Gene deletion of inositol hexakisphosphate kinase 1 reveals inositol pyrophosphate regulation of insulin secretion, growth, and spermiogenesis

Abstract

Inositol pyrophosphates, also designated inositol diphosphates, possess high-energy β-phosphates that can pyrophosphorylate proteins and regulate various cellular processes. They are formed by a family of inositol hexakisphosphate kinases (IP6Ks). We have created mice with a targeted deletion of IP6K1 in which production of inositol pyrophosphates is markedly diminished. Defects in the mutants indicate important roles for IP6K1 and inositol pyrophosphates in several physiological functions. Male mutant mice are sterile with defects in spermiogenesis. Mutant mice are smaller than wild-type despite normal food intake. The mutants display markedly lower circulating insulin.

Numerous inositol phosphates regulate biological functions with the best characterized, inositol-1,4,5-trisphosphate (IP₃), releasing intracellular calcium (1). Recent attention has focused on the inositol pyrophosphates, also designated inositol diphosphates, of which the best studied are diphosphoinositol pentakisphosphate [5-PP-I(1,2,3,4,6)P₅, here designated IP₇] and bisdiphosphoinositol tetrakisphosphate ([PP]₂-IP₄, IP₈) (2–4). These compounds are synthesized by a family of three inositol hexakisphosphate (IP₆) kinases (IP6Ks) (5–7). Inositol pyrophosphates have been implicated in a variety of physiologic functions including apoptosis (8, 9), endocytosis (10, 11), telomere length maintenance (12, 13), and chemotaxis (14). Additionally, another form of IP₇, tentatively identified as 4/6-PP-IP₅, is formed by the Vip1 enzyme in yeast and is implicated in the regulation of cell morphology, cell growth, and phosphate homeostasis (15, 16). Shears et al. have established that the mammalian Vip1 ortholog physiologically synthesizes IP₈, which is involved in osmotic regulation (17, 18). Inositol pyrophosphates may exert their functions in two ways, by binding to proteins or via phosphorylation. IP₇ can directly bind cytosolic proteins, such as the cyclin–cyclin-dependent kinase (CDK)-CDK inhibitor (CKI) complex, required for phosphate homeostasis in yeast (19). IP₇ can also compete with membrane phosphoinositides for binding to phosphoinositide-binding modules such as PH domains, as observed during chemotactic regulation in Dictyostelium (14). In addition, IP₇ physiologically phosphorylates proteins in a nonenzymatic fashion, analogous to S-nitrosylation, whereby the β-phosphate is transferred from IP₇ to previously phosphorylated proteins; hence, IP₇ pyrophosphorylates proteins (20, 21).

The various mammalian IP6Ks serve diverse functions. IP6K2 regulates apoptotic cell death (8, 9, 22, 23). IP6K1 has been implicated in the disposition of insulin and glucose. A putative disruption of the IP6K1 gene has been described in a family with type 2 diabetes (24). Illies et al. recently showed that selective depletion of IP6K1 by RNA interference in pancreatic beta cells impairs insulin secretion (25). To elucidate the physiologic role of IP6K1, we created mice with a targeted deletion of the exon encoding the C-terminal catalytic domain. We report that these mice display defects in insulin disposition, spermiogenesis, and growth.

Results and Discussion

We generated mice with a targeted deletion of the coding sequence of IP6K1 exon 6 (Fig. 1a). This sequence encodes amino acids 265–433, comprising the C-terminal catalytic domain of the protein. Male and female heterozygous mice carrying the IP6K1 knockout allele, were mated to produce wild-type (IP6K1^+/+) and IP6K1 knockout (IP6K1^−/−) littermates. RT-PCR analysis of tissues from IP6K1^−/− mice, using primers flanking exons 5 and 6, reveals loss of transcript in this region (Fig. 1b). As a control, we observe that the analogous region of mRNA encoding IP6K2 is preserved. Protein sequence comparison of IP6K1 with IP₃ 3-kinase A and inositol polyphosphate multikinase (IPMK), two other members of the IP kinase family whose crystal structures have been resolved (26–28), indicates that exon 6 encodes the C-terminal subdomain of IP6K1, which is required for ATP binding and catalytic activity. By inference, if the remainder IP6K1 mRNA (exons 1–5) is transcribed and stable, the resulting truncated form of IP6K1 would not possess any catalytic activity.

Fig. 1.

