Compartmentation of Spermidine in Neurospora crassa*

The polyamines putrescine, spermidine, and sperm- ine are multivalent cations that bind to anionic cell constituents such as nucleic acids. Their distribution between free and bound states within the cell is not known. Such knowledge would be important in rela-tion to the negative control of polyamine synthesis. We report a tracer experiment in which [ “C ]ornithine was added to logarithmically growing Neurospora crassa mycelia. The amount and the specific radioactivity of the three polyamines thereafter suggested that new molecules of spermidine were made preferentially from new molecules of putrescine, and that new mole- cules of spermine were made from new molecules of spermidine. The extent of mixing of new [‘“CI- and resident [laC]spermidine indicated that 70% or more of the resident spermidine was sequestered, and not immediately accessible to spermine synthase. Cell fractionation revealed that about 28% of the cellular sper- midine was vacuolar, and nonexchangeable with [‘“C] spermidine added at the time of cell breakage. We suggest that the remainder of sequestered spermidine is bound strongly to anionic sites in the cell, and is relatively inactive in the control and synthesis of polyamines. The polyamines (putrescine, spermidine, and and initial enzyme the synthetic pathway, ornithine decarboxylase (EC 4.1.1.17), are widely distributed in nature (1). as

The polyamines (putrescine, spermidine, and spermine) and in initial enzyme of the synthetic pathway, ornithine decarboxylase (EC 4.1. 1.17), are widely distributed in nature (1). Elevated activities of ornithine decarboxylase are almost universally associated with rapid cell growth in organisms as diverse as Escherichia coli, fungi, and mammals (2)(3)(4)(5). The control of ornithine decarboxylase by polyamines is not direct, as a rule, and various indirect mechanisms have been proposed (5)(6)(7)(8)(9). Much uncertainty surrounds these mechanisms, not only because the effect of polyamines is indirect, but also because the pool sizes of polyamines are poorly correlated with the manifestation of control. For instance, ornithine decarboxylase activity of Neurospora cells with substantial polyamine pools becomes elevated after a brief interruption of polyamine synthesis brought about by ornithine deprivation.' Similarly, in other systems, very small amounts of added, exogenous polyamines have major negative effects upon ornithine decarboxylase activity, despite the existence of large endogenous pools in the same cells (10,ll).
* This work was supported by American Cancer Society Grant BC-366. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  IT To whom reprint requests should be addressed. R. H. Davis and G. N. Krasner, unpublished observations.
Most of the polyamines, which are highly charged cations at cellular pH values, are probably bound to anionic cell constituents such as ribosomes, DNA, and membranes (3,12,131, and these sequestered pools may be inactive in control (14). While this possibility is supported by many polyaminebinding studies in vitro (e.g. Ref. 15), direct evidence that polyamines are sequestered in vivo is scarce (16,17). Because cell disruption alters the ionic environment and leads to rapid equilibration of labeled (exogenous) and unlabeled (endogenous) polyamines among cell fractions (e.g. Ref. 18), no definite conclusions can be drawn about the diffusible state or the location of polyamines in vivo.
The present work demonstrates through tracer metabolism in living cells that most of the endogenous spermidine, the predominant polyamine in Neurospora, is sequestered from the metabolic reactions which produce and use it. Moreover, the vacuole of Neurospora is shown to sequester some of the polyamines of Neurospora in a nonexchangeable form. This work extends previous work done on putrescine sequestration (19) and vacuolar polyamines (20) of Neurospora.

RESULTS
Compartmentation of Spermidine in Vivo-Previous studies have shown that exponentially growing mycelia of Neurospora crmsa rapidly take up traces of [14C]ornithine from the medium and incorporate it into polyamines (21)(22)(23). We wished to use [14C]ornithine to label the spermidine pool and to observe the rate and pattern by which the spermine acquired label. Compartmentation of spermidine woul goo1 be indicated (i) if the specific radioactivity of new spermine molecules synthesized during the labeling period was significantly greater than the specific radioactivity of the total, acidextractable spermidine pool; and (ii) if spermine became labeled more quickly than predicted according to isotope dilution by the resident spermidine pool. Such observations would indicate that label flowing from ornithine into spermine was bypassing the resident spermidine pool. Similar experiments have revealed the vacuolar compartmentation of ornithine in Neurospora (21) and the sequestration of putrescine (19).
