A Dictyostelium Mutant with Reduced Lysozyme Levels Compensates by Increased Phagocytic Activity*

Lysozymes are bacteria-degrading enzymes and play a major role in the immune defense of animals. In free-living protozoa, lysozyme-like proteins are involved in the digestion of phagocytosed bacteria. Here, we purified a protein with lysozyme activity from Dictyostelium amoebae, which constitutes the founding member, a novel class of lysozymes. By tagging the protein with green fluorescent protein or the Myc epitope, a new type of lysozyme-containing vesicle was identified that was devoid of other known lysosomal enzymes. The most highly expressed isoform, encoded by the alyA gene, was knocked out by homologous recombination. The mutant cells had greatly reduced enzymatic activity and grew inefficiently when bacteria were the sole food source. Over time the mutant gained the ability to internalize bacteria more efficiently, so that the defect in digestion was compensated by increased uptake of food particles.

The lysosome is the most potent degradative organelle within the eukaryotic cell. It contains hydrolytic enzymes that fulfill essential functions. In humans many mutations affecting its constituents lead to the diseased state, often with dramatic consequences. Mutations of this sort fall into two classes. The first class affects proteins involved in trafficking to and from the lysosome and/or regulating its integrity by modulating fusion and fission events. Among these proteins are Rabs, proteins belonging to a family that act as molecular timers (1) controlling virtually every trafficking step in the cell (2) and the LYST proteins, which when mutated cause enlarged lysosomes manifested as Chediak-Higashi Syndrome (3). The second class of mutations affects lysosomal enzymes with metabolic roles. These enzymes degrade macromolecules that the cell wants to dispose of, producing small metabolites that are useful building blocks or provide energy if completely degraded in the cytosol and mitochondria. Tragically, many of the lysosomal enzymes are exoenzymes, i.e. they digest macromolecules from their ends. If one of these enzymes is lacking, degradation proceeds up to the residue that cannot be removed, and as a consequence, all steps that ensue will be blocked. This leads to the accumulation of partially degraded intermediates within the lysosomes culminating either in the formation of toxic compounds or blowing up the volume of the lysosome until it sterically interferes with vital cellular activities. Depending on the cell type that is most seriously harmed, be it neurons, muscle, or leukocytes, patients will suffer from mental retardation, cardiac failure, or immunodeficiency (4,5).
In contrast, free-living amoebae are unlikely to develop lysosomal storage diseases because their endocytic pathway differs from that of mammalian cells. In mammalian cells, the lysosome is a dead-end of vesicular trafficking. In most cell types, it accumulates non-degradable material and retains it for the lifetime of the cell and organism. In amoebae, endocytosed cargo transits through the cell. Any material that cannot be degraded because specific enzymes are lacking is released from the cell by exocytosis after an hour. Despite these differences, amoebae such as Dictyostelium can serve as bona fide model systems to study lysosomal function in vivo. Interference with Rabs (6, 7) and LYST proteins (8,9) produces phenotypes in Dictyostelium that are comparable with those observed in mammalian cells, indicating that many lysosomal constituents are functionally conserved in evolution. On the other hand, a deficiency of cathepsin D leads to profound defects in mice (10) and sheep (11) but is tolerated well in Dictyostelium (12).
More recently mice have been generated that specifically lack the M isoform of lysozyme, which constitutes the most prominent bacteriolytic activity in the airways (13). Although homozygous mutants increase the level of expression of the lysozyme P isoform to compensate for their deficiency, they remain much more sensitive to infections of the lung (14). Lysozymes are not only found in animals but are also present in numerous phylogenetically diverse organisms such as plants, fungi, bacteria, and bacteriophages (15). Several different classes of theses enzymes have been described revealing that structurally diverse proteins fulfill the function of degrading the peptidoglycan of bacteria by splitting a 1,4-linkage between N-acetylmuramic acid and N-acetylglucosamine (16). Despite the fact that Dictyostelium has been regarded as a good model for elucidating molecular mechanisms underlying phagocytosis of cells from higher organisms, e.g. mammalian defensive cells, nothing is known about the molecular armament that it uses to efficiently eliminate the enormous number of bacteria phagocytosed for nutrition. We have isolated a protein with lysozyme activity from Dictyostelium amoebae. The protein is the first antimicrobial polypeptide characterized in this protozoan organism at the molecular level, and it appears to be a member of a previously unrecognized lysozyme family. The enzyme is stored in a unique organelle, and disruption of its gene had a dramatic effect on the phenotype.

