Proc. Natl. Acad. Sci. USAVol. 83, pp. 1817-1821, March 1986Genetics

Sialidosis and galactosialidosis: Chromosomal assignment of twogenes associated with neuraminidase-deficiency disorders

(sialidase/gene mapping/complementation analysis)

0. THOMAS MUELLER, W. MICHAEL HENRY, LINDA L. HALEY, MARY G. BYERS, ROGER L. EDDY,AND THOMAS B. SHOWSDepartment of Human Genetics, Roswell Park Memorial Institute, New York State Department of Health, Buffalo, NY 14263

Communicated by Victor A. McKusick, November 14, 1985

ABSTRACT The inherited human disorders sialidosis andgalactosialidosis are the result of deficiencies of glycoprotein-specific a-neuraminidase (acylneuraminyl hydrolase, EC3.2.1.18; sialidase) activity. Two genes were determined to benecessary for expression of neuraminidase by using human-mouse somatic cell hybrids segregating human chromosomes.A panel of mouse RAG-human hybrid cells demonstrated asingle-gene requirement for human neuraminidase and allowedassignment of this gene to the (pter-_q23) region of chromo-some 10. A second panel of mouse thymidine kinase (TK)-deficient LM/TK-human hybrid cells demonstrated thathuman neuraminidase activity required both chromosomes 10and 20 to be present. Analysis of human neuraminidaseexpression in interspecific hybrid cells or polykaryocytesformed from fusion of mouse RAG (hypoxanthine/guaninephosphoribosyltransferase deficient) or LM/TK- cell lineswith human sialidosis or galactosialidosis fibroblasts indicatedthat the RAG cell line complemented the galactosialidosisdefect, but the LM/TK- cell line did not. This eliminates therequirement for this gene in RAG-human hybrid cells andexplains the different chromosome requirements of these twohybrid panels. Fusion of LM/TK- cell hybrids lacking chro-mosome 10 or 20 (phenotype 10+,20- and 10-,20+) andneuraminidase-deficient fibroblasts confirmed by complemen-tation analysis that the sialidosis disorder results from amutation on chromosome 10, presumably encoding theneuraminidase structural gene. Galactosialidosis is caused by amutation in a second gene required for neuraminidase expres-sion located on chromosome 20.

Several inherited diseases have been found to be associatedwith a deficiency of glycoprotein-specific N-acetyl-a-neuraminidase activity (acylneuraminyl hydrolase, EC3.2.1.18; sialidase). These disorders are typically classified asthe sialidoses, which have only a neuraminidase deficiency,and the galactosialidoses, which have a coexistent deficiencyof P-galactosidase (1, 2). The sialidosis disorder, originallytermed lipomucopolysaccharidosis (3), includes several var-iants with different degrees of clinical severity, including anadult-onset form known as sialidosis type I and the infantile-onset variant known as mucolipidosis I or sialidosis type II (4,5). The galactosialidosis disorder, which has also beentermed the Goldberg Syndrome (6), GM1 gangliosidosis type4 (7), the cherry-red-spot-myoclonus syndrome with demen-tia (2), and the juvenile-onset form of sialidosis type II (8), isalso clinically heterogeneous (9, 10).The primary defect in the sialidoses is thought to be a

mutation in the neuraminidase structural gene. Obligateheterozygotes show a gene-dosage effect and have approxi-mately half the normal neuraminidase levels (11, 12). Obligate

galactosialidosis heterozygotes have not consistently showna reduction in either ,B-galactosidase or a-neuraminidase (2,9, 13-16). Complementation studies from a number of labo-ratories show that the primary defect is different from thatcausing single deficiencies of P-galactosidase (GM1-gangliosidosis) or neuraminidase (7, 17-19). The residual3-galactosidase activity in galactosialidosis fibroblasts exists

entirely as a monomer with an absence of a large multimerform found in normal cells (20). The mutation causinggalactosialidosis also causes a decrease in turnover time dueto an increased susceptibility of,-galactosidase to proteo-lytic degradation (20, 21). The P-galactosidase deficiency, butnot the neuraminidase deficiency, can be corrected by theaddition of protease inhibitors (21-24). This could be accom-plished also with the addition of a "corrective factor"concentrated from cell culture medium, which has charac-teristics of a glycoprotein (23-25). Recent evidence suggeststhat the mutation causing galactosialidosis results in anabsence of a 32-kDa glycoprotein in anti-p-galactosidaseimmunoprecipitates, and this protein was postulated to be theprimary defect in galactosialidosis (23).We present evidence for the chromosomal assignment of

two genes required for the expression of human neuramini-dase in somatic cell hybrids. In addition, our evidenceindicates that the sialidosis and galactosialidosis diseases arecaused by mutations in these genes located on chromosomes10 and 20, respectively. A portion of this work has beenpresented in abstract form (26).

