The function of spermine


  • Anthony E. Pegg

    Corresponding author
    1. Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, Hershey, PA
    • Address correspondence to: Anthony E. Pegg, Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, PO Box 850, Hershey, PA 17033, USA. Tel: +1-717-531-8152. Fax +1-717-531-5157; E-mail:

    Search for more papers by this author


Polyamines play important roles in cell physiology including effects on the structure of cellular macromolecules, gene expression, protein function, nucleic acid and protein synthesis, regulation of ion channels, and providing protection from oxidative damage. Vertebrates contain two polyamines, spermidine and spermine, as well as their precursor, the diamine putrescine. Although spermidine has an essential and unique role as the precursor of hypusine a post-translational modification of the elongation factor eIF5A, which is necessary for this protein to function in protein synthesis, no unique role for spermine has been identified unequivocally. The existence of a discrete spermine synthase enzyme that converts spermidine to spermine suggest that spermine must be needed and this is confirmed by studies with Gy mice and human patients with Snyder-Robinson syndrome in which spermine synthase is absent or greatly reduced. In both cases, this leads to a severe phenotype with multiple effects among which are intellectual disability, other neurological changes, hypotonia, and reduced growth of muscle and bone. This review describes these alterations and focuses on the roles of spermine which may contribute to these phenotypes including reducing damage due to reactive oxygen species, protection from stress, permitting correct current flow through inwardly rectifying K+ channels, controlling activity of brain glutamate receptors involved in learning and memory, and affecting growth responses. Additional possibilities include acting as storage reservoir for maintaining appropriate levels of free spermidine and a possible non-catalytic role for spermine synthase protein. © 2014 IUBMB Life 66(1):8–18, 2014


The polyamine, spermine, was first reported as a component of seminal plasma by Leeuwenhoek in 1678 [1]. Polyamines are present in virtually all living cells and a wide variety of polyamines are formed in nature according to the species. Mammals produce only spermine, spermidine, and their precursor, the diamine putrescine (Fig. 1A). These polyamines play important roles in many cellular processes including the regulation of transcription and translation, control of the activity of ion channels, modulation of kinase activities, effects on the cell cycle, protection from oxidative damage, the maintenance of membrane structure/function, and contributing to nucleic acid structure and stability. These functions have been studied extensively and current reviews describe them in detail [2-5]. However, biochemical studies of polyamine function have usually shown that either spermidine or spermine, or even their precursor putrescine, can fulfill these functions.

Figure 1.

Spermine synthesis and metabolism. Panel (A) shows the biosynthetic aminopropyltransferase reactions leading to spermine which are effectively irreversible and the oxidative catabolic reactions converting spermine back to spermidine and putrescine. Panel (B) shows the reaction catalyzed by spermine synthase.

One important exception to this is that only spermidine can be used for the formation of hypusine [Nε-(4-amino-2-hydroxybutyl)lysine], a post-translational modification of the protein eIF5A [3]. This modification is required for the essential functions of eIF5A, which include a key role in translation elongation particularly at consecutive proline residues [6]. Spermidine is the substrate for the enzyme deoxyhypusine synthase and this is well established to be an essential role for this polyamine in eukaryotes. Spermidine is produced by spermidine synthase, which catalyzes the reaction of decarboxylated S-adenosylmethionine (dcAdoMet) and putrescine to generate spermidine and 5′-methylthioadenosine (MTA) (Fig. 1B). Gene deletions of either of the enzymes forming the substrates for this reaction ornithine decarboxylase (ODC) and S-adenosylmethionine decarboxylase (AdoMetDC) lead to an inability to produce spermidine and are lethal in mice at very early embryonic stages [7, 8].

In contrast, there is no known function that is exclusive to spermine. Vertebrates, arthropods, plants, yeast, and some other organisms all contain a specific spermine synthase (SpmSyn) for the formation of spermine [9]. SpmSyn catalyses the transfer of an aminopropyl group from dcAdoMet to the N1-position of spermidine forming spermine (Fig. 1B). It is highly specific for its substrate spermidine and for the end of the substrate to which the aminopropyl group is attached. The retention of the SpmSyn gene (SMS) suggests that spermine must have an important function but deletions of this gene in yeast and in plants do not affect growth under normal conditions [9]. SpmSyns are not present in nematode worms, trypanosomes, most Eubacteria, and Archea.

