Cell, Vol.

63, 225-233, October 5, 1990, Copyright 0 1990 by Cell Press

The Hematopoietic Growth Factor KL Is Encoded by

the SI Locus and Is the Ligand of the c-kit
Receptor, the Gene Product of the W Locus
Eric Huang: Karl No&a: David R. Beier,t* the transition from BFU-E to CFU-E), and tissue mast cells
Tang-man Chu,’ Jochen Buck,5 Hans-Werner Lahm,ll (Russell, 1970; Kitamura et al., 1978; Nocka et al., 1989).
Daniel Wellner,’ l Philip Leder,t and Peter Besmer’ in gametogenesis c#/Wfunction is required during early
* Molecular Biology Program development as well as in postnatal oogenesis and sper-
§lmmunology Program matogenesis (Mintz and Russell, 1957; Orr-Urtreger et al.,
Sloan Kettering institute 1990; Manovaet al., 1990; V. Sorrentino, personal commu-
and Cornell University Graduate School nication). The defect W mutations exert is cell autono-
of Medical Sciences mous; in agreement with this property, there is evidence
New York, New York 10021 for c-kit RNA expression in cellular targets of W mutations
*Howard Hughes Medical Institute (Nocka et al., 1989; Orr-Urtreger et al., 1990). The recent
and Department of Genetics characterization of the molecular lesions of several mu-
Harvard Medical School tant alleles indicated that they are loss-of-function muta-
Boston, Massachusetts 02115 tions that disrupt the normal activity or expression of the
II F. Hoffmann-LaRoche Ltd. c-kit receptor (Nocka et al., 1989; Tan et al., 1990; Reith
CH-4005 Basel, Switzerland et al., 1990; Nocka et al., 1990a).
l * Department of Biochemistry Mutations at the steel (S/) locus on chromosome 10 of
Cornell University School of Medicine the mouse result in phenotypic characteristics that are
New York, New York 10021 very similar to those seen in mice carrying W mutations,
i.e., they affect hematopoiesis, gametogenesis, and mel-
Summary anogenesis (Bennett, 1956; Russell, 1979; Silvers, 1979).
Many alleles are known at the SI locus; they are semi-
Mutations at the steel locus (SI) of the mouse affect dominant mutations, and the different alleles vary in their
the same cellular targets as mutations at the white effects on the different ceil lineages and their degree of
spotting locus(W), which Is allelic with the c-kit proto- severity (Silvers, 1979; Russell, 1979). The original SI al-
oncogene. We show that KL, a hematopoietic growth lele is a severe mutation. S//S/ homozygotes are deficient
factor obtained from conditioned medlum of BALB/c in germ cells, are devoid of coat pigment, and die perina-
3T3 fibroblasts that stimulates the proliferation of tally of macrocytic anemia (Bennett, 1956; Sarveila and
mast cells and early erythroid progenitors, specifically Russell, 1956). Mice homozygous for the Sld allele, ai-
binds to the c-kit receptor. The predicted amino acid though viable, have severe macrocytic anemia, lack coat
sequence of isolated KGspecific cDNA clones sug- pigment, and are sterile. Both S/l+ and SId/+ heterozy-
gests that KL is synthesized as an integral transmem- gotes have a diluted coat color and a moderate macrocytic
brane protein. Linkage analysis maps the KL gene to anemia but are fertile, although their gonads are reduced
the SI locus on mouse chromosome 10, and KL se- in size. In contrast to Wmutations, SI mutations are not ceil
quences are deleted in the genome of the SI mouse. autonomous and are thought to be caused by a defect in
These results indicate that the SI locus encodes the the microenvironment of the targets of these mutations
ligand of the c-kit receptor, KL. (Mayer and Green, 1966; McCulloch et al., 1965; Dexter
and Moore, 1977). Because of the parallel and com-
Introduction plementary characteristics of mice carrying SI and W mu-
tations, we and others had previously hypothesized that
The proto-oncogene c-kit is the normal cellular counter- the S/gene product is the iigand of the c-kit receptor (Rus-
part of the oncogene v-kit of the HZ4 feline sarcoma virus sell, 1979; Chabot et al., 1988).
(Besmer et al., 1986). c-kit encodes a transmembrane The identification of the iigand for c-kit and the determi-
tyrosine kinase receptor that is a member of the piatelet- nation of its relationship to the murine SI locus is of great
derived growth factor (PDGF) receptor subfamily and is interest because of the pieiotropic effects it has on the
the gene product of the murine white spotting (w) locus different cell types that express c-kit. insight about cell
(Qiu et al., 1986; Yarden et al., 1967; Majumder et al., types that may produce the c-kit iigand had been derived
1968; Chabot et al., 1988; Geissier et al., 1988; Nocka et from understanding c-kit/W as well as SI function in vivo.
al., 1989). W mutations typically affect hematopoiesis, W7wY mice lack both connective tissue and gastrointesti-
melanogenesis, and gametogenesis during embryonic nal mucosa mast cells, indicating a function for c-kit in
development as well as in adult life (Russell, 1979; Silvers, mast cell development. in vitro, bone marrow-derived
1979). The c-kit/W gene functions in immature progenitor mast cells (BMMC) require interleukin 3 (IL-3), and con-
cell populations and in more mature ceil types of the three nective tissue mast cells isolated from the peritoneal
cell lineages. In the hematopoietic system Wmutations af- cavity (CTMC) require both IL-3 and IL-4 for proliferation
fect the stem ceil compartment, erythroid precursors (e.g., in vitro (Yung et al., 1981; Tsuji et al., 1990). Both BMMC
and CTMC can be maintained by coculture with 3T3 fibro-
*Present address: Division of Genetics, Department of Medicine, blasts in the absence of IL-3 (Levi-Schaffer et al., 1986).
Brigham and Women’s Hospital, Boston, Massachusetts 02115. BMMC from mice homozygous for severe W mutations,
 10 20 30

KEIXGNPVT D N V K D I T K L V A N L P N D Y M I T L N Y V A C M XVLP

5' CG TC G C 3'
aagcttCATMTGTMMGACATTACMMClGGlGGCMATCTTCCMATGACTATATGATMCCCTCMTTACGTGGCCGCMTGggatcc
T A C A C

cgccaagcttGATMTGTAAAAGATATTAC 3'
5' CCGGCC 3' TTMTACAGCGGCCGTACcctaggggcc
T A G G T T T 5’
C c c c
A A A

Figure 1. N-Terminal Amino Acid Sequence of KL and Deduction of the Corresponding Nucleic Acid Sequence by PCR
Top line: N-terminal amino acid sequence (residues lO-36)of KL. Middle line: Nucleotide sequences of three cDNAs obtained by cloning the 101
bpPCRproduct(~rightpanel)intoM13andsubsequentsequencedetermination.Bottomline:sequencesofthe degenerate senseandantisense
primers used for first-strand cDNA synthesis and PCR. Right panel: Derivation of cDNAs corresponding to the N-terminal amino acids 10-36 of
KL by RT-PCR. One microgram of pcly(A)+ RNA from BALB/c 3T3 cells was used as template for cDNA synthesis and subsequent PCR amplifica-
tion in combination with the two degenerate oligonucleotide primers. Electrophoretic analysis of the 101 bp PCR product in agarose is shown.

