predominant phosphoprotein of 100 kDa and a minor species of 90 kDa. The appearance of these two bands is inhibited by the inclusion of immunizing peptide in the immunoprecipitation reaction. In addition, the bands are not observed when cells are transfected with an antisense version of the construct. Additional phosphorylated species of 75 kDa or less are seen to some extent in all lanes, indicating that they represent phosphorylated immunoglobulins or other co-immunoprecipitating contaminants. The 100 kDa band is most likely the autophosphorylated ltk protein. This is supported by our observation of a protein of the same size in anti-ltk immunoprecipitates of similarly -
-
2915
J.J.Krolewski and R.Dalla-Favera
4'''o 10
.-
N&&11,21S
..
.
A.
77vf
--ooofidk.
Fig. 4. In vitro translation of RNAs transcribed from ltk cDNA templates. A schematic at the bottom of the figure illustrates the structure of the Itk cDNA templates and the corresponding sense orientation in vitro transcripts. The ltk cDNA and its predicted protein domains are as indicated in Figure 1; in addition, the position of the sequence corresponding to the synthetic peptide (C-peptide) used to generate the anti-ltk antiserum is shown. These transcripts were translated in vitro in the presence of [35S]methionine, and the protein products were immnoprecipitated, boiled, electrophoresed on a 10% SDS-polyacrylamide gel and autoradiographed. The template plasmids used to synthesize the RNA transcripts are indicated across the top of the autoradiograph; below, the lanes (for both templates) contain the following: lane A, in vitro translation products (not immunoprecipitated); lane B, translation products immunoprecipitated with 1:100 dilution of the anti-ltk antiserum; lane C, translation products plus synthetic peptide (200 14M) immunoprecipated as in lane B; lane D, translation products immunoprecipitated with a 1:500 dilution of an irrelevant antiserum (anti-actin). The position of co-electrophoresed, prestained molecular weight markers is shown on the right side of the figure. Immunoprecipitations (lanes B-D) were performed with 5-fold more translation reaction mixture than was loaded in lane A. Two- to three-fold more RNA was used in the pB-LTK-AX translation mixture than in the pB-LTK-3.0 reaction; however, it is apparent from the results that the CTG codon is more efficient at initiating translation than the ATG in the full length construct. Minor bands seen in lanes A and B are apparently ltk related as they are efficiently competed with the immunizing peptide.
transfected cells which were metabolically labeled with [35S]methionine (data not shown). Further, if the metabolic labeling is carried out in the presence of tunicamycin, which blocks N-linked glycosylation, a 97 kDa band is seen in place of the 100 kDa one (data not shown). Taken together, these data suggest that, in vivo, the 3.0 kb ltk cDNA encodes a glycoprotein with an apparent molecular weight of 100 kDa. We are not sure of the origin of the 90 kDa minor band. One possibility is that it represents a processed version of the 100 kDa form. Other members of the insulin receptor subfamily of PTKs (Hanks et al., 1988) including the insulin and insulin-like growth factor I receptors (Ullrich et al., 1986), the met proto-oncogene product (Giordano et al., 1989) and the Drosophila sevenless gene product (Simon et al. 1989) are composed of two polypeptide chains, linked by disulfide bonds, generated by the cleavage of the primary translation product. By analogy, the ltk product may be similarly processed in a reaction which is carried out inefficiently in the transfected COS-1 cells. 2916
We also carried out the same transfection and immune complex kinase assay experiments with our ltk construct resembling the 2.2 kb murine cDNAs. In this case a 78 kDa phosphoprotein was observed, which is somewhat larger than a 69 kDa protein identified by Bernards and de la Monte (1990) in a similar set of transfection experiments using the murine Itk cDNA. The absence of the 78 kDa band in samples from the cells transfected with 3.0 kb cDNA supports our in vitro translation data (Figure 4), indicating that the ATG at position 171 in Figure 1 is the only initiation codon utilized in vivo. -
Discussion We have cloned and sequenced a 3.0 kb ltk cDNA which encodes a receptor-type protein tyrosine kinase containing the key structural and functional domains that define this class of proteins: a leader signal sequence, an extracellular domain, a hydrophobic membrane spanning domain and a
The Itk encodes receptor tyrosine kinase
CY -1'1*
construct
peptide
+
+
x* +
Fig 5. Immune complex kinase assay of lysates from COS-1 cells transfected with ltk cDNAs. COS-1 cells were transfected with the SV40-based expression vector p91023, or derivatives containing the following coding sequence constructs: the 3.0 kb ltk cDNA in the sense orientation (Itk 3.0 S); the 3.0 kb cDNA in the antisense orientation (ltk 3.0 AS); a 2.2 kb ltk cDNA, beginning at the XnaIII site at position 862 in Figure 1 and designed to resemble the previously isolated murine cDNAs (see text) (ltk AX S); and the antisense version of the same 2.2 kb cDNA (Itk AX AS). Lysates from cells containing the various constructs were immunoprecipitated with anti-ltk antiserum in the presence (+) or absence (-) of the immunizing peptide. The immunoprecipitates were incubated with [-y-32P]ATP, boiled and electrophoresed on a 10% SDS-polyacrylamide gel. The gel was fixed, soaked in KOH to hydrolyze most serine and threonine phosphates, dried and autoradiographed for 1 h at -70°C with an intensifying screen. The position of co-electrophoresed pre-stained molecular weight standards is shown on the right. The -65 kDa band can be seen in all lanes upon longer exposure of the gel.
