An Allelic Series of Mutations in the Kit ligand Gene of Mice. II. Effects of Ethylnitrosourea-Induced Kitl Point Mutations on Survival and Peripheral Blood Cells of Kitl^Steel Mice

Corresponding author: M. A. Bedell, B416 Life Sciences, University of Georgia, Athens, GA 30602-7223., bedell{at}arches.uga.edu (E-mail)

Communicating editor: C. KOZAK

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The ligand for the Kit receptor tyrosine kinase is Kit ligand (Kitl; also known as mast cell growth factor, stem cell factor, and Steel factor), which is encoded at the Steel (Sl) locus of mice. Previous studies revealed that Kitl^Sl mutations have semidominant effects; mild pigmentation defects and macrocytic, hypoplastic anemia occur in heterozygous mice, and more severe pigmentation defects and anemia occur in homozygotes. Lethality also occurs in mice homozygous for severe Kitl^Sl mutations. We describe the effects of seven new N-ethyl-N-nitrosourea (ENU)-induced Kitl^Sl mutations and two previously characterized severe Kitl^Sl mutations on pigmentation, peripheral blood cells, and mouse survival. Mice heterozygous for each of the nine mutations had reduced coat pigmentation and macrocytosis of peripheral blood. In the case of some of these mutations, however, red blood cell (RBC) counts, hemoglobin concentrations, and hematocrits were normal in heterozygotes, even though homozygotes exhibited severely reduced RBC counts and lethality. In homozygous mice, the extent of anemia generally correlates with effects on viability for most Kitl^Sl mutations; i.e., most mutations that cause lethality also cause a more severe anemia than that of mutations that allow viability. Interestingly, lethality and anemia were not directly correlated in the case of one Kitl^Sl mutation.

MUTATIONS at the Steel (Sl) and Dominant White Spotting (W) loci of mice identify two genes essential for the development of hematopoietic cells, germ cells, and melanocytes (reviewed by BESMER et al. 1993 ; LEV et al. 1994 ). The W locus encodes Kit, a type III receptor tyrosine kinase, and the Sl locus encodes Kitl (also known as mast cell growth factor, stem cell factor, and Steel factor), which is the only known ligand for Kit and is a member of the short-chain subgroup of helical cytokines (JIANG et al. 2000 ; ZHANG et al. 2000 ). While Kit is expressed on the surface of hematopoietic cells, germ cells, and melanocytes, Kitl is expressed by various cells that support the survival, proliferation, and differentiation of the former cell types. Interestingly, recent evidence suggests that the Kitl/Kit signaling pathway may operate differently in the different cell types (JORDAN et al. 1999 ; BLUME-JENSEN et al. 2000 ; KISSEL et al. 2000 ).

A large collection of Kitl^Sl mutations exists and offers a powerful genetic resource for understanding the in vivo functions of the Kitl/Kit signaling pathway. Importantly, different Kitl^Sl mutations produce phenotypes that are graded with respect to severity; i.e., some Kitl^Sl mutations produce very severe phenotypes while other Kitl^Sl mutations produce very mild phenotypes. While severe mutations are critical to understanding the consequences of the near or complete absence of function of a particular gene, identification and characterization of milder mutations may reveal requirements during later developmental stages (SCHUMACHER et al. 1996 ). In Kitl^Sl mutants, homozygous null mutations cause prenatal or perinatal lethality with severe effects on numbers of hematopoietic cells, germ cells, and melanocytes. On the other hand, homozygous hypomorphic Kitl^Sl mutations allow viability but have milder effects on each cell type. Gene dosage is important to Kitl function, as all Kitl^Sl mutations (with the exception of one extinct allele) are semidominant (MOUSE GENOME DATABASE 2002) and this is likely to be due to haplo-insufficiency (BEDELL et al. 1996A ). The best-known hematopoietic defects in Kitl^Sl mutants are specific for stem cells, erythroid cells, mast cells, and megakaryocytes (BESMER et al. 1993 ; LEV et al. 1994 ). Kitl^Sl mutants have a macrocytic, hypoplastic anemia resulting from an increased volume of red blood cells (RBCs) and a reduced number of RBCs (RUSSELL 1979 ). Recent evidence indicates that, in addition to these classically defined targets, other cell types are defective in Kit^W and Kitl^Sl mutants (HUIZINGA et al. 1995 ; RODEWALD et al. 1995 ; MOTRO et al. 1996 ; LAKY et al. 1997 ).

In this report we describe the effects of seven new ethylnitrosourea (ENU)-induced Kitl^Sl mutations and two previously characterized Kitl^Sl mutations on pigmentation, peripheral blood cells, and survival of mice. We describe the molecular defects associated with the ENU-induced mutations in the accompanying article (RAJARAMAN et al. 2002 , this issue). Our analysis of the effects of these mutations on survival and peripheral blood cells reveals a graded effect, with five mutations having very severe effects and two mutations having milder effects. However, with some Kitl^Sl mutations, there is no direct relationship between severity of heterozygous and homozygous phenotypes nor is there a direct relationship between severity of effects on different blood-cell parameters.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mice:
The generation and molecular characterization of the seven ENU-induced Kitl^Sl mutations (Kitl^Sl-30R, Kitl^Sl-31R, Kitl^Sl-22R, Kitl^Sl-28R, Kitl^Sl-42R, Kitl^Sl-39R, and Kitl^Sl-36R) studied in this article have been described in the accompanying article (RAJARAMAN et al. 2002 ). Each mutant allele contains a point mutation, and the positions of the Kitl sequences affected in each are shown in Table 1. The ENU-induced Kitl^Sl mutations were made congenic on a common strain background by backcrossing to C3H/Rl for >20 generations and subsequently to C3H/HeNCR for at least 5 generations. Two previously characterized mutations, Kitl^Sl-gb and Kitl^Sl-d, were also used in this study (Table 1). Kitl^Sl-gb contains an 120-kb deletion whose proximal breakpoint is located 60 kb upstream of the Kitl transcription unit and whose distal breakpoint is located within the 3' untranslated region of Kitl (BEDELL et al. 1996A ). Kitl^Sl-gb mice were originally obtained from the MRC Radiobiology Unit (Chilton, Didcot, UK) and have been maintained on a C3H/HeNCR background for >20 generations. Kitl^Sl-d contains a 4-kb intragenic deletion that removes the transmembrane and cytoplasmic domains of Kitl (BRANNAN et al. 1991 ; FLANAGAN et al. 1991 ). Kitl^Sl-d mice were originally obtained from the Jackson Laboratory (Bar Harbor, ME) on a C57BL/6J background and the Kitl^Sl-d mutation was made congenic on the C3H/HeNCR strain by backcrossing for >12 generations. All strains are currently maintained in a pathogen-free colony at the University of Georgia.

Survival studies:
To generate homozygous mutant mice or compound heterozygous mice, mice heterozygous for each mutant allele were intercrossed. Each about-to-deliver female and each litter were examined daily until postnatal day 18 (P18) and the numbers of pups of each genotype recorded. Homozygous Kitl^Sl mice are readily identified at birth by their runted size and pallor due to anemia. A subset of presumed homozygous mutants for each mutant allele was subjected to molecular genotyping (see below). In every case, the genotype assigned by phenotype was confirmed.