MATERIALS AND METHODSFibroblasts and Hybrid Cells. The fibroblasts used were

GM1718, derived from an infantile-onset variant of sialidosistype II (mucolipidosis I), and GM806, derived from a subjectwith an early-onset form of galactosialidosis. Hybrid cellswere made by fusing normal human cells or cells withbalanced chromosomal translocations with either mouseLM/TK- [thymidine kinase (TK) deficient] or mouse RAG(hypoxanthine/guanine phosphoribosyl transferase defi-cient) cells and selecting clones in hypoxanthine/aminopter-in/thymidine selection medium as described (27). The pres-ence of human chromosomes was determined in each hybridclone by scoring for previously mapped human enzymemarkers and by karyotype analysis with trypsin/Giemsabanding (28).

Neuraminidase Assays. Hybrids used for chromosomallocalization studies were analyzed for human neuraminidaseexpression as freshly harvested cultures due to the lability ofthis enzyme to freezing. Hybrid harvests were scored forneuraminidase and analyzed karyotypically on the samepassage. Hybrid cultures were rinsed in Dulbecco's phos-phate-buffered saline and scraped from flasks with a rubber

Abbreviation: TK, thymidine kinase.

1817

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 10

, 202

0

Proc. Natl. Acad. Sci. USA 83 (1986)

policeman. Cell pellets were homogenized gently in colddistilled water by using a Dounce type homogenizer with a

Teflon pestle. Neuraminidase activity was determined by a

procedure modified from that of Warner and O'Brien (29).Homogenates (20 ,ul) were incubated with 5 .ul of 1 M sodiumacetate (pH 4.2) and 5 ALI of 7.1 mM 4-methylumbelliferyl-a-neuraminide (Koch-Light Laboratories, Colnbrook, U.K.) at370C for 30 min. Reactions were terminated by adding 2 ml of0.85 M glycine (pH 10), and the fluorescence was determinedwith an Aminco fluorimeter. The inclusion of bovine serum

albumin at 1 mg/ml or 1 mM CaCl2 in assays or thepreparation of homogenates in the presence of 1 mMphenylmethylsulfonyl fluoride had no effect on activity.Protein concentration was measured by using a modificationof the Folin phenol procedure (30).Complementation Analysis. Each of the two cultures used

in complementation experiments was seeded at half theconfluent cell density, and the mixture was cultivated over-night. These cultures were fused by using 42% (wt/vol)polyethylene glycol 1000 (Koch-Light Laboratories) contain-ing 7% (vol/vol) dimethyl sulfoxide (31), and the resultingpolykaryocytes were enriched by sedimentation velocityunder sterile conditions with a Sta-Put apparatus (JohnsScientific, Toronto) as described (19). This polykaryocytefraction was cultured for 10-14 days, and the neuraminidaseactivity was determined along with cultures of each parentalcell type and a cocultivated mixture. In complemnentationexperiments involving human fibroblasts and RAG orLM/TK- cells, the polykaryocytes were cultivated inhypoxanthine/aminopterin/thymidine selection medium inorder to inhibit growth of the unfused mouse parental cells.A significant increase in the neuraminidase activity of thepolykaryocyte fraction over that of the cocultivated mixtureand parental cell activity range, as determined with Student'st test (at the 1% confidence level), was assumed to indicatecomplementation.