EXPERIMENTAL PROCEDURES
Purification of Lysozyme-The AX2 strain of Dictyostelium discoideum was cultured at 22°C axenically (17). Amoebae in late-logarithmic phase were harvested, sedimented at 320 ϫ g at 4°C for 5 min, and washed 2 times in Sörensen's 17 mM sodium/potassium phosphate, pH 6.0. Freshly harvested and washed amoebae (2,5 ϫ 10 9 cells) were lysed by three cycles of freezing and thawing in 5 volumes of 10 mM sodium acetate, pH 4.5, supplemented with complete proteinase inhibitor mixture (Roche Applied Science). The lysate was centrifuged at 150,000 ϫ g at 4°C for 1 h. Subsequent purification steps were carried out at 10°C, and samples were kept on ice. The 150,000 ϫ g supernatant was applied to a chromatography column (Econo column 2.5 ϫ 20 cm, Bio-Rad) manually filled with a Bio-Gel matrix (Bio-Gel A-0.5m; Bio-Rad). The column was equilibrated with 10 mM sodium acetate, pH 4.5. Adsorbed protein was eluted by washing the column with the same buffer (350 ml) and with 100 mM sodium acetate, pH 4.5 (200 ml). Finally, the column was washed with 500 mM sodium acetate, pH 4.5 (200 ml). Fractions with lysozyme activity were pooled and loaded on a Resource S cation exchange column (1 ml; Amersham Biosciences) equilibrated with 10 mM sodium acetate, pH 4.5. Adsorbed protein was eluted by first washing the column with the same buffer (5 ml), then by use of a gradient from 0 to 300 mM NaCl (25 ml) and finally by a wash with 1 M NaCl. Each fraction was tested for lysozyme activity.
Protein Analysis-Protein concentration was determined using the microbicinchoninic acid reagent assay (Pierce) and hen egg lysozyme as standard (18). Tricine 1 SDS-PAGE was performed according to Schägger and von Jagow (19). For N-terminal sequencing proteins were reduced with dithiothreitol and alkylated with iodoacetamide, subjected to Tricine SDS-PAGE and subsequent semidry electroblotting using a polyvinylidene difluoride membrane (Problott; Applied Biosystems), and analyzed using a gas-phase protein sequencer (model 437A; Applied Biosystems). For the detection of glycosylation proteins were analyzed after blotting onto nitrocellulose membrane using a glycan detection assay according to the manufacturer's instructions (Roche Applied Science). For mass spectrometric analysis, samples were dialyzed against 5% acetic acid and mixed with the same volume of saturated trans-sinapinic acid in acetonitrile, 0.1% trifluoroacetic acid as UV-absorbing matrix. About 10 l of the samples were spotted on a stainless steel probe tip and dried at room temperature. Measurements were performed using a Bruker Relex MALDI-TOF. Ions were formed by laser desorption at 337 nm using an N 2 laser. Spectra were recorded with an acceleration voltage of 35 kV in the linear mode. The mass spectrometer was calibrated with bovine serum albumin and carbonic anhydrase from bovine erythrocytes (both from Sigma) as external standards.
Enzyme Assays-Lysozyme activity was measured as described previously (20). In all assays hen egg lysozyme (Sigma) served as a control. To monitor lysozyme activity during purification, column fractions were analyzed using the lysoplate technique (21). For quantitative measurements of lysozyme activity, cell wall degradation of lyophilized Micrococcus luteus (Sigma) was determined turbidimetrically essentially according to Shugar (22). Degradation of cell walls of Staphylococcus aureus (Sigma) was tested in the same assay. To determine the effect of pH and ionic strength on lysozyme activity, the turbimetric assay was adjusted according to Dobson et al. (23). Various buffers were prepared as specified by Miller and Golder (24). They covered a pH range from 3.5 to 7.5, their ionic strength varying between 0.005 and 0.2. The occurrence of reducing N-acetyl amino sugars was tested as described by Reissig et al. (25). Chitinolytic activity was assayed using chitin glycol (Sigma) as a substrate (26). Activity against viable bacteria was determined in a microdilution susceptibility assay as published previously (27).