however, do not proliferate upon coculture with fibroblasts

in the absence of IL-3 (Fujita et al., 1988; Tan et al., 1990). ,......_..................._.. *p._........ ~..
MKKTPTYlITCIVLQ 15
These results indicated that there is a function for the c-kit GCGGTGCCTTTCCTTATGAAGAAW\CACAAACTTGGATTATCACTTGCATTTATCTTC~

receptor in mature mast cells and suggested that the

LLLfNPLVKT'KEI&GNPVTD 35
ligand of the c-kit receptor is produced by fibroblasts. CTGCTCCTATTTAATCCTCTCGTCAAAACCAAGGAGATC
CGGGAATCCTGTGACTGAT
We have recently purified a hematopoietic growth factor, NVKDITKLVANLPNOYMITL 55
AATGTAAAAGACATTACACTGGTGGCAAATCTTCCACCCTC
designated KL, from conditioned medium of BALBlc 3T3
fibroblasts, which has the biological properties expected 75

of the c-kit ligand (No&a et al., 1990b). KL was purified DLSLSLTTLLDKFSNISEGL 95

based on its ability to stimulate the proliferation of BMMC CAATTATCACTCAGCTTGACTACTCTTCTGGACAAGTTCTC~TATTTCT~GGCTTG
from normal mice but not from W mutant mice in the ab- SNVSIIDKLGKIVDDLVLCW Q 115
AGTAATTACTCCATCATAGACAAACTTGGGAAAATAGTGGATGACCTCGTGTTA
ATG
sence of IL-3. The purified factor stimulates the prolifera-
EENAPKNIKESPKRPETRSF 135
tion of BMMC and CTMC in the absence of IL-3 and there- GAAGAAAACGCACCGAAGAATATAAAAGAATCTCCGAAGAGGCCAGAAACTAGATCCTTT
fore appears to play an important role in mature mast TPEEFFSIFNRSIDAFKDFM 155
cells. In regards to the anticipated function of c-kit in ACTCCTGAAGAATTCTTTAGTATTTTCAATI\GATCCATTGATGCCTTT~GGACTTTATG

erythropoiesis, KL was shown to facilitate the formation I75

of erythroid bursts (day 7-14 BFU-E) in combination with
RVSVTKPFMLPPVAASSLRN 195
erythropoietin. KL has a molecular mass of 30 kd and a AGAGTCAGTGTCACAAAACCATTTATGTTACCCCCTGTTGCAGCCAGCTCCCTTAGG~T
pl of 3.8; it is not a disulfide-linked dimer, although the DSSSSNRKAAKSPEDSGLDW 215
GACAGCAGTAGCAGTAATAGG~GCCGCAAAGTCCCCT~~CTCGGGCCTAC~TGG
characteristics of KL upon gel filtration indicate the forma-
, . . . .._......._........__.___ T,,S_...................._....
tion of noncovalently linked dimers under physiological TAMALPALISLVIGFAFGAL 235
conditions. ACAGCCATGGCATTGCCGGCTCTCATTTCGCTTGTAATTGGCTTTGCTTTTG~GCCTTA
We have now defined the relationship of KL to the c-kit I
VWKKKDSSLTRAVENIQINE 255
receptor, and we show that KL is the ligand of c-kit based TACTGGAAGAAGAAACAGTCAAGTCTTACMGGGGCAGTTGA
on binding and cross-linking experiments. N-terminal pro- EDNEISMLDDKEREF 270
tein sequence of KL was used to derive KL-specific cDNA GAGGATAATGAGATAAGTATGCTGCAACAGAAAGAGAGAGAATTT

clones. We have used these cDNA clones to investigate

the relationship of the KL gene to the SI locus, and we
demonstrate that KL is encoded by the SI locus. SP YY y y TMS
c i
Results c

Isolation and Characterization of Murine cDNAs

Figure 2. NucleotldeSequence and Predicted Amino Acid Sequence
Encoding the Hematopoietic Growth Factor KL
of the 1.4 kb KL cDNA Clone
The KL protein was purified from conditioned medium
The predicted amino acid sequence ofthelong open reading frame
from BALB/c 3T3 cells by a series of chromatographic isshownabovethenucleotldesequenceusingthesingle-letteramino
steps including anion exchange and reverse-phase HPLC acid code. The numbers at right refer to amino acids, with rnethionine
as described elswhere (No&a et al., 1990b). The se- (nucleotlles 16-18)being number 1. The potential N-terminalsignal
quence of the N-terminal 40 amino acids of KL was deter- sequence (SP)and the transmembrane domain (TMS) are indicated
with dashed lines above the sequence, and cysteine residues in the
mined to be KEIXGNPVTDNVKDITKLVANLPNDYMITL- extracellular domain are circled. A schematic of the predicted protein
NYVAGMXVLP To derive a nondegenerate homologous structureisindiiedbelow. N-linked glycesylatlon sltesandtheloca-
hybridization probe, fully degenerate oligonucleotide prim- tion of the N-terminal peptide sequence (Pep. Seq.) are indicated.
;k$ Ligand Is SI Gene Product