catalytic tyrosine kinase domain. The sequences in the cDNA appear to be present in the mature mRNA and the expression of the cDNA, in vitro as well as in vivo, results in the biosynthesis of an ltk specific protein of 100 kDa, consistent with the utilization of the large open reading frame predicted from the cDNA sequence. Additionally, we demonstrated ltk specific kinase activity in immunoprecipitates from cells transfected with Itk cDNAs. We have not provided definitive proof, in the form of phosphoamino acid analysis, that the kinase activity is directed at tyrosine residues. However, the predicted amino acid sequence contains the hallmarks which differentiate tyrosine from serine/threonine kinases (Hanks et al., 1988) and the phosphoproteins produced were resistant to alkali, strongly suggesting that ltk is a tyrosine kinase. The putative extracellular domain does not show homology with any known protein, nor does it appear to contain the cysteinerich or immunoglobulin-like motifs present in some members of the receptor class of PTKs (Ullrich and Schlessinger, 1990). It does, though, seem to be sufficiently large to act as a receptor for a polypeptide ligand, as evidenced by the
fact that the keratinocyte growth factor binds to a smaller (-240 amino acid) extracellular domain contained in a recently described receptor PTK (Miki et al., 1991). In contrast to the protein deduced from our 3.0 kb cDNA, the protein which is initiated at the CTG of the 2.2 kb murine ltk cDNA lacks a signal sequence. While this seems to rule out a role for this putative ltk protein as a cell surface receptor, it could conceivably function intracellularly. For instance, the cDNAs for similar receptor-like PTKs lacking the signal sequence have been identified among the multiple forms of the fibroblast growth factor receptor found in hepatocytes (Hou et al., 1991). Most of our evidence supports the notion that the 2.2 kb cDNA of Ben-Neriah and Bauskin (1988) and Bernards and de la Monte (1990) is a short or aberrant clone. First, the reading frame of the shorter murine cDNA remains open and homologous to the human sequence 35 amino acids upstream of the CTG start codon identified by Bernarda and de la Monte (1990), until an inframe stop codon is encountered. Then, at this point, where the two sequences diverge completely, the sequences characteristic of the 3' end of an intron can be identified in the murine cDNA. Second, ribonuclease protection data indicate that the 5' end of our ltk cDNA overlaps the transcriptional start sites. Thus, if the 2.2 kb murine cDNA represents mature ltk mRNA sequences, then either the cDNA is short or a separate promoter must exist for this mRNA. Third, we have thus far observed only a 3.1 kb ltk mRNA. In this regard, however, it should be noted that other groups have observed two Itk mRNAs, of 3.0 and -2.5 kb, supporting the possibility that there are two alternatively spliced ltk mRNAs. Our attempts to clarify this issue completely, by identifying the native ltk protein(s), have been unsuccessful. The definitive experimental proof of the structure of the ltk cDNA(s) and protein(s) awaits the identification of cells expressing higher levels of this protein and/or technical improvements in out detection methods. Keeping these caveats in mind, we believe that our data establish the existence of a cell surface receptor form of the ltk protein tyrosine kinase. The availability of an ltk cDNA encoding a putative cell surface receptor provides us with a tool to search for the putative ligand of this receptor-type PTK. We are currently attempting to confirm preliminary findings that ltk expression is restricted to hematopoietic and neural tissues (Bernards and de la Monte, 1990; Maru et al., 1990). If we can identify tissues in which ltk functions, we may be able to exploit the autocrine transforming activity of receptor tyrosine kinases and their growth factor ligands (Miki et al., 1991) as a means of identifying the ltk ligand. -