Generation and characterization of IP6K1 knockout mice. (a) Schematic representation of the mouse IP6K1 gene (Ihpk1) with coding regions shaded black, noncoding regions in white, and the exon number indicated on top. A close-up view of exons 5 and 6 in the wild-type, targeted and knockout alleles is shown. The knockout allele loses the splice site, coding region, and a portion of the 3′UTR of exon 6. (b) RT-PCR analysis of tissues from wild-type and IP6K1 knockout adult male mice. For both IP6K1 and IP6K2, the forward PCR primer was directed against exon 5 and the reverse primer against the coding region of exon 6. IP6K1 and IP6K2 cDNA in expression plasmids (pCMV) were used as positive controls. (c) (Upper) IP₆ kinase activity was assayed in extracts prepared from MEFs. Products (IP₇ and IP₈) derived from IP₆ are expressed as a percentage of IP₆ added to the reaction mix. Data are expressed as mean ± SE of duplicate determinations from two cell lines of each genotype. (Lower) To ensure that equal amounts of total protein were used for IP₆ kinase assays, the whole cell, nuclear, and cytosolic MEF extracts were quantified by Western blotting with the indicated antibodies. (d) IP₇ levels present in two MEF cell lines of each genotype were determined by equilibrium labeling with [³H]inositol and resolution of labeled inositol phosphates by HPLC. The HPLC fractions encompassing the IP₇ peak are shown. Levels of IP₇ in each cell line were normalized against the amount of [³H]inositol incorporated into the lipid fraction.

The relative importance of the three IP₆ kinases in generating IP₇ in different tissues has not been well characterized. We monitored the conversion of [³H]IP₆ to IP₇ and IP₈ in extracts of mouse embryonic fibroblasts (MEFs) prepared from IP6K1 mutant and wild-type embryos (Fig. 1c). In whole cell and cytosolic extracts, IP₇ and IP₈ formation are virtually undetectable in the knockout MEFs, whereas a major decrease is evident in nuclear extracts. In intact MEFs we also observed a marked diminution of IP₇ levels after incubation of the cells with [³H]inositol (Fig. 1d). Thus, in MEF cell lines, IP6K1 is the predominant source of IP₇ synthesis.

Mating studies reveal a major abnormality in male fertility (Table 1). No pregnancies emerge when homozygous knockout males are mated with wild-type or homozygous knockout females. However, normal rates of pregnancy and litter size are observed when the same females are paired with heterozygous males, suggesting that oogenesis is normal in IP6K1 knockout females. Offspring from wild-type, heterozygous, or knockout females paired with heterozygous males follow expected Mendelian ratios of inheritance (data not shown). Histologic examination of the testes and epididymis of IP6K1^−/− males reveals very few advanced spermatids in the seminiferous tubules and no sperm in the epididymis (Fig. 2). Thus, IP6K1 evidently plays an important role in spermiogenesis, the final stage of spermatogenesis during which round spermatids develop into mature, motile spermatozoa.

Table 1.

IP6K1 knockout male mice are sterile, whereas females are fertile^*

Genotype		Rate of pregnancy	Litter size
Male	Female
−/−	+/+	0/5	0
−/−	−/−	0/5	0
+/−	+/+	4/5	8.8 ± 1
+/−	−/−	4/5	7.8 ± 0.9
+/−	+/−	7/12	8.5 ± 1

*Three-month-old males and females of the specified genotype were paired for 7 days. The number of pregnant females and the number of pups in each litter were counted. Litter sizes are expressed as mean ± SE for each female.

Fig. 2.

Spermiogenesis is impaired in male mice lacking IP6K1. Histological analysis of testes and epididymides from wild-type and IP6K1 knockout adult male mice are shown. Cross-sections of the testes [wild-type (a and b) and knockout (e and f)] and epididymides [wild-type (c and d) and knockout (g and h)] were imaged at ×40 and ×100 magnification as indicated.

Male mutant mice are smaller than wild-type with a 15–20% reduction in weight evident consistently from 4 to 10 weeks (Fig. 3a) and up to 5 months (data not shown), despite normal food intake (Fig. 3b). Female knockout mice also display decreased body weight compared with their wild-type littermates (data not shown). We monitored plasma growth hormone levels, which are the same in mutant and wild-type mice (data not shown).

Fig. 3.