Genotyping Kitl^Sl mutants:
Methods for genotyping were based on PCR amplification of genomic DNA from mouse tissues and are summarized in Table 1. A sequence polymorphism in the 5'-flanking region of Kitl was used for genotyping alleles that arose on non-C3H chromosomes [including Kitl^Sl-30R, Kitl^Sl-28R, Kitl^Sl-42R, and Kitl^Sl-39R (RAJARAMAN et al. 2002 ) and Kitl^Sl-d, which arose on a DBA/2J chromosome (BRANNAN et al. 1991 )]. This polymorphism consists of a 6-bp insertion located 259 bp 5' to the Kitl transcription initiation site (BEDELL et al. 1996B ) that is present in C3H DNA but is absent from C57BL/6J, C57BL/10 and 101/Rl DNA, and DBA/2J (W. S. DAVIS and M. A. BEDELL, unpublished results). Portions of the cloned Kitl^Sl-gb deletion breakpoint (BEDELL et al. 1996A ) were sequenced (data not shown), and oligonucleotide primers that span the breakpoint were used for PCR amplification of genomic DNA. Three of the ENU-induced mutations (Kitl^Sl-31R, Kitl^Sl-22R, and Kitl^Sl-36R) arose on C3H chromosomes (RAJARAMAN et al. 2002 ), and allele-specific genotyping methods were developed (Table 1). In the Kitl^Sl-22R allele, the mutation abolishes a DdeI site and restriction fragment length polymorphism (RFLP) analysis was used for genotyping. For Kitl^Sl-31R and Kitl^Sl-36R genotyping, allele-specific PCR amplification methods were developed.

Peripheral blood analysis:
Newborn (P1) mice were anesthetized by hypothermia and euthanized, and peripheral blood was collected using heparinized capillary tubes. The blood was diluted in PBS, and RBCs were counted using a hemacytometer. At P24-P25, juvenile mice were euthanized using CO₂ asphyxiation, and peripheral blood was collected by cardiac puncture. Blood from four to eight mice of each genotype was analyzed. Complete blood cell analysis was performed using a Celldyne 3500 hematology analyzer. The parameters evaluated are RBC counts; hemoglobin concentration; mean corpuscular volume (MCV), which is the average volume of individual RBCs; mean corpuscular hemoglobin (MCH), which is the average hemoglobin concentration in RBCs and is calculated from hemoglobin and RBC values; mean corpuscular hemoglobin concentration (MCHC), which is the ratio of the hemoglobin concentration to the average RBC volume and is calculated from hemoglobin and hematocrit values; hematocrit, which is the percentage of whole blood made up of RBCs and calculated from RBC and MCV values; counts of platelets; the mean platelet volume (MPV), which is the average volume of individual platelets; and counts of white blood cells (WBC), segmented neutrophils, and lymphocytes. Note that all of these parameters, except MCH, MCHC, and hematocrit, are determined directly, while the latter three parameters are calculated values.

Statistical analysis:
For each Kitl^Sl mutation, the observed numbers of homozygous mutant and compound heterozygous mice at P1 were compared against the expected numbers of these mice using a chi-square test for significance. Survival curves for mice of different genotypes were calculated and compared using Prism software (GraphPad Software, San Diego); the Kaplan-Meier method was used to calculate fractional survival at each time point and the log-rank test with calculation of two-sided P values was used to do pairwise comparisons of survival curves. The survival curves of homozygous mutant mice were compared against the survival curves of control mice (Kitl⁺/Kitl⁺ and heterozygous mice) and of homozygous null (Kitl^Sl-gb/Kitl^Sl-gb) mice, and the survival curves of compound heterozygous mice were compared against the survival curves of Kitl^Sl-d hemizygous (Kitl^Sl-gb/Kitl^Sl-d) mice. Values for peripheral blood analysis were evaluated using unpaired, two-tailed t-tests using Prism software. For each parameter, the values for heterozygotes and homozygotes for each mutant allele were compared against that of Kitl⁺/Kitl⁺, Kitl^Sl-gb/ Kitl⁺, and Kitl^Sl-gb/Kitl^Sl-gb mice. In addition, comparisons were made between pairs of values for mice homozygous for viable alleles, i.e., Kitl^Sl-36R/Kitl^Sl-36R vs. Kitl^Sl-39R/Kitl^Sl-39R, Kitl^Sl-36R/Kitl^Sl-36R vs. Kitl^Sl-d/Kitl^Sl-d, and Kitl^Sl-39R/Kitl^Sl-39R vs. Kitl^Sl-d/Kitl^Sl-d.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Pigmentation of Kitl^Sl mutant mice:
All seven of the new Kitl^Sl mutants described here arose from progeny of mice derived from the specific locus test using ENU as mutagen (RUSSELL et al. 1982 ). Although the Kitl^Sl locus is not one of the loci used in the specific locus test, the new Kitl^Sl mutants were apparent because of their mild pigmentation defects in heterozygous mice (Table 1). However, the severity of the heterozygous pigmentation defect does vary somewhat, with two of the viable mutations having less of an effect than that which is characteristic of lethal mutations, and the Kitl^Sl-39R mutation having the mildest effect of all the mutations. In all mice homozygous for viable Kitl^Sl mutations and in compound heterozygotes between each of the ENU-induced mutations and Kitl^Sl-d, white coats were observed (data not shown). Occasionally, small pigmented patches were seen on the heads of Kitl^Sl-39R/Kitl^Sl-39R mice (not shown). Overall, these observations indicate that all the Kitl^Sl mutations described here have semidominant effects on pigmentation.

Survival of homozygous mutant mice:
Previous studies revealed that the majority of mice homozygous for severe Kitl^Sl mutations die during late gestation with severe anemia (SARVELLA and RUSSELL 1956 ; RUSSELL 1979 ). To determine the effects of the ENU-induced Kitl mutations on prenatal or perinatal survival, we examined the ratios of genotypes in P1 mice born to intercrosses of mice heterozygous for each mutant allele. According to Mendelian segregation, 0.25 of the total number of mice born to these intercrosses should be homozygous mutant. If the ratio of homozygous mutant mice is significantly <0.25 for a given allele, then lethality must be occurring either prior to birth or within a few hours after birth.

To establish the survival pattern for the null condition at Kitl and, by comparison, to determine whether any of the ENU-induced mutations might be null functionally, we examined progeny of intercrosses of mice heterozygous for the smallest complete Kitl deletion, Kitl^Sl-gb (BEDELL et al. 1996A ). Since 209 wild-type and heterozygous mice were born to Kitl^Sl-gb/Kitl⁺ intercrosses, 70 Kitl^Sl-gb/Kitl^Sl-gb mice would have been expected in the absence of any lethality to the latter (Table 2). However, only 32 Kitl^Sl-gb/Kitl^Sl-gb mice were observed, indicating that 54% of the Kitl^Sl-gb/Kitl^Sl-gb mice die either before birth or immediately following birth. Furthermore, none of the observed homozygous null mice survived beyond P2 (Fig 1, solid black line in each of B–I). On embryonic day 14.5 (E14.5), however, the expected ratio of Kitl^Sl-gb/Kitl^Sl-gb embryos was observed (data not shown). Thus, Kitl^Sl-gb/Kitl^Sl-gb mice on the C3H strain background die between E14.5 and P2, indicating that the Kitl^Sl null phenotype on this background is pre- or perinatal lethality.