66-
u-

91- I)
l&3-
21.s-

sihw 125
Figure 3. Identification of KL-Specific RNA Transcripts in BALB/c 3T3 Figure 4. SDS-PAGE Analysis of KL
Cell RNA by Northern Blot Analysis
(A) Silver staining of KL. (B) Autoradiography of ‘*%KL.
Poly(A)+ RNA (4 kg) from BALB/c 3T3 cells was electrophoretically
separated, transferred to nitrocellulose, and hybridized with 32P-
labeled 1.4 kb KL cDNA. The migration of 18s and 28s ribosomal
RNAs is indicated. transcript of 6.5 kb and two minor transcripts of 4.6 and 3.5
kb were identified on a blot containing poly(A)+ RNA by
using the 1.4 kb KL cDNA as a probe. Identical transcripts
were detected by using an end-labeled oligonucleotide de
ers corresponding to amino acids 10-16 (sense primer) rived from the N-terminal protein sequence (not shown).
and 31-36 (antisense primer) provided with endonuclease This result then indicates that KL is encoded by a large
recognition sequences at their 5’ ends were synthesized mRNA that is abundantly expressed in BALE/c 3T3 cells.
as indicated in Figure 1A. A cDNA corresponding to the
KL mRNA sequences that specify amino acids 10-36 of The Soluble Form of KL Is a Ligand of the
KL was obtained by using the reverse transcriptase modifi- c-kit Receptor
cation of the polymerase chain reaction (RT-PCR). pOly(A)+ The fibroblast-derived hematopoietic growth factor KL
RNA from BALE/c 3T3 cells was used as template for had been shown to facilitate the proliferation of primary
cDNA synthesis and PCR amplification in combination bone marrow mast cells and peritoneal mast cells and to
with the degenerate oligonucleotide primers. display erythroid burst-promoting activity. To determine if
The amplified DNA fragment was subcloned into M13, KL is the ligand of the c-kit receptor, we first thought to
and the sequences of three inserts were determined. The demonstrate specific binding of KL to cells that express
sequence in between the primers was found to be unique high levels of the c-kit protein: mast cells (BMW) and NW
and to specify the correct amino acid sequence (Figure 1). 1+r2cells expressing the c-kit cDNA. KL was labeled to high
An oligonucleotide (49 nucleotides) corresponding to the specific activity with 1251 by using the modified chlora-
unique sequence of the PCR products was then used to mine T method (Stanley and Guilbert, 1961). Analysis of
screen a hgtll mouse fibroblast cDNA library. A 1.4 kb the labeled material by SDS-PAGE showed a single band
clone was obtained that, in its 3’ half, specifies an open of 26-30 kd (Figure 4), and mast cell proliferation assays
reading frame that extends to the 3’ end of the clone and indicated that the labeled material had retained its bio-
encodes 270 amino acids (Figure 2). The first 25 amino logical activity. Binding of increasing concentrations of
acids of the KL amino acid sequence have the character- ‘*%KL to NIH ~2 cells expressing the c-kit cDNA, NIH ~2
istics of a signal sequence. The N-terminal peptide se- control cells, normal BMMC, and WW, WI+, and WV/WV
quence that had been derived from the purified protein BMMC at 4% was measured. The results shown in Figure
(amino acids 26-65) follows the signal sequence. A hydro- 5 indicated binding of labeled KL to NIH IJ.I~c-kit cells and
phobic sequence of 21 amino acids (residues 217-237) fol- to +I+, WI+, and WY/WY mast cells, but not to NIH ~2
lowed at its carboxyl end by postively charged amino control cells or WV mast cells. The WY mutation is the re-
acids has the features of a transmembrane segment. In sult of a missense mutation in the kinase domain of c-kit
the sequence between the signal peptide and the trans- that impairs the in vitro kinase activity but does not affect
membrane domain, four potential N-linked glycosylation the expression of the c-kit protein on the cell surface
sites and four irregularly spaced cysteines are found. A (Nocka et al., 199Oa). By contrast, W results from a dele-
C-terminal segment of 33 amino acids follows the trans- tion due to a splicing defect that removes the transmem-
membrane segment without reaching a termination signal brane domain of the c-kit protein; the protein therefore is
(end of clone). The KL amino acid sequence therefore has not expressed on the cell surface (Nocka et al., 199Oa).
the features of a transmembrane protein: an N-terminal Furthermore, binding of 1251-KL could be competed with
signal peptide, an extracellular domain, a transmembrane unlabeled KL and with two different anti-c-kit antisera (not
domain, and a C-terminal intracellular segment. shown). These results indicated binding of 1251-labeled
RNA blot analysis was performed to identify KL-specific KL to cells that express c-kit on their cell surface.
RNA transcripts in BALE/c 3T3 cells (Figure 3). A major To obtain more direct evidence that KL is the ligand of
 Cdl
226

A. Figure 5. Binding of l*WKL to Mast Cells and

c-kit-Expressing ~2 Cells

+I+ ,--
(A) NIH qQ/c-kit cells containing
expression vector and expressing
the pLJ c-M
a high level

i
t

,,,f’ /1
.--,~
,‘ey
VN
.’
4”
/
of c-kit protein.
(6) Mast cells derived from bone marrow of

7G‘Or :/,’<’
+I+ or WY/W’ adult mice or fetal liver cells of

/ +/+ W/W or a normal littermate control (W/+ or

,’I +I+).

8 L’ ,,:‘I
5 ,*;
/
I / / / /

wlw
0, 1
0 .4 .0 1.2 -1 .6
KL (nrr) KL (na)

the c-kit receptor, we determined if receptor-ligand com- dicated three species: one at approximately 30 kd, repre-
plexes could be purified by immunoprecipitation with c-kit senting KL coprecipitated but not cross-linked to c-kit; one
antisera. This experiment requires that a KL-c-kit com- at 180-190 kd, corresponding to a covalently linked c-kif-
plex be stable and not be affected by the detergents used KL monomer-monomer complex; and a high molecular
for the solubilization of the c-kit receptor. Precedent for weight structure that is at the interface between the sepa-
such properties of receptor-ligand complexes derives rating and stacking gels (Figure 6B). Molecular structures
from the closely related macrophage colony&imulating of similar size were observed if the cell lysates were sepa-
factor (CSF-1) receptor and PDGF receptor systems (Sherr rated directly on SDS-PAGE without prior immunoprecipi-
et al., 1985; Heldin et al., 1989). 1251-KL was bound to tation (not shown). Following precipitation with nonim-
receptors on BMMC by incubation at 4%. Upon washing mune serum, no l~l-labeled molecules were observed.
to remove free 1251-KL, the cells were solubilized by us- The formation of the high molecular weight structures was
ing the Triton X-100 lysis procedure and precipitated with dependent on the incubation of KL with mast cells and
anti-v-kit and anti-c-kit rabbit sera conjugated to protein was not observed by cross-linking KL with itself. Taken to-
A-Sepharose. ‘25l-KL was retained in immunoprecipi- gether, these results provide evidence that KL specifically
tates obtained by incubation with anti-kit sera but not with binds to the c-kit receptor and is a ligand of c-kit.
nonimmune controls, as shown by the analysis of the im-
mune complexes by SDS-PAGE (Figure 6A), where recov- Mapping of KL to the SI Locus
ery of intact 1251-KL was demonstrated from the samples To test whether KL is encoded at the SI locus, we used
containing the immune complexes prepared with anti-kit recombination analysis to determine the map position of
sera. KL with respect to a locus that is tightly linked to SI. This
To further characterize the c-kit-KL receptor-ligand com- locus is the site of the transgene insertion in the trans-
plexes, we determined whether KL could be cross-linked genie line TG.EB (Schmidt et al., 1988). We have deter-
to c-kit. BMMC were incubated with ‘*%KL, washed, and mined that genomic sequences cloned from the insertion
treated with the cross-linker disuccinimidyl suberate. Cell site map 0.8 f 0.8 CM from SI (D. Ft. B. and I? L., submit-
lysates were then immunoprecipitated with anti-v-kit an- ted). This therefore represents the closest known marker
tiserum and analyzed by SDS-PAGE. Autoradiography in- to SI.

A. 0. Figure 6. Coprecipitation and Cross-Linking

of 1251-KL with the c-kit Receptor on Mast Cells
!.f T 4 12% 7.6% (A) Coprecipitation of KL with normal rabbit se-
m rum (NM) or two anti-c-kit rabbit antisera (a-v-
la _.-_ ___-
kit and a-c-kit).
bb- mo- -KL+oK (B) Cross-linking of KL to c-kit with disuc-
43- 116 - mc-
-KL+cK cinimidyl suberate. SDS-PAGE analysis was
WI-
31 - ss-
on either 12% or 7.5% polyacrylamide gels.
116 - Cross-linked species are labeled “KL + cK.”
21.5- 07 -
43-
~2: Ligand Is SI Gene Product

Table 1. Mapping of the Position of the KL Gene by Linkage Analysis

Using an Interspecific cross

Progeny

Probe Nonrecombinant Recombinant A.

1.4 kb KL cDNA 66 SP B6 SP
TIS DralSal B6 SP SP 66

32 20 0 1

n = 53 % recombination = 1.9 2 1.9

The concordance of inheritance of C57BlW (B6) or M. spretus (Sp)

alleles in progeny of an interspecific cross (see Experimental Proce-
dures) was determined by scoring for Taql RFLPs of the KL 1.4 kb
cDNA probe and TIS DraBa/ (a probe from a transgene insertion site
that is tightly linked to SI; see Results). Percent recombination was
0.
calculated according to Green (1961).

