Characterization of a DL-dityrosine-containing macromolecule from yeast ascospore walls.

We have shown previously that the outer layers of yeast ascospore walls contain dityrosine and that this amino acid is a major component of the cross-linked peptides present in the spore wall (Briza, P., Winkler, G., Kalchhauser, H., and Breitenbach, M. (1986) J. Biol. Chem. 261, 4288-4294). We now present evidence that dityrosine is located in the outermost layer and that it is in the DL-configuration. Although the proteins (peptides) of the spore wall are insoluble, the macromolecule containing dityrosine can be solubilized by partial acid hydrolysis of spore walls. Analysis of this macromolecule indicates that it contains more than 50 mol% dityrosine and a very limited number of other amino acids. Interestingly, part of the dityrosine of spore walls is present in the DL-configuration. We speculate that not only the high degree of cross-links in the outermost layer but also the D-configuration of part of the alpha-C-atoms of dityrosine could contribute to the spores' resistance to lytic enzymes.

I I Wien A-1090, Austria . We have shown previously that the outer layers of yeast ascospore walls contain dityrosine and that this amino acid is a major component of the cross-linked peptides present in the spore wall (Briza, P., Winkler, G., Kalchhauser, H., and Breitenbach, M. (1986) J. Biol. Chem. 261,[4288][4289][4290][4291][4292][4293][4294]. We now present evidence that dityrosine is located in the outermost layer and that it is in the DL-configuration.
Although the proteins (peptides) of the spore wall are insoluble, the macromolecule containing dityrosine can be solubilized by partial acid hydrolysis of spore walls. Analysis of this macromolecule indicates that it contains more than 50 mol % dityrosine and a very limited number of other amino acids. Interestingly, part of the dityrosine of spore walls is present in the DL-COnfigUration.
We speculate that not only the high degree of cross-links in the outermost layer but also the D-COnfiguration of part of the a-C-atoms of dityrosine could contribute to the spores' resistance to lytic enzymes.
The ascospore wall of budding yeast, Saccharomyces cereuisiae, is a multilayered structure (l-3) biosynthesized during ascus maturation, the last phase of yeast sporulation. Morphologically, the spore wall outer layers have no counterpart in the vegetative cell wall (3). We showed that the amino acid dityrosine is a sporulation-specific component of the outer layers (2) and that the second outer layer is made up of chitosan (3). Here we present evidence that dityrosine is located only in the outermost layer of the spore wall.
Dityrosine has been shown to occur in hard and protective biological structures such as insect cuticle resilin (4) and the hard fertilization membrane of the sea urchin egg (5). In the latter system, a specific peroxidase was described, which is responsible for the oxidative cross-linking of 2 tyrosine residues of a precursor polypeptide (6). In the yeast ascospore wall, the dityrosine-containing layer serves purposes similar to those of other biological structures containing dityrosine. It seems to be essential for the spores' resistance to lytic enzymes but not for its viability (2). We have investigated the properties of a dityrosine-containing macromolecule of the spore wall outermost layer. The chemical structure of this macromolecule seems to be different from those of other dityrosine-containing biological structures. First, the dityrosine content of the spore wall macromolecule is much higher than that of comparable biological structures (5), pointing to a highly cross-linked macromolecular network. Second, the same macromolecule contains a very limited set of different amino acids. Third, some of the dityrosine c&-atoms exhibit the D-configuration in vivo. The implications of these findings for our understanding of the structure and biosynthesis of the dityrosine-containing macromolecule will be discussed.

RESULTS
Dityrosine Is Located in the Outermost Layer of Yeast Ascospore Walls-To test for the presence of dityrosine on the surface of yeast spores, spores were purified on a Percoll gradient as described earlier (2,11) and treated with [Y] potassium iodide. This procedure labels phenolic residues present on the surface of cells. Mid-log phase vegetative cells were collected by centrifugation, washed, and similarly surface labeled with lZ51. In both experiments, the same number of cells or spores (3.3 x 107) was used and treated with exactly the same amount of lZ51 (50 j&i). The total amount of radioactivity incorporated was measured by liquid scintillation counting. The spores were found to bind 60 times more '9 than vegetative cells. Both labeled preparations were then treated with Pronase, and the kinetics of the liberation of Ylabeled material was monitored.
80% of the iodine incorporated on t,he surface of intact yeast spores was stable to proteolytic digestion whereas more than 80% of the '7 bound on the surface of vegetative yeast cells was liberated by Pronase digestion (Fig. 1).