Gene Disruption-The sequence encoding AlyA was amplified from AX2 genomic DNA using primers bearing an EcoRI recognition sequence in front of the start codon (5Ј-CCGGAATTCAAAATGAGAAT-TGCTTTCTTTTTACTTGTTTTAGCCG-3Ј) and a BglII site after the last codon specifying an amino acid (5Ј-GGAAGATCTTTCGAGAT-CAAAACCATTACAGGTCTTTTCAGC-3Ј), cleaved with the corresponding enzymes, and ligated into the same sites of pIC19H. After deleting a SalI site in the polylinker of this plasmid, which is also recognized by HincII, the lysozyme sequence was cut at its remaining HincII site, and a blunted cassette conferring blasticidin resistance, isolated via BamHI and HindIII digestion from plasmid pUCBsrDel-Bam (28), was inserted. The resulting vector was cleaved with SmaI and NruI situated in the polylinker region to release the plasmid backbone required for replication and selection in Escherichia coli. The remaining linear 2.5-kilobase fragment containing 560 bp of lysozyme homologous sequences upstream and 550 bp downstream of the resistance cassette was transfected into Dictyostelium AX2 cells by electroporation, and the resulting clones were isolated by limited dilution into multiwell plates. To identify clones that carry a disrupted lysozyme gene, we performed PCR on preparations of genomic DNA (29) using primers (5Ј-GTGGTGCTCTCTATTTCACAGC-3Ј and 5Ј-CTTGTTCAACCCACATGGCTGG-3Ј) that flank 85 bp of the coding sequence containing two introns of different lengths which distinguish between the genes encoding lysozyme isoforms.
Expression of Lysozyme and Tagged Variants-To rescue possible defects in a lysozyme mutant, EcoRI and HindIII enzymes were used to cut out lysozyme from pIC19H, and the gene was inserted into the expression vector pDEX-RH (30) cut with the same restriction enzymes. To allow for detection of the transgene, a Myc epitope sequence cloned in pIC20R (31) was cut out using SmaI and AccI enzymes, blunted by filling in the two-base overhang, and ligated into the HincII site within the lysozyme sequence of the pDEX-RH expression construct. To monitor the distribution of lysozyme, a GFP fusion was constructed. The lysozyme coding sequence from the ATG up to the cleavage site of the prodomain was amplified by PCR using a cDNA library in phage (courtesy of Peter Devreotes) as a template. The product was digested at the EcoRI and BglII cleavage sites that flanked the PCR primers 5Ј-C-CGGAATTCAAAATGAGAATTGCTTTCTTTTTACTTGTTTTAGCC-G-3Ј and 5Ј-GGAAGATCTTGTAACAGTTGCTGTGATAAGGAAATGA-TCAC-3Ј and inserted into pDd-A15-gfp (32) as modified by (33), so that GFP replaces the C-terminal prodomain of lysozyme.
Endocytosis and Cell Growth-The uptake of particles and fluid phase was measured as described previously (34,35). To determine the capacity of cells to degrade bacteria in vivo, we used the assay established by Maselli et al. (36), except that we used E. coli expressing the green fluorescent protein (37) to measure the decrease in intracellular fluorescence after phagocytosis. To assay the growth properties of cells on bacterial lawns, we used M. luteus, Klebsiella aerogenes, and E. coli as substrates. These bacterial species were pre-grown on SM plates, harvested in Sörensen's phosphate buffer (as above), and mixed with about 80 cells of the appropriate Dictyostelium strain, and plaque diameter was measured after 5 days of growth on SM agar plates. In this assay no dramatic growth differences were observed between Gram-negative and Gram-positive bacteria.