To map KL with respect to the transgene insertion site,

we employed interspecific mapping analysis utilizing
crosses of C57BLEJ mice with mice of the species Mus
spretus. This strategy exploits the observation that restric-
tion fragment length polymorphisms (RFLPs) for cloned Figure 7. RFLP Analysis of Taql-Digested DNA from SW+ and S//S/
DNA are observed much more frequently between mice Mice
of different species than between different inbred labora- The S/allele from C3HeB/FeJ a/a Cd S/ /-/m mice was introduced into
tory strains (Avner et al., 1988). Linkage between the 1.4 a C57BU6J background, and progeny of a C57BLW SW x S?
kb KL cDNA probe and TIS D&Sal, a probe from the cross were evaluated.
(A) Hybridization of the 1.4 kb KL cDNA probe to DNA from two non-
transgene insertion site, was assessed by scoring for con- anemic (lanes S/I+) and two anemic (lanes S//S/j mice. No hybridiza-
cordance of inheritance of their respective C57BU8J or M. tion to the DNA from the S//S/ mice was detected.
spretus alleles. These could be easily distinguished by (B) Hybridization of the same blot to TIS D&Sal, a probe that is tightly
analyzing RFLPs that are revealed by Taql restriction linked to S/(see text). This probe identifies a 4 kb CSHeB/FeJ-derived
digests (not shown). The results of this linkage analysis allele and a 2 kb C57BU6J allele in the SIcV+ heterozygotes and
only the CBHeB/FeJ-derived allele in the SIc3H/S/c3H homozygotes.
are shown in Table 1. Only one recombinant was found in
53 progeny. This corresponds to a recombination percent-
age of 1.9 f 1.9. Since this value is very close to the
genetic distance measured between the transgene inser- Discussion
tion site and SI, this result is consistent with the notion that
KL maps to the SI locus. The discovery of allelism between the c-kit proto-onco-
We also examined the locus identified by KL in mice that gene and the murine W locus revealed the pleiotropic
carry the original S/mutation (Sarvella and Russell, 1958). functions of the c-kit receptor in development and in the
For this purpose, we took advantage of our observation adult animal. Furthermore, it provided the first genetic
that the transgene insertion site locus is polymorphic in system of a transmembrane tyrosine kinase receptor in a
inbred strains, and can be utilized to determine the geno- mammal. Mutations at the S/locus and at the c-kit/Wlocus
type at S/during fetal development (D. Ft. 8. and P L., sub- affect the same cellular targets. Because of the comple-
mitted). C57BU8J mice that carry the SI mutation main- mentary and parallel properties of these mutations, we
tained in the C3HeB/FeJ strain were generated by mating, had proposed that the ligand of the c-kit receptor is en-
and Fl progeny carrying the SI allele were intercrossed coded by the SI locus.
(C57BU8J SlcV+ x SIc3HI+). Homozygous SIISI prog- The experiments reported here provide evidence that
eny from this mating are anemic and are homozygous for the SI gene encodes the ligand of the c-kit receptor. The
a C3HeB/FeJ-derived RFLP at the transgene integration evidence for this conclusion is as follows. Based on the
site (Figure 7). Nonanemic mice are either heterozygous knowledge of the function of the c-kit receptor in mast cells
S/l+ or wild type, and are heterozygous for the C3HeB/ and in erythropoiesis, we had identified and purified a
FeJ- and C57BU8J-derived polymorphisms or are homo- putative ligand of the c-kit receptor designated KL (No&a
zygous for the C57BU8J polymorphism, respectively. et al., 1990b). We have now demonstrated specific binding
When genomic DNA from S/l+ and S/IS/ mice was ana- of KL to the c-kit receptor, as evidenced by the binding of
lyzed using the 1.4 kb KL cDNA probe, no hybridization to KL to cells expressing a functional c-kit receptor and the
the homzygous S//S/ DNA was observed (Figure 7). It thus formation of a stable complex between KL and the c-kit
appears that the locus that encodes the KL protein is protein. We have derived KL-specific cDNA clones and
deleted in the SI mutation. This finding further supports shown that KL maps to the SI locus on mouse chromo-
the notion that KL is the product of the SI gene. some 10. In addition, we demonstrated that KL sequences
are deleted in the genome of the SI mouse. Taken together, IL-4, which are thought to be mediators of allergic and in-
these results suggest that KL is encoded by the SI locus flammatory responses (Stevens and Austen, 1989). In the
and is the ligand of the c-kit receptor, thus providing a mo- stem cell compartment the affected populations possibly
lecular basis for the SI defect. include the spleen colony-forming units (CFU-S), which
The amino acid sequence predicted from the nucleotide produce myeloid colonies in the spleen of lethally irradi-
sequence of the KL cDNA clone suggests that KL is syn- ated mice, as well as cells with long-term repopulation
thesized as an integral transmembrane protein. The struc- potential for the various cell lineages (McCulloch et al.,
tural features of the primary translation product of KL 1984; Russell, 1979; Harrison, 1980; Barker and McFarland,
therefore are akin to those of CSFl. CSF-1 is synthesized 1988). It will now be of interest to determine the effect of
as a transmembrane molecule, which is processed by pro- KL on the self-renewal or the differentiation potential of he-
teolytic cleavage to form asoluble product that is secreted matopoietic stem cell populations in vitro, possibly in com-