IP6K1 knockout mice display growth retardation and altered insulin disposition. (a) The body weight of wild-type and IP6K1 knockout male littermates, measured once per week, is expressed as mean ± SE (n = 6–8). Two-way ANOVA shows significant difference between curves, P < 0.01. (b) The average daily food intake of 6-week-old male mice is expressed as mean ± SE (n = 6–8). There is no significant difference in food intake between wild-type and knockout mice (Student's t test, P > 0.1). (c) Serum insulin levels were measured in wild-type and IP6K1 knockout adult male mice, either after fasting for 4 or 16 h (overnight). Data are mean ± SE (n = 5–10). (d) Insulin tolerance test was performed by 0.75 unit/kg body weight i.p. insulin injection in adult male mice. Data are mean ± SE (n = 4). Two-way ANOVA indicates that the curves for insulin tolerance are significantly different (P < 0.05) between IP6K1 knockout and wild-type mice.

Because of prior evidence implicating inositol pyrophosphates in insulin disposition, we monitored plasma insulin and detected a striking 65–70% reduction in the mutant mice (Fig. 3c). This data correlates with the observations of Illies et al. (25) that RNA silencing of IP6K1 in pancreatic beta cells inhibits exocytosis of insulin-containing granules from the readily releasable pool. Therefore, loss of IP6K1 in mice most likely lowers insulin secretion from pancreatic beta cells and consequently lowers the levels of circulating insulin.

Because long-term deficiency of insulin might elicit supersensitivity of the insulin-response system, we evaluated insulin tolerance (Fig. 3d). The mutant mice display a greater diminution of blood glucose levels in response to insulin than wild-type mice, consistent with higher glucose uptake by tissues in response to insulin.

We wondered whether the apparent supersensitivity to insulin compensates for the impaired insulin secretion and, so, monitored blood glucose, glucose tolerance, and levels of glycosylated hemoglobin (Fig. 4). Mutant mice display normal blood glucose levels whether fed or fasted, manifest a normal glucose tolerance curve, and have normal levels of glycosylated hemoglobin.

Fig. 4.

Mice lacking IP6K1 are not diabetic despite low insulin levels. (a) Tail vein blood glucose measured in ad libitum-fed or 16-hour-fasted adult male mice. Data are mean ± SE (n = 7 or 8). (b) Glucose tolerance test by 2 g/kg body weight i.p. glucose injection in adult male mice. Data are mean ± SE (n = 6 or 7). (c) Measurement of glycosylated hemoglobin (HbA1c) in whole blood from adult male and female mice. Data are mean ± SE (n = 5).

How might the phenotypic abnormalities observed in the mutant mice relate to the known functions of IP6K1? Inositol pyrophosphates have been implicated in processes that regulate general metabolism, protein synthesis, cell growth, and cell death. Thus, IP₇ regulates telomere elongation (12, 13), several targets of IP₇ pyrophosphorylation participate in ribosomal biogenesis (4, 21), IP₇ levels fluctuate during the cell cycle (29), and IP6K2 plays a prominent role in apoptotic cell death (8, 9). Some of these actions may be relevant to the phenotype we have observed. There is abundant evidence for a role of IP₇ in vesicular trafficking. Yeast with deletion of IP6K manifests striking defects in vesicular endocytosis and vacuolar morphology (10, 11, 30). Clathrin-associated proteins that regulate endocytosis and exocytosis bind IP₆ and IP₇ with selective high affinity (11, 31, 32). The β3 subunit of the adaptor protein AP3 is pyrophosphorylated by IP₇ (20). The abnormalities of the IP6K1 mutant mice may be related to processes in which vesicular secretion is prominent. Insulin release occurs directly through vesicular secretion. The defects in spermiogenesis could result from multiple factors, including vesicular trafficking events or the secretion of regulatory endocrine and paracrine factors (33–35). We do not have insights into the molecular mechanisms underlying the growth defects of the mutant mice, but insulin itself participates in growth dynamics, and impairment of insulin secretion can lead to growth deficiency (36). Disruption of components of the insulin-signaling pathway can result in infertility and growth retardation (37–39). Defects in insulin secretion after IP6K1 deletion may be linked to the regulation of Ca²⁺-dependent activator protein for secretion (CAPS), a protein required for Ca²⁺-triggered exocytosis of insulin (40) and which possesses motifs for IP₇-mediated pyrophosphorylation.

In summary, the phenotype of IP6K1-deleted mice establishes that cellular actions of inositol pyrophosphates, including regulation of vesicular trafficking, impact organismic functions such as hormonal disposition, growth, and fertility.

Materials and Methods

Generation and Maintenance of IP6K1 Knockout Mice.