RESULTS

Somatic Cell Hybrid Mapping of Neuraminidase. The pres-ence of human neuraminidase activity in human-mousesomatic cell hybrids that segregate human chromosomescould be determined on the basis of a quantitative assay.Mouse RAG and LM/TK- parental cells, which were used tomake somatic cell hybrids, had low levels of neuraminidaseactivity (<8% of that of normal human cells) under the assayconditions used (Table 1). Therefore, the human activity in ahybrid cell could be quantitated by the amount of activity inexcess of the mouse cell activity levels. Hybrid cells werefound to have a large variation in neuraminidase activitybecause of the different numbers of the chromosome(s)encoding neuraminidase genes present in each hybrid (Table2). Hybrid cells with activity close to that of mouse parentalcells (within one standard deviation of the mean: 4.1 ± 3.1nmol/hr per mg for RAG hybrids and 4.9 ± 1.5 nmol/hr per

Table 1. Neuraminidase activity of human fibroblasts andmouse cells

4-Methylumbelliferyl-neuraminidase activity

Cell cultures n Mean SD RangeHuman fibroblastsNormal 60 66.3 45.5 17.6-189Sialidosis 9 1.6 1.7 0.0- 5.3Galactosialidosis 15 7.7 4.5 0.8- 17.0

Mouse cellsLM/TK- 22 4.9 1.5 2.6- 8.7RAG 22 4.1 3.1 0.0- 10.3

Table 2. Human-mouse hybrid cell neuraminidase activity

RAG Neuraminidase LM/TK- Neuraminidasehybrids Activity Score hybrids Activity Score

ATR-22 6.0 - ICL-6 10.8 +DUA-1 CH 2.0 - NSL-16 5.1 -DUA-5 BA 4.0 - TSL-2 10.3 +DUM-13 29.8 + TSL-2 CF 3.7 -DUM-23 36.3 + TSL-6F 6.0 -JSR-2 6.7 - VTL-6 10.3 +JSR-6C 24.5 + VTL-11 4.1 -JSR-6D 21.2 + VTL-13 4.4 -JWR-22H 57.3 + VTL-14 4.5 -JWR-26C 25.3 + VTL-15 5.8 -RAS-M3 26.0 + VTL-16 10.6 +REW-4 8.0 - VTL-17 4.3 -REW-8I C4 5.0 - VTL-18 4.4 -REW-11 5.5 - VTL-19 3.2 -REX-11 BF 20.5 + VTL-21 4.2 -REW-13 57.0 + VTL-22 5.0 -REW-15 34.8 + VTL-23 3.4 -REX-26 6.3 - WIL-2 3.5 -REX-33 5.0 - WIL-2 C 3.9 -REX-57 BB 5.0 - WIL-6 10.4 +SIR-1 4.0 - WIL-8 12.9 +XER-8 21.3 + WIL-8S 9.6 +XER-11 27.6 + WIL-8Y 11.9 +XER-15 21.3 + WIL-li 14.5 +XTR-1 27.0 + WIL-13 4.0 -XTR-1 BD 35.7 + WIL-14 5.7 -XTR-3 BH 20.8 + WIL-14 C 6.1 -XTR-22 26.3 +

The neuraminidase activity of hybrid cells is the mean of at leastthree separate harvests and is expressed in nmol/hr per mg ofhomogenate protein. Neuraminidase scores are based on statisticallysignificant increases over the activity of parental mouse cells (Table1). Hybrid clone XTR-3 BH contains del(10)(q23--qter) (deletion inchromosome 10 in region q23--qter).

mg for LM/TK- hybrids) were scored negative for humanactivity. Hybrid cells with neuraminidase activity well abovethis range (>20 nmol/hr per mg for human-RAG hybrids and>9 nmol/hr per mg for human-LM/TK- hybrids) werescored as positive (Table 2). Hybrid cells with activity levelsbetween these upper and lower thresholds were not consid-ered in discordance calculations. These hybrids were elimi-nated because of the variabilities associated with thenonquantitative scoring for human enzyme markers presentat low levels in hybrid cells. The presence of humanneuraminidase in hybrid cells was also confirmed qualita-tively by mobility differences between human and mouseenzymes following cellulose acetate electrophoresis (unpub-lished data).