Data Base Searches and Accession Numbers-A computer-assisted homology search was performed using the Internet BLAST and FASTA searches in the nucleic acid data base of the National Center for Biotechnology Information (NCBI) and the Dictyostelium protein data base (dictybase.org/db/cgi-bin/blast.pl), resulting in sequence information for alyA (AAM08434), alyB (AAM08432), alyC (AAO051440), alyD (DDB0217279 and DDB016810), DdLysC1, DdLysC2, and DdLysC3 (lysozymes of D. discoideum similar to c-type lysozymes; U66523, DDB0189256, DDB0205537, respectively). Dictyostelium genes coding for DdLys T41 and DdLysT42 (lysozymes of D. discoideum similar to T4 phage-type lysozymes; DDB0217501 and DDB0167824, respectively) and DdLysEh1 and DdLysEh2 (lysozymes of D. discoideum similar to Entamoeba histolytica; DDB0167552 and DDB0219864, respectively) were also identified in the Dictyostelium genome. The amino acid sequences of lysozymes of diverse organisms used in the phylogenetic analysis were retrieved from the Protein Entrez data base of NCBI and listed below with their respective abbreviations in Fig. 7  Protein alignments were performed using the software ClustalX (42), and phylogenetic relationships were analyzed using PAUP 4.0 (43). Bootstrap analysis (1000 replicates) was added to test tree robustness.

RESULTS
Purification of a Novel Lysozyme and Identification of the alyA Gene-We purified a protein with lysozyme activity from axenically cultured Dictyostelium amoebae, exploiting its adsorption to a Bio-Gel matrix. The main part of active material found in acid extracts of amoebae was only weakly retained by the column. It was eluted as a single protein peak with the starting buffer. Final purification was achieved by cation exchange chromatography and yielded homogeneous material with a molecular mass of ϳ12 kDa as judged by SDS-PAGE (Fig. 1A). The purified protein degraded M. luteus cell walls optimally at low ionic strength between pH values ranging from 4.5 to 6 (Fig. 1B). A colorimetric test for detecting reducing ends of N-acetyl amino sugars within a preparation of degraded cell walls was positive and comparable to hen egg lysozyme. Because the protein did not exert chitinase activity in the detection system used (data not shown), these results together support a true N-acetylmuraminidase (lysozyme) activity of the protein. Purified lysozyme was relatively stable; at pH 5.5, 10-min incubations at 50°C did not reduce the activity, but in contrast to hen egg white lysozyme, its activity decreased to half when the enzyme was incubated at 60°C for 10 min. The enzyme displayed antibacterial activity against viable bacteria; the protein inhibited the growth of the Grampositives Bacillus subtilis and Bacillus megaterium at 2.5 g/ ml. The growth of E. coli, representing Gram-negatives, was not inhibited at 10 g/ml. Neither viable S. aureus nor cell walls of this bacterium were affected by the enzyme at this concentration (data not shown).
The purified lysozyme was subjected to N-terminal sequencing, yielding a single amino acid sequence up to residue 25 (YSCPKPCYGNMCCSTSPNNQYYLTD). This sequence was back-translated into nucleic acid sequence and used to search a cDNA data base (44). A cDNA clone, SSF469, was retrieved, the deduced amino acid sequence of which perfectly matched the experimentally determined amino acid sequence of lysozyme. Using this clone we could identify the corresponding gene on chromosome 2 in the data base from the Dictyostelium genome project (45) and named it alyA (for amoeba lysozyme ). The information on genomic structure and amino acid sequence is summarized in Fig. 2.
The N-terminal tyrosine residue of the mature enzyme is preceded by a stretch of 19 mostly hydrophobic amino acid residues that may serve as a signal peptide necessary for the transport of the protein into the endoplasmic reticulum membrane. MALDI-TOF analysis of the purified AlyA protein yielded a molecular mass of 12.652 kDa. This value confirms the molecular mass estimated from SDS-PAGE and is in good agreement with a mature peptide of 119 residues that would have a calculated molecular mass of 12.633 kDa provided that the 12 cysteine residues participate in disulfide bonding. No glycosylation of the purified protein was experimentally evident as judged by a glycan detection assay. The only putative N-glycosylation site in the primary translation product is at Asn 149 , beyond the C terminus of the mature protein. Neither the mature protein nor its precursor molecule reveal significant similarity to any protein of other organisms reported so far, in particular not to other bacteriolytic enzymes.