View larger version (41K): In this window In a new window Download PPT slide   Figure 1. Survival curves of homozygous mutant and compound heterozygous mice carrying KitlSl mutations. For each allele, heterozygous mice were mated and the resulting progeny observed every day after birth until P18. The values at P0 represent the sum of the numbers of dead and living homozygous mutant or compound heterozygous mice observed at P1. For subsequent ages, the fractional survival and SEM at each age were calculated using the Kaplan-Meier method. (A) The red line is the survival curve for 942 wild-type mice and heterozygotes combined from all intercrosses and the other survival curves are for mice homozygous for each of the KitlSl mutant alleles. (B–H) The survival curves for mice carrying each of the ENU-induced KitlSl mutant alleles. (I) The survival curve for KitlSl-d. The lines and symbols used in the graphs are as follows: solid lines with solid circles, homozygous mice; dashed lines with open circles, compound heterozygous mice (KitlSl-X/KitlSl-d, where X represents any of the ENU alleles); solid black lines, KitlSl-gb/KitlSl-gb; dashed black lines with open symbols, KitlSl-gb/KitlSl-d).

With four of the ENU-induced mutations (Kitl^Sl-31R, Kitl^Sl-22R, Kitl^Sl-28R, and Kitl^Sl-36R), only 61–71% of the expected homozygous mutant P1 mice were observed (Table 2) and these proportions are significantly (P < 0.05) below expectations. These results indicate that the Kitl^Sl-22R, Kitl^Sl-28R, Kitl^Sl-31R, and Kitl^Sl-36R alleles have reduced activity for prenatal or perinatal survival. In contrast, the observed numbers of P1 mice homozygous for three of the ENU-induced mutations (Kitl^Sl-30R, Kitl^Sl-42R, and Kitl^Sl-39R) are not significantly different (P > 0.05) from expectation (Table 2), indicating that the effects of these mutations on prenatal and perinatal survival are milder than those of the null allele.

We examined the postnatal survival of mice homozygous for each of the Kitl^Sl mutations until P18, the age at weaning (Table 3 and Fig 1). The survival of these mutant mice was compared to the survival of control mice (red line in Fig 1A), which consisted of nearly 1000 wild-type and heterozygous siblings segregating from all the intercrosses of each mutant allele. In crosses involving each mutant allele, the expected numbers of heterozygous mice were observed at P18 (data not shown), indicating that there was little or no postnatal lethality of heterozygous mice. While 83% of control mice survived to P18, none of the mice homozygous for Kitl^Sl-30R, Kitl^Sl-31R, Kitl^Sl-22R, Kitl^Sl-28R, or Kitl^Sl-42R survived beyond P7 (Fig 1), and the survival curves of the homozygous mutants were highly significantly different from those of the control mice (Table 3). Thus, these five alleles are classified as homozygous lethal alleles. To determine whether these lethal alleles behave as null alleles with respect to postnatal viability, the survival curves of mice homozygous for each of the lethal alleles were compared to those of the Kitl^Sl-gb/Kitl^Sl-gb mice (solid lines in B–F of Fig 1). Interestingly, mice homozygous for the Kitl^Sl-22R, Kitl^Sl-28R, or Kitl^Sl-42R mutations displayed postnatal survival curves significantly different (P < 0.05, Table 3) from those of the homozygous null mutants. In particular, the survival curves of Kitl^Sl-22R/Kitl^Sl-22R mice are highly significantly different (P < 0.0001) from those of Kitl^Sl-gb/Kitl^Sl-gb mice. These results suggest that although these three mutations cause severe effects on survival, the alleles may be mildly hypomorphic because they allow a slightly prolonged survival time compared to the null allele. In comparison, the postnatal survival curves of mice homozygous for the Kitl^Sl-30R or Kitl^Sl-31R mutations are not significantly different (P > 0.05) from those of homozygous null mutants, indicating that these mutations are likely to be null functionally.

In contrast to mice homozygous for lethal mutations, the majority of Kitl^Sl-39R/Kitl^Sl-39R and Kitl^Sl-36R/Kitl^Sl-36R mice survived beyond P7, with 73 and 63% survival to P18, respectively, compared to 83% survival of control mice (Fig 1A). Thus, both of these alleles are classified as homozygous viable alleles. However, the survival of Kitl^Sl-36R/Kitl^Sl-36R mice is significantly less than that of control mice (P = 0.024), while the survival of Kitl^Sl-39R/Kitl^Sl-39R mice is not significantly different (P = 0.126) from that of control mice (Table 3). For Kitl^Sl-36R/Kitl^Sl-36R mice, the decreased viability is restricted to the period before P7 (see Fig 1A). This early lethality in Kitl^Sl-36R/Kitl^Sl-36R mice is consistent with observations made on P1 mice (see above and Table 2), where the ratio of Kitl^Sl-36R/Kitl^Sl-36R mice was less than expected. Thus, the Kitl^Sl-36R mutation affects perinatal and juvenile viability in some homozygous mice, but has less of an effect on homozygous mice surviving longer than 1 week.

Additive effect of some Kitl^Sl mutations on survival of compound heterozygous mice:
During the course of our studies with the ENU-induced Kitl^Sl mutations, experiments with Kitl^Sl-d mice revealed that the latter mutation exerts gene dosage effects on mouse survival (Fig 1I). This gene dosage effect was observed when survival of Kitl^Sl-d/Kitl^Sl-d mice (pink line in Fig 1I), which carry two copies of the Kitl^Sl-d allele, was compared with survival of Kitl^Sl-gb/Kitl^Sl-d mice (dashed black line in Fig 1I), which carry only one copy of the Kitl^Sl-d allele. In Kitl^Sl-d/Kitl^Sl-d mice, the expected numbers of P1 mice were observed (Table 2), and 82% of them survived to P18 (Fig 1I). However, we observed only 65% of the expected number of Kitl^Sl-gb/Kitl^Sl-d P1 mice (Table 2) and the postnatal survival curve of these mice is intermediate between that of Kitl^Sl-gb/Kitl^Sl-gb mice and Kitl^Sl-d/Kitl^Sl-d mice (Fig 1I). Thus, hemizygosity for Kitl^Sl-d causes an intermediate phenotype for postnatal survival.

The gene dosage effects observed with the Kitl^Sl-d allele provided the basis for a second test for activity of the ENU-induced Kitl^Sl gene products, namely, whether a given allele could exert an additive effect when in trans with Kitl^Sl-d. To accomplish this, mice heterozygous for each ENU-induced mutation were crossed with Kitl^Sl-d/Kitl⁺ mice to generate compound heterozygous mice (i.e., Kitl^Sl-X/Kitl^Sl-d, where X stands for any ENU-induced mutation). The survival of the compound heterozygotes was then determined and compared to that of Kitl^Sl-gb/Kitl^Sl-d mice (see Table 3 and Fig 1). If the test mutation is null functionally, then the survival curve of the compound heterozygotes should be identical to that of Kitl^Sl-gb/Kitl^Sl-d mice. If the test mutation is hypomorphic, then it should exert an additive effect with Kitl^Sl-d and the survival curve of the compound heterozygote would be shifted to the right of the Kitl^Sl-gb/Kitl^Sl-d survival curve. To validate this test, we first determined whether the two homozygous viable alleles (Kitl^Sl-39R and Kitl^Sl-36R) could exert an additive effect with the Kitl^Sl-d allele. As expected, the survival curves of Kitl^Sl-39R/Kitl^Sl-d and Kitl^Sl-36R/Kitl^Sl-d mice shifted to the right (dashed, colored lines in Fig 1G and Fig H, respectively) and are highly significantly different (P < 0.0001, Table 3) from the Kitl^Sl-gb/Kitl^Sl-d survival curve. In contrast, none of the lethal alleles (Kitl^Sl-30R, Kitl^Sl-31R, Kitl^Sl-22R, Kitl^Sl-42R, and Kitl^Sl-28R) exhibited a significant additive effect with Kitl^Sl-d (Table 3 and dashed colored lines in Fig 1, B–F). Thus these alleles are likely to be null, or nearly null, for activity required for prenatal and postnatal survival. Although the survival curve for Kitl^Sl-22R/Kitl^Sl-d mice was not statistically different from that of Kitl^Sl-gb/Kitl^Sl-d mice (P = 0.0719, Table 3), examination of these curves (dashed colored line in Fig 1D) suggests a trend toward increased survival in the former that is consistent with the enhanced survival of some Kitl^Sl-22R/Kitl^Sl-22R mice. Thus, from the analysis of homozygous mutant and compound heterozygous mice, Kitl^Sl-22R mutation appears to be mildly hypomorphic for postnatal survival.