The structure of human SpmSyn has been determined in complexes with substrates and products and its mechanism of action is well understood [10]. hSMS encodes a 366 amino acid protein, which is active only as a homodimer (Figs. 2A and 2B). Each monomer unit contains an active site that binds dcAdoMet and spermidine in the correct position to allow transfer to the aminobutyl- end of spermidine. It is noteworthy that human SpmSyn has a low activity compared to other well-characterized aminopropyltransferases and its kcat value can be increased 10-fold by mutations altering the active site [11].

Figure 2.

Structure of SpmSyn. Panel (A) shows the enzymatically active dimer of human SpmSyn. The two chains are in yellow and magenta with bound molecules of substrates shown in sphere representation with carbon atoms in gray. Panel (B) shows a monomer of SpmSyn with the dimerization interface in magenta. Panel (C) compares the structures of the monomers of SpmSyn, spermidine synthase (labelled SpdSyn) and AdoMetDC. Panel (D) shows the topology diagrams for these proteins. In these panels, the N-terminal domain, connecting loop, central domain and C-terminal domain in SpmSyn are in blue, yellow, orange, and greencyan, respectively. Spermidine synthase domains are color-coded similarly except for the first two b strands that structurally align with long inter-domain loop of SpmSyn, which are shown are in yellow, and the three C-terminal helices that are not present in SpmSyn, which are shown in red. The bound substrates are shown in sphere representation, with carbon atoms in gray. The AdoMetDC is in purple. This figure was originally published in the Journal of Biological Chemistry. Wu, H., Min, J., Zeng, H., McCloskey, D. E., Ikeguchi, Y., Loppnau, P., Michael, A. J., Pegg, A. E., and Plotnikov, A. N. Crystal structure of human spermine synthase: implications of substrate binding and catalytic mechanism. J. Biol. Chem. 2008, 283, 16135–16146

© The American Society for Biochemistry and Molecular Biology


A number of mammalian cell lines that lack SpmSyn activity and have no detectable spermine have been grown in culture. These include embryonic stem cells from mouse strain 129/SvJ [12], primary skin fibroblasts from Gy mice [13], and immortalized embryonic fibroblasts [14]. These cells grew normally but showed differences in response to drugs, oxidative stress, and UV radiation. With a few exceptions, these results are in agreement with studies in which specific inhibitors of SpmSyn were used to reduce spermine content (see refs. [15, 16] and references therein).

However, alterations in SMS have more striking effects on whole animal physiology. Mutations in the human gene hSMS cause Snyder-Robinson syndrome (SRS) with intellectual disability and a variety of other alterations that are described below [17, 18]. The loss of the mSMS gene in mice due to a deletion of part of the X chromosome in Gy mice produces a severe phenotype and is lethal in some strains [19, 20]. These findings focus attention on the possible unique role of spermine in critical areas in brain function, cellular communications, protection from stress, and other processes specific to animals.

Gy Mice

Embryonic stem cells with targeted disruption in the mSMS gene did not lead to the production of viable mice suggesting that SpmSyn plays an essential role in mammalian development in the mouse strain 129/SvJ. However, a viable strain of mouse on the mixed B6C3H background that has a deletion of part of the X chromosome as a result of exposure to radiation was described in 1986 [21]. Male mice with this deletion were termed Gyro (usually abbreviated to Gy) because they exhibited a circling behavior along with other neurological alterations. They also have a number of other significant abnormalities including greatly reduced size, poor bone development, a short life span with a tendency to sudden death, sterility, and deafness [22]. In 1998, it was discovered that the X chromosomal deletion in Gy mice eliminated the SMS gene [19, 22]. Tissues from these mice have virtually undetectable levels of spermine but most have an increased content of spermidine [19, 22]. In fact, in many tissues the increase in spermidine was greater than the normal spermine content. Therefore, the total polyamine content was increased. As described below, the changes in polyamine content and most of the phenotypic alterations were reversed by the transgenic expression of SpmSyn [23]. These results provide strong support for the essential role of spermine (or conceivably SpmSyn protein) in mammalian physiology.