Phenotypic Instability of aly Knockouts-To analyze the in vivo function of this novel protein, we disrupted the alyA gene. Therefore, we electroporated Dictyostelium cells in the presence of a linear DNA fragment in which genomic regions of alyA flanked a cassette conferring blasticidin resistance to transformed cells. As soon as clones appeared on the primary Petri dish, we re-cloned them by limited dilution in liquid growth medium. In the first attempt to isolate a mutant lacking alyA, we screened about 100 clones by PCR analysis of genomic DNA and obtained a single mutant cell line. Cell homogenates from this clone 74 had a lysozyme activity that was reduced to about 50% of wild-type levels. When plated on a bacterial lawn, the mutant initially formed plaques where the food bacteria were consumed that grew only to half the size measured for the wild-type strain. This phenotype, however, was not stable. Plating experiments repeated over the course of 2 weeks revealed that the diameter of plaques increased in the alyA mutant until no difference was seen to wild-type cells. In contrast, the enzyme activity remained low. Because of this phenotypic instability, we repeated the transformation to obtain another mutant. In the second attempt, one mutant was found among about 150 clones analyzed. The following analysis focuses on this clone, 138. PCR analysis using a primer pair that binds to both alyA and alyB genes (see below) revealed both genes in genomic DNA from wild-type cells, whereas only a product corresponding to alyB was detected in the mutant (Fig. 3A). The mutant carried the blasticidine resistance cassette precisely within the alyA locus (Fig.  3B). As seen previously in strain 74, the knock-out mutant alyA138 had a total lysozyme activity corresponding to about 40% of wild-type cells and maintained this level over more than 2 months (Fig. 3C). When the plaque diameters of the alyA138 strain were measured and compared with wild-type values, they were found to increase from 60% to 90% within about 10 days, steadily rising to 150% after 40 days and finally stabilizing at a constant value corresponding to 180% of wild-type cells (Fig. 3D). Despite these changes, repeated PCR analyses did not reveal any alteration at the alyA locus (data not shown). All of the subsequent experiments were performed using cells from this stage.
To test whether the increased plaque size was due to an AlyA-independent increase in the capacity to degrade bacteria, we fed the cells with GFP-expressing bacteria and measured the disappearance of their fluorescence signal. Unexpectedly, the alyA138 mutant performed as efficiently as wild-type cells (Fig. 4A), although it was still reduced in lysozyme activity. To find out whether the increased plaque size was caused by an increased efficiency of phagocytosis, we measured the internalization of an indigestible particle, fluorescently labeled yeast. Indeed, the rate of particle uptake by alyA138 mutant cells was almost 2-fold higher than the rate of wild-type phagocytosis (Fig. 4B), providing a reasonable explanation for the increased  Fig. 2). The upper band corresponding to the alyA gene is missing in the mutant. The remaining band was cut out and sequenced and, thus, identified as a product of a second isoform, the alyB gene. The intron positions are conserved, but their length varies. Intron 4 is 90 bp in alyA and 81 bp in alyB. Intron 5 is 125 bp in alyA and 79 bp in alyB. Together, the sizes of alyA and alyB introns amount to a 55-bp difference detected in the PCR reaction. M, molecular mass standards. B, PCR using a primer corresponding to a sequence coding for the N terminus of the signal peptide and a primer that is located within the blasticidin resistance cassette only amplifies a product from alyA138 mutant DNA. C, enzymatic activity of lysozyme was measured in wild-type cells (wt, filled symbols, set to 100%) or the alyA138 mutant (open symbols) at intervals after transformation, and the turnover rates were plotted over time (days). It did not change over the 2-month period shown and remained constant thereafter. D, plaque diameter of the alyA138 mutants increased with time (days). Single cells were seeded together with lawns of bacteria onto agar plates, and the diameter of the zone cleared by phagocytosis was measured on day 5 after inoculation. Wild-type values were set to 100%. plaque diameter. Because the rate of fluid-phase uptake was unaltered (Fig. 4C), the phenotype observed in the alyA138 mutant is not caused by a general change in the endocytic properties of these cells.