This experiment demonstrated that vegetative yeast cells contain relatively little iodinatable protein on their surfaces. The proteins that could be labeled were readily degraded by Pronase.
In contrast, yeast spores contain much larger amounts of an iodinatable substance on their surface. The labeled substance was stable to proteolytic digestion and could not be solubilized by boiling spores in SDS-P-mercaptoethanol. These properties suggest that the '251-labeled material is the dityrosine-containing macromolecule described below and that this macromolecule is present on the surface of the spore.
Properties of the Surface Layer-Both intact spores and purified spore walls were used for the experiments described here. Treatment with a number of commercially available enzyme preparations (Pronase, pepsin, trypsin, Glusulase, zymolyase) did not lyse or kill yeast spores. The same treatment failed to liberate soluble compounds containing dityrosine (as monitored by its fluorescence) from spores or purified spore walls. Treatment of intact spores with the enzymes mentioned above also resulted in no detectable morphological changes (monitored by electron microscopy; data not shown). However, the inner layers of purified spore walls could be digested with Glusulase or zymolyase, leading to very thin "walls" that seemed to consist only of layers 1 and 2 but still retained the shape of the spore walls as well as their dityrosine component.
Analysis of a Dityrosine-containing Macromolecule from Spore Walls-Although treatment of intact spores or purified spore walls with proteases, Glusulase, or zymolyase did not release any soluble compounds containing dityrosine, we found that partial acid hydrolysis of spore walls (4 N HCl, 110 "C, 10 min) liberated a water-soluble macromolecular fraction.
The fluorescence spectrum of this fraction was nearly identical with that of dityrosine. A fluorescent macromolecule with very similar properties was liberated by treatment of spore walls with trifluoromethylsulfonic acid, a reagent believed to cleave specifically the glycosidic bonds of sugar residues (12).
The fluorescent macromolecule present in the acid hydrolysate of spore walls was partially purified by ethanol precipitation and reversed phase HPLC (Fig. 2). Two fluorescent peaks were resolved. Peak 1 was sharp and free of polysaccharides. Peak 2 (Fig. 2) was very broad and contained most of the fluorescent macromolecular material along with soluble chitosan (3). The molecular mass of both fluorescent substances from peak 1 and peak 2 was nonhomogeneous, producing a broad smear on SDS-polyacrylamide gel electrophoresis ranging from about 10 to about 100 kDa (Fig. 3A). Upon analytical isoelectric focusing, the fluorescent material of the two peaks behaved similarly, forming a reproducible pattern of about seven bands (Fig. 3B). The most likely interpretation of a regular pattern such as this one is that the macromolecule contains regular repeats of an ionizable substructure. The substance constituting peak 2 contained a Coomassie-positive contaminant (soluble chitosan and minor amounts of protein). The substance constituting peak 1 was free of any nonfluorescent contaminating chitosan or protein. Therefore, the amino acid composition of peak 1 is characteristic of the dityrosine-containing macromolecule. The amino acid compositions of the partially purified fluorescent macromolecules of peaks 1 and 2 ( Fig. 2) were determined by HPLC (2) and compared with those of intact spore walls and vegetative cell walls (Table I) A, SDS-polyacrylamide gel electrophoresis and B, analytical isoelectric focusing of dityrosine-containing macromolecules of spore walls (peak 1 of Fig. 2). enriched nearly 3-fold in the macromolecule constituting peak 1 when compared with the total amino acid content of the spore wall. The macromolecule constitutes about 50% (by weight) of the proteinaceous material of the spore wall or about 6% of the dry weight of the spore wall.
As shown in Table I, the dityrosine-containing macromolecule (peak 1) consists of very few different amino acids. The most abundant amino acid is dityrosine, followed by glycine, glutamic acid, alanine, and lysine. We do not know at present if there are other, non-amino acid components present in that macromolecule. Despite its proteinaceous nature, the macromolecule, like the surface layer of intact spores, was not degraded by the proteases and lytic enzymes mentioned above. This fact together with the very unusual amino acid composition of the macromolecule raised the question of whether this molecule is a polypeptide at all. We investigated the infrared spectrum of the substance constituting peak 1 (Fig.  4). The spectrum confirmed the absence of polysaccharides and the presence of dityrosine. It did not show the diagnostic peaks of standard trans-amide bonds at 1640 and 1520-1530 cm-' (13). The substance constituting peak 1 shows prominent bands at 1685 and 1435 cm-'. These bands do occur in the infrared spectrum of free dityrosine. They are also char-  Fig. 2; B, material of peak 2 of Fig. 2 acteristic of c&amide bonds (13) as observed, for instance, in diketopiperazines, and at present we cannot exclude the presence of such &-amide bonds in the dityrosine-containing macromolecule of peak 1. The chromophore of dityrosine seemed to be relatively undisturbed in the macromolecule, as both the UV and the fluorescence excitation and emission spectra of the macromolecule were very similar to those of free dityrosine (Fig. 5). They also showed exactly the same pH dependence (data not shown). This showed that the bonding of dityrosine in the macromolecule cannot occur via the phenolic hydroxyl groups. Free phenolic hydroxyl groups are essential for the observed fluorescence, as only the phenolate form of the molecule . Therefore, dityrosine is probably bound to the macromolecule via its amino and/or carboxylate groups.
