Spermine Synthesis Is Required for Normal Viability, Growth, and Fertility in the Mouse*

Spermidine is essential for viability in eukaryotes but the importance of the longer polyamine spermine has not been established. Spermine is formed from spermidine by the action of spermine synthase, an aminopropyltransferase, whose gene (SpmS) is located on the X chromosome. Deletion of part of the X chromosome that include SpmS in Gy mice leads to a striking phenotype in affected males that includes altered phosphate metabolism and symptoms of hypophosphatemic rickets, circling behavior, hyperactivity, head shaking, inner ear abnormalities, deafness, sterility, a profound postnatal growth retardation, and a propensity to sudden death. It was not clear to what extent these alterations were due to the loss of spermine synthase activity, since this chromosomal deletion extends well beyond the SpmS gene and includes at least one other gene termed Phex. We have bred the Gy carrier female mice with transgenic mice (CAG/SpmS mice) that express spermine synthase from the ubiquitous CAG promoter. The resulting Gy-CAG/SpmS mice had extremely high levels of spermine synthase and contained spermine in all tissues examined. These mice had a normal life span and fertility and a normal growth rate except for a reduction in body weight due to a loss of bone mass that was consistent with the observation that the derangement in phosphate metabolism is due to the loss of the Phex gene and was not restored. These results show that spermine synthesis is needed for normal growth, viability, and fertility in male mice and that regulation of spermine synthase content is not required.

Polyamines are ubiquitous cellular components that play many roles in cellular physiology (1). Bacteria such as Escherichia coli can grow without polyamines albeit at a reduced rate (2) and a greater susceptibility to oxidative damage (3). However, eukaryotes require spermidine (4). Deletion of either the ornithine decarboxylase gene (5) or the S-adenosylmethionine decarboxylase gene (6) prevents synthesis of spermidine and is lethal at very early embryonic stages in mice. It is possible that this requirement is due to the fact that spermidine is needed as a precursor of hypusine, a post-translational modification of the protein eIF-5A (7)(8)(9). This modification has been shown to be essential for viability in yeast (10,11). Polyamines clearly have multiple functions in addition to serving as precursors of hypusine, and these may also provide essential functions particularly in higher eukaryotes. Known functions of polyamines include: interactions with DNA and RNA that can modulate chromatin stability and gene expression; direct effects on activity of transcription factors, enzymes, and phospholipid membranes; anti-oxidant effects; and interactions with ion channels (12)(13)(14)(15)(16). Polyamines may also either promote or prevent apoptosis (17,18). In most cases where polyamine functions unrelated to hypusine have been studied in vitro or in cultured cells, both spermidine and spermine are effective. This raises the question as to why spermine is formed.
Aminopropyltransferases are very widely distributed and the basic active site motif has been identified in structural studies (19). Despite their similarity in this motif, there are two totally distinct aminopropyltransferases in eukaryotes, spermidine synthase and spermine synthase, and there is no overlap in specificity (20,21). Spermine is normally a substantial fraction of the total polyamine pool in such organisms. Although it seems most unlikely that the capacity to synthesize spermine has been so highly conserved without a specific function for spermine, the specific functions of spermine remain elusive, and yeast mutants that lack spermine synthase are viable and grow at a normal rate (22). Similarly, cultured mammalian cells that lack spermine due to either the use of inhibitors of spermine synthase or inactivation of the spermine synthase gene (SpmS) are viable and grow normally (23)(24)(25)(26)(27). In the more complex environment of a multicellular organism, it remains possible that spermine is essential due to a requirement in some specialized cell type or to a need for spermine under certain environments.
Although one obvious way to examine this question is to specifically inactivate the SpmS gene, it has not yet been possible to obtain knockout mice with such a deletion. Attempts to generate mice that lack spermine synthase by standard gene deletion techniques have not been successful, since on the 129/SvJ background the mice lacking SpmS were not viable (26). Mice lacking spermine synthase have been identified as a result of exposure to mutagens. The male Gyro (Gy) mouse, which is on the B6C3H background, has a chromosomal deletion of part of the X chromosome in the Xp22.1 region, which spans 160 -190 kb. This includes most of SpmS and at least one other gene termed Phex (28 -30). The Phex gene product is known to regulate phosphate metabolism in a way in which loss of a single copy is dominant; both male Gy and female XGy mice have symptoms of hypophosphatemia. In addition to this, the male Gy mice have a variety of other phenotypic changes that include inner ear abnormalities, deafness, hyperactivity, circling behavior, sterility, a profound growth retardation, and a propensity to sudden death (Refs. 31 and 32 and references therein). Gy mice survive only on the B6C3H background; attempts to transfer the gene deletion to other strains have not been successful. There was a complete absence of spermine in tissues of the male Gy mice (24,28,29). These results suggest that the absence of spermine may be the cause of some or all of the alterations in male Gy mice described above.