Unfortunately, the deletion in the X chromosome in Gy mice extends into another gene Phex (also referred to as PEX) [19, 22]. The Phex gene product is involved in phosphate metabolism and it is very likely to be an important factor in the poor bone development of Gy mice, which were originally proposed as a model for hypophosphatemic rickets [21, 24]. Other mouse strains with deletions of Phex that do not extend into mSMS or with targeted inactivating mutations in Phex cause hypophosphatemia and changes in bone but do not cause any of the other symptoms in Gy mice and provide much better models for bone disease [24]. The results of these comparisons focus attention on the possible role of SpmSyn in the causation of the changes that are unique to the Gy mice [20].

Direct evidence for an important developmental role of SpmSyn was provided by studies in which female Gy carriers were bred with transgenic male mice expressing a SpmSyn transgene (CAG-SpmS) from a ubiquitous strong promoter, that is integrated into an autosome [25]. Some of the resulting transgenic male offspring contained the Gy deletion but had significant SpmSyn activity [23, 26]. These mice had normal hearing, life span, testicular development, and fertility and were only slightly smaller than their littermates that lacked the Gy deletion. Their bone development was improved significantly but was still impaired due to the Phex deletion.

It is noteworthy that the absence of spermine in tissues of Gy mice occurs not only in the absence of any attempts to limit spermine in the diet but even when these mice are fed on diets low in spermidine and supplemented with high levels of spermine [14, 19, 20, 26]. This is quite surprising in view of the well-documented presence of systems for the uptake of polyamines at both the gastrointestinal and cellular membranes [27, 28]. It is possible that the high levels of spermidine suppress the activity of the transport system or that the lack of spermine during early development affects the expression of the GI uptake system. Even direct injection of spermine into the blood stream resulted in only a very modest rise in tissue spermine levels although the doses that could be administered were small owing to the acute toxicity resulting from such spermine treatments [26].

Another change in metabolites that occurs in Gy mice is a large increase in the content of dcAdoMet although levels of S-adenosylmethionine (AdoMet) itself were not altered. The dcAdoMet nucleoside is usually present in very low amounts (<5% of that of AdoMet) but its content was strikingly increased (by 10- to 150-fold depending on the tissue) in Gy mice. This increase is due to an elevation in the activity of AdoMetDC, which is powerfully repressed by spermine [16], with an additional effect due to the loss of spermine synthesis since spermine formation uses two molecules of dcAdoMet rather than one used for spermidine [29].

Importantly, attempts to transfer the Gy phenotype to other gene backgrounds such as C57Bl/6J by back-crossing have been unsuccessful with no viable offspring [21, 22, 24]. Presumably, some genes present in the B6C3H strain are expressed at sufficient levels to allow the survival of Gy mice. These are currently unknown and may throw light on the functions of spermine. It should be possible to identify such genes since crosses of Gy females with male CAG-SpmS mice on the inbred B6 background lead to fertile Gy/SpmS males.

Snyder-Robinson Syndrome

Schwartz and coworkers have provided a series of important papers indicating a critical role for the SMS gene in human development. In 2003 they discovered that SRS (OMIM: 309583) was linked to a mutation that altered a splice site in the hSMS sequence, which led to a truncation of the encoded SpmSyn at position 111 and an inactive protein [17]. This mutation did not completely eliminate SpmSyn activity since some correct splicing still occurred. SpmSyn activity in cultured fibroblasts or lymphoblasts for SRS patents had about 12% of the activity in cells from unaffected family members. Spermine content in these cells was reduced only moderately but there was a substantial increase in spermidine and decrease in the spermine:spermidine ratio (Table 1). Determinations of this ratio can therefore be used to detect SRS and a convenient highly sensitive way to measure this using liquid chromatography tandem mass spectrometry has been reported [30]. As in Gy mice, there was also an increase in dcAdoMet in cells derived from patients with SRS (1–3 pmol/mg protein in patients-derived samples compared to 0.29–0.55 in controls). AdoMet content was not altered [29].