The gene encoding mouse IP6K1, Ihpk1, is located on chromosome 9 and has 6 exons, of which exon 1 is exclusively noncoding (Fig. 1a). The start codon is located in exon 2, and the stop codon is located in exon 6. Exon 6 is composed of 3,318 bp, of which the first 510 bp encode the C-terminal region of IP6K1 (amino acids 265–433), and the remaining 2,808 bp give rise to the 3′ untranslated region (UTR) of the mRNA. Our strategy was to delete the coding region of exon 6, including the exon 6 splice site. The resulting transcript would lack exon 6 (including the 3′UTR) and therefore be unstable, resulting in a complete knockout. Alternatively, cryptic splicing may occur between exon 5 and the remainder of exon 6, possibly generating a stable mRNA lacking the exon 6 coding sequence, resulting in a catalytically inactive, truncated protein lacking the C-terminal subdomain.

IP6K1^+/− mice were generated at Ozgene. The targeting construct was based on the sequence of the C57BL/6 strain Ihpk1 gene (Ensembl gene ID: ENSMUSG00000032594). A LoxP site was inserted between exons 5 and 6, 295 bp upstream of the exon 6 splice site. A phosphoglycerine kinase (PGK) Neo cassette flanked by FLP recombinase target (FRT) sequences and another LoxP site was inserted 1,464 bp downstream of the stop codon in exon 6. The targeting vector was electroporated into 129SV/J ES cells, and Neomycin resistant ES cells were microinjected into C57BL/6 blastocysts and implanted into pesudo-pregnant female mice. The resulting chimeric mice were crossed with knockin C57BL/6 mice carrying Cre recombinase driven by a PGK promoter to generate heterozygous mice carrying the IP6K1 knockout allele and Cre recombinase (IP6K1^+/−/Cre). These mice were interbred and F₁ heterozygous mice lacking Cre (IP6K1^+/−) were subsequently interbred to generate wild-type (IP6K1^+/+) and IP6K1 knockout (IP6K1^−/−) littermates used for phenotypic and metabolic studies.

All mice were maintained on a 129SV-C57BL/6 mixed background. Animal care and experimentations were approved by the Johns Hopkins University Animal Care and Use Committee. Mice were housed in a 12-h light/12-h dark cycle, at an ambient temperature of 22°C, and fed standard rodent chow.

Genotyping and RT-PCR Analysis.

Mice were genotyped by PCR analysis of genomic DNA from tail biopsies. The detailed genotyping strategy is described in the supporting information (SI) Materials and Methods.

RT-PCR analysis was used to assess whether the IP6K1 transcript was missing in the knockout mice. RNA was prepared from tissues obtained from IP6K1 knockout mice and their wild-type littermates by using TRIZOL reagent (Invitrogen). cDNA was prepared by using Oligo(dT) primers (Invitrogen) and the Omniscript reverse transcriptase kit (Qiagen). PCR was performed by using cDNA templates and primers flanking exons 5 and 6 of IP6K1 (see SI Fig. 5). IP6K2 primers were used as a positive control to assess the quality of cDNA.

Generation of MEF Cell Lines.

Wild-type (IP6K1^+/+) and mutant (IP6K1^−/−) E14 embryos were isolated from a single heterozygous female that had been paired with a heterozygous male. The head and organs were removed, and the remaining carcasses were minced and incubated in trypsin to obtain single cells. These primary fibroblasts were transfected with a plasmid encoding the SV40 Large T antigen (Addgene plasmid 9053: pSG5 Large T), and passaged five times at split ratios of 1:10 to yield independent, immortalized MEF cell lines obtained from separate embryos. Two cell lines of each genotype (IP6K1^+/+ and IP6K1^−/−) were used for experiments. Cells were cultured in DMEM (Invitrogen) supplemented with 10% FBS (Gemini Bio-Products), l-glutamine (2 mM; Invitrogen), and penicillin (100 units/ml)/streptomycin (100 μg/ml) (Invitrogen).

Measurement of IP₆ Kinase Activity.

Extracts were prepared from MEFs by sonication in the presence of 50 mM Hepes (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 1% Triton X-100, protein phosphatase inhibitor mixture (Sigma), and protease inhibitor mixture (Sigma). Nuclear and cytosolic fractions from MEFs were obtained by using a nuclear/cytosol fractionation kit (BioVision). Protein content in the extracts was quantified by using the Bradford assay.