In the experiment, 10 p~ [14C]ornithine (4200 cpm/nmol) was added to exponentially growing cells. The radioactivity was wholly taken up by 20 min. The amounts and specific radioactivities of the polyamines were determined at intervals ~ * Portions of this paper (including "Experimental Procedures," part of "Results," Figs. 1 and 2, and Tables 1-111) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 82M-3265, cite authors, and include a check or money order for $8.40 per set of photocopies.Ful1 size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.  Table I) are discussed in the Miniprint.
One objective method of estimating spermidine sequestration is to compare the specific radioactivity of new molecules of spermine (and, thus, of their actual precursor pool) with the specific radioactivity of total, extractable spermidine for a short interval. The results of such a comparison are shown in Fig. 3 for all three polyamines and their extractable precursors. On the right, total spermidine is compared to new spermine. Over the interval of 0-10 min, the average specific radioactivity of spermidine was 62.5 cpm/nmol, while that of new spermine was 210. Thus, in this period, only 62.5/210, or 30% of the spermidine, was available as a spermine precursor. This is a maximal estimate, because during 10 min, there is an opportunity for equilibration of the sequestered and the spermine-precursor pools, as we have noted in previous applications of this method (21). Nevertheless, it demonstates that at least 70% of the spermidine pool is sequestered from use as a spermine precursor. The pattern of the new spermine and total spermidine curves in later times (Fig. 3, right) is consistent with this interpretation. Label increases in total spermidine, while the new spermine ultimately comes to be made from less radioactive molecules. The data suggest that increasing numbers of labeled spermidine molecules are becoming sequestered, and spermine is made, in the 80-120-min interval, from newer, less radioactive molecules of spermidine.
A second method of estimating spermidine sequestration is to compare the specific radioactivities of new spermidine (Fig.  3, middle) and new spermine (Fig. 3, right). The first point of the curves (1750 and 210 cpm/nmol for spermidine and spermine, respectively) show that, as radioactivity moves from new spermidine to new spermine, it is diluted 8.3-fold by unlabeled, endogenous spermidine. The amount of spermine made each generation is very small: 0.31 nmol/mg dry weight. To dilute this 8.3-fold requires only 7.3 X 0.31 = 2.3 nmol of unlabeled spermidine. This is 2.3/18.2, or 12.6% of the total spermidine pool. Thus, on this basis, about 87% of the spermidine pool fails to participate in spermine synthesis.
Thus, according to these calculations, over 70% of the cellular spermidine pool is sequestered from the enzyme which uses it in spermine synthesis. Both methods use single initial values of specific radioactivity, one of which (spermine) is Spermidine Compartmentation in Cell Fractions-We wished to associate a t least some of the sequestered spermidine with one or more cell fractions. To do so, we disrupted cells in conditions which preserve many organelles, and isolated organelles by differential and gradient centrifugation. In order to control for redistribution of spermidine during cell breakage and fractionation, [14C]spermidine was added to cells before cell breakage. In these experiments, only 8% of the total spermidine was sedimentable at 20,000 x g. Sorbitolsucrose density gradient centrifugation of this organellar pellet revealed that, while spermidine was associated with both mitochondria and the vacuoles, only the vacuoles were relatively free of the added [14C]spermidine tracer (Fig. 4). About 2% of cellular spermidine was found in vacuoles; the recovery of vacuoles was estimated to be 7%, judging from the arginine content of this fraction. Assuming proportionate loss of vacuolar arginine and vacuolar spermidine, and knowing that 98% of cellular arginine is in vacuoles (24), we consider 28% of cellular spermidine to reside in (or bound in nonexchangeable form to) vacuoles. This is only a minor portion of the 70-87% of spermidine judged to be sequestered by the metabolic experiments.