Table 1. Mutations in SpmSyn leading to SRS
MutationActivityaProteina, bSpermine:spermidine ratioaReference
  1. a

    Measured in cultured cells (fibroblasts or lymphoblasts).

  2. b

    Measured by immunoblotting using a specific antiserum.

  3. c

    The G56S, V132G, I150T mutants have been expressed as recombinant proteins and were found not to dimerize and to be unstable and to have activity less than 4% of control recombinant human SpmSyn (Pegg, unpublished). In silico modeling also indicates that these mutations are likely to disrupt the structure [34].

  4. At least two other mutations including N148R cause SRS but details of the resulting SpmSyn have not yet been published (18).

Splice variant12%10%0.74 [17]
G56S<1%<5%c0.37 [31]
V132G<2%<5%c0.33 [32]
I150T<1%<5%c0.39 [33, 34]
Y328Cc. 1%20%0.53 [35]
G67X<1%Not measured0.39 [36]

Although SRS remains a very rare disease, a number of additional mutations in SMS leading to it have now been identified [18, 30-36] (Table 1). These mutations can occur in any of the domains of SpmSyn shown in Fig. 2D. All lead to a drastic reduction in SpmSyn activity and protein (Table 1). Studies of these patients have refined the phenotype imparted by SRS [18]. It is an X-linked intellectual disability syndrome combined with a number of other clinical features including osteoporosis, kyphoscoliosis, speech abnormalities, asthenic habitus with diminished muscle mass, hypotonia, facial dysmorphism, a mild short stature, and a high incidence of seizures. Ambulatory difficulties, which vary in severity, are seen in most cases. Genital abnormalities and renal changes have been reported in some individuals.

In all cases, cells from affected individuals show a large reduction of SpmSyn activity and a decrease in the spermine:spermidine ratio (0.3–0.7 compared to normal values of 1.6–1.9) (Table 1). Most of the mutations lead to a large reduction in the amount of SpmSyn protein. This instability may be related to a decrease in the ability to form dimers (see Figs. 2A and 2B), which also contributes to a loss of activity since dimerization is essential for activity [10].

An interesting feature of these studies is that the severity of the phenotype varies significantly according to the family studied. The most severe cases are those imparted by the G56S and N148R changes to the SpmSyn sequence [18]. This severity is not clearly related to the magnitude of the reduction in SpmSyn activity or the spermine:spermidine ratio since other mutations also cause similar changes (Table 1). Also, although the Y328C mutation is associated with a very mild phenotype [35] and this alteration does leave some residual SpmSyn activity and protein, there is an equal or greater amount in cells from the individuals in the original family with the splice site variation [17], who are more severely affected. One plausible explanation for the differences in severity in various SRS families is the activity of other genes that alter the requirement for spermine as noted with the Gy mice.

Possible Functions of Spermine

To Act as a Storage Mechanism to Ensure Adequate Access to Spermidine and Putrescine

Spermine is the most strongly basic polyamine and the great majority of the cellular spermine is bound to acidic sites on macromolecules. Release of spermine from these sites is likely to induce the polyamine back-conversion/catabolic pathways that generate spermidine (Fig. 1A). Two such pathways are known: spermine oxidase (SMO), which converts spermine into spermidine and 3-aminopropanal; and acetylpolyamine oxidase (APAO), which converts N1-acetylspermine into spermidine and N-acetyl-3-aminopropanal [37-39]. Both of the pathways are inducible. The APAO pathway is regulated by changes in the amount of spermidine/spermine-N1-acetyltransferase (SSAT), which forms the substrate for APAO. SSAT is rapidly and profoundly increased in response to elevations in the free polyamine concentration [37]. SMO also increases significantly in response to various stimuli [38]. Mice with targeted disruption of either SSAT or SMO genes are viable with only minor phenotypes but a double knock out has not yet been described and it is also possible that these mice may develop more severe phenotypes when subjected to stress. Many stress-related situations are likely to lead to the degradation of RNA and the subsequent release of bound polyamines such as spermine [39].