In each hybrid the human chromosome content wasdetermined karyotypically and by scoring for the presence ofpreviously mapped human enzyme markers. The correlationbetween the presence of human neuraminidase and eachhuman chromosome is shown in Tables 2 and 3, expressed asthe discordance frequency. Instances where chromosomeswere present in <15% of the 30 metaphase spreads analyzedfor each hybrid were not considered in calculating discord-ance percentages in Table 3. The findings from a panel of 28RAG-human hybrids indicated that the expression ofneuraminidase was completely concordant only with chro-mosome 10 and the enzyme markers glutamic-oxaloacetictransaminase 1 (GOTI) and adenosine kinase (ADK), whichhad been mapped previously to chromosome 10 (32, 33). Forall other chromosomes, the expression of neuraminidase andthe chromosome was discordant in four or more hybrids

1818 Genetics: Mueller et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 10

, 202

0

Proc. Natl. Acad. Sci. USA 83 (1986) 1819

Table 3. Segregation of human neuraminidase gene in human-mouse somatic cell hybridsChromosome

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 XRAG hybrids scoredNeur (+)/chrom (+) 8 5 13 9 9 13 10 7 3 16 10 10 4 11 7 8 12 12 9 11 12 4 10Neur(+)/chrom(-) 4 9 3 8 8 4 3 11 13 0 4 7 14 6 8 10 5 4 8 6 4 13 2Neur (-)/chrom(+) 0 1 2 3 1 1 4 3 0 0 3 1 5 4 0 2 3 1 2 0 4 3 5Neur(-)/chrom (-) 7 10 7 7 8 9 7 7 7 10 8 8 3 6 10 8 7 9 6 10 5 4 3

Discordant hybrids, % 21 40 20 41 35 19 29 50 57 0 28 31 73 37 32 43 30 19 40 22 32 67 35

LM/TK- hybrids scoredNeur (+)/chrom(+) 3 2 4 5 5 5 5 4 0 8 6 3 3 5 3 0 8 5 3 9 6 0 5Neur (+)/chrom(-) 6 5 4 4 4 2 4 5 9 0 3 6 6 4 6 9 0 3 5 0 3 8 3Neur(-)/chrom(+) 1 3 6 4 7 3 3 10 0 6 5 8 4 5 8 3 15 5 4 0 9 2 4Neur(-)/chrom (-) 17 15 10 13 10 14 14 7 17 11 13 7 14 11 9 14 2 11 14 16 9 16 14

Discordant hybrids, % 26 32 42 31 42 21 27 58 35 24 30 58 37 36 54 46 60 33 35 0 44 38 27The number of hybrids expressing neuraminidase (Neur) and each chromosome (chrom) concordantly, (+)/(+) or (-)/(-), or discordantly,

(+)/(-) or (-)/(+), is shown for each chromosome. Human neuraminidase was scored (+) or (-) in each hybrid cell (Table 2) based on theactivity thresholds as discussed in the text.

(Table 3). This indicates that a gene necessary for neur-aminidase expression is encoded on human chromosome 10.Further, the presence of neuraminidase activity in hybridXTR-3-BsAg H with a deletion in chromosome 10 in theq23-*qter region indicated that this gene is located in the(pter-*q23) region of chromosome 10. A second hybrid panelconsisting of 27 human-LM/TK- somatic cell hybrids indi-cated that human neuraminidase expression was apparentlyconcordant with the presence of human chromosome 20(Table 3). For all other chromosomes, the discordancefrequency was 21% or more, including six hybrids thatdemonstrated discordance with chromosome 10 (24% discord-ance, explained below). This suggests that another genecontrolling the expression of neuraminidase is encoded onhuman chromosome 20.

Expression of Neuraminidase Deficiencies in Somatic CellHybrids. To explain this apparent discrepancy in the mappingresults from these two hybrid panels, we constructed hybridcells using sialidosis and galactosialidosis fibroblasts as thehuman parental cells and tested for human neuraminidaseexpression. Sialidosis and galactosialidosis fibroblasts areknown to be caused by mutations in different genes, sinceneuraminidase is complemented in polykaryocytes formedfrom a sialidosis-galactosialidosis fibroblast fusion experi-ment (17-19). Hybrid cells formed from fusion ofmouse RAGwith galactosialidosis fibroblasts (GAR hybrids) have signif-icant levels of neuraminidase activity in those hybrids re-taining chromosome 10 (Table 4). The activity in each GARhybrid varied somewhat according to the numbers of chro-mosome 10 present, but chromosome 20 was not necessaryfor expression. This indicates that the neuraminidase gene onchromosome 10 is functionally normal in this galactosialido-sis subject and also that RAG cells have a gene or proteinsimilar to that defective in this disorder. Therefore, inhuman-RAG hybrid cells made with normal human cells,human neuraminidase can be expressed in the absence of thisgene, and mapping analysis detected only a single require-ment for chromosome 10 (Table 3).Hybrid cells formed from the fusion of sialidosis fibro-

