Proc. Natl. Acad. Sci. USA Vol. 83, pp. 1817-1821, March 1986 Genetics and : Chromosomal assignment of two genes associated with -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. SHOWS Department 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 and galactosialidosis heterozygotes have not consistently shown galactosialidosis are the result of deficiencies of - a reduction in either ,B-galactosidase or a-neuraminidase (2, specific a-neuraminidase (acylneuraminyl hydrolase, EC 9, 13-16). Complementation studies from a number of labo- 3.2.1.18; sialidase) activity. Two genes were determined to be ratories show that the primary defect is different from that necessary for expression of neuraminidase by using human- causing single deficiencies of P-galactosidase (GM1- mouse somatic cell hybrids segregating human chromosomes. ) or neuraminidase (7, 17-19). The residual A panel of mouse RAG-human hybrid cells demonstrated a 3-galactosidase activity in galactosialidosis fibroblasts exists single-gene requirement for human neuraminidase and allowed entirely as a monomer with an absence of a large multimer assignment of this gene to the (pter-_q23) region of chromo- form found in normal cells (20). The mutation causing some 10. A second panel of mouse thymidine kinase (TK)- galactosialidosis also causes a decrease in turnover time due deficient LM/TK-human hybrid cells demonstrated that to an increased susceptibility of,-galactosidase to proteo- human neuraminidase activity required both chromosomes 10 lytic degradation (20, 21). The P-galactosidase deficiency, but and 20 to be present. Analysis of human neuraminidase not the neuraminidase deficiency, can be corrected by the expression in interspecific hybrid cells or polykaryocytes addition of protease inhibitors (21-24). This could be accom- formed from fusion of mouse RAG (hypoxanthine/guanine plished also with the addition of a "corrective factor" phosphoribosyltransferase deficient) or LM/TK- cell lines concentrated from cell culture medium, which has charac- with human sialidosis or galactosialidosis fibroblasts indicated teristics of a glycoprotein (23-25). Recent evidence suggests that the RAG cell line complemented the galactosialidosis that the mutation causing galactosialidosis results in an defect, but the LM/TK- cell line did not. This eliminates the absence of a 32-kDa glycoprotein in anti-p-galactosidase requirement for this gene in RAG-human hybrid cells and immunoprecipitates, and this protein was postulated to be the explains the different chromosome requirements of these two primary defect in galactosialidosis (23). hybrid panels. Fusion of LM/TK- cell hybrids lacking chro- We present evidence for the chromosomal assignment of mosome 10 or 20 (phenotype 10+,20- and 10-,20+) and two genes required for the expression of human neuramini- neuraminidase-deficient fibroblasts confirmed by complemen- dase in somatic cell hybrids. In addition, our evidence tation analysis that the sialidosis disorder results from a indicates that the sialidosis and galactosialidosis diseases are mutation on chromosome 10, presumably encoding the caused by mutations in these genes located on chromosomes neuraminidase structural gene. Galactosialidosis is caused by a 10 and 20, respectively. A portion of this work has been mutation in a second gene required for neuraminidase expres- presented in abstract form (26). sion located on chromosome 20. MATERIALS AND METHODS Several inherited diseases have been found to be associated with a deficiency of glycoprotein-specific N-acetyl-a- Fibroblasts and Hybrid Cells. The fibroblasts used were neuraminidase activity (acylneuraminyl hydrolase, EC GM1718, derived from an infantile-onset variant of sialidosis 3.2.1.18; sialidase). These disorders are typically classified as type II ( I), and GM806, derived from a subject the sialidoses, which have only a neuraminidase deficiency, with an early-onset form of galactosialidosis. Hybrid cells and the galactosialidoses, which have a coexistent deficiency were made by fusing normal human cells or cells with of P-galactosidase (1, 2). The sialidosis disorder, originally balanced chromosomal translocations with either mouse termed lipomucopolysaccharidosis (3), includes several var- LM/TK- [thymidine kinase (TK) deficient] or mouse RAG iants with different degrees of clinical severity, including an (hypoxanthine/guanine phosphoribosyl transferase defi- adult-onset form known as sialidosis type I and the infantile- cient) cells and selecting clones in hypoxanthine/aminopter- onset variant known as mucolipidosis I or sialidosis type II (4, in/thymidine selection medium as described (27). The pres- 5). The galactosialidosis disorder, which has also been ence of human chromosomes was determined in each hybrid termed the Goldberg Syndrome (6), GM1 gangliosidosis type clone by scoring for previously mapped human enzyme 4 (7), the cherry-red-spot- syndrome with demen- markers and by karyotype analysis with trypsin/Giemsa tia (2), and the juvenile-onset form of sialidosis type II (8), is banding (28). also clinically heterogeneous (9, 10). Neuraminidase Assays. Hybrids used for chromosomal The primary defect in the sialidoses is thought to be a localization studies were analyzed for human neuraminidase mutation in the neuraminidase structural gene. Obligate expression as freshly harvested cultures due to the lability of heterozygotes show a gene-dosage effect and have approxi- this enzyme to freezing. Hybrid harvests were scored for mately halfthe normal neuraminidase levels (11, 12). Obligate neuraminidase and analyzed karyotypically on the same passage. Hybrid cultures were rinsed in Dulbecco's phos- The publication costs of this article were defrayed in part by page charge phate-buffered saline and scraped from flasks with a rubber payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: TK, thymidine kinase.