Resistance to Reactive Oxygen Species and to Other Stress

Polyamines play an important role in the protection from reactive oxygen species in bacteria, yeast, and mammalian cells [[38-41] and references therein]. Polyamines have been shown to act as free radical scavengers, to quench singlet molecular oxygen and shield phage and microbial DNA from oxidative damage [42-44]. They also mediate defense from oxidative damage by stimulating the synthesis of protective gene products such as superoxide dismutase, heat shock proteins, and cell cycle regulators [44, 45]. Although spermidine has some activity in this respect, spermine is considerably more effective. The increased spermidine in skin fibroblasts from Gy mice does impart resistance to H2O2 [13] but detailed comparison of the ability to resist such damage in immortalized Gy cell cultures supplemented with polyamines demonstrated clearly that spermine was more potent [41].

Other stress situations may also be moderated by spermine. The SMS gene has been shown to play a role in protecting plants from heat-stress by increasing the expression of heart shock genes [46]. The SMS gene also imparted a modest increase in resistance to salt stress in Arabidopsis thaliana [47] and spermine accumulation is associated with resistance to drought [48].

Regulation of Ion Channels

Several types of ion channels that play critical roles in mammalian physiology are profoundly influenced by polyamines. These include: the inwardly-rectifying potassium (Kir) channels, which control membrane potential and potassium homeostasis in many cell types; glutamate receptors including N-methyl-D-aspartate (NMDA), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), and kainate receptors that mediate excitatory synaptic transmission in the mammalian brain; and other related channels affecting intracellular calcium signaling, Na+ transport, and some connexin-linked gap junctions (refs. [2, 4, 49, 50].

Intracellular polyamines modulate the intrinsic gating and rectification of Kir channels by directly occupying the ion channel pore. Extracellular polyamines stimulate NMDA receptors increasing the size of the NMDA receptor currents. Several effects including “glycine-independent” stimulation (at saturating concentrations of glycine) and a “glycine-dependent” stimulation at low concentrations of glycine, which the alters the affinity of the NMDA receptor for glycine have been reported. At least two discrete binding sites in these tetrameric receptors have been identified.

The polyamine-medicated effects on Kir and NMDA channels are fairly well understood at the molecular level. Detailed structural and biophysical studies of the binding of polyamines to these proteins have been published. These include work on the specificity for individual polyamines of both binding and the strength of the resulting effect. Although both spermidine and spermine can be active in these respects they differ in potency and affinity with spermine being considerably more effective. Thus, the correct spermine:spermidine ratio may be necessary for appropriate physiological activity and increased levels of spermidine unable to compensate for the loss of spermine. Alterations of activity of ion channels could therefore explain a number of the features of both human SRS and Gy mice.

Gy mice have greatly impaired hearing. There was a total loss of the endocochlear potential and distortion product otoacoustic emission testing showed no difference between the response of Gy mice and the noise floor [26]. One possible explanation for this loss is an alteration in the activity of cochlear lateral wall-specific Kir4.1 channel that is needed to develop this potential. This is one of the channels in which strong inward rectification by spermine is known to occur.

Similarly, Kir channels in myocytes, whose inward rectification is firmly established to be polyamine dependent, are needed to maintain cardiac electrical activity. Malfunction of these channels can lead to cardiac arrhythmias and this is a plausible explanation for the propensity of Gy mice to sudden death. Electrocardiograms in Gy mice indicate abnormal cardiac electrical activity [20]. The steepness of rectification of Kir channels was reduced in cardiac myocytes from Gy mice confirming a role for spermine in controlling rectification [51].

Altered activity of brain glutamate receptors in response to the reduction or loss of spermine may be the source of the intellectual impairment in SRS patients and the behavioral changes in Gy mice. The NMDA receptors, a subtype of glutamate receptors, are critical for induction of synaptic plasticity, needed for memory and learning. Their activity is profoundly influenced by extracellular spermine, which is more effective than other polyamines [4]. The brain has the highest level of SpmSyn in all the tissues examined [25]. Although the extent to which this contributes to extracellular levels of spermine is not yet well understood, it has been shown recently that spermine increases glial intercellular communication by interacting with glial gap junctions. The glial syncitium propagates spermine through these gap junctions and increase astrocytic coupling [50].