blasts-LM/TK- cells have no detectable human neuramin-idase activity, whether human chromosomes 10, 20, or both10 and 20 were present (Table 4). Further, no increases inneuraminidase activity could be detected in enrichedpolykaryocytes formed from fusion of these cells (Table 5).Hybrid cells from the fusion of LM/TK--galactosialidosisfibroblasts were not successfully constructed. However,complementation analysis using enriched polykaryocytesfrom fusion of LM/TK--galactosialidosis fibroblasts did not

demonstrate any increases in neuraminidase activity (Table5), indicating that LM/TK- cells, unlike RAG cells, lack thegene or protein that complements the galactosialidosis de-fect. Since LM/TK- cells apparently complement neither thesialidosis nor the galactosialidosis mutations, there is arequirement for both of these genes before humanneuraminidase can be expressed in LM/TK- hybrid cellsconstructed from normal human cells (Table 3). AllLM/TK--human hybrid cells in Table 3 that are scoredpositive for neuraminidase have both chromosomes 10 and20. All six hybrid clones in which the expression ofneuraminidase and chromosome 10 was discordant containedchromosome 10 but had no detectable human neuraminidaseactivity. All of these hybrid clones lack human chromosome20, explaining the absence of human neuraminidase activity.Chromosomal Assignment of Sialidosis and Galactosialidosis

Mutations. Complementation analysis also was used to de-termine whether either of the genes on chromosomes 10 and20 was mutated in the sialidosis and galactosialidosis disor-ders. Two hybrid clones from the LM/TK--human hybridpanel were identified that have the chromosomal phenotypes10+,20- and 10-,20+. These hybrids were used in comple-

Table 4. Neuraminidase expression in human-mouse hybrid cellsmade with human neuraminidase-deficient cells

Hybrid Neuraminidaseclone Activity Score Chromosome 10 Chromosome 20

Galactosialidosis fibroblast-RAGGAR-2 29.8 (+) +GAR-3 45.0 (+) + +GAR-9 27.1 (+) + +GAR-12 15.0 (w) +GAR-15 17.5 (+/w) + +GAR-16 18.7 (+ /w) +

Sialidosis fibroblast-LM/TK-SIL-1 0.9 (-) +SIL-2 1.2 (-) -SIL-3 1.0 (-) +SIL-5 2.6 (-) + +SIL-6 3.6 (-) + +

Neuraminidase activity is the average of three separate harvestsand is expressed in nmol/hr per mg of homogenate protein. Scoringfor human neuraminidase and chromosomes (+, w for weak, and -)was determined according to the thresholds established for Tables 2and 3.

Genetics: MueUer et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 10

, 202

0

Proc. Natl. Acad. Sci. USA 83 (1986)

Table 5. Complementation analysis of human neuraminidase deficiencies

4-Methylumbelliferyl-neuraminidase activity

Fusion experiments Cocultivated Fused Complementation

Human-mouse cellsSialidosis FB-mouse LM/TK- 7.5 5.8Galactosialidosis FB-mouse LM/TK- 4.5 4.5

Human-hybrid cells10+,20- hybrid only (n = 7) 2.3- 6.110-,20+ hybrid only (n = 5) 7.3-11.7Sialidosis FB x 10+,20- hybrid 0.1 10.9 +Sialidosis FB x 10-,20+ hybrid 3.1Galactosialidosis FB x 10+,20- hybrid 2.8 4.5Galactosialidosis FB x 10-,20+ hybrid 10.7 18.8 +

Complementation scoring was determined based on a highly significant increase (P < 0.01) in neuraminidase activity ofthe fused cell cultures compared with the range of appropriate parental cell activities and of cocultivated mixtures usingStudent's t test. FB, fibroblasts.