Whole cell, nuclear, or cytosolic extracts were used in an IP₆ kinase activity assay. Extracts (35 μg total protein) were incubated in assay buffer [20 mM Tris·HCl (pH 7.4), 6 mM MgCl₂, 1 mM DTT], in the presence of [³H]IP₆ (100 μM, 150,000 dpm, Perkin–Elmer Life Sciences), ATP (5 mM), Creatine kinase (40 units/ml), and phosphocreatine (10 mM), at 37°C for 1 h. Assays were terminated by the addition of Na₂EDTA (5 mM), followed by perchloric acid (0.6 M), and quenched as described previously (2). [³H]IP₆ and its products, [³H]IP₇ and [³H]IP₈, were resolved by HPLC with a Partisphere SAX column (Whatman), with a gradient generated by mixing Buffer A (1 mM Na₂EDTA) and Buffer B [Buffer A plus 1.3 M (NH₄)₂HPO₄ (pH 3.85)] as follows: 0–5 min, 0% B; 5–10 min, 0–30% B; 10–60 min, 30–100% B; 60–75 min, 100% B. One-milliliter fractions were collected, and the radiolabeled fractions were identified by using a liquid scintillation counter. When comparing IP₆ kinase activity from extracts prepared from different cell lines, total, nuclear, or cytoplasmic protein content was normalized by Western blotting by using antibodies against β-tubulin (Upstate Biotechnology), p84/N5 (GeneTex), or lactate dehydrogenase (Chemicon International), respectively.

Analysis of Inositol Phosphate Content in Cells.

MEFs were plated at a density of 2 × 10⁶ cells per 10-cm dish, then labeled with 200 μCi (1 Ci = 37 GBq) [³H]inositol (Perkin–Elmer Life Sciences) for 5 days. Soluble inositol phosphates were extracted from labeled cells as described previously (41). Inositol incorporated into lipids was measured by extracting the remaining cell pellet with 0.1 M NaOH, 0.1% Triton X-100 overnight at room temperature with shaking and counting a fraction of the solubilized material in a liquid scintillation counter. The [³H]-labeled inositol phosphates were resolved by HPLC as described above. Soluble inositol phosphate levels were normalized against total lipid inositol content for each cell line.

Phenotypic Analysis of Mice.

Histological analysis of the testes and epididymis was performed by the Phenotyping Core in the Department of Molecular and Comparative Pathobiology, Johns Hopkins University. Eleven-week-old male mice (IP6K1^+/+ and IP6K1^−/− littermates) were killed with carbon dioxide, and the dissected tissue specimens were fixed in 10% neutral buffered formalin for 24 h before trimming for histology processing. Paraffin-embedded tissue sections were stained with hematoxylin/eosin according to standard protocols.

To monitor weight gain, male mice (n = 6–8) were weighed once per week. To measure daily food intake, singly housed, 6-week-old male mice were given a preweighed amount of rodent chow (150 g), and the remaining food (to the nearest 0.1 g) was weighed once a day, at 1500 h, for the next 3 days. The average amount of food consumed per mouse per day was calculated as mean ± SE (n = 6–8). The mice drank water ad libitum throughout the study.

Metabolic Analysis.

To measure serum insulin levels, blood was collected by facial vein puncture from 4- to 6-month-old male mice (n = 5–10), after they were starved either for 4 h or overnight (16 h). Serum was obtained by centrifugation at 10,000 × g for 10 min at 4°C, and insulin levels were measured by using the Insulin (Mouse) Ultrasensitive EIA kit (Aplco Diagnostics). Glycosylated hemoglobin (HbA1c) levels were measured in whole blood collected in K₂EDTA-coated tubes by facial vein puncture from 4- to 6-month-old male and female mice by using the A1CNow⁺ monitor (Metrika).

Blood glucose levels were measured from tail vein bleeding of overnight (16 h) fasted or ad libitum-fed 5 month old male mice (n = 7 to 8) by using an Ascensia Contour blood glucose meter and test strips (Bayer). For glucose tolerance tests, mice were fasted for 4–6 h and injected i.p. with 2 g/kg body weight glucose. Blood glucose was measured at 0, 15, 30, 60, 90, and 120 min by tail vein bleeding. For insulin tolerance tests, mice were fasted for 4 h and were given 0.75 units/kg body weight human regular insulin (Sigma) i.p. Blood glucose measurements were obtained from tail veins at 0, 15, 30, 60, 90, and 120 min postinjection.

Statistics.

The data are presented as mean ± SEM. All statistical analyses were performed by using GraphPad Prism. Differences between two groups were assessed by unpaired Student's t test. Multiple parameters were analyzed by ANOVA test. P < 0.05 was considered significant.