(Kawasaki et al., 1985; Rettenmier and Roussel, 1988). bination with other hematopoietic growth factors, in order
Presumably, like CSF-1, KL is also synthesized as a cell to identify the stage(s) where c-/rit/KL functions in stem
surface molecule that may be processed to form a soluble cells. Another possible function for c-kit might be to facili-
protein. The protein we had purified from conditioned tate the transition from noncycling to cycling cells (McCul-
medium of BALE/c 3T3 cells then would represent the loch, 1970). The increased radiation sensitivity of MSP
soluble form of KL that was released from the cell mem- and of WTWYmice might suggest such a role in stem cell
brane form by proteolytic cleavage. Although the post- dynamics; furthermore, the related PDGF receptor is known
translational processing and expression of the KL protein to promote entry into the cell cycle.
have not yet been characterized, a cell surface-bound In gametogenesis the W and SI mutations affect the
form of KL may mediate the cell-cell interactions pro- proliferation and the survival of primordial germ cells, and
posed for the proliferative and migratory functions of the their migration from the yolk sac splanchnopleure to the
c-kit/W receptor system. In agreement with the notion of genital ridges during early development. In postnatal gam-
a cell membrane-associated form of KL, a soluble c-kit etogenesis c-kit expression has been detected in imma-
receptor-alkaline phosphatase fusion protein has been ture and mature oocytes and in spermatogonia A and B
shown to bind to the cell surface of BALBlc 3T3 cells but as well as in interstitial tissue (Orr-Urtreger et al., 1990;
not to fibroblasts derived from S//S/ mice (Flanagan and K. Manova, K. N., I? B., and Ft. Bachvarova, unpublished
Leder, 1990). data; V. Sorrentino, personal communication). In melano-
A most significant aspect of the identification of the genesis c-kitlKL presumably functions in the proliferation
ligand of the c-kit receptor lies in the fact that it will facili- and migration of melanoblasts from the neural crest to the
tate the investigation of the pleiotropic functions of c-kit. periphery in early development as well as in mature mela-
In the hematopoietic system c-/&/W mutations affect the nocytes. The availability of KL may now facilitate in vitro
erythroid and mast cell lineages, and an effect on the stem studies of the function of the c-kit receptor in these cell
cell compartment has been inferred as well. In erythroid systems.
cell maturation c-kitlKL plays an essential role, and this is The microenvironment in which c-/&expressing cells
best seen by the anemia of mutant animals. Furthermore, function is defective in SI mutant mice and is the pre-
the number of CFU-E in fetal livers from W&V and S//SP sumed site where the c-kit ligand is produced. Because
animals is repressed, whereas the number of BFU-E re- of the extrinsic nature of the mutation, the precise identity
mains normal, suggesting that c-IritlKL facilitates the pro- of the cell types that produce KL in vivo is not known. In
gression from the BFU-E to the CFU-E stage of differentia- vitro systems that reproduce the genetic defect of the W
tion (Chui et al., 1978; Nocka et al., 1989). In this regard, and the SI mutations, however, have shed some light on
KL has been shown to stimulate the proliferation and dif- this question. In the long-term bone marrow culture sys-
ferentiation of BFU-E (day 7) as well as earlier erythroid tem, S//SP adherent cells are defective but the nonadher-
multipotential precursors in bone marrow, which appear at ent hematopoietic cells are not, and in the mast cell-fibro-
later times in culture (day 14-20) (Nocka et al., 1990b; blast coculture system S//SP fibroblasts are defective but
K. N., unpublished data). the mast cells are not (Dexter and Moore, 1977; Fujita et
An essential role for c-kitlKL in the proliferation, differ- al., 1989). The results from these in vitro systems then
entiation, and/or survival of mast cells in vivo has been in- would suggest that hematopoietic stromal cells and em-
ferred because of the absence of mast cells in W and SI bryonic and connective tissue fibroblasts produce KL.
mutant mice (Kitamura et al., 1978; Kitamura and Go, The BALBlc 3T3 cell line, which is of embryonic origin, ex-
1978; Kitamura and Fujita, 1989). The precise stage(s) at presses significant levels of KL and was the source for its
which c-kitlKL function is required in mast cell differentia- purification. Knowledge of KL-expressing cell types may
tion is not known. The in vitro derivation of BMMC from help to evaluate if there is a function for c-kit in the diges-
bone marrow or fetal liver does not require c-kif/KL func- tive tract, the nervous system, the placenta, and certain
tion since BMMC can be generated with comparable effi- craniofacial structures, sites where c-kit expression has
ciency from both normal and W mutant mice (Yung and been documented (No&a et al., 1989; Orr-Urtreger et al.,
Moore, 1982). Our demonstration of proliferation of BMMC 1990; Manova and Bachvarova, personal communica-
and connective tissue-type mast cells in response to KL tion). No SI or W phenotypes are known to be associated
indicates a role for c-kitlKL at multiple stages in mast cell with these cell systems.
proliferation and differentiation independent of IL-3 and Interspecific backcrosses were used to establish close
c-kit Ligand Is SI Gene Product
231