To examine this possibility, we have bred the XGy carrier female mice with transgenic male CAG/SpmS mice from lines 8 and 21 (33). These mice express spermine synthase under the control of a composite cytomegalovirus-immediate early enhancer/chicken ␤-actin promoter, which was designed to provide ubiquitous expression (34 -36). Assays of the spermine synthase activity in CAG/SpmS lines 8 and 21 confirmed that there was a high level of expression of the transgene in many different organs and that this was maintained for at least one year (33). The male Gy-CAG/SpmS mice resulting from crossing these lines with female XGy mice contained spermine (with a slightly higher spermine:spermidine ratio than present in controls), and the normal life span and fertility were restored. They also grew at a normal rate, although their body weight was reduced slightly compared with controls. This reduction is probably due to the hypophosphatemia, which was not affected by the SpmS transgene. These results establish that, in mammals, spermine synthesis is needed for normal viability, growth, and fertility.

EXPERIMENTAL PROCEDURES
Materials-All chemicals, unless noted, were purchased from Sigma. Oligodeoxyribonucleotides used as primers were synthesized in the Macromolecular Core Facility, Pennsylvania State University College of Medicine. The reagents used for PCR genotyping, Taq DNA polymerase, and an Ultrapure dNTP set were purchased from Promega (Madison, WI) and Amersham Biosciences, respectively. 35 S-Labeled decarboxylated AdoMet was synthesized from L-[ 35 S]methionine (PerkinElmer Life Sciences) (24,37).
Polyamine and Enzyme Analysis-Polyamine content was determined by HPLC using an ion pair reverse phase high performance liquid chromatography separation method with post-column derivatization using o-phthalaldehyde as described previously (38). Spermine synthase assays were carried out by measuring the production of [ 35 S]methylthioadenosine from 35 S-labeled decarboxylated AdoMet in the presence of the appropriate amine acceptor as described previously (24,37).
Mice-Heterozygous Gy mice were a kind gift from Dr. R. A. Meyer, Jr. of the Department of Orthopaedic Surgery, Carolinas Medical Center, Charlotte, NC. Other Gy mice were purchased from The Jackson Laboratory (Bar Harbor, ME). XGy carrier females were bred with B6C3H males from The Jackson Laboratory. Female offspring were retained and the heterozygous females carrying the Gy deletion required for breeding identified by phosphate analyses of blood plasma (31). The Gy males resulting from the breeding of these mice were identified by PCR using genomic DNA isolated from the tails (using the DNeasy tissue kit, Qiagen, Inc., Valencia, CA). The 5Ј sense primer used was 5Ј-GGTGTTGCTGGACCTTCAGA-3Ј, and the 3Ј antisense primer was 5Ј-CCCAGTACTGTCTTGACTCA-3Ј. The amplified product at the predicted size of 631 bp was absent in Gy mice and present in controls. A band corresponding to 144 bp, which is the size derived from an inactive pseudogene, was present in both normal and Gy males (24,28).
The CAG/SpmS mice were generated by DNA microinjection of fertilized B6D2F2 oocytes using a 3.5-kb fragment released by SalI and BamHI digestion from plasmid pCAG-hSpmSyn, which was constructed by replacing the insert encoding green fluorescent protein in plasmid pCX-EGFP (34,35) with the human spermine synthase cDNA (33). Genomic DNA was isolated from the tails as above and subjected to PCR analysis to identify mice bearing the transgene. For identification of the CAG/SpmS, the 5Ј sense primer used was 5Ј-TTCGGCTTCT-GGCGTGTGAC-3Ј, which corresponds to a sequence in the actin promoter region and the 3Ј antisense primer was 5Ј-CCAGTACTG-TCCTGACTC-3Ј, which corresponds to nucleotides 300 -317 in the spermine synthase coding sequence. A 440-bp fragment was produced from the transgene. Phosphate Assays-The plasma level of phosphate was measured by a colorimetric assay. Blood from a retro-orbital bleed was collected in heparinized microcentrifuge tubes and spun at 14,000 rpm (15,000 ϫ g) for 15 min, and then plasma was prepared for analysis. Aliquots of 20 l of plasma were mixed with 1.35 ml of 10% trichloroacetic acid and centrifuged at 14,000 rpm (15,000 ϫ g) for 10 min. One ml of supernatant was placed into a clean glass tube and mixed with 2 ml of 2% ascorbic acid, 0.5% ammonium molybdate, and 1.2 N H 2 SO 4 . Samples were then incubated at 37°C for 90 min and cooled for 10 min at room temperature, and absorbance was measured on a spectrophotometer at 820 nm.