Materials and methods cDNA library construction and screening Double stranded cDNA was prepared by the method of Gubler and Hoffman (1983) using poly(A)+ selected RNA (Aviv and Leder, 1972) from the human neuroblastoma cell line SK-N-MC (American Type Culture Collection) and an Itk gene specific synthetic oligonucleotide primer corresponding to the reverse complement of positions 1876-1910 of the sequence in Figure 1. After the addition of EcoRI adaptors (Promega) the cDNA was ligated into the EcoRI site of XgtlO and packaged in vitro and the resulting phage plaques were screened with a restriction fragment from the 5' end of the previously isolated clone Xl (Krolewski et al., 1990) (see panel A, Figure 1) essentially as described (Firmbach-Kraft et al., 1990). A second gene specific cDNA library was prepared in the same manner, using a primer corresponding to the reverse complement of positions 1111 - 1145 followed by screening with a probe from the 5' end of X2.
2917
J.J.Krolewski and R.Dalla-Favera DNA subcloning, nucleotide sequencing and sequence analysis Phage insert fragments, and derivative subfragments, prepared by digestion with appropriate restriction enzymes and purification on low melting point agarose gels, were cloned into the Bluescript-KS(+) plasmid vector. Plasmid DNA was prepared by the alkaline lysis/polyethylene glycol procedure (Sambrook et al., 1989), sequencing reactions were performed using the Sequenase kit (United States Biochemical) and the resulting products were electrophoresed on 6% urea-polyacrylamide sequencing gels, fixed, dried and autoradiographed. Overlapping subclones were employed to sequence the entire cDNA in both directions, at least once. Occasional areas of band compression were resolved by sequencing with 7-deaza-guanosine derivatives. Nucleotide sequence data were collected and analyzed using the Microgenie software package (Queen and Korn, 1984). Hydrophobicity analysis, according to the algorithm of Kyte and Doolittle (1982), was performed using the DNA Strider software package (Marck, 1988) and a search for secretory signal sequences was done by the method of von Heijne (1986) using release 6.01 of the PC/Gene software package (Intelligenetics).
microcentrifuge and washed with 150 mM NaCI/50 mM Tris-HCI pH 7.4/5 mM EDTA/0.5% Triton X-100. To perform the immune complex kinase assay, the immunoprecipitate was washed further in a high salt buffer (500 mM NaCl) and then with 20 mM HEPES pH 7.4/15 mM MnCl2/0. 1% NP40. Following the addition of [-y-32P]ATP (25 ItCi; 3000 Ci/mmol) the immunoprecipitates were incubated for 30 min at 25°C and then washed for a final time in 20 mM HEPES pH 7.4. SDS sample buffer (with 2-mercaptoethanol) was added, and the immunoprecipitates were boiled, electrophoresed on 10% SDS-polyacrylamide gels, fixed, dried and autoradiographed (Harlow and Lane, 1988). Prior to drying, gels containing [35S]methionine were treated with the fluor En3Hance (New England Nuclear), while gels containing samples from the immune kinase assay were soaked in 1 N KOH at 55 °C to hydrolyze most of the phosphates on serine and threonine.
Ribonuclease protection analysis Total RNA was prepared from cultured cell lines by the guanidinium isothiocyanate/CsCl method (Chirgwin et al., 1979). Unifonnly 32P-labeled antisense RNA probes were synthesized from appropriate linearized plasmid templates using bacteriophage RNA polymerases as per the manufacturer's instructions (Promega), hybridized to 50 Ag of total RNA and then digested with a mixture of ribonucleases A and Ti. The resulting products were denatured, electrophoresed on a 6% urea-polyacrylamide sequencing gel and autoradiographed for 5 days at -70°C with intensifying screens.