Rescue of Phenotypic Defects-In the case of unstable phenotypes, it is particularly important to check whether the reexpression of the product corresponding to the disrupted gene restores the phenotype to the wild-type-like character. To this end we inserted the entire coding region of alyA including the introns into an expression vector and produced stably transformed cell lines in the alyA138 mutant background. RT-PCR analysis of various clones generally revealed a more intense band corresponding to the alyA gene (data not shown), which reflects the high copy number integration of the plasmid into the genome. In clone 2.4 phagocytosis was reduced to approximately wild-type levels (Fig. 4B), plaque size returned to normal (Fig. 4D), and the lysozyme activity in homogenates was also partially rescued (Fig. 4E).
Because we failed to produce a polyclonal antibody in chicken, which was suitable for the localization of the enzyme in cells, we resorted to analyzing the distribution of genetically tagged AlyA proteins bearing an Myc epitope in the middle of the mature enzyme or a GFP extension at its C terminus (Fig.  5A). In the former construct the enzyme prodomain was maintained, whereas the GFP tag eliminated the cleavage signal, replacing the C-terminal prodomain. These constructs were individually transformed into the alyA mutant background. The presence of the transgene was verified by PCR (Fig. 5B), and the protein was detected in Western blots (Fig. 5C). For all rescued clones we established that the endogenous copy of alyA was still disrupted (data not shown). Although expression of each of the tagged lysozymes increased the enzymatic activity of cell extracts only mildly (Fig. 5D), the efficiency of phagocytosis approached wild-type levels (Fig. 5E), indicating that the hybrid molecules were sorted to proper destination in the cells.
Localization of the alyA Product-Both tagged proteins localized to numerous small vesicles distributed throughout the cytoplasm. Because the GFP moiety replaced the prodomain in the hybrid construct, we compared the subcellular distributions of AlyA-GFP and Myc-tagged AlyA directly in cells transformed with a mixture of both plasmids. Fig. 6A documents that the majority of GFP-containing vesicles are also loaded with the Myc-tagged enzyme. The perinuclear ring and the peripheral structures labeled with the anti-Myc-antibody alone represent the endoplasmic reticulum (data not shown), suggesting that the Myc epitope in the middle of the AlyA protein reduces the efficiency of protein folding. To investigate the identity of the labeled vesicles, we checked the AlyA-GFPexpressing strain for the distribution of three classes of lysosomal enzymes known to reside in different types of lysosomes. the decrease of intracellular fluorescence was measured over time. B, the rate of phagocytosis is increased in the alyA138 mutant and can be rescued to wild-type-like levels upon re-expression of alyA (open squares). Cells were incubated together with fluorescent yeast particles, and the intracellular fluorescence signal (relative units) originating from phagocytosed particles was measured at 15-min intervals. r.u., relative units. C, fluid-phase uptake is unaltered in the alyA138 mutant. Cells were allowed to internalize nutrient medium containing fluorescently labeled dextran, and intracellular fluorescence (relative units) was recorded over time. D, re-expression of AlyA reverts the plaque size to wild-type-like dimensions. Conditions were as for Fig. 3D. Surprisingly, neither enzymes bearing the common-antigen 1, a mannose-6-SO 4 -containing oligosaccharide (Fig. 6B), nor enzymes characterized by the N-acetylglucosamine 1-phosphate modification co-distribute with AlyA (Fig. 6C). Vesicles containing crystal-forming esterases are also devoid of AlyA-GFP (Fig. 6D). We, therefore, conclude that the enzyme resides in a novel type of vesicle in Dictyostelium. To test whether these vesicles deliver their contents into phagosomes, AlyA-GFPexpressing cells were fed with synthetic particles. One of the rare events, where internalized beads are surrounded by a rim of fluorescence, is shown in Fig. 6E. More frequently we saw fluorescent granules accumulating in the vicinity of a bead as if docked to the phagosome membrane (also visible in Fig. 6E). We assume that the protein is rapidly degraded when it encounters the degradative enzymes of phagosome but is rather stable in the storage vesicles. A, schematic drawing of the tagged constructs. signal indicates the N-terminally cleaved signal peptide for translocation into the endoplasmic reticulum. The active mature enzyme is followed by the C-terminally cleaved prodomain (pro). Myc and GFP tags were inserted in the middle of the mature protein (catalytic) or replacing pro, respectively. B, PCR of Myc and GFP rescues produced in the alyA138 background. In clones bearing the Myc epitope within the coding sequence (Myc) the PCR-fragment was up-shifted in size relative to the alyA band from wild-type genomic DNA (AX2) because the construct was based on the genomic clone. In cells expressing the GFP-hybrid protein (GFP) the PCR fragment was smaller than the genomic alyA product because in this case the fusion protein was constructed using the