linkage between the KL gene, the SI locus, and the trans- Yung and Moore, 1962). BALB/c 3T3 cells (Aaronson and Todaro, 1966)
gene insertion locus Tg.fB on mouse chromosome 10. A were obtained from Paul O’Donnell (Sloan Kettering Institute) and were
grown in Dulbecco’s modified MEM supplemented with 10% calf se-
similar approach had previously been used to map the rum, penicillin, and streptomycin.
Tg.fB locus in the vicinity of S/(D. Ft. 8. and l? L., submit-
ted). The finding that the KL coding sequences are de- Purification and Amino Acid Sequence Determination of KL
leted in the original SI allele supports the identity of the KL was purified from conditioned medium of BALB/c 3T3 ceils by using
a mast cell proliferation assay as described elsewhere (NO&I et al.,
SI locus with the KL gene. The size of the deletion in the 1990b). Conditioned medium was then concentrated 100. to 200-fold
SI allele at this time is not known. It will be important to with a Peilicon ultrafiltration apparatus followed by an Amicon stirred
determine whether it affects neighboring genes as well. ceil. The concentrate was then chromatographed on Blue Agarose
The lack of KL coding sequences in the SI allele indi- (Bethesda Research Laboratories, Gaithersburg, MD), and the fiow-
through, which contained the active material, wasconcentrated in dial-
cates that this allele is a KL null mutation. When homozy-
ysis tubing with polyethylene glycol 6000 and then fractionated by gel
gous for the SI allele, most mice die perinatally of macro- filtration chromatography on an A0154 Ultrogel (LKB, Rockland, MD)
cytic anemia, and rare survivors lack coat pigmentation column. The biological activity eiuted as a major and a minor peak, cor-
and are devoid of germ cells (Bennett, 1956). This pheno- responding to 55-70 kd and 30 kd, respectively. The fractions of the
type closely parallels that of severe c-M/W loss-of-func- main peak were pooled, dialyzed, and fractionated by FPLC on a
DEAEdPW column with an NaCl gradient. The activityeluted at 0.11 M
tion mutations, in agreement with the ligand-receptor NaCl from the FPLC column. Peak fractions were pooled and sub-
relationship of KL and c-kit. Although differences exist be- jected to HPLC with a semipreparative Cl8 column and an ammonium
tween SIISI and IYW homozygotes, e.g., in germ cell de- acetate-n-propanol gradient. The active material eluted at 30% n-pro-
velopment, SI may have a more pronounced effect, and in panol from the semipreparative Cl6 column was diluted I:1 and re-
hematopoiesis SI may cause a more severe anemia; how- chromatographed by using an analytical Cl6 column. A single peak of
activity eiuted again at 30% n-propanol, which corresponded to a ma-
ever, it is not known if these differences are a result of jor peak of absorbance (280 nm) in the eluant profile. Similar results
different strain backgrounds or are possibly effects of the were obtained by using a C4 column with H20 and acetonitriie con-
SI deletion on neighboring genes (Bennett, 1956). taining 0.1% TFA as solvents. N-terminal amino acid sequence was de-
The original W mutation is an example of a c-kit null mu- termined on an Applied Biosystems 477A on-line PTH amino acid ana-
lyzer (Hewick et al., 1981).
tation (No&a et al., 199Oa). When heterozygous with the
normal allele, WI+ mice typically have a ventral spot but iodination
no coat dilution and no effects on hematopoiesis and KL was iodinated with chloramine T with modifications of the method
gametogenesis. The weak heterozygous phenotype of of Stanley and Guilbert (1981). Briefly, the labeling reaction contained
tV+ mice is in contrast to the phenotype of heterozygous 200 ng of KL. 2 nmol of chloramine T, 10% dimethyl sulfoxide, and
0.02% polyethylene glywl 8000, in a total volume of 25 pl in 0.25 M
S/l+ mice, which have moderate macrocytic anemia and
phosphate buffer (pH 6.5). The reaction was carried out for 2 min at
a diluted coat pigment in addition to a ventral spot and 4OC and stopped by the addition of 2 nmol of cysteine and 4 pM KI.
gonads that are reduced in size. Thus 50% gene dosage KL was then separated from free Nai by gel filtration on a PDlO column
of KL is limiting and is not sufficient for normal function (Pharmacia). iodinated KL was stored for up to 2 weeks at 4OC.
of the c-kit receptor, yet 50% dosage of the c-kit receptor
does not appear to be limiting in most situations. Binding Assay
Binding buffer contained RPM1 1640 medium, 5% BSA (Sigma), 20
The c-kit receptor system functions in immature progen- mM HEPES (pH 7.5), and NaNa. Binding experiments with nonadher-
itor cell populations as well as in more mature cell types ent ceils were carried out in 96-well tissue culture dishes with 2 x lo5
in hematopoiesis, gametogenesis, and melanogenesis. cells per well in a volume of 100 pl. Binding experiments with ~2 ceils
Severe SI or W mutations may block the development of were carried out in 24-well dishes in a volume of 300 ~1. Cells were
equilibrated in binding buffer 15 min prior to the addition of competitor
these cell lineages, and therefore a function for the c-kit or labeled KL. To determine nonspecific binding, unlabeled KL or anti-
receptor in more mature cell populations would not be evi- c-kit rabbit serum was added in a IO-fold excess 30 min prior to the ad-
dent. Sland W mutations in which c-kitlKL function is only dition of ‘*VKL. Ceils were incubated with 1251-KL for QO min, and
partially impaired often reveal effects in more mature cell nonadherent cells were pelleted through 150 pl of FCS. Cell pellets
were frozen and counted.
populations. Numerous weak SI alleles are known. Their
phenotypes, e.g., in gametogenesis and melanogenesis, tmmunoprctcipttatlon and Cross-Linking
will be of great value in the elucidation of the pleiotropic BMMC were incubated with ‘=i-KL under standard binding conditions
functions of the c-kit receptor system. and washed in FCS and then in PBS at 4OC. Ceils were iysed as previ-
ously described (Nocka et al., 1989) in 1% Triton X-100,20 mM Tris (pH
Ex~rlmentai Procedures 7.4), 150 mM NaCI. 20 mM EDNA. 10% glycerol, and protease inhibitors
phenytmethylsulfonyl fluoride (1 mM) and ieupeptin (20 &ml). Lysates
Mice and Tiaaue Culture were immunoprecipitated with normal rabbit serum, or c-kit-specific
WBB6 +I+, C57BL/W, C57BU6.l WV+, WBB6 ‘191+, CZHeB/FeJ 8/a sera raised by immunization of rabbits with a fragment of the v-kit tyro-
Ca’SIHm, and M. spretus mice were obtained from The Jackson Lab- sine kinase domain (Majumder et al., 1988) or the murine c-kit ex-
oratory (Bar Harbor, ME). For the interspecific cross, female C57BU6J pressed from a cDNA in a recombinant vaccinia virus (No&a et al.,
and male M. spretus mice were mated; progeny of this cross were 1QQOa). For coprecipitation experiments, immunoprecipitates were
scored for inheritance of C57BU6J or M. spretus alleles as described washed three times with wash A (0.1% Triton X-100,20 mM Tris [pH 7.41,
in Results. (C57BU6J x M. spretus)Fl female offspring were back- 150 mM NaCI, 10% glycerol), solublized in SDS sample buffer, and
crossed with C57BU6J males. analyzed by SDS-PAGE and autoradiography. For cross-linking ex-
Mast cells were grown from the bone marrow of adult +I+, WY/W”, periments, cells were incubated with disuccinimidyi suberate (0.25
and W+ mice and IMW fetal liver of day 14-15 fetuses in RPM 1640 mg/ml) in PBS for 30 min at 4OC, washed in PBS, and lysed as de-
medium supplemented with 10% fetal calf serum (FCS), conditioned scribed above. Washing conditions following precipitation were as fol-
medium from WEHI- cells, nonessential amino acids, sodium pyru- lows: one time in wash B (50 mM Tris, 500 mM NaCI, 5 mM EDTA, 0.2%
vate, and P-mercaptoethanol (RPMi-Compiete) (NO&R et al., 1QQOa; Triton X-100). three times in wash C (50 mM Tris, 150 mM NaCI, 0.1%
Cell
232