Effects of Kitl^Sl mutations on RBCs of newborn mice:
Previous studies revealed that Kitl^Sl mutations cause a macrocytic, hypoplastic anemia that is apparent during embryogenesis and continues into adulthood (RUSSELL 1979 ). Like the pigmentation phenotype, the anemia phenotype in Kitl^Sl mutants is semidominant; i.e., the number of RBCs is mildly reduced in heterozygous mice but markedly reduced in homozygous mice. Because mice homozygous for severe Kitl^Sl mutations die either pre- or perinatally with severe anemia, it is likely that the anemia is the cause of the lethality. If so, then mice homozygous for each lethal Kitl^Sl mutation should exhibit effects on RBCs that are more severe than those of mice homozygous for each viable mutation. Furthermore, the effect on heterozygous mice would be expected to parallel the effect on homozygous mice; i.e., mutations that cause lethality when homozygous would be expected to have a heterozygous phenotype that is more severe than that of mutations that allow viability when homozygous. We tested this by examining RBC counts in P1 mice that were heterozygous or homozygous for each of the nine Kitl^Sl mutations and by studying peripheral blood parameters in P24-P25 mice that were heterozygous for each of the nine Kitl^Sl mutations and in mice that were homozygous for each of the three viable Kitl^Sl mutations.

In general, the mean RBC counts were reduced in P1 Kitl^Sl mice when the mutations were heterozygous or homozygous (Fig 2). In comparison to mean RBC counts of 4.1 ± 0.1 x 10⁹ cells/ml in P1 Kitl⁺/Kitl⁺ mice, the corresponding values for heterozygous Kitl^Sl mutants ranged from 2.7 ± 0.2 x 10⁹ cells/ml (66% of wild type) in P1 Kitl^Sl-39R/Kitl⁺ mice to 3.4 ± 0.2 x 10⁹ cells/ml (83% of wild type) in P1 Kitl^Sl-22R/Kitl⁺ mice. With the notable exception of Kitl^Sl-22R, each lethal and viable Kitl^Sl mutation resulted in RBC counts in heterozygous P1 mice that are significantly different (P < 0.02) from wild-type values. Given that Kitl^Sl-22R is a lethal mutation when homozygous, it is surprising that the mean RBC counts for P1 Kitl^Sl-22R/Kitl⁺ mice are not significantly different (P = 0.08) from the P1 wild-type mean. The mean RBC counts for P1 mice heterozygous for viable mutations (Kitl^Sl-39R, Kitl^Sl-36R, and Kitl^Sl-d) are not significantly different from the 2.9 ± 0.1 x 10⁹ cells/ml (71% of wild type) observed in P1 Kitl^Sl-gb/Kitl⁺ mice (P = 0.50, 0.26, and 0.08, respectively), and they are not significantly different from each other (Kitl^Sl-39R/+ vs. Kitl^Sl-36R/+, P = 0.13; Kitl^Sl-39R/+ vs. Kitl^Sl-d/+, P = 0.03; Kitl^Sl-36R/+ vs. Kitl^Sl-d/+, P = 0.82). Thus, the severity of the anemia phenotype in P1 heterozygous mice does not correlate directly with severity of the survival phenotype in homozygous mice (see Table 4). Furthermore, the heterozygous anemia phenotype does not correlate with the pigmentation defects in older heterozygous mice (see Table 4), as juvenile Kitl^Sl-22R/Kitl⁺ mice have a pigmentation defect that is more severe than that of Kitl^Sl-39R/Kitl⁺ mice (Table 1) even though the former mice are less anemic at birth than the latter (Fig 2).

View larger version (42K): In this window In a new window Download PPT slide   Figure 2. Peripheral RBC counts in newborn KitlSl mutant mice. Blood was collected from euthanized P1 mice and RBCs counted using a hemacytometer. The mean and SEM are shown for each genotype. Each set of values for heterozygous or homozygous mutants was compared against the values of wild-type mice using unpaired t-test: (**) P = 0.002–0.02; (***) P < 0.002. Solid bar, wild type; open bars, alleles that cause lethality to homozygous mice; shaded bars, alleles that allow viability to homozygous mice; bar with lines, hemizygous KitlSl-d mice (KitlSl-gb/KitlSl-d). The dashed horizontal line is the mean value for wild-type mice.

With P1 homozygous mice, RBC counts in all Kitl^Sl mutants are significantly reduced (P < 0.0001) relative to wild-type values (Fig 2). Moreover, unlike the situation in P1 heterozygous mice, there is a direct correlation between lethality and RBC counts in P1 mice homozygous for each of the nine mutations. While mean RBC counts in P1 Kitl^Sl-gb/Kitl^Sl-gb mice were 0.71 ± 0.04 x 10⁹ cells/ml (17% of wild type), mean RBC counts in P1 mice homozygous for the other lethal mutations ranged from 0.71 ± 0.07 x 10⁹ cells/ml (17% of wild type) in Kitl^Sl-31R/Kitl^Sl-31R mice to 0.82 ± 0.06 x 10⁹ cells/ml (20% of wild type) in Kitl^Sl-22R/Kitl^Sl-22R mice. Importantly, none of the values for homozygous lethal Kitl^Sl mutations are significantly different (P > 0.2) from that of Kitl^Sl-gb/Kitl^Sl-gb mice. In comparison, mean RBC counts in Kitl^Sl-39R/Kitl^Sl-39R, Kitl^Sl-36R/Kitl^Sl-36R, and Kitl^Sl-d/Kitl^Sl-d mice were 1.4 ± 0.2 x 10⁹ cells/ml (34% of wild type), 1.7 ± 0.1 x 10⁹ cells/ml (41% of wild type), and 1.3 ± 0.1 x 10⁹ cells/ml (32% of wild type), respectively. Each of these values for homozygous viable mutations is significantly different (P < 0.02) from that of P1 Kitl^Sl-gb/Kitl^Sl-gb mice. Although the mean in P1 Kitl^Sl-36R/Kitl^Sl-36R mice is not significantly different from that of Kitl^Sl-39R/Kitl^Sl-39R mice (P = 0.18), it is significantly different from that of Kitl^Sl-d/Kitl^Sl-d mice (P = 0.003). Thus, the RBC counts found in P1 mice homozygous for viable mutations are significantly higher than those found in mice homozygous for lethal mutations, and the Kitl^Sl-36R mutation has the mildest effect of all mutants on RBC counts when in the homozygous condition. In conclusion, in P1 homozygous mice, the lethal mutations behave as null alleles with respect to RBC counts while the viable mutations are hypomorphic with respect to RBC counts.