Histology and Immunohistochemistry-Tissues were fixed in methyl Carnoy's solution (60% methanol, 30% chloroform, and 10% acetic acid) for 24 h. The fixed testes were embedded in paraffin wax, and 5-m sections were stained with hematoxylin and eosin for histological analysis.
Staining for PCNA 1 was carried out as follows. After deparaffinization and rehydration, sections were incubated with 3% hydrogen peroxide to quench endogenous peroxidases. Antigen retrieval was performed by heating to 92°C in antigen unmasking solution (Vector H-3300) three times with 5-min incubations in between in a microwave at 20% power. Slides were cooled to room temperature for at least 30 min. They were then blocked with nonspecific serum (10% goat serum in phosphate-buffered saline containing 0.5% bovine serum albumin, Vector S-1000) for 1 h. The slides were then incubated with mouse anti-PCNA monoclonal antibody (1:500; Santa Cruz Biotechnology catalog number SC-56) in phosphate-buffered saline containing 0.5% bovine serum albumin overnight at 4°C. They were then incubated with biotinylated goat anti-mouse IgG polyclonal antibody (1:500; Dako catalog number E0433) for 45 min followed by Streptavidin-horseradish peroxidase (1:1000; Dako catalog number P0397). Immunostaining was detected with AEC peroxidase substrate kit (Vector SK-4200) and counter-stained with hematoxylin QS (Vector H-3404). Slides were dehydrated and mounted with permount (Cytoseal-60). Digital photographs were taken with a Nikon-TMS microscope and Nikon Coolpix 950 camera.

Restoration of Polyamines in Gy
Mice-In agreement with previous reports (24,28,29), tissues of male Gy mice had no spermine synthase activity (Table I) and contained insignificant amounts of spermine (Table II). Spermine was restored in the tissues in the Gy-CAG/SpmS mice (Table II). Although the amount of spermine synthase that is produced from the CAG-driven transgene is greatly in excess of that present in normal mice (Table I), the spermine levels in the Gy-CAG/ SpmS mice were only slightly elevated in most tissues (Table II).
Tissues from the Gy mice have a large increase in spermidine compared with normal mice, which is consistent with the in- 1 The abbreviation used is: PCNA, proliferating cell nuclear antigen. ability to further metabolize this polyamine (Table II). This increase was completely prevented in the Gy-CAG/SpmS mice in which there was frequently a small decrease in spermidine compared with the normal mice. This decrease in spermidine and small increase in spermine results in a statistically significant rise in spermine:spermidine ratio in most tissues of the Gy-CAG/SpmS mice compared with normal mice. This alteration in ratio was also seen in the CAG/SpmS mice compared with normals as reported previously (33).
Although the results shown in Table II were for polyamines measured at 1 month, the alterations in polyamine metabolism described above were also seen in mice that were 1 year old (results not shown). None of the tissues from Gy mice of this age contained spermine; the spermine levels in the Gy-CAG/ SpmS mice were similar to control values, and the rise in spermidine in the Gy mice was completely prevented in the Gy-CAG/SpmS mice.
Survival, Growth, and Behavior of Gy and Gy-CAG/SpmS Mice-The Gy mice had a very poor survival over the first 10 weeks of life with sudden death occurring in more than 70% of the population. In contrast, the Gy-CAG/SpmS mice showed no mortality over this time period (Fig. 1 and results not shown) and were kept for 1 year with no difference in survival from normal mice (Fig. 1). The poor viability of the Gy mice is reflected in the numbers of Gy mice that survive to the time of weaning. In previous studies in our laboratory after breeding XGy carrier females with normal males, less than the expected 25% of offspring surviving to weaning were Gy males (24). This difference was also seen in the current experiments and it was corrected in the Gy-CAG/SpmS mice. At 21 days of age, there were 17% Gy males (significantly less than the expected 25%) and 26% Gy-CAG/SpmS mice in 91 male offspring from the breeding of XY females with CAG/SpmS males. In contrast to the Gy mice that were much smaller than the corresponding normal males, the Gy-CAG/SpmS were not noticeably different from normal mice or CAG/SpmS mice for the first 4 weeks of life (Fig. 2). At later ages, the Gy-CAG/SpmS were slightly smaller than the control or CAG/SpmS mice (Fig.  2), but at all times their body weight was more than twice that of Gy mice of the same age. The small reduction in body weight is likely to be related to hypophosphatemia in these mice. The phosphate level, which is significantly decreased in Gy mice, was not increased in the Gy-CAG/SpmS mice (Table III). The hypophosphatemia in Gy mice leads to rickets/osteomalacia (31,32). The femur length in the 12-month-old Gy mice was significantly reduced, and this decrease was not fully reversed in the Gy-CAG/SpmS mice.