Optical Configuration of Spore Wall Dityrosine-Amino acid analysis of a hydrolysate of yeast spore walls yielded two closely related peaks of dityrosine in about equal amounts, which have been assigned to the mesoand racemic forms of dityrosine, respectively (2). Here we demonstrate that the meso-dityrosine is present in vivo and is not an artifact of acid hydrolysis.
First, LL-dityrosine was not measurably converted to the meso-form under the conditions used for acid hydrolysis of yeast spore walls (6 N HCl, 110 "C, overnight, i.e. "standard conditions").
Second, trypsin was oxidatively cross-linked by performic acid treatment and hydrolyzed under standard conditions. The dityrosine recovered was entirely in the LL-form. Third, we isolated spore wall dityrosine under the mildest possible conditions (6 N HCl, 40 "C, overnight). The relative amount of meso-dityrosine recovered was the same as under standard conditions. Fourth, the residual dityrosine present in our dityrosine-less mutant ditlO1 was analyzed after standard hydrolysis and purification.
The mutant contained nearly exclusively the LL-dityrosine peak. Similarly, spores of a mutant auxotrophic for glucosamine (15) contained nearly exclusively the LL-dityrosine peak (Fig. 6). These experimental results support the notion that DL-dityrosine does occur in vivo in wild-type spores.
As a more direct test for possible racemization during hydrolysis of spore walls we measured the incorporation of "H into spore wall dityrosine using tritiated HCl (7) ( Table  II). The samples used for the tritium measurements were also checked for the presence of the two stereoisomeric forms of dityrosine by HPLC analysis. Phenolic compounds do show a small amount of 3H incorporation even in the absence of racemization (7). Table II shows the amount of background 3H incorporation into LL-dityrosine under standard hydrolysis conditions without racemization.
Hydrolyzing spore walls under standard conditions showed no significant increase in 3H incorporation over the background level. Using LL-dityrosine and a catalyst (benzaldehyde) in tritiated 80% acetic acid (lo), a 15-fold increase in "H incorporation and a DL-dityrosine peak (HPLC) were found, indicating partial racemization. This experiment showed that the system used was sen-  sitive enough to detect even moderate amounts of racemization during acid hydrolysis. LL-Dityrosine in 80% acetic acid, with no added catalyst, was not racemized and incorporated little tritium. Spore walls were hydrolyzed in 80% acetic acid, the peak ratio of the two forms of dityrosine was normal, and little 3H was incorporated (Table II). As there was no increase in 3H incorporation into dityrosine (i.e. no racemization) when spore walls were hydrolyzed, we conclude that meso-dityrosine exists in uiuo in the dityrosine-containing macromolecule. The dityrosine of the spore wall is not an equilibrium mixture of the two stereoisomeric forms. As shown by the circular dichroism measurements (Fig. 7), spore wall dityrosine contains a preponderance of L-configurated cY-C-atoms. (The equilibrium mixture should show no circular dichroism signals.) In other words, spore wall dityrosine was mostly in the DL-and LL-form with very small amounts of the DDconfigurated form.

DISCUSSION
The data presented here support a model for the yeast ascospore wall in which the outermost layer is of importance for the resistance of the spore against enzymatic attack. Ultrastructural analysis of wild-type spores shows that this layer is very thin. The iodination experiments indicate that dityrosine is exposed on the surface of the spore. Further evidence that the dityrosine-containing macromolecule is located exclusively in the surface layer of yeast spores was provided by the electron microscopy of spores of the yeast mutant ditl02 recently isolated in our laboratory.3 This mutant produced spores that were almost completely devoid of dityrosine and sensitive to Glusulase yet viable and lacking only the outermost layer of the spore wall. This suggests that only the outermost layer contains dityrosine and confers resistance to Glusulase. The glucosamine auxotroph gcnl(15) produced spores that lacked both outer layers and were sensitive to Glusulase but were viable. The results of the iodination experiment and of electron microscopy of mutant spores lacking dityrosine show that the dityrosine of yeast spore walls is located in the surface layer.