We thank Andre Bernards and Michael Ittman for helpful discussions and Robert Lee for assistance with some of the DNA sequencing. We also acknowledge Ulla Beauchamp and Nick Pileggi for the preparation of synthetic oligonucleotides and peptides, respectively. The anti-human actin antiserum was a gift of Chloe Bulinski and the cell line CLL was provided by Giorgio Inghirami. J.J.K. is a Special Fellow of the Leukemia Society of America. This work is partially supported by NIH grant CA44029 to
Antiserum production A synthetic peptide corresponding to an amino-terminal cysteine residue followed by the 11 amino acids at the carboxy terminus of the predicted Itk protein sequence (Figure 1, panel B) was linked to keyhole limpet hemocyanin (Calbiochem), using the crosslinking reagent m-maleimidobenzoic acid-N-hydroxysuccinimide ester (Pierce Chemicals), essentially as described (Doolittle, 1986). The resulting conjugate was used to immunize rabbits (Pocono Rabbit Farms). Serum from an 8 week bleed was used for the experiments shown.
References
In vitro transcription and translation Plasmid subclones from the original phage isolates were joined together in a series of ligation reactions to produce a 3.0 kb clone, corresponding to the entire ltk cDNA in Figure 1, in the Bluescript-KS(+) plasmid vector. Sense orientation RNA transcripts were synthesized in vitro off a linearized version of this full length template plasmid (pB-LTK-3.0) or a 5' truncated version beginning at the XnaIII restriction site at position 862 (pB-LTKAX), using bacteriophage RNA polymerases as per the manufacturer's instructions (BRL). To produce molecules more closely resembling eukaryotic mRNAs, and thereby enhance the translational efficiency, the synthetic cap analogue 7-methyl-guanosine-P3-5' adenosine triphosphate (Boehringer Mannheim) was included in the transcription reaction mixture. The RNA was partially purified and 1-3 ytg was used to program protein synthesis in a rabbit reticulocyte lysate (Promega) in the presence of [ 5S]methionine, as per the supplier's instructions.
COS- 1 monkey cell transfection A derivative of the ltk cDNA containing the putative coding region and some of the flanking untranslated regions, which extends from nucleotide position 126 to 2622 (see Figure 1), was linked to EcoRl adaptors and cloned into the EcoRl site of the eukaryotic plasmid expression vector p91023(B) (Wong et al., 1985). A second derivative, extending from the XmaIII site at position 862 to position 2622 was similarly constructed, and these two plasmids, along with their respective antisense versions and a vector control, were introduced into subconfluent cultures of COS- 1 monkey cells using Lipofectin (BRL) according to the manufacturer's instructions. After 3 days the transfected cells were scraped from the dishes, washed and lysed in the presence of 1 % NP-40. Nuclei and cellular debris were pelleted for 20 min in a microcentrifuge and the supernatant was recovered and frozen at -20°.
Immunoprecipitation and immune complex kinase assay Portions of the in vitro translation reaction mixtures or the NP-40 lysates of the transfected COS-1 cells were immunoprecipitated by the addition of anti-ltk antiserum at a dilution of 1:100. In some cases, immunoprecipitation was performed with an anti-actin antiserum (1:500) as a control reaction. The mixture (200 Al) was incubated on ice for 1 h and then 50 y1 of 50 mg/mi slurry of Protein A-sepharose beads was added. After 15 min at 4°C with gentle rocking, the immunoprecipitate was collected in a
2918
Acknowledaements
R.D.-F.
Aviv,H. and Leder,P. (1972) Proc. Natl Acad. Sci. USA, 69, 1408- 1412. Ben-Neriah,Y. and Bauskin,A.R. (1988) Nature, 333, 672-676. Bernards,A. and de la Monte,S.M. (1990) EMBO J., 9, 2279-2287. Bottaro,D.P., Rubin,J.S., Faletto,D.L., Chan,A.M.-L., Kmiecik,T.E., Vande Woude,G.F. and Aaronson,S.A. (1991) Science, 251, 802-804. Cantley,L.C., Auger,K.R., Carpenter,C., Duckworth,B., Graziani,A., Kapeller,R. and Soltoff,S. (1991) Cell, 64, 281-302. Chirgwin,J.M., Przybyla,A.E., MacDonald,R.J. and Rutter,W.J. (1979) Biochemistry, 18, 5294-5499. Doolittle,R.F. (1986) Of URFs and ORFs: a Primer on How to Analyze Denived Amino Acid Sequences. University Science Books, Mill Valley, CA.