Gy mice were originally named for their hyperactivity and circling behavior (31,32). This trait, which is more easily observed at older ages, was also completely abolished in the Gy-CAG/SpmS mice.
Fertility of Gy and Gy-CAG/SpmS Mice-The Gy mice were sterile, as reported previously (31,32). Histological examination of the seminiferous tubules in these mice revealed a large decrease in the number of meiotic and postmeiotic cells (Fig. 3). The proportion of cells staining positive for PCNA was increased in every layer of the seminiferous tubules in the Gy mice indicating that the cells remained at spermatogonia and early stage primary spermatocytes and rarely progressed to spermatozoa (Fig. 4). These changes in the testicular histology and PCNA staining were totally reversed in the Gy-CAG/SpmS mice (Figs. 3 and 4). Consistent with this, these mice were as fertile as normal male mice. Using the Gy-CAG/SpmS line, breeding with normal B6C3H female mice or XGy female mice or CAG/SpmS female mice was readily accomplished. As expected, all of the female offspring from these crosses had at least one Gy gene. There were no viable GyGy females from the cross between Gy-CAG/SpmS and XGy, but there were 3/16 female offspring that had the GyGy-CAG/SpmS genotype. This suggests that the absence of spermine synthase activity may be lethal in female mice in the mixed B6C3H/B6D2 background.
Comparison of Gy-CAG/SpmS-8 and Gy-CAG/SpmS-21 Mice-Lines 8 and 21 of the CAG/SpmS mice, which were derived from two different founder animals, had very similar levels of spermine synthase with a high degree of overexpression in all lines (33). The studies for which results are shown above were carried out primarily with line 8 but identical results were obtained with regard to restoration of polyamine content, growth, survival, and fertility when line 21 was used. DISCUSSION Characterization of the CAG/SpmS mice has indicated that despite the very large increase in spermine synthase activity in many tissues, there was only a modest change in cellular polyamines with a consistent but small rise in the spermine:spermidine ratio (33). These results indicate that, as long as a certain critical level is exceeded, the absolute level of spermine synthase activity is not a critical factor in maintaining cellular polyamines. Other regulatory factors, which have not currently been identified, are sufficient to maintain polyamine levels. The finding that the Gy-CAG/SpmS mice have relatively similar levels of polyamines to normal littermates in all of the tissues examined (Table II) is therefore consistent with previous studies. The even closer similarity in polyamine content between the Gy-CAG/SpmS and CAG/SpmS mice is not surprising, since the additional spermine synthase contributed by the endogenous gene in the latter strain is minimal compared with the high level supplied from the transgene (Table I and Ref. 33). The CAG enhancer/promoter construct was designed to give ubiquitous expression of transgenes by combining a chicken ␤-actin promoter and a cytomegalovirus enhancer sequence. Studies with mice in which it was used to produce green fluorescent protein show general expression (34 -36). Our characterization of the CAG/SpmS mice has indicated expression of the transgene in all tissues examined, although the spermine synthase activity in the latter was greatest in heart and muscle, which may relate to the ␤-actin promoter in the CAG construct (33). Therefore it is reasonable to postulate that the capacity to produce spermine is restored in all tissues of the Gy-CAG/SpmS mice and that this restoration is related to the normal survival, fertility, behavior, and growth of these mice.
Transgenic restoration of spermine synthase in the Gy-CAG/ SpmS synthase mice effectively reversed all of the physical deficiencies and circling behavior in the Gy mice in parallel to the restoration of tissue spermine and a reduction in the excess spermidine present in these mice. It is interesting that, despite the presence of spermine in the diet and the known existence of uptake mechanisms for polyamines in mice, tissues from Gy mice contain no spermine. This may be related to repression of the transport mechanism by the high level of spermidine in the Gy mice or to a lack of development of the transport system in these mice. Our studies indicate that spermine synthase (and presumably the spermine product) is essential for normal development and fertility. The fact that spermine was not required for growth of mammalian cells in culture may be related to the environment and the limited cells tested (23)(24)(25)(26)(27). The striking effect of the availability of spermine on the testis, which is likely to account for the infertility of the Gy mice, provides evidence that this polyamine may be needed in some specialized cell types. In this context it is noteworthy that: (a) transgenic mice overexpressing ornithine decarboxylase show an age-related decrease in testicular function (39) and that (b) testicular germ cells express antizyme-3, a novel form of antizyme that regulates polyamine synthesis in a stage-related manner (40).