Experiments with the fluorescent stains fluorescein isothiocyanate-concanavalin A and primulin (3) allow some inferences to be made about the permeability of the surface layer. Both fluorescein isothiocyanate-concanavalin A and primulin showed little reactivity toward intact spores but readily stained spore wall preparations, indicating that the surface layer was impermeable to those molecules in intact spores. Both the intact surface layer and the soluble, dityrosinerich macromolecule described here are resistant to the attack of all proteases tested and to the enzyme mixtures Glusulase and zymolyase. Mutant spores that lack only the outermost layer of the spore wall and are devoid of dityrosine are readily lysed by Glusulase or zymolyase. The experiments described here indicate that the molecular structure of the surface layer protects the spores against lytic enzymes. The model suggested by these experimental results is that of a dityrosinerich, two-dimensionally cross-linked (and therefore dense and ' P. Briza, M. Breitenbach, A. Ellinger, and J. Segall, manuscript in preparation. insoluble) macromolecule covering the surface of the spore. What are the chemical composition and structure of this macromolecule?
We speculate that the dityrosine-rich macromolecule may be similar in structure to a prokaryotic peptidoglycan. This is based on the following observations.
(i) The amino acid composition, aside from dityrosine, of the purified macromolecule (Table I) is similar to that of a peptidoglycan (16). The macromolecule contains dityrosine, glycine, alanine, glutamic acid, and lysine whereas typical bacterial peptidoglycans contain the amino acids glycine, alanine, glutamic acid, and lysine (or diaminopimelic acid, depending on the species).
(ii) D-Amino acids do occur in the macromolecule.
We have shown here that about 50% of the spore wall dityrosine is DLconfigurated in vivo, and we have recently obtained evidence (to be published elsewhere4) that D-alanine and D-glutamic acid, which are well known components of bacterial peptidoglycans, do occur in the yeast spore wall. On the other hand, aspartic acid and valine are moderately abundant in the spore wall, are not part of the dityrosine-containing macromolecule (Table I), and exist exclusively in the L-configuration.
(iii) The dityrosine-containing macromolecule is closely associated with a structural polymer consisting of /3(1,4)linked glucosamine residues. This is the chitosan of the second outer layer of the spore wall (3), which is hard to separate from the dityrosine-containing macromolecule and may be covalently bound to it in Go. The corresponding structure in the prokaryotic cell wall is the /3(1,4)-linked polymer of Nacetylglucosamine and N-acetylmuramic acid. However, the structure of the dityrosine-rich macromolecule also exhibits one feature different from a peptidoglycan structure. Dityrosine cross-links have not been found in bacterial cell walls. They have been found in extremely small quantity in a bacterial spore wall, possibly due to an oxidative artifact during workup (17).
It would be interesting to know the molecules to which dityrosine is bound in the macromolecule, but only indirect evidence is available. The chromophore of dityrosine is undisturbed in the macromolecule, indicating that it carries free phenolic hydroxyl groups. Classical trana-amide bonds are absent (shown by infrared spectroscopy), dityrosine is set free by relatively mild acid treatment, and the only other known compounds constituting the macromolecule are glycine, alanine, glutamic acid, and lysine.
It has been shown here that the spore wall dityrosine in uiuo consists of the DL-and LL-forms.
To the best of our knowledge this is the first report on a D-amino acid in a eukaryotic cell or spore wall structure. We propose here that the D-configuration of part of the dityrosine molecules contributes to the spores' resistance to proteases. As D-amino acids have been found infrequently in eukaryotic cells and little is known about their biogenesis or function, we discuss here the published examples of D-aII'Iin0 acids in eukaryotes and possible biosynthetic routes for the DL-dityrosine-containing macromolecule of the yeast spore wall.
The cyanelles of the eukaryotic alga Cycmophoru paradoxa elaborate a peptidoglycan containing D-amino acids (18). These organelles are intermediates between free living photosynthetic bacteria and true chloroplasts. They have retained part of their bacterial wall structure. Free D-alanine and other D-amino acids as well as their malonyl derivatives exist in higher plants (19). However, their function and biosynthesis are unknown at present. The neuropeptide of frog skin, dermorphin, contains D-alanine (20), is very probably biosynthe-' M. Espelund, R. Ferris, and M. Breitenbach, manuscript in preparation.