Other classes of glutamate receptors such as AMPA and kainate receptors are also regulated by spermine. Kainate receptors are involved in synaptic plasticity and have been linked to epilepsy-like seizure activity and alterations in their activity could contribute to the tendency to seizures reported in some SRS patients [36].

Bone Development

The occurrence of changes in bone development including osteoporosis and kyphoscoliosis are a feature seen in almost all SRS patients. Gy mice have profound defects in bone that have been ascribed to hypophosphatemia associated with the loss of the Phex gene. As expected, these defects are not fully reversed by transgenic expression of SpmSyn, but there was a statistically significant increase in femur length when the Gy/CAG-SpmS mice were compared to their Gy littermates [23]. This suggests that spermine may be needed for bone development. Such a role for polyamines in osteogenesis is supported by laboratory studies. The expression of runx-2 (a transcription factor known to be involved in early stages of osteogenic differentiation) and osteopontin was stimulated by addition of spermine to adipose tissue-derived mesenchymal stem cells [52, 53]. At later times, alkaline phosphatase activity was increased. Osteogenic differentiation in response to 1,25-dihydroxyvitamin-D3 increased SSAT in these cells. Spermine addition also increased beta-catenin expression and activation leading to induction of the mature osteoblast commitment factor Osterix [53] and that polyamines promoted osteogenic differentiation of human bone marrow-derived mesenchymal stem cells up-regulating osteogenic genes and SSAT, whilst down-regulating adipogenic genes [54].

Inhibition of Lipid Formation

Adipogenesis is accompanied by an increase in spermidine and is blocked by inhibitors of ODC unless exogenous spermidine is provided [55]. Comparison of the effects of selective inhibitors of spermidine synthase and SpmSyn on 3T3-L1 cells showed that the spermine:spermidine ratio was inversely correlated with lipid accumulation [56]. SSAT activity was induced by stimuli for adipogenesis, which was impaired by inhibitors of APAO. The ability to overcome the effects of ODC inhibition was also seen when non-metabolizable analogs of spermidine were used [57]. These results suggest that spermidine is the key polyamine for fat formation and the interconversion of spermine and spermine may regulate lipid levels. Recent mechanistic studies reveal that spermidine prevents translocation of the nuclear RNA-binding protein HuR to the cytoplasm, increases phosphatase PP2A activity via alteration in the binding of ANP32 (acidic nuclear phosphoprotein 32) and increases levels of the transcription factor CCAAT/enhancer-binding protein beta promoting adipogenesis [57]. Detailed information on the fat content of SRS patients and Gy mice have not been reported but these results are consistent with the asthenic body mass phenotype seen in SRS and the greatly reduced mass of Gy mice.

Anabolic Effects

Gy mice are very small and males with SRS have a reduced growth rate, a moderate reduction in stature, diminished body bulk, and hypotonia due to impaired muscle development. These effects are consistent with a large body of literature showing that polyamines are needed for gene expression and protein synthesis [2-5]. Polyamines play multiple roles in protein synthesis in addition to the activation of eIF5A by the hypusine modification described above. There are no clear laboratory studies indicating the specific role for spermine in gene expression and protein synthesis but spermine binds more tightly to nucleic acids than spermidine and it is easy to see how a change in the spermine:spermidine ratio could influence these processes and thus growth.

Reproductive Functions

Genital abnormalities including low testicular volume and undescended testes have been reported in some patients with SRS although their potential fertility is unknown. Gy mice have a profound defect in the testicular morphology. There is an almost complete absence of mature spermatozoa and most of the germ cells located in the seminiferous tubules remain at the spermatogonia or early stage primary spermatocyte stages [23]. These changes were totally corrected by normalization of the polyamine profile via transgenic expression of SpmSyn. It is possible that neuro-derived effects on the endocrine system are responsible for the abnormal testicular development since there was also a reduction in the number of Leydig cells in the testes but there is experimental evidence showing that a normal testicular polyamine content is essential for reproductive function. Testicular polyamine levels are tightly controlled and transgenic overexpression of ODC led to infertility in mice [58]. At present the specific reproductive role of polyamines is not understood. Although there is a testis-specific form of antizyme (antizyme3) which is needed for sperm development, it has now been shown that this gene may also encode an activity regulating phosphatase activity and affect fertility in this way rather than via an effect on polyamines [59].