The analysis of other cell fractions (see Miniprint) reveals that spermidine associated with mitochondria and ribosomes has the same specific radioactivity as the spermidine in supernatants from which it came. Therefore, no conclusion can be drawn regarding specific associations of spermidine with these organelles in uiuo.

DISCUSSION
According to our tracer experiments, both spermidine and its precursor putrescine (19) are sequestered in uiuo in some fashion. The degree of sequestration is on the order of 70-87%. This figure is based on the extent to which resident polyamines fail to mix with newly labeled molecules of polyamine before the latter are used in a subsequent reaction. To our knowledge, these are the first demonstrations of polyamine sequestration with tracers in living cells. We can account for about one-third of the sequestered spermidine as the vacuolar pool, which is nonexchangeable even in vitro Compartmentation i n Neurospora with added [14C]spermidine. This is a finding directly comparable with that of Seiler and Deckhardt (25) who found nonexchangable polyamines in synaptosomal membranes of the rat brain. The remainder of the polyamine sequestration cannot be accounted for by our cell fractionation experiments. Two mechanisms besides organellar compartmentation could account for the in vivo tracer data. The first is that enzymes of polyamine synthesis are aggregated such that products of one reaction tend not to diffuse before becoming substrates of the subsequent reaction (26-28). This would prevent new molecules from mixing fully with resident polyamines. However, in all systems investigated to date, spermidine and spermine synthetases are distinct enzymes (29, 30). Nevertheless, aggregation of polyamine enzymes may prevail, and must be retained as a hypothesis to explain our data. The second possible mechanism of sequestration is that binding of resident spermidine (and putrescine) to cell constituents is sufficiently strong to prevent extensive mixing of new, freely diffusing molecules with the bound fraction. This phenomenon may be magnified by a distribution of spermidine to the nucleus and to ribosomes, if they offer sites which are less accessible than others to free diffusion. (Even less basic molecules have been found to be disposed in gradients in bacteria (34,35).) Ribosomes bind spermidine and spermine well, particularly at low ionic strength (15, 31-33). Spermine appears to bind to ribosomes at sites that cannot be displaced by acridine orange; rRNA does not have such sites (15). The nucleus also may, owing to its membrane, offer barriers to exchange between DNA-bound and cytosolic polyamines (16, 17). Thus "sequestered," nonvacuolar polyamines may be those bound to nucleic acids.
The detection of "bound" and "free" pools of spermidine and putrescine may explain several unusual features of polyamine metabolism in various organisms. Cells treated with the ornithine decarboxylase inhibitor, a-difluoromethylornithine, become depleted of polyamines and develop a greatly increased rate of polyamine transport into the cell (36, 37). It may well be, as Seppanen et al. (37) recognized, that the increased rate reflects the emergence of unsaturated polyamine-binding sites during polyamine depletion. A second phenomenon which is widely observed (5, 38-41) is that treatment of cells of various species with hypotonic medium is sufficient to induce elevation of ornithine decarboxylase activity. This may reflect a normal response of the enzyme's negative control system to the greater binding of polyamines (and, thus, their withdrawal) brought about by lowering the ionic strength. Finally, in diverse cell types, treatment with low levels of putrescine or spermidine have large regulatory effects despite far larger internal pools of these molecules (10, 11). In this case, the added polyamines may be more freely diffusible (and, thus, more effective in control) as they enter the cell than the resident, bound polyamines. The total radioactlvity ~n putresclne. spermldlne and aperrnlne (Fly.