mentation experiments with sialidosis and galactosialidosisfibroblasts (Table 5). The 10-,20+ hybrid has weak levels ofhuman neuraminidase activity and was not included in theinitial assignment studies (Tables 2 and 3), according to theestablished criteria discussed above. This hybrid clone maycontain a low amount of chromosome 10, which was notdetected by enzyme marker measurement. It was neverthe-less useful for these complementation studies because scor-ing was based on activity increases over that measured incocultivated mixtures of parental cells. Both of these hybridclones were made with normal human cells, thereforecomplementation in these fusions is expected if the hybridcell supplies the chromosome containing the mutated gene inthe neuraminidase-deficient (sialidosis or galactosialidosis)fibroblasts. The 10+,20- hybrid-sialidosis fibroblast fusionresulted in a significant increase in neuraminidase activity,suggesting that sialidosis fibroblasts have a functionallynormal chromosome 20. The 10-,20+ hybrid-sialidosis fibro-blast fusion resulted in no change in neuraminidase activityover control cultures, confirming that the gene on chromo-some 10 is mutated in sialidosis. In the two experimentsinvolving galactosialidosis fibroblasts, the fusion with the(10-,20+) hybrid resulted in a significant increase in activity,indicating that galactosialidosis fibroblasts have a normalgene on chromosome 10. The fusion between the galacto-sialidosis fibroblast and the 10+,20- hybrid resulted in nocomplementation, confirming that chromosome 20 is mutatedin this disorder. Therefore, the inferred chromosomal phe-notype for sialidosis is 10m,20+ and for galactosialidosis is10+,20m.

DISCUSSIONGenetic complementation analysis of human diseases asso-ciated with a deficiency of glycoprotein-specific neuramini-dase activity indicated that there are apparently two distinctgenetic variants (17-19). We have determined that there aretwo genes required for the expression of this enzyme insomatic cell hybrids located on chromosomes 10 and 20.Although one of the two hybrid cell panels used demonstrat-ed a single-gene requirement for chromosome 10, it wasshown by complementation analysis that the mouse RAGgenome substitutes or complements one of the two necessarygenes (the gene on chromosome 20). In a second hybrid clonepanel made with mouse LM/TK- cells, the data indicate thatthere was no apparent contribution by the mouse genome,and neuraminidase expression was shown to require thepresence of both chromosomes 10 and 20. Although thesedata also can be interpreted to indicate a single-gene require-

ment located on chromosome 20, this chromosome is notsufficient for neuraminidase expression, as demonstratedwith hybrid cells made by fusion of sialidosis and LM/TK-cells (Table 4). Complementation analysis findings indicatedthat sialidosis fibroblasts have a mutation in the gene onchromosome 10 and that the gene on chromosome 20 isapparently normal (Table 5). Sialidosis-LM/TK- hybridcells retaining chromosome 20 do not have any detectablehuman neuraminidase activity, indicating that a functionalchromosome 10 gene is also required for neuraminidaseactivity. The different contributions of the two parentalmouse cell lines are most likely a result of the accumulationof different mutations during mutagenesis and selection forhypoxanthine/guanine phosphoribosyl transferase-deficientor TK- phenotypes. The two-chromosome requirement ob-served in human-LM/TK- cell hybrids also suggests areason for the relatively low neuraminidase activity observedin LM/TK- hybrid cells scored as positive, relative to that ofRAG-human hybrid cells scored positive (Table 2). Humanchromosomes in hybrid cells are rarely present at theirdiploid number (60 homologues out of the 30 metaphasespreads that were analyzed), and hybrid cells with as few as15% of possible homologues (9/60) were scored as positive.Therefore, the low levels of human neuraminidase activity inLM/TK- hybrid cells with both chromosomes scored posi-tive reflect the low numbers of hybrid cells containingsufficient numbers of both chromosomes.We also have shown that the deficiency of neuraminidase