Firmbach-Kraft,I., Byers,M.B., Shows,T., Dalla-Favera,R. and Krolewski,J.J. (1990) Oncogene, 5, 1329-1336. Giordano,S., Ponzetto,C., DiRenzo,M.F., Cooper,C.S. and Comoglio,P.M. (1989) Nature, 339, 155-156. Gubler,U. and Hoffman,B.J. (1983) Gene, 25, 263-269. Hanks,S.K., Quinn,A.M. and Hunter,T. (1988) Science, 241, 42-52. Harlow,E. and Lane,D. (1988) Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Hou,J., Kan,M., McKeehan,K., McBride,G., Adams,P. and McKeehan,W.L. (1991) Science, 251, 665-668. Kozak,M. (1987) Nucleic Acids Res., 15, 8125-8148. Krolewski,J.J., Lee,R., Eddy,R., Shows,T.B., and Dalla-Favera,R. (1990) Oncogene, 5, 277-282. Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105-132. Lee,P.L., Johnson,D.E., Cousens,L.S., Fried,V.A. and Williams,L.T. (1989) Science, 245, 57-60. Lewin,B. (1990) Genes IV. 4th ed. Oxford University Press, New York, NY. Lupu,R., Colomer,R., Zugmaier,G., Sarup,J., Shepard,M., Slamon,D. and Lippman,M.E. (1990) Science, 249, 1552-1555. Marck,C. (1988) Nucleic Acids Res., 16, 1829-1836. Marth,J.D., Overell,R.W., Meier,K.E., Krebs,E.G. and Perlmutter,R.M. (1988) Nature, 332, 171-173. Maru,Y., Hirai,H. and Takaku,F. (1990) Oncogene Res., 5, 199-204. Matsui,T., Heidaran,M., Miki,T., Popescu,N., La Rochelle,W., Kraus,M., Pierce,J. and Aaronson,S. (1989) Science, 243, 800-804. Miki,T., Fleming,T.P., Bottaro,D.P., Rubin,J.S., Ron,D. and Aaronson,S.A. (1991) Science, 251, 72-75. Queen,C. and Korn,L.J. (1984) Nucleic Acids Res., 12, 581-599. Resh,M.D. (1990) Oncogene, 5, 1437-1444. Rudd,C.E., Trevillyan,J.M., Dasgupta,J.D., Wong,L.L. and Schlossman,S.F. (1988) Proc. Natl Acad. Sci. USA, 85, 5190-5194. Ruta,M., Burgess,W., Givol,D., Epstein,J., Neiger,N., Kaplow,J., Crumley,G., Dionne,C., Jaye,M. and Schlessinger,J. (1989) Proc. Natl Acad. Sci. USA, 86, 8722-8726. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A
The Itk encodes receptor tyrosine kinase Laboratory Manul. 2nd edn. Cold Spring Harbor Laboratoxy Press, Cold Spring Harbor, NY. Shaw,A.S., Amrein,K.E., Hammond,C., Stern,D.F., Sefton,B.M. and Rose,J.K. (1989) Cell, 59, 627-636. Sherr,C.J., Rettenmier,C.R., Roussel,M.F., Look,T.A. and Stanley,E.R. (1985) Cell, 41, 665-676. Simon,M.A., Bowtell,D.D. and Rubin,G.M. (1989) Proc. Natl Acad. Sci. USA, 86, 8333-8337. Sudol,M. (1991) In Kumar,R. (ed.) 7he Molecular Basis of Human Cancer. Futura Publishing Co., New York, NY, in press. Ullrich,A. and Schlessinger,J. (1990) Cell, 61, 203-212. Ullrich,A. et al. (1984) Nature, 309, 418-422. Ullrich,A. et al. (1985) Nature, 313, 756-761. Ullrich,A. et al. (1986) EMBO J., 5, 2503-2512. Veillette,A., Bookman,M.A., Horak,E.M. and Bolen,J.B. (1988) Cell, 55, 301-308. von Heijne,G. (1986) Nucleic Acids Res., 14, 4683-4690. Witte,O.N. (1990) Cell, 63, 5-6. Wong,G.G. et al. (1985) Science, 228, 810-815. Yarden,Y. et al. (1986) Nature, 323, 226-232. Received on May 3, 1991
2919