Effects of Spermine on Maturation of the Gut and Immune System

Spermine has been implicated in the maturation of the gut and the immune system [60, 61]. Although no mechanistic studies explaining this effect have been reported, it is possible that delayed or incomplete maturation may contribute to the defects is seen in male Gy mice and SRS males. However, spermine in maternal milk is an important factor in promoting this development [61, 62] and female carriers of SMS mutations do not have significant spermine deficiency due to the active allele.

Possible Other Functions for SpmSyn

The N-terminal domain of human SpmSyn that forms the bulk of the dimerization interface has strong structural homology to the proenzyme form of AdoMetDC [9, 10] (Figs. 2C and 2D). This domain is present in all vertebrates, arthropods, sea urchins, sea anemones, and some other species. No other aminopropyltransferase has such a domain and it is not present in the SpmSyns of yeast or plants in which its gene is not essential for viability. It is possible that it has a role in facilitating interactions with other proteins and that such interactions are another essential function of SpmSyn.

Comparison of Alterations in Polyamine Content and in Phenotypes between Gy Mice and SRS Patients

It is not unexpected that there are would be some substantial differences in the changes seen due to spermine synthase deficiency in SRS human and Gy mice but there are many similarities outlined in Table 2.

Table 2. Comparison of SRS and Gy mice
PhenotypeGy male miceaSRS malesb
  1. a

    From refs. 20, 23, 26.

  2. b

    From refs. 18, 35, 36.

Intellectual disabilityClear behavioral abnormalitiesUniversal
Poor muscular developmentLow muscle massUniversal
HypotoniaLikely but not measuredUniversal
Asthenic body bulkVery low body weightUniversal
Short statureVery smallCommon but moderate
Bone abnormalitiesYes but complicated by Phex deletionUniversal
Facial asymmetryNot seen in miceCommon
Ambulatory difficultiesCircling behavior/Poor balanceCommon
SeizuresNot reported but could contribute to short life spanCommon
Propensity to sudden death/ cardiac arrhythmiaUniversalNot reported
Hearing difficultiesTotally deafNot reported
Genital abnormalitiesSterile. Abnormal Leydig cells and testicular/sperm developmentSome have testicular abnormalities
Speech abnormalitiesDifficult to assess in miceUniversal

Brain related alterations that may be attributable to the altered response at glutamate receptors are seen in both cases. Intellectual disability is a major feature of SRS and the Gy mice clearly have neurological abnormalities although it is difficult to assess their mental impairment since the circling behavior and poor balance seen in these mice makes physiological experimentation on learning very hard to carry out.

A major difference is seen in effects that may be due to changes in Kir channel activity. Both the deafness and the abnormal cardiac electrical activity that can lead to sudden death are seen in all Gy mice but are not reported in SRS patients. It is possible that the relevant Kir channels in humans are less sensitive to alterations in the spermine:spermidine ratio. There is however good evidence that polyamines play a role in human hearing. A reversible high-frequency hearing loss is a known side effect of exposure to 2-difluoromethylornithine, a drug that inhibits ODC reducing spermidine content with little effect on spermine. It is used for treatment of African sleeping sickness, unwanted facial hair growth, and shows promise for cancer chemoprevention (see ref. 26).

An important factor, which may contribute to the differences in the observed phenotypes, is that spermine is virtually absent from all the tissues of Gy mice. It has not been possible to obtain detailed tissue information on SRS patients but all the studies on cells derived from these patients (fibroblasts and lymphoblasts) show a significant spermine content. This occurs even in cells from patients having mutations that either completely inactivate SpmSyn and/or cause the protein to be extremely unstable resulting in a total loss of immunoreactive protein on Western blots of cell extracts. The uptake of spermine from the culture medium by these cells is ruled out since fibroblasts from Gy cells grown in the same medium do not contain any spermine. It therefore appears that SRS cells have some other means to produce spermine.