The transgenic provision of spermine synthase and thus of spermine very effectively reversed the tendency to sudden death in the Gy mice, which is particularly marked in the first 8 weeks of life (Fig. 1). Although we do not know the exact cause of this death, a strong possibility is cardiac arrhythmias caused by dysfunction of potassium channels (41). Rectification of the strong inward rectifier Kir channels is known to be brought about by polyamines (42). Although both spermidine and spermine are able to produce such rectification, spermine is more potent, and reduced inward rectification was demonstrated in myocytes from Gy mice (43). A neurological role for spermine is indicated by the ability of the spermine synthase transgene to correct the characteristic circling behavior of Gy mice. Interactions of spermine with ion channels such as the ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and Nmethyl-D-aspartate receptors (15) are a plausible explanation for these effects. It has also been reported that Gy mice have inner ear deficiencies leading to deafness (31,32). We did not investigate these properties, but the Gy-CAG/SpmS mice would be useful for this purpose.
Our results show clearly that the elimination of the Phex gene in Gy mice is not involved in either the infertility or the poor survival of these mice. These alterations can be reversed by the provision of spermine synthase activity, but this had no effect on the loss of phosphate, which is regulated by Phex. This finding is consistent with studies in two other mouse strains where inactivations of Phex are known, Ska1 (a point mutation) (44) and Hyp, which has a deletion that removes part of the 3Ј end of the Phex and does not extend into SpmS (45). Although both male and female mice from all three strains have symptoms of hypophosphatemia and osteomalacia due to the lack of Phex, they do not show the other alterations seen in Gy mice including severe growth reduction, infertility, and poor survival. The reduction in body weight seen in the Gy-CAG/ SpmS compared with control or CAG/SpmS mice (Fig. 2) is similar to that seen in Hyp mice (32). The much greater differ-   ence in body weight between the Gy and the Gy-CAG/SpmS mice provides direct evidence for a requirement for spermine for normal mammalian growth. Although this effect on cell growth could be related to protection from oxidative stress (46 -48), another possibility is that spermine is needed for the optimal expression of genes involved in growth. Polyamines are known to regulate both transcription and translation of many genes (14,49,50), and the comparison of Gy and Gy-CAG/SpmS mice by gene array analysis and proteomic techniques may be useful to identify these genes. The use of other transgenic mice expressing lower levels of spermine synthase from tissue-specific promoters may allow the key cell types and level of spermine needed for normal mammalian function to be determined.
Male Gy mice survive only on the B6C3H background, and the Gy deletion could not be transferred to the C57Bl/6J strain (32). Hyp mice lacking Phex are viable on this strain. The genetic factors allowing survival in the absence of spermine in B6C3H are unknown. The ability to maintain the Gy chromosomal deletion in the Gy-CAG/SpmS crosses would allow further study of this question. Although the total numbers studied were small, it is noteworthy that no female GyGy mice were obtained in the offspring of the cross between B6C3H XGy females and Gy-CAG/SpmS male mice. Female mice that had the GyGy-CAG/SpmS genotype and male mice that were Gy were present in these offspring. It is therefore possible that spermine synthase deficiency is even more deleterious in females than in males.
Recently, it was reported that Snyder-Robinson syndrome, an X-linked mental retardation disorder, is caused by a deficiency in spermine synthase. This deficiency is due to a mutation in a splice site sequence that causes a large reduction but does not totally abolish enzyme activity. This reduction produces a decrease in spermine and a large rise in the spermidine:spermine ratio (51). In addition to mild-to-moderate mental retardation, affected males have a variety of other symptoms including childhood hypotonia, facial asymmetry, thin habitus, osteoporosis, and kyphoscoliosis. The Gy mouse is clearly not an accurate model for this disorder, and this could be due to species differences or, more likely, to the complete absence of spermine synthase in the Gy mice. However, our observations that unregulated expression of a spermine synthase transgene can reverse the Gy phenotype in mice suggest that attempts to increase spermine by dietary manipulations, drug treatment, or gene therapy may be successful in preventing Snyder-Robinson syndrome.