activity in two subjects with sialidosis and galactosialidosis isassociated with a loss in function of the genes on chromo-somes 10 and 20, respectively. It has not been determinedwhether this can be generalized for all subjects with thesediseases, although no genetic heterogeneity has thus far beenidentified within sialidosis or galactosialidosis (17, 19). Theindirect evidence from the literature (i.e., a gene-dosageeffect in obligate heterozygotes and kinetic alterations in theenzyme ofhomozygotes) suggests that the sialidosis disorderis caused by a mutation in the neuraminidase structural gene.We presume, therefore, that the gene on chromosome 10encodes the structural enzyme protein. Oohira et al. (34)reported that there was a suggestion of linkage between alocus causing a case of sialidosis II (single neuraminidasedeficiency) with theHLA complex located on chromosome 6.Their analysis, however, was based on only seven informa-tive family members, and this may not be sufficient data toestablish linkage.Our data also indicate that the galactosialidosis disorder is

caused by a loss in function of the gene located on chromo-some 20. Evidence from complementation analysis and from

1820 Genetics: MueUer et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 10

, 202

0

Proc. Nati. Acad. Sci. USA 83 (1986) 1821

the lack of a demonstrated gene dosage effect in tissues ofobligate heterozygotes suggests that the defect is not in thestructural gene of either neuraminidase or 13-galactosidase.d'Azzo et al. (23) have proposed that the primary defect ingalactosialidosis is an absence of a 32-kDa "protectiveprotein," which forms part of the P-galactosidase andneuraminidase enzyme complex. This protein was found tobe encoded on chromosome 22 (35). The function of thechromosome 20 gene and its relationship to the protectiveprotein remains to be determined.The galactosialidosis mutation is complemented or cor-

rected by one of the two mouse cell lines (RAG) that wereused to make cell hybrids. The cocultivated mixture ofRAGand galactosialidosis cells also had a slight increase in activityover that expected by averaging the parental cells' activities.This could be due to diffusion of the postulated correctivefactor or protective protein from the hybrid cells, resulting inpartial correction of the activity of galactosialidosis fibro-blasts as noted (23-25). Most human structural gene muta-tions are not corrected by the mouse genome, althoughcorrection of the lysosomal enzyme deficiencies in I-celldisease in man-mouse hybrid cells was demonstrated (36).This disease is caused by a deficiency of a Golgi enzyme,N-acetylglucosaminyl phosphotransferase, involved in intra-cellular lysosomal enzyme transport. This suggests that thedefective gene in galactosialidosis also may be involved inposttranslational processing or regulation of the affectedenzymes. This postulated regulatory gene would function tomodulate or coordinate the activity of P-galactosidase andneuraminidase and would have a key role in the initial stepsof glycoprotein catabolism.

This work was supported in part by grants HD 051% andGM 20454from the National Institutes of Health and 1-935 from the March ofDimes.

1. Okada, S., Kato, T., Miura, S., Yabuuchi, H., Nishigaki, M.,Kobata, A., Chiyo, H. & Furuyama, J. (1978) Clin. Chim. Acta86, 559-562.

2. Wenger, D. A., Tarby, T. J. & Wharton, C. (1978) Biochem.Biophys. Res. Commun. 82, 589-595.

3. Spranger, J., Wiedemann, H. R., Tolksdorf, M., Graucob, E.& Caesar, R. (1%8) Z. Kinderheilkd. 103, 285-306.

4. Cantz, M., Gehler, J. & Spranger, J. (1977) Biochem. Biophys.Res. Commun. 74, 732-738.

5. Durand, P., Gatti, R., Cavalieri, S., Borrone, C., Tondeur, M.,Michalski, J.-C. & Strecker, G. (1977) Helv. Paediatr. Acta 32,391-400.

6. Goldberg, M. F., Cotlier, E., Fichenscher, L. G., Kenyon,K., Enat, R. & Borowsky, S. A. (1971) Arch. Intern. Med.128, 387-398.

7. Galjaard, H., Hoogeveen, A., Keijzer, W., de Wit-Verbeek,H. A., Reuser, A. J. J., Ho, M. W. & Robinson, D. (1975)Nature (London) 257, 60-62.

8. Lowden, J. A. & O'Brien, J. S. (1979) Am. J. Hum. Genet. 31,1-18.

9. Andria, G., Del Giudice, E. & Reuser, A. J. J. (1978) Clin.Genet. 14, 16-23.

10. Miyatake, T., Yamada, T., Suzuki, M., Pallmann, B.,Sandhoff, K., Ariga, T. & Atsumi, T. (1979) FEBS Lett. 97,257-259.