The relationship between extent of anemia and survival is extended further by examination of the gene dosage effects of the Kitl^Sl-d mutation. The mean P1 RBC counts in Kitl^Sl-d/Kitl^Sl-gb mice were 0.9 ± 0.1 x 10⁹ cells/ml (22% of wild type), which is significantly different from that of Kitl^Sl-d/Kitl^Sl-d mice (P < 0.02) but is not significantly different from that of Kitl^Sl-gb/Kitl^Sl-gb mice (P = 0.12). This intermediate effect on RBC counts is consistent with the intermediate effect seen in the survival of Kitl^Sl-d/Kitl^Sl-gb mice (see above and Fig 1I). However, the relationship between viability and extent of anemia does not extend to all Kitl^Sl mutations. Although Kitl^Sl-36R exerts the mildest effect on RBC counts of the three viable mutations (Fig 2), it is the only viable mutation that causes significant lethality to homozygous mice prior to and during the first week after birth (see Fig 1, G–I, and above). If perinatal or juvenile lethality occurs in some Kitl^Sl-36R/Kitl^Sl-36R mice because of severe anemia, then a range of RBC counts would have been observed in individual mice of this genotype. However, all six Kitl^Sl-36R/Kitl^Sl-36R mutants sampled had very similar RBC counts (not shown), none of which were below the values seen in Kitl^Sl-39R/Kitl^Sl-39R and Kitl^Sl-d/Kitl^Sl-d mutants. Thus, the cause of lethality in some Kitl^Sl-36R/Kitl^Sl-36R mice during the first week after birth does not appear to be due to severe anemia.

Effects of Kitl^Sl mutations on peripheral blood of juvenile mice:
We examined the effects on peripheral blood in P24-P25 mice, each of which was heterozygous for one of the nine Kitl^Sl mutations or homozygous for one of the three viable Kitl^Sl mutations. Significant differences between wild-type and mutant mice were observed for RBC counts, hemoglobin, MCV, MCH, MCHC, and hematocrit (Fig 3). However, no differences were observed between any mutant and wild-type mice for MPV, WBC counts, platelet counts, or lymphocytes (data not shown). Counts of segmented neutrophils (data not shown) were marginally reduced in Kitl^Sl-31R/Kitl⁺, Kitl^Sl-39R/Kitl⁺, and Kitl^Sl-39R/Kitl^Sl-39R mice; however, the values are not significantly different from wild-type values (P = 0.032, P = 0.021, and P = 0.041, respectively). While neutrophil counts in Kitl⁺/Kitl⁺ mice were 0.9 ± 0.1 x 10⁶ cells/ml, these values in Kitl^Sl-36R/Kitl^Sl-36R mice were 0.5 ± 0.1 x 10⁶ cells/ml. Although the neutrophil counts for Kitl⁺/Kitl⁺ and Kitl^Sl-36R/Kitl^Sl-36R mice are significantly different (P = 0.01), there are no previous reports of neutrophil defects in Kitl^Sl mutants. Because the effect on neutrophils was marginal and restricted to a specific allele, the relevance of these data is uncertain.

View larger version (73K): In this window In a new window Download PPT slide   Figure 3. Peripheral blood cell analysis in P24-P25 KitlSl mutant mice. Blood was collected from euthanized mice and analyzed using a Celldyne 3500 hematology analyzer. The mean and SEM are shown for each genotype. (A) RBCs. (B) Hemoglobin. (C) MCV. (D) MCH. (E) MCHC. (F) Hematocrit. Each set of values for heterozygous or homozygous mutants was compared against the values of wild-type mice using unpaired t-test: (**) P = 0.002–0.02; (***) P < 0.002. Solid bar, wild type; open bars, alleles that cause lethality to homozygous mice; shaded bars, alleles that allow viability to homozygous mice. The dashed horizontal line is the mean value for wild-type mice.

Consistent with the semidominant pigmentation defects (Table 1), every heterozygous Kitl^Sl mutation caused a significant effect on at least one peripheral blood parameter in P24-P25 mice (Fig 3). Interestingly, the severity of the heterozygous effect on peripheral blood did not relate directly to the severity of the heterozygous pigmentation phenotype or to the severity of the homozygous survival phenotype (Table 4). In comparison to RBC counts of 6.9 ± 0.1 x 10⁹ cells/ml in wild-type mice, RBC counts in Kitl^Sl-30R/Kitl⁺, Kitl^Sl-28R/Kitl⁺, and Kitl^Sl-42R/Kitl⁺ mice are significantly reduced (Fig 3A), ranging from 5.5 ± 0.1 x 10⁹ to 6.0 ± 0.1 x 10⁹ cells/ml (80–87% of wild type). None of these values is significantly different from RBC counts of 5.8 ± 0.1 x 10⁹ cells/ml (84% of wild type) in Kitl^Sl-gb/Kitl⁺ mice. However, in Kitl^Sl-31R/Kitl⁺ and Kitl^Sl-22R/Kitl⁺ mice, RBC counts are not significantly different from wild type (6.4 ± 0.2 x 10⁹ and 6.5 ± 0.3 x 10⁹ cells/ml, respectively; Fig 3A). The lack of effect on RBC counts in Kitl^Sl-31R/Kitl⁺ and Kitl^Sl-22R/Kitl⁺ mice is unexpected, given the clear pigmentation defects in these mice (see Table 1) and the severe effects on survival and P1 RBC counts of Kitl^Sl-31R/Kitl^Sl-31R and Kitl^Sl-22R/Kitl^Sl-22R mice (Fig 1 and Fig 2). Furthermore, neither the hemoglobin concentrations (Fig 3B) nor the hematocrit (Fig 3F) was affected in P24-P25 Kitl^Sl-31R/Kitl⁺ and Kitl^Sl-22R/Kitl⁺ mice. Similarly, neither hemoglobin concentrations nor hematocrits of Kitl^Sl-28R/Kitl⁺, Kitl^Sl-39R/Kitl⁺, Kitl^Sl-36R/Kitl⁺, and Kitl^Sl-d/Kitl⁺ mice were affected even though these mice had significantly reduced RBC counts.

The only parameters for which all nine P24-P25 heterozygous mutants had significant defects are MCV (Fig 3C) and MCH (Fig 3D). While the mean MCV of P24-P25 wild-type mice is 56.2 ± 0.4 fl, the mean MCV of P24-P25 heterozygous mutants ranged from 60.6 ± 0.3 fl (108%) in Kitl^Sl-36R/Kitl⁺ mice to 63.6 ± 0.6 fl (113%) in Kitl^Sl-42R/Kitl⁺ mice. For MCH, the mean of wild-type mice is 18.3 ± 0.2 pg, and the mean of heterozygous mutants ranged from 19.4 ± 0.2 pg (106%) in Kitl^Sl-22R/Kitl⁺ mice to 21.0 ± 0.1 pg (115%) in Kitl^Sl-28R/Kitl⁺ mice. Notably, the mean MCV (Fig 3C) and MCH (Fig 3D) of P24-P25 Kitl^Sl-31R/Kitl⁺ and Kitl^Sl-22R/Kitl⁺ mice are significantly increased relative to P24-P25 wild-type mice even though the RBC counts of these mice are not significantly different from that of wild-type mice (Fig 3A). However, it should be noted that the MCH is calculated by dividing the hemoglobin concentration by RBC counts. Since the MCV is increased in all mutants, the elevated MCH simply means that the hemoglobin concentration is higher because the cells are larger. Consistent with this, most of the MCHC values (Fig 3E) in mutant mice are not significantly different from those of wild-type mice (32.7 ± 0.3 g/dl). MCHC is calculated from hemoglobin concentrations, RBC counts, and MCV and therefore takes the volume of the cells into consideration. Thus, in mice heterozygous for each of the Kitl^Sl mutations, the increased hemoglobin concentration is proportional to the increased MCV, and the anemia is classified as macrocytic and normochromic.