Triton X-100, 0.1% SDS, 5 mM EDTA), and one time in wash D (10 mM Refemnces
Tris, 0.1% Triton X-100).
Aaronson, S. A., and Todaro, G. (1968). Development of 3T3 like lines
cDNA Syntheeis, PCR Ampllflcetlon (RT-PCR), and from BALB/c mouse embryo cultures: transformation susceptibility to
Sequencs Determinetion SV40. J. Cell. Physiol. 72, 141-148.
The RT-PCR amplification was carried out essentially as described Altus, M. S., Bernstein, S. E., Russell, E. S., Carsten, A. L., and Upton,
(Tan et al., 1990). For cDNA synthesis, 1 l.rg of poly(A)+ RNA from con- A. C. (1971). Defect extrinsic to stem cells in spleens of steel anemic
fluent SALB/c 3T3 cells in 25 ul of 0.05 M his-HCI (pH 6.3) 0.075 M mice. Proc. Sot. Exp. Biol. Med. 738, 985-988.
KCI, 3 mM MgCIs, 10 mM dithiothreitol, 208 uM dNTPs, and 25 U of
Avner, i?, Amar, L., Dandolo, L., and Guenet, J. L. (1988). Genetic anal-
RNAsin (Promega) was incubated with 50 pmol of antisense primer
ysis of the mouse using interspecific crosses. Trends Genet. 4, 18-23.
and 50 U of Moloney murine leukemia virus reverse transcriptase at
409c for 30 min. Another 50 U of reverse transcriptase was added, and Barker, J. E., and McFarland, E. C. (1988). Hemopoietic precursor cell
incubation was continued for another 30 min. The cDNA was amplified defects in nonanemic but stem cell-deficient WulwU mice. J. Cell.
by bringing up the reaction volume to 50 pl with 25 ul of 50 mM KCI, Physiol. 135, 533-538.
10 mM Tris-HCI (pH 8.3) 1.5 mM MgCls, 0.01% (w/v) gelatin, and 200 Bennett, D. (1956). Developmental analysis of a mutant with pleiotropic
uM dNTPs, adding 50 pmol of sense primer and 2.5 U of Taq DNA poly effects in the mouse. J. Morphol. 98, 199-234.
merase, and amplifying for 25-30 cycles in an automated thermal cy- Besmer, P., Murphy, F! C., George, P C., Qiu, F., Bergold, P J., Leder-
cler (Perkin-Elmer Cetus). The amplified fragments were purified by man, L., Snyder, H. W., Brodeur, D., Zuckerman, E. E., and Hardy,
agarose gel electrophoresis, digested with the appropriate restriction W. D. (1986). A new acute transforming feline retrovirus and relation-
enzymes, and subcloned into M13mp18 and M13mp19 for sequence ship of its oncogene v-kit with the protein kinase gene family. Nature
analysis (Sanger et al., 1977). 320. 415-421.
Chabot, B., Stephenson, D. A.. Chapman, V. M., Besmer, P, and Bern-
cDNA Isolstlon and Sequencing stein, A. (1988). The proto-oncogene c-&it encoding a transmembrane
A mouse 3T3 fibroblast Igtll cDNA library obtained from Clontech was tyrosine kinase receptor maps to the mouse W locus. Nature 335,
used in this work. Screening in duplicate was done with Escherichia 68-89.
coli Y1090 as a host bacterium (Sambrook et al., 1989); B’end-labeled
Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J.
oligonucleotide was used as a probe. Hybridization was in 6x SSC at
(1979). Isolation of biologically active ribonucleic acid from sources en-
63OC, and the final wash of the filters was in 2x SSC, 0.2% SDS at
63oC. Recombinant phage were digested with EcoRl and the inserts riched in ribonuclease. Biochemistry 18, 5294-5299.
subcloned into Ml3 for sequence analysis. The nucleotide sequence Chui, D. K., Liao, S-K., and Walker, K. (1978). Fetal erythropoiesis in
of these cDNAs was determined, on both strands and with overlaps, steel mutant mice. Ill. Defect in differentiation from BFU-E to CFU-E
by the dideoxy chain termination method of Sanger et al. (1977) by during early development. Blood 51, 539-547.
using synthetic oligodeoxynucleotides (17mers) as primers. Dexter, T. M., and Moore, M. A. S. (1977). In vitro duplication and cure
of hematopoietic defects in genetically anaemic mice. Nature 289,
DNA and RNA Analysis 412-414.
Genomic DNA was prepared from tail fragments, digested with restric- Feinberg, A. P., and Vogelstein, B. (1983). A technique for radiolabeling
tion enzymes, electrophoreticallyfractionated, and transferred to nylon DNArestriction fragments to high specific activity. Anal. Biochem. 732,
membranesasdescribed elsewhere (D. R. B. and l? L., submitted). For 6-13.
hybridization, the 1.4 kb KL cDNA and TIS Dra/.Sa/ (a probe derived
Flanagan, J. G., and Leder, P (1990). The kit ligand: a cell surface mol-
from the transgene insertion site in the transgenic line TG.EB [Schmidt
ecule altered in steel mutant fibroblasts. Cell 83, this issue.
et al., 1988)) were used as probes.
BALB/c 3T3 cells were homogenized in guanidinium isothiocyanate, Fujita, J., Nakayama, H., Onoue, H., Kanakura, Y., Nakano, T., Asai,
and RNA was isolated according to the method of Chirgwin et al. H., Takeda, S., Honjo, T, and Kitamura, Y. (1988). Fibroblast dependent
(1979). Total cellular RNA (10 ug) and poly(A)+ RNA were fractionated growth of mouse mast cells in vitro: duplication of mast cell depletion
in 1% agarose-formaldehyde gels and transferred to nylon mem- in mutant mice of w genotype. J. Cell. Physiol. 134, 78-84.
branes (Nytran, Schleicher & Schuell); prehybridization and hybridiza- Fujita, J., Onoue, H., Ebi, Y., Nakayama, H., Kanakura, Y., and
tion were performed as previously described (Lehrach et al., 1978; Kitamura, Y. (1989). In vitro duplication and in ho cure of mast-cell
No&a et al., 1989). The 1.4 kb KL cDNA labeled with [32P]phosphate deficiency of S//SP mutant mice by cloned 3T3 fibroblasts. Proc. Natl.
was used as a probe for hybridization (Feinberg and Vogelstein, 1983). Acad. Sci. USA 88, 2886-2891.
Geissler, E. N., McFarland, E. C., and Russell, E. S. (1981). Analysis
of pleiotropism at the dominant white-spotting (W) locus of the house
mouse: a description of ten new W alleles. Genetics 97; 337-361.
K. Nocka and E. Huang contributed equally to the work reported here.
Geissler, E. N., Ryan, M. A., and Housman, D. E. (1988). The dom-
This paper is dedicated to the memory of Dorothea Bennett, who was
inant-white spotting (IV) locus of the mouse encodes the c-kit proto-
the first to provide a comprehensive description of the developmental
oncogene. Cell 55. 185-192.
defects of the SI mouse. We are particularly indebted to Dr. Robert
Peterson for his help with the labeling of KL. We thank Easter Chiu and Green, E. L. (1981). Genetics and Probability in Animal Breeding Ex-
Urs Rothlisberger for excellent technical assistance. We would like to periments (Oxford: Oxford University Press).
thank Drs. Ulrich Haemmerling, William Hayward, Prabir Ray, Jimmy Gregory, C. J., and Eaves, A. C. (1976). Three stages of erylhropoietic
Tan, and Elizabeth Lacy for their interest, for helpful discussions, and progenitor cell differentiation distinguished by a number of physical
for use of their facilities. This work was supported by grants from the and biologic properties. Blood 57, 527-537
American Cancer Society (MV 246D), the National Cancer Institute
Harrison, D. E. (1960). Competitive repopulation: a new assay for long-
(CA 32926 and CA 16599 to I? B.), and the E. I. DuPont de Nemours
term stem cell functional capacity. Blood 55, 77-61.
Co., Inc. (to P L.). K. N. is the recipient of a postdoctoral fellowship
from the NIH. Heldin, C.-H., Ernlund, A., Rorsman C., and Rdnnstrand, L. (1969). Di-
The costs of publication of this article were defrayed in part by the merization of B-type platelet-derived growth factor receptors occurs af-
payment of page charges. This article must therefore be hereby ter ligand binding and is closely associated with receptor kinase activa-
marked “advertisement” in accordance with 18 USC Section 1734 tion J. Biol. Chem. 284, 8905-8912.
solely to indicate this fact. Hewick, R. M., Hunkapiller, M. W., Hood, L. E., and Dryer, W. J. (1981).
A gas-liquid solid phase peptide and protein sequenator. J. Biol. Chem.
Received September 10, 1990. 258, 7990-7887.
c-kit Ligand Is SI Gene Product
233