DISCUSSION
The only lysozymes of protozoan origin characterized at the molecular level to date are from E. histolytica (20,46), a parasitic amoeba colonizing the human small intestine. Entamoeba is thought to have diverged from the evolutionary path to higher organism virtually at the same time as Dictyostelium (47), and it was tempting to think that Dictyostelium lysozymes are more closely related to the Entamoeba lysozymes than to enzymes from other organisms. We purified the major protein with lysozyme activity from extracts of Dictyostelium amoebae. The protein was characterized and N-terminally sequenced, and a subsequent data base search identified a gene product with unknown function and allowed the elucidation of the entire primary structure of the protein. The enzyme described here is not related to any known lysozyme, and we suggest viewing it as the founding member of a new lysozyme class.
Several other sequences for putative gene products that may have lysozyme-like properties can be extracted from the Dictyostelium data base (Fig. 7A). A recent survey of the genome reveals the existence of 11 lysozyme genes that potentially code for lysozymes belonging to four classes; two are of the bacteriophage T4 type, and three are homologous to the C-type lysozymes. Dictyostelium also has two genes that may code for proteins closely related to the unconventional lysozymes first described from E. histolytica (20,46). Searching with the sequence of the AlyA protein turns up three additional members of this so far unique lysozyme gene family. The predicted protein sequences of ALY B and ALY C are characterized by 12 and 22 exchanged residues, respectively. The fourth and most divergent member of the proteins encoded by the aly family, ALY D, carries a 77-residue insertion in the middle of the molecule, which consists almost exclusively of serines and glycines. The remaining sequence is only about 40% identical to AlyA. In summary, no other organism is as well equipped with lysozyme genes as Dictyostelium.
Many of the animals investigated so far appear to possess a single type of lysozyme only; multiple isoforms of a single lysozyme type can be found when the protein has also been recruited as a digestive enzyme (23). The only other organisms known to date bearing genes potentially coding for more than one type of lysozyme are D. melanogaster and C. elegans. They possess c-type and i-type lysozymes or E. histolytica-type and i-type lysozymes, respectively (16,46). Phylogenetic analysis revealed that the newly recognized AlyA-D described here represent a novel type of lysozymes (Fig. 7B). It is reasonable to assume that the amoebic lysozymes, as in other phagocytic cells, are constituents of an intracellular armament that fulfill a digestive function to exploit bacteria for nutrition. In addition, these enzymes may prevent uncontrolled microbial growth within the digestive vacuoles.
Irrespective of whether AlyA is tagged with a Myc epitope in the middle of the molecule or with a GFP moiety replacing the C-terminal prodomain, the enzyme accumulates in the same type of small vesicle (Fig. 6A). This observation indicates that the prodomain is not required for proper sorting of AlyA, an observation that has been made with other lysosomal enzymes as well (48,49). Because the N-terminal signal peptide is likely to be cleaved immediately after entry into the endoplasmic reticulum, the mature enzyme must encompass all the signals required for correct subcellular targeting.
Interestingly, AlyA is concentrated in a specific sort of vesicle devoid of other lysosomal enzymes (Fig. 6). This finding increases the number of primary lysosomes to four, because it has been convincingly shown that mannose-6-SO 4 -modified enzymes and N-acetylglucosamine 1-phospate-modified proteins do not coexist in the same vesicle (39), and the number and morphology of esterosomes (40,50) identifies the AlyA-containing organelles as a novel class of lysosomes. A possible reason for sorting AlyA away from the majority of proteases lies in the sensitivity of the molecule; it is very short-lived in total cell extracts at acidic conditions and becomes increasingly stable during purification. Indeed, immunofluorescence experiments reveal that bead-bearing phagosomes only rarely contain detectable steady state levels of the enzyme (Fig. 6), but its acidic pH optimum is in good agreement with the pH values found within Dictyostelium endosomes.