The most likely source of this spermine is from a very weak activity of human spermidine synthase to use spermidine rather than putrescine as a substrate. Some aminopropyltransferases such as that from Thermotoga maritima are relatively non-specific and can use a number of amine acceptors of varying length. Spermidine synthases from other sources including E. coli, yeast, and various mammals are much more specific. A structural basis for this specificity has been determined based on crystallographic studies of the human spermidine synthase and human SpmSyn complexed with substrates [10, 63]. These enzymes have significant structural similarity (Figs. 2C and 2D) but the active site pocket in human spermidine synthase is sterically restricted to prevent spermidine binding with a bulky tryptophan residue (Trp28) being particularly important in this respect. This provides a great deal of discrimination in favor of the smaller substrate (more than 100-fold). However spermidine synthase is present in much larger amounts than SpmSyn [in mice approximately 25 times in heart; 400 times in liver, and 40 times in brain [25, 64]] Thus, even 0.5% activity would provide a considerable capacity to produce spermine, particularly, since levels of dcAdoMet, the other substrate for the reaction, are increased.

This argument would explain the presence of spermine in the absence of the specific SpmSyn in human cells but the absence of spermine in mouse cells lacking SpmSyn remains unclear. It is certainly possible that mouse spermidine synthase is more specific than the human equivalent but such a comparison has not been made and it is not easy to measure such very low levels of activities with non-preferred substrates. Unfortunately, a crystal structure for the mouse spermidine synthase is not yet reported. There are only 16 amino acid differences out of 302 total between the human and mouse enzymes and the Trp28 residue is conserved. However, several altered residues could limit this space even more either directly or by slightly changing the position of Trp28. Transgenic mice in which human spermidine synthase is expressed from a ubiquitous promoter are available [64] and cells derived from crosses of these mice with the Gy mice could be used to test this hypothesis.

Conclusions and Possible Treatments for SRS

It seems essential to use laboratory animal studies to determine which of the altered parameters (spermine deficiency; elevated spermidine; decreased spermine:spermidine ratio; increased dcAdoMet; and loss of SpmSyn protein) is responsible for the phenotypic changes seen in conditions where SpmSyn is reduced or absent. Another key area for further research is the possible presence of genes that limit the effect of the loss of SpmSyn. As described above, the existence of such genes is indicated by the effect of genetic background on the viability of Gy mice and is consistent with the differences in severity seen in SRS patients despite similar levels of loss of SpmSyn. High throughput DNA sequencing of human patients and carriers as well as selective breeding of the Gy/SpmS mice should allow identification of such genes.

Early intervention to normalize the spermine:spermidine ratio may be necessary to treat SRS. The increase in understanding of the phenotype of SRS and improved assays for SRS may allow for timely interventions if adequate treatments could be devised. Use of diets rich in spermine is the obvious and simplest potential remedy. However, based on the results with Gy mice where spermine uptake from the diet was minimal even with diets highly enriched for spermine, this approach may not be useful unless spermine is formulated in a way to facilitate transport across cell membranes.

Other approaches may be possible to restore SpmSyn activity using modern molecular biological and pharmacological techniques Modeling techniques may be used to design compounds that may restore activity/stability to the SRS protein containing point mutations [65]; the splice site variant may be approachable by use of compounds enhancing the correct splicing [66] generating the wild type form of the protein; and there has been some interest in compounds that allow the insertion of an amino acid at the stop codon [67] such as in the G67X mutant. These approaches have the disadvantage that they may each be limited to treatment of a particular mutation causing loss of SpmSyn activity and thus impractical for a very rare condition. A more global approach using gene therapy to restore a normal SpmSyn using the mutants with increased activity [11] to achieve sufficient activity even with low expression may have more potential if suitable vectors to provide expression in a wide range of tissues is needed. Some support for this approach is provided by the studies with the CAG-SpmS vector where transgenic expression restored the normal phenotype in Gy mice even though SpmSyn production was unregulated using a generic promoter [23].


The author is most grateful to Dr. Charles Schwartz for reading this manuscript and for sharing his discovery of the genetic basis of Snyder-Robinson syndrome. Author also thanks many other investigators who worked in the laboratory or collaborated with studies of spermine synthesis and function. There is a very large literature on polyamine research and the author apologizes to many individuals whose important contributions are not cited with specific references. More detailed citations can be found in the review articles listed in the bibliography.