As expected, the homozygous effects of each of the viable mutations (Kitl^Sl-39R, Kitl^Sl-36R, and Kitl^Sl-d) on RBC counts, hemoglobin, MCV, MCH, MCHC, and hematocrit are significantly greater than the corresponding heterozygous effects of these mutations (Fig 3). While RBC counts in wild-type P24-P25 mice were 6.9 ± 0.1 x 10⁹ cells/ml, RBC counts in Kitl^Sl-36R/Kitl^Sl-36R, Kitl^Sl-39R/Kitl^Sl-39R, and Kitl^Sl-d/Kitl^Sl-d mice were 4.6 ± 0.2 x 10⁹ cells/ml (67% of wild type), 3.9 ± 0.1 x 10⁹ cells/ml (56% of wild type), and 3.1 ± 0.1 x 10⁹ cells/ml (45% of wild type), respectively (Fig 3A). The effects on hemoglobin concentration (Fig 3B) and hematocrit (Fig 3F) in each homozygous viable mutant were of approximately the same magnitude as for RBC counts. For RBC counts, hematocrit, and hemoglobin, pairwise comparisons between the three homozygous mutants revealed that the Kitl^Sl-d/Kitl^Sl-d mutants have values for each of these three parameters significantly lower (P < 0.01 in each case) than those of the other two viable mutants. Moreover, the RBC count in Kitl^Sl-36R/Kitl^Sl-36R mice is significantly higher than that in Kitl^Sl-39R/Kitl^Sl-39R and Kitl^Sl-d/Kitl^Sl-d mice (P = 0.004 and P < 0.0001, respectively) but neither the hematocrit nor the hemoglobin concentration in the former mice is significantly different from that in either of the latter mice. Nonetheless, with respect to homozygous viable mutations, the trend for RBC counts, hematocrit, and hemoglobin concentration in P24-P25 mice is the same as for RBC counts in P1 mice; i.e., Kitl^Sl-36R/Kitl^Sl-36R mutants have the mildest effect, and Kitl^Sl-d/Kitl^Sl-d mutants have the most severe effect. Surprisingly, this trend was not observed with MCV. While the MCV of wild-type P24-P25 mice was 56.2 ± 0.4 fl, MCV of Kitl^Sl-39R/Kitl^Sl-39R, Kitl^Sl-36R/Kitl^Sl-36R, and Kitl^Sl-d/Kitl^Sl-d mice was 73.0 ± 0.5 fl (130%), 69.7 ± 0.7 fl (124%), and 71.4 ± 2.8 fl (127%), respectively. In particular, there is no difference (P = 0.6) in the MCV of Kitl^Sl-36R/Kitl^Sl-36R mice and Kitl^Sl-d/Kitl^Sl-d mice, even though the latter have only 67% of the RBC counts of the former (4.6 ± 0.2 x 10⁹ vs. 3.1 ± 0.2 x 10⁹, P < 0.0001). Interestingly, mice homozygous for either Kitl^Sl-39R or Kitl^Sl-d mutations had significantly increased MCHC (34.0 ± 0.2 and 36.0 ± 1.3 g/dl, respectively, compared to 32.7 ± 0.3 g/dl in wild-type mice, P < 0.01 for each pair). This suggests that hemoglobin concentration in these mutant mice is not proportional to the increased volume of RBCs, and the anemia could be classified as macrocytic and hyperchromic.

Relative effects of Kitl^Sl mutations:
The heterozygous and homozygous phenotypes observed for the seven ENU-induced Kitl^Sl mutations and Kitl^Sl-d are summarized in Table 4. The alleles were considered functionally null if their phenotype was similar to that of Kitl^Sl-gb. Four of the alleles (Kitl^Sl-30R, Kitl^Sl-31R, Kitl^Sl-28R, and Kitl^Sl-42R) clearly behave as null alleles, while one mutation (Kitl^Sl-22R) appears to be a strong hypomorph and three mutations (Kitl^Sl-39R, Kitl^Sl-36R, and Kitl^Sl-d) are clearly hypomorphic. While most of the Kitl^Sl mutations cause comparable effects on different aspects of the phenotype, the Kitl^Sl-31R and Kitl^Sl-22R mutations are unusual in that they do not cause any detectable effects on RBC counts in heterozygous mice.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this report we describe the effects of seven ENU-induced Kitl^Sl mutations on survival and peripheral blood cells of mice and compare the severity of these effects to those caused by two previously characterized Kitl^Sl mutations (Kitl^Sl-gb and Kitl^Sl-d). Careful analysis of the survival curves and extent of anemia of homozygous mutant mice has revealed that five of the seven new mutants are functionally null (or nearly so) and two of the seven new mutants are hypomorphic. There does not seem to be a relationship between phenotypic severity and type of mutation, as four of the Kitl^Sl missense mutants are null while one is hypomorphic and one of the Kitl^Sl truncation mutants is null while the other is hypomorphic. Although we do not know how each of the Kitl^Sl mutations described here affects Kitl function, it is likely that the mutations affect some aspect of Kitl conformation, processing, localization, or binding to Kit because none of the mutations affect steady-state levels of Kitl mRNA (RAJARAMAN et al. 2002 ). In this regard, the Kitl^Sl-22R mutation is particularly interesting. Several aspects of the phenotypic analysis suggest that the Kitl^Sl-22R mutant is hypomorphic, yet the L54P substitution, which is in the second -helical domain of Kitl, would be expected to have very severe effects on Kitl structure. Further structural studies of all of the mutants should provide useful information about structural requirements for Kitl function.

Three mutant alleles used in this study (Kitl^Sl-42R, Kitl^Sl-36R, and Kitl^Sl-d) are expected to produce truncated S-Kitl proteins with varying extents of C-terminal deletions (see Table 1 and RAJARAMAN et al. 2002 ). The Kitl^Sl-36R mutant potentially encodes two isoforms: Kitl^Sl-36R-A, which is 147 aa, and Kitl^Sl-36R-B, which is 96 aa with an additional 25 aa out of frame (RAJARAMAN et al. 2002 ). Since the Kitl^Sl-36R-B isoform contains the identical 96 N-terminal aa as Kitl^Sl-42R (RAJARAMAN et al. 2002 ), and the present studies indicate that the latter is null functionally, it is likely that the former is also null functionally. This is consistent with in vitro studies indicating that Kitl activity is abolished by deletions that remove more than the N-terminal 142 aa (NISHIKAWA et al. 1992 ). However, we cannot exclude the possibility that the Kitl^Sl-36R-B isoform is a gain-of-function mutant because of the abnormal C-terminal sequences. Regardless of which Kitl^Sl-36R isoform is biologically active, both are deleted for more C-terminal sequences than Kitl^Sl-d. It might therefore be expected that the latter mutation would have more severe effects on function. However, comparison of Kitl^Sl-d and Kitl^Sl-36R mutant mice reveals that all aspects of the heterozygous and homozygous RBC phenotype are less severe in Kitl^Sl-36R mutants than in Kitl^Sl-d mutants. Further studies of the mutant proteins encoded by each of these alleles will be necessary to understand these phenotypic observations.