Kawasaki, E. S., Ladner, M. B., Wang, A. M., Van Arsdell, J., Warren, defects contain distinct point mutations in the kinase domain of the
M. K., Coyne, M. Y., Schweickart, V. L., Lee, M. T, Wilson, K. J., Boos- c-kit receptor. Genes Dev. 4, 390-400.
man, A., Stanley, E. R., Ralph, R, and Mark, D. F. (lQ65). Molecular Rettenmier. C. W., and Roussel, M. F (1966). Differential processing of
cloning of a complementary DNA encoding human macrophage spe- colony-stimulating factor 1 precursors encoded by the two human
cific colony stimulating factor (CSF-1). Science 230, 291-296. cDNAs. Mol. Cell. Biol. 8, 5026-5034.
Kitamura, Y., and Fujita, J. (1969). Regulation of mast cell differentia- Russell, E. S. (1970). Abnormalities of erythropoiesis associated with
tion. Bio-Essays 70, 193-196. mutant genes in mice. In Regulation of Hematopoiesis, A. S. Gordon,
Kitamura, Y., and Go, S. (1916). Decreased production of mast cells in ed. (New York: Appleton), pp. 649-675.
S//.9’ anemic mice. Blood 53. 492-497. Russell, E. S. (1979). Hereditary anemias of the mouse: a review for
Kitamura, Y., Go, S., and Hatanaka. K. (1978). Decrease of mast cells geneticists. Adv. Genet. 20, 357-459.
in WWV mice and their increase by bone marrow transplantation. Sambrook, J., Fritsch, E. F, and Maniatis, T. (1969). Molecular Clon-
Blood 52, 447-452. ing: A Laboratory Manual, Second Edition (Cold Spring Harbor, New
Lehrach, H., Diamond, D., Wozney, J. M., and Boedtker, H. (1977). York: Cold Spring Harbor Laboratory Press).
RNA molecular weight determinations by gel electrophoresis under Sanger, F., Nicklen, S., and Coulson, A. R. (1977). DNA sequencing
denaturing conditions-a critical reexamination. Biochemistry 16, with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74,
4743. 54634467.
LeviSchaffer, F., Austen, K. F., Gravallese, i? M., and Stevens, R. L. Sawella, P A., and Russell, L. B. (1956). Steel, a new dominant gene
(1966). &culture of interleukin-3 dependent mouse mast cells with fi- in the house mouse. J. Hered. 47, 123-126.
broblasts results in a phenotypic change of the mast cells. Proc. Natl. Schmidt, E. V., Pattengale, P K., Weir, L., and Leder, P (1966). Trans-
Acad. Sci. USA 83, 6465-6468.
genie mice bearing the human c-myc gene activated by an immuno-
Majumder, S., Brown, K., Qiu, F.-H., and Besmer, P (1966). c-kirprotein, globulin enhancer: a pm-B cell lymphoma model. Proc. Natl. Acad.
a transmembrane kinase: identification in tissues and characteriza- Sci. USA 85, 6047-6051.
tion. Mol. Cell. Biol. 8, 4696-4903. Sherr, C. J.. Rettenmier, C. W., Sacca, R., Roussel, M. F, Look, A. T.,
Manova, K., No&a, K., Besmer, l?, and Bachvarova, R. F (1990). and Stanley, E. R. (1965). The c-fms proto-oncogene product is related
Gonadal expression of c-kit encoded at the W locus of the mouse. De- to the receptor for the mononuclear phagocyte growth factor, CSFI.
velopment, in press. Cell 47, 665-676.
Mayer, T. C., and Green, M. C. (1966). An experimental analysis of the Silvers, W. K. (1979). White-spotting, patch and rump-white. In the Coat
pigment defect caused by mutations at the Wand SI loci in mice. Dev. Colors of Mice: A Model for Gene Action and Interaction (New York:
Biol. 78, 62-75. Springer-Verlag), pp. 206-241.
McCulloch, E. A. (1970). Control of hematopoiesis at the cellular level. Stanley, E. R., and Guilbert, L. J. (1981). Methods for the purification,
In Regulation of Hematopoiesis, A. S. Gordon, ed. (New York: Apple- assay, characterization and target cell binding of a colony stimulating
ton), pp. 649-675. factor (CSF-1). J. Immunol. Meth. 42, 253-264.
McCulloch, E. A., Siminovitch, L., and Till, J. E. (1964). Spleen-colony Stevens, R. L., and Austen, K. F (1969). Recent advances in the cellu-
formation in anemic mice of genotype w. Science 144, 644-646. lar and molecular biology of mast cells. Immunol. Today 70, 361-366.
McCulloch, E. A., Siminovitch, L., Till, J. E., Russell, E. S., and Bern- Tan, J. C., No&a, K., Ray, P, Traktman, R, and Besmer, P (1990). The
stein, S. E. (1965). The cellular basis of the genetically determined dominant W42 spotting phenotype results from a missense mutation in
hemopoietic defect in anemic mice of genotype S/IS/d. Blood 26, the c-kit receptor kinase. Science 247, 209-212.
399-410. Tsuji, K., Nakahata, T, Takagi, M., Kobayashi, T, Ishiguro, A., Kikuchi,
Mintz, B., and Russell, E. S. (1957). Gene induced embryological T., Naganuma, K., Koike. K., Miyajima. A., Arai, K., and Akabane, T
modifications of primordial germ cells in the mouse. J. Exp. Zool. 134, (1990). Effects of interleukin-3 and interleukin-4 on the development of
207-237. connective tissue-type mast cells: interleukin-3 supports their survival
Nocka, K., Majumder, S., Chabot, B., Ray, F!, Cervone, M., Bernstein, and interleukin-4 triggers and supports their proliferation synergisti-
A., and Besmer, P. (1969). Expression of c-kit gene products in cellular cally with interleukin-3. Blood 75, 421-427.
targets of W mutations in normal and W mutant mice- evidence for an Yarden, Y., Kuang, W. J.. YangFeng. T., Coussens, L., Munemitsu, S.,
impaired c-kit kinase in mutant mice. Genes Dev. 3, 816-826. Dull, T J., Chen, E., Schlessinger, J., Fran&e, U., and Ullrich. A.
No&a, K., Tan, J., Chiu, E., Chu, T. Y., Ray, P, Traktman, P, and Bes- (1967). Human proto-oncogene c-kit: a new cell surface receptor tyro-
mer, P (IQQOa). Molecular bases of dominant negative and loss of func- sine kinase for an unidentified ligand. EMBO J. 6, 3341-3351.
tion mutations at the murine c-kit/white spotting locus: Wn, WY, W47, Yung, Y. P., and Moore, M. A. S. (1962). Long term in vitro culture of
K! EMBO J. 9, 1605-1613. murine mast cells: Ill. Discrimination of mast cell growth factor and
Nocka, K., Buck, J., Levi, E., and Besmer, I? (IQQOb). Candidate ligand granulocyte-CSF J. Immunol. 729, 1256-1261.
for the c-kit transmembrane kinase receptor: KL, a fibroblast-derived Yung, Y. f?, Eger, R. Tertian, G., and Moore, M. A. S. (1961). Long term
growth factor stimulates mast cells and erythroid progenitors. EMBO in vitro culture of murine mast cells. II. Purification of a mast cell growth
J. 9, 3267-3294. factor and its dissociation from TCGF J. Immunol. 727, 794-799.
Orr-Urtreger, A., Avivi, A., Zimmer. Y., Givol, D., Yarden. Y., and Lonai,
P (1990). Developmental expression of c-/rit, a proto-oncogene en- GenBenk Accesslon Number
coded by the W locus. Development 709, 911-923. The accession number for the sequence reported in this paper is
Park, L. S., Friend, D., Gillis, S., and Urdal, D. L. (1966). Characteriza- M38511.
tion of the cell surface receptor for granulocyte-macrophage colony-
stimulating factor. J. Biol. Chem. 261, 4177-4163. Note Added In Proof
Qiu, F, Ray, P., Brown, K., Barker, R E., Jhanwar, S., Ruddle, F H., and
Besmer, I? (1966). Primary structure of c-kin relationship with the CSF Although the first two authors contributed equally to this work, the or-
IIPDGF receptor kinase family-oncogenic activation of v-kit involves der of authorship should be K. No&a followed by E. Huang.
deletion of extracellular domain and C-terminus. EMBO J. 7, 1003-
1011.
Reith, A. D., Rottapell, R., Giddens, E., Brady, C., Forrester, L., and
Bernstein, A. (1990). Wmutant mice with mild or severe developmental