Spermidine Compartmentation in Neurospora
appearance Of label in spermlne (Fig. 21 beglns before 10 minutes, and -The quantitatively lnterpretable becaue the rate of entry of label Into reaches a Constant rate by 20 minutes. Unfortunately, the curve 18 not spermidine is not constant. Thus, while new spermine molecules achieve constant speclfie radioactivity quickly, the time it take6 to do so cannot be used 1. B It has been previously 12111 to calculate the Size of the dynamic subpool of spermldme which gives r l~e to it. The constant rate of appearance Of radioactive spermine, in fact, obscures the compensatory changes during the first 40 minutes in the entry Of label into the spermidine p o d (diminishing with tme: eee below) and the number Of turnovers of this pool lincreasing wlth time). Nevertheless, the tlme required for l a b e l to appear in spermine 1s very short, g l v e n the large spermdine pool and the low rate of synthesis of Its product (Table 11. adequacy of the data, and to suggest ponslble modele Of polyarnlneintermediste channeling. In the left panel, new putrescine is more radloactlve than ornithine iqttially. because the cytosol IS belng preferentially labelled by [ Clornithine uptake: most Of the cellular acnlthme is unlabelled and sequestered in the vacuole. After 10 mln, however, new putrescine becomes lese radioactive than ornlthlne because the Cytosol has been flushed with endogenously eynthesired orwthine. and the remaining labelled ornithme is now l a r g e l y in the vacuole ( 2 5 , 3 3 , 3 1 1 . It is useful to analyze Fig. 3  remained zn the Boluble fractLons iTable 11. EXP. 1). Only 88 of the total spermidine vas found 1" the organellar pellet l 2 0 . 0 0 0 x g pellet]. However, the organellar spermidine had a speclfx radioactlvlty only 37% that Of spermldine in the 20,000 x g supernatant. Subsequent fractionation Table I1 and Fig. 4. The vacuolar pellet contahned 2 5 1 of the endogenous of the organellar pellet by denslty gradlent centrlfugatlon 1s shown in o c 9 d n e l l a r spermidine, but less than 1% Of the added I 4C1spermidine.
The speciflc radmactivity of vacuolar spermldlne was only 1.7% the Mitochondria were located in this gradient by assay for S U C C~~X C specific radioactivlty of spermidme at the top of the gradient.
dehydrogenase activity (Fig. 4 1 . The nltochondrial fractions contained a Substantla1 amount of speraldine but It had equ~llbrated to a large extent with erogenous spermidine. The peak mitochondrial fraction had spermdine of a speclf~c radioactivity comparable to the sample zone.
During fractxonatlon most Of the endogenous and exogenous spermdine spermidine as indicated by the tracer experiment. NUmOrOUS reporte vacuolar spermidine cannot account for all of the sequestered suggest thit a likely site for specmidine is the ribosome (12, 15, 181.
T P~ spermidine content of ribosomes and its equilibration with added cellular RNA vas found in the 100.000 I p pellet, whereas only 201 of the I Clspermldlne was examined (Table 1111. nore than 40% of the total spermldine was aeboclated Vith this fraction. In sharp contrast to the exogenous spermidine Occurled. The Specific radtoactivies of the results obsecved for the ~acuole, complete equilibration of endogenous and spermidine of the supernatant and pellet Of the 20,000 I p centrifugatlon step were also s m l l a r . nore spermidine was asmciated with the organellar pellet In thlS experiment than in Exp. 1 of Table 11. Thls reflects the lover OSmOtiC strength of the fraatlonation buffer In the experiment Of Table I11 which led to lysla and lose of many vacuoles (201. Anether difference between t4$ fractionation schemes of Table I1 and Ill vas the necessary for efficient ribosome iSOlat10n 1521.
In fact, less than 3% Of ~n c l u s~o n of 2 mn Hg for the experlaent ln Table 111. Th16 was the total RNA was pelletable at 100,000 x g m Exp. 1 Of Table I1 16% In ~x p . 21 compared to 42% in Table 3 . The data I" Table I1 ale0 indicated complet; equilibration of 6pernidlne ~n the 100.000 x g supernatant and pellet fractions. sperrnidrne thus cannot be shown to be sequestered on r~b o n o m e s m rirn on the bas16 of these data.