The ENU-induced Kitl^Sl mutations described here cause mild pigmentation defects when present in the heterozygous condition and severe pigmentation defects in the homozygous or compound heterozygous condition. In addition, none of the heterozygous Kitl^Sl mutations caused a more severe effect on any aspect of the mutant phenotype described here than that of Kitl^Sl-gb. This indicates that none of these Kitl^Sl mutations acts in a dominant-negative manner, despite the facts that Kitl is known to function as a dimer (JIANG et al. 2000 ; ZHANG et al. 2000 ) and that some Kit^W mutants act in a dominant-negative fashion (NOCKA et al. 1990 ). Together, these observations conform to the well-known semidominant effects of Kitl^Sl mutations due to haploinsufficiency (BEDELL et al. 1996A ). Moreover, all of the Kitl^Sl mutations described here cause an increase in MCV and MCH when in the heterozygous condition, whereas other aspects of peripheral blood, such as RBC counts, hematocrit, and hemoglobin concentration, are normal in some heterozygous mutants (see Fig 3). Thus, increased MCV, MCH, and pigmentation are the most sensitive aspects of the Kitl^Sl semidominant phenotype. Whether the differences in effects on RBC counts and pigmentation reflect true differences between Kitl signaling in erythroid progenitors and melanoblasts or different thresholds for Kitl activity in the two cell types remains to be determined. Although the basis for the increased MCV is not known, it would be expected to relate directly to effects on RBC numbers. This is not always the case; for example, Kitl^Sl-d/Kitl^Sl-d mice have fewer RBCs than Kitl^Sl-36R/Kitl^Sl-36R mice, but the MCV in both mutants is the same (Fig 3A and Fig D).

If lethality to Kitl^Sl mice was always caused by anemia, then each of the lethal Kitl^Sl mutations would be expected to cause a more severe anemia than that of each of the viable Kitl^Sl mutations. The results described in this report are in basic agreement with this hypothesis. For example, P1 mice homozygous for each of the six lethal mutations (Kitl^Sl-gb, Kitl^Sl-30R, Kitl^Sl-31R, Kitl^Sl-22R, Kitl^Sl-42R, and Kitl^Sl-28R) have nearly identical RBC counts that are all significantly lower than the RBC counts in P1 mice homozygous for each of the viable mutations (Kitl^Sl-39R, Kitl^Sl-36R, and Kitl^Sl-d; Fig 2). However, in P1 mice, RBC counts in homozygous lethal Kitl^Sl mutants are only slightly less than that in hemizygous Kitl^Sl-d mice (see Fig 2). Although none of the latter mice survive until weaning, their survival time is significantly greater than that of mice homozygous for other lethal Kitl^Sl mutations (Fig 1I). In contrast, homozygous Kitl^Sl-d mice have a nearly normal survival curve (Fig 1I), and their RBC counts are only slightly higher than that in hemizygous Kitl^Sl-d mice. Thus, if Kitl^Sl-induced lethality is caused by severe RBC hypoplasia, these results suggest that the threshold for P1 RBC counts that allows survival to weaning may be between the mean values seen for hemizygous Kitl^Sl-d mice (0.9 x 10⁹ cells/ml, 22% of wild type) and homozygous Kitl^Sl-d mice (1.3 x 10⁹ cells/ml, 32% of wild type).

Although the Kitl^Sl-36R mutation allows survival in the majority of homozygous mice, a significant number of Kitl^Sl-36R/Kitl^Sl-36R mice die either perinatally or within the first week of birth. This could be explained if P1 Kitl^Sl-36R/Kitl^Sl-36R mice had large variations in RBC counts, with death occurring to mice with low RBC counts and survival occurring in mice with higher RBC counts. However, we did not observe such variation; in fact, RBC counts in all Kitl^Sl-36R/Kitl^Sl-36R P1 mice tested were very similar and are significantly higher than those seen in Kitl^Sl-d/Kitl^Sl-d mice. Therefore the perinatal and juvenile lethality of Kitl^Sl-36R/Kitl^Sl-36R mice does not seem to be reflected in the severity of the anemia. It is possible that some aspect of RBC function, rather than RBC numbers, is more drastically affected in a subset of Kitl^Sl-36R/Kitl^Sl-36R mice. Alternatively, development or function of some other cell type may be contributing to the lethality of Kitl^Sl-36R/Kitl^Sl-36R mice. In this regard, it is interesting to note that the absence of Kitl/Kit signaling in the enteric nervous system causes a lethal condition in mice (MAEDA et al. 1992 ; HUIZINGA et al. 1995 ). Whether such intestinal defects contribute to the lethality of the present Kitl^Sl mutants remains to be determined.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: Horizon Molecular Medicine, Norcross, GA 30071.
³ Present address: University Program in Genetics, Duke University, Durham, NC 27710.

	ACKNOWLEDGMENTS

We thank Drs. Neal Copeland and Nancy Jenkins for their support during the early stages of these experiments and for providing resources for embryo rederivation. We are grateful to Kent Wood and Melissa McNeely for excellent technical assistance. We also thank Drs. Eugene Rinchik and David Williams and two anonymous reviewers for their comments on the manuscript and Dr. Daniel Promislow for advice with statistical analysis. The mutants were generated in research jointly sponsored by the Office of Biological and Environmental Research, USDOE, at the Oak Ridge National Laboratory, managed by UT-Batelle, LLC, under contract DE-AC05-00OR22725, and by the National Institute of Environmental Health Sciences under interagency agreement no. Y1-ES-8048/0524-I119-A1. The material in this manuscript is based on work supported by the National Science Foundation under grant no. IBN-9728428 and the University of Georgia Research Foundation.

Manuscript received February 27, 2002; Accepted for publication June 12, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BEDELL, M. A., L. S. CLEVELAND, T. N. O'SULLIVAN, N. G. COPELAND, and N. A. JENKINS, 1996a  Deletion and interallelic complementation analysis of Steel mutant mice. Genetics 142:935-944.[Abstract]

BEDELL, M. A., N. G. COPELAND, and N. A. JENKINS, 1996b  Multiple pathways for Steel regulation suggested by genomic and sequence analysis of the murine Steel gene. Genetics 142:927-934.[Abstract]

BESMER, P., K. MANOVA, R. DUTTLINGER, E. J. HUANG, and A. PACKER et al., 1993  The kit-ligand (steel factor) and its receptor c-kit/W: pleiotropic roles in gametogenesis and melanogenesis. Development (Suppl.): 125–137.