11. Thomas, G. H., Tipton, R. E., Ch'ien, L. T., Reynolds, L. W.& Miller, C. S. (1978) Clin. Genet. 13, 369-379.

12. Mueller, 0. T. & Wenger, D. A. (1981) Clin. Chim. Acta 109,313-324.

13. Thomas, G. H., Goldberg, M. F., Miller, C. S. & Reynolds,L. W. (1979) Clin. Genet. 16, 323-330.

14. Suzuki, Y., Sakuraba, H., Potier, M., Akagi, M., Sakai, M. &Beppu, H. (1981) Hum. Genet. 58, 387-389.

15. Maire, I. & Nivelon-Chevallier, A. (1981) J. Inherited Metab.Dis. 4, 221-223.

16. Loonen, M. C. B., Reuser, A. J. J., Visser, P. & Arts, W. F.M. (1984) Clin. Genet. 26, 139-149.

17. Hoogeveen, A. T., Verheijen, F. W., d'Azzo, A. & Galjaard,H. (1980) Nature (London) 285, 500-502.

18. Swallow, D. M., Hoogeveen, A. T., Verheijen, F. W. &Galjaard, H. (1981) Ann. Hum. Genet. 45, 105-112.

19. Mueller, 0. T. & Shows, T. B. (1982) Hum. Genet. 60, 158-162.20. Van Diggelen, 0. P., Schram, A. W., Sinnott, M. L., Smith,

P. L., Robinson, D. & Galjaard, H. (1981) Biochem. J. 200,143-151.

21. Van Diggelen, 0. P., Hoogeveen, A. T., Smith, P. J., Reuser,A. J. J. & Galjaard, H. (1982) Biochim. Biophys. Acta 703,69-76.

22. Suzuki, Y., Sakuraba, H., Hayashi, K., Suzuki, K. & Imahori,K. (1981) J. Biochem. (Tokyo) 90, 271-273.

23. d'Azzo, A., Hoogeveen, A., Reuser, A. J. J., Robinson, D. &Galjaard, H. (1982) Proc. Natl. Acad. Sci. USA 79, 4535-4539.

24. Strisciuglio, P., Creek, K. E. & Sly, W. S. (1984) Pediatr. Res.18, 167-171.

25. Hoogeveen, A., d'Azzo, A., Brossmer, R. & Galjaard, H.(1981) Biochem. Biophys. Res. Commun. 103, 292-300.

26. Mueller, 0. T., Henry, W. M., Haley, L. L., Byers, M. G.,Eddy, R. E. & Shows, T. B. (1984) Am. J. Hum. Genet. 36,205S (abstr.).

27. Shows, T. B., Eddy, R., Haley, L., Byers, M., Henry, M.,Fujita, T., Matsui, H. & Taniguchi, T. (1984) Somatic CellMol. Gen. 10, 315-318.

28. Shows, T. B., Brown, J. A., Haley, L. L., Byers, M. G.,Eddy, R. L., Cooper, E. S. & Goggin, A. P. (1978) Cytogenet.Cell Genet. 21, 99-104.

29. Warner, T. G. & O'Brien, J. S. (1979) Biochemistry 18,2783-2787.

30. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall,R. J. (1951) J. Biol. Chem. 193, 265-275.

31. Norwood, T. H., Ziegler, C. J. & Martin, G. M. (1976) So-matic Cell Genet. 2, 263-270.

32. Shows, T. B. (1974) Cytogenet. Cell Genet. 13, 143-145.33. Klobutcher, L. A., Nichols, E. A., Kucherlapati, R. S. &

Ruddle, F. H. (1976) in Human Gene Mapping 3: Birth De-fects, Original Article Series, ed. Bergsma, D. (Karger, Basel,Switzerland), Vol. 12, pp. 171-174.

34. Oohira, T., Nagata, N., Akaboshi, I., Matsuda, I. & Naito, S.(1985) Hum. Genet. 70, 341-343.

35. Sips, H. J., de Wit-Verbeek, H. A., Westerveld, A. &Galjaard, H. (1985) Hum. Genet. 69, 340-344.

36. Champion, M. J. & Shows, T. B. (1977) Nature (London) 270,64-66.

Genetics: MueUer et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 10

, 202

0