When alyA, the gene encoding the major lysozyme isoform, is disrupted, the total lysozyme activity of cells extracts is reduced to 40% of wild-type cells as measured by the degradation of bacterial cell walls in vitro (Fig. 3C). In support of these findings, the initial ability of cells to form plaques on a lawn of bacteria is reduced to a similar degree (Fig. 3D). Taken together these two observations clearly support a role for AlyA in efficient degradation of phagocytosed bacteria, which is perfectly consistent with its well known activity to cleave the bacterial peptidoglycan, the major constituent of the cell wall of Gram-positives. As peptidoglycan is also a constituent of the cell wall of Gram-negative bacteria, it is not surprising to find the same initial results when alyA mutant cells grow on lawns of E. coli and K. aerogenes.
However, when the alyA-deficient cells are cultivated for prolonged times, a surprising observation can be made. The efficiency of plaque formation increases gradually until it reaches wild-type levels, finally exceeding these values by almost 2-fold (Fig. 3D). The reasons for these phenotypic changes are not trivial, because the enzymatic activity of lysozyme in both strains remains constantly at 40% of wild-type level in vitro (Fig. 3C). Most importantly, alyA mutant cells attain wild-type properties after re-expression of wild-type and hybrid AlyA transgenes (Figs. 4 and 5), indicating that the changes observed in the mutant over time are not the consequence of an accumulation of a number of second site mutations. Instead, the compensatory events leading to increased phagocytosis in the mutant are more likely to involve metabolic changes, regulation within networks of gene expression, or epigenetic phenomena.
In principle, compensation that leads to the increased plaque size in the Dictyostelium alyA mutants can be of two kinds. Either the ability to degrade bacteria is enhanced or the rate of phagocytosis is augmented. The first possibility can be ruled out because when the mutants were measured for their ability to attack GFP-expressing bacteria, they perform similar to wild-type cells (Fig. 4A). On the contrary, when the uptake of non-degradable yeast particles is measured, the rate of phagocytosis for all strains corresponds to the behavior observed in the plaque-forming assay (Figs. 4 and 5).
The rate of phagocytosis depends on the dynamics of the actin cytoskeleton, and many Dictyostelium mutants affected in actin-binding proteins show a reduced rate of particle uptake (51). Interestingly, a number of mutants exist that are characterized by an increased rate of phagocytosis (52)(53)(54). When the expression levels of a variety of these proteins were analyzed by Western blotting of mutant cell extracts, no differences to the wild type were detected (data not shown). On the other hand, it may be informative to look in more detail into signal-transducing cascades. In particular, overexpression of the small GTPases RacC or Rap1 is known to result in increased phagocytosis, whereas fluid-phase uptake remains unaffected (55,56). Although these proteins could possibly signal to the cytoskeleton, it remains to be investigated how they would perceive the level of AlyA within the compartments of the secretory and/or endocytic pathway. FIG. 7. Multiple sequence alignment and phylogenetic tree of lysozymes. A, multiple sequence alignment using ClustalX. The amino acid sequences of the ALY isoforms A-D are shown as proenzymes. The other amino acid sequences of putative lysozymes of D. discoideum, namely c-type (DdLysC1-3), phage-type (DdLysT41-2), and Entamoeba-type lysozymes (DdLysEh1-2), were found in databases and are aligned with representative homologues of the respective class of lysozymes from other organisms. B, phylogenetic tree of lysozymes from different species drawn and analyzed with PAUP 4.0. Data on sequences are given under "Experimental Procedures." An unrooted distance tree (neighbor joining) is presented. Numbers at the nodes represent bootstrap proportions on 1000 replicates; values greater than 70% are depicted only. The ALY family members are marked by bold letters. The invertebrate lysozymes are considered a monophyletic family, and therefore, the entire branch is shaded.