BLUME-JENSEN, P., G. JIANG, R. HYMAN, K. F. LEE, and S. O'GORMAN et al., 2000  Kit/stem cell factor receptor-induced activation of phosphatidylinositol 3'-kinase is essential for male fertility. Nat. Genet. 24:157-162.[Medline]

BRANNAN, C. I., S. D. LYMAN, D. E. WILLIAMS, J. EISENMAN, and D. M. ANDERSON et al., 1991  Steel-Dickie mutation encodes a c-kit ligand lacking transmembrane and cytoplasmic domains. Proc. Natl. Acad. Sci. USA 88:4671-4674.[Abstract/Free Full Text]

FLANAGAN, J. G., D. C. CHAN, and P. LEDER, 1991  Transmembrane form of the kit ligand growth factor is determined by alternative splicing and is missing in the Sl^d mutant. Cell 64:1025-1035.[Medline]

HUANG, E. J., K. H. NOCKA, J. BUCK, and P. BESMER, 1992  Differential expression and processing of two cell associated forms of the kit-ligand: KL-1 and KL-2. Mol. Biol. Cell 3:349-362.[Abstract]

HUIZINGA, J. D., L. THUNEBERG, M. KLUPPEL, J. MALYSZ, and H. B. MIKKELSEN et al., 1995  W/kit gene required for interstitial cells of Cajal and for intestinal pacemaker activity. Nature 373:347-349.[Medline]

JIANG, X., O. GUREL, E. A. MENDIAZ, G. W. STEARNS, and C. L. CLOGSTON et al., 2000  Structure of the active core of human stem cell factor and analysis of binding to its receptor kit. EMBO J. 19:3192-3203.[Medline]

JORDAN, S. A., R. M. SPEED, and I. J. JACKSON, 1999  Deficiency of Trp53 rescues the male fertility defects of Kit(W-v) mice but has no effect on the survival of melanocytes and mast cells. Dev. Biol. 215:78-90.[Medline]

KISSEL, H., I. TIMOKHINA, M. P. HARDY, G. ROTHSCHILD, and Y. TAJIMA et al., 2000  Point mutation in kit receptor tyrosine kinase reveals essential roles for kit signaling in spermatogenesis and oogenesis without affecting other kit responses. EMBO J. 19:1312-1326.[Medline]

LAKY, K., L. LEFRANCOIS, and L. PUDDINGTON, 1997  Age-dependent intestinal lymphoproliferative disorder due to stem cell factor receptor deficiency: parameters in small and large intestine. J. Immunol. 158:1417-1427.[Abstract]

LEV, S., J. M. BLECHMAN, D. GIVOL, and Y. YARDEN, 1994  Steel factor and c-kit protooncogene: genetic lessons in signal transduction. Crit. Rev. Oncog. 5:141-168.[Medline]

MAEDA, H., A. YAMAGATA, S. NISHIKAWA, K. YOSHINAGA, and S. KOBAYASHI et al., 1992  Requirement of c-kit for development of intestinal pacemaker system. Development 116:369-375.[Medline]

MOTRO, B., J. M. WOJTOWICZ, A. BERNSTEIN, and D. VAN DER KOOY, 1996  Steel mutant mice are deficient in hippocampal learning but not long-term potentiation. Proc. Natl. Acad. Sci. USA 93:1808-1813.[Abstract/Free Full Text]

MOUSE GENOME DATABASE, 2002 Mouse Genome Informatics web site. The Jackson Laboratory, Bar Harbor, ME (http://www.informatics.jax.org/).

NISHIKAWA, M., A. TOJO, K. IKEBUCHI, K. KATAYAMA, and N. FUJII et al., 1992  Deletion mutagenesis of stem cell factor defines the C-terminal sequences essential for its biological activity. Biochem. Biophys. Res. Commun. 188:292-297.[Medline]

NOCKA, K., J. C. TAN, E. CHIU, T. Y. CHU, and P. RAY et al., 1990  Molecular bases of dominant negative and loss of function mutations at the murine c-kit/white spotting locus: W³⁷, W^v, W⁴¹ and W. EMBO J. 9:1805-1813.[Medline]

RAJARAMAN, S., W. S. DAVIS, A. MAHAKALI-ZAMA, H. K. EVANS, and L. B. RUSSELL et al., 2002  An allelic series of mutations in the Kit ligand gene of mice. I. Identification of point mutations in seven ethylnitrosourea-induced Kitl^Steel alleles. Genetics 162:331-340.[Abstract/Free Full Text]

RODEWALD, H. R., K. KRETZSCHMAR, W. SWAT, and S. TAKEDA, 1995  Intrathymically expressed c-kit ligand (stem cell factor) is a major factor driving expansion of very immature thymocytes in vivo. Immunity 3:313-319.[Medline]

RUSSELL, E. S., 1979  Hereditary anemias of the mouse: a review for geneticists. Adv. Genet. 20:357-459.[Medline]

RUSSELL, W. L., P. R. HUNSICKER, G. D. RAYMER, M. H. STEELE, and K. F. STELZNER et al., 1982  Dose-response curve for ethylnitrosourea-induced specific-locus mutations in mouse spermatogonia. Proc. Natl. Acad. Sci. USA 79:3589-3591.[Abstract/Free Full Text]

SARVELLA, P. A. and L. B. RUSSELL, 1956  Steel, a new dominant gene in the house mouse. J. Hered. 47:123-128.[Free Full Text]

SCHUMACHER, A., C. FAUST, and T. MAGNUSON, 1996  Positional cloning of a global regulator of anterior-posterior patterning in mice. Nature 384:648.[Medline]

ZHANG, Z., R. ZHANG, A. JOACHIMIAK, J. SCHLESSINGER, and X. P. KONG, 2000  Crystal structure of human stem cell factor: implication for stem cell factor receptor dimerization and activation. Proc. Natl. Acad. Sci. USA 97:7732-7737.[Abstract/Free Full Text]

This article has been cited by other articles:

				B. Billemont, H. Izzedine, O. Rixe, L. Graf, S. Gillessen, and W. Korte Macrocytosis Due to Treatment with Sunitinib N. Engl. J. Med., September 27, 2007; 357(13): 1351 - 1352. [Full Text] [PDF]

				M. Puttini, A. M. L. Coluccia, F. Boschelli, L. Cleris, E. Marchesi, A. Donella-Deana, S. Ahmed, S. Redaelli, R. Piazza, V. Magistroni, et al. In vitro and In vivo Activity of SKI-606, a Novel Src-Abl Inhibitor, against Imatinib-Resistant Bcr-Abl+ Neoplastic Cells Cancer Res., December 1, 2006; 66(23): 11314 - 11322. [Abstract] [Full Text] [PDF]

				C.L. Kennedy and M.K. O'Bryan N-ethyl-N-nitrosourea (ENU) mutagenesis and male fertility research Hum. Reprod. Update, May 1, 2006; 12(3): 293 - 301. [Abstract] [Full Text] [PDF]

				A. Mahakali Zama, F. P. Hudson III, and M. A. Bedell Analysis of Hypomorphic KitlSl Mutants Suggests Different Requirements for KITL in Proliferation and Migration of Mouse Primordial Germ Cells Biol Reprod, October 1, 2005; 73(4): 639 - 647. [Abstract] [Full Text] [PDF]

				M. A. Bedell and A. M. Zama Genetic Analysis of Kit Ligand Functions During Mouse Spermatogenesis J Androl, March 1, 2004; 25(2): 188 - 199. [Full Text] [PDF]

				S. Chandra, R. Kapur, N. Chuzhanova, V. Summey, D. Prentice, J. Barker, D. N. Cooper, and D. A. Williams A rare complex DNA rearrangement in the murine Steel gene results in exon duplication and a lethal phenotype Blood, November 15, 2003; 102(10): 3548 - 3555. [Abstract] [Full Text] [PDF]

<