1817 Downloaded by guest on September 30, 2021 1818 Genetics: Mueller et al. Proc. Natl. Acad. Sci. USA 83 (1986)

policeman. Cell pellets were homogenized gently in cold Table 2. Human-mouse hybrid cell neuraminidase activity distilled water by using a Dounce type homogenizer with a Teflon pestle. Neuraminidase activity was determined by a RAG Neuraminidase LM/TK- Neuraminidase procedure modified from that of Warner and O'Brien (29). hybrids Activity Score hybrids Activity Score Homogenates (20 ,ul) were incubated with 5 .ul of 1 M sodium ATR-22 6.0 - ICL-6 10.8 + acetate (pH 4.2) and 5 ALI of 7.1 mM 4-methylumbelliferyl-a- DUA-1 CH 2.0 - NSL-16 5.1 - neuraminide (Koch-Light Laboratories, Colnbrook, U.K.) at DUA-5 BA 4.0 - TSL-2 10.3 + 370C for 30 min. Reactions were terminated by adding 2 ml of DUM-13 29.8 + TSL-2 CF 3.7 - 0.85 M glycine (pH 10), and the fluorescence was determined DUM-23 36.3 + TSL-6F 6.0 - with an Aminco fluorimeter. The inclusion of bovine serum JSR-2 6.7 - VTL-6 10.3 + albumin at 1 mg/ml or 1 mM CaCl2 in assays or the JSR-6C 24.5 + VTL-11 4.1 - preparation of homogenates in the presence of 1 mM JSR-6D 21.2 + VTL-13 4.4 - phenylmethylsulfonyl fluoride had no effect on activity. JWR-22H 57.3 + VTL-14 4.5 - Protein concentration was measured by using a modification JWR-26C 25.3 + VTL-15 5.8 - of the Folin phenol procedure (30). RAS-M3 26.0 + VTL-16 10.6 + Complementation Analysis. Each of the two cultures used REW-4 8.0 - VTL-17 4.3 - in complementation experiments was seeded at half the REW-8I C4 5.0 - VTL-18 4.4 - confluent cell density, and the mixture was cultivated over- REW-11 5.5 - VTL-19 3.2 - night. These cultures were fused by using 42% (wt/vol) REX-11 BF 20.5 + VTL-21 4.2 - polyethylene glycol 1000 (Koch-Light Laboratories) contain- REW-13 57.0 + VTL-22 5.0 - ing 7% (vol/vol) dimethyl sulfoxide (31), and the resulting REW-15 34.8 + VTL-23 3.4 - polykaryocytes were enriched by sedimentation velocity REX-26 6.3 - WIL-2 3.5 - under sterile conditions with a Sta-Put apparatus (Johns REX-33 5.0 - WIL-2 C 3.9 - Scientific, Toronto) as described (19). This polykaryocyte REX-57 BB 5.0 - WIL-6 10.4 + fraction was cultured for 10-14 days, and the neuraminidase SIR-1 4.0 - WIL-8 12.9 + activity was determined along with cultures of each parental XER-8 21.3 + WIL-8S 9.6 + cell type and a cocultivated mixture. In complemnentation XER-11 27.6 + WIL-8Y 11.9 + experiments involving human fibroblasts and RAG or XER-15 21.3 + WIL-li 14.5 + LM/TK- cells, the polykaryocytes were cultivated in XTR-1 27.0 + WIL-13 4.0 - hypoxanthine/aminopterin/thymidine selection medium in XTR-1 BD 35.7 + WIL-14 5.7 - order to inhibit growth of the unfused mouse parental cells. XTR-3 BH 20.8 + WIL-14 C 6.1 - A significant increase in the neuraminidase activity of the XTR-22 26.3 + polykaryocyte fraction over that of the cocultivated mixture The neuraminidase activity of hybrid cells is the mean of at least and parental cell activity range, as determined with Student's three separate harvests and is expressed in nmol/hr per mg of t test (at the 1% confidence level), was assumed to indicate homogenate protein. Neuraminidase scores are based on statistically complementation. significant increases over the activity of parental mouse cells (Table 1). Hybrid clone XTR-3 BH contains del(10)(q23--qter) (deletion in RESULTS chromosome 10 in region q23--qter). Somatic Cell Hybrid Mapping of Neuraminidase. The pres- ence of human neuraminidase activity in human-mouse mg for LM/TK- hybrids) were scored negative for human somatic cell hybrids that segregate human chromosomes activity. Hybrid cells with neuraminidase activity well above could be determined on the basis of a quantitative assay. this range (>20 nmol/hr per mg for human-RAG hybrids and Mouse RAG and LM/TK- parental cells, which were used to >9 nmol/hr per mg for human-LM/TK- hybrids) were make somatic cell hybrids, had low levels of neuraminidase scored as positive (Table 2). Hybrid cells with activity levels activity (<8% of that of normal human cells) under the assay between these upper and lower thresholds were not consid- conditions used (Table 1). Therefore, the human activity in a ered in discordance calculations. These hybrids were elimi- hybrid cell could be quantitated by the amount of activity in nated because of the variabilities associated with the excess of the mouse cell activity levels. Hybrid cells were nonquantitative scoring for human enzyme markers present found to have a large variation in neuraminidase activity at low levels in hybrid cells. The presence of human because of the different numbers of the chromosome(s) neuraminidase in hybrid cells was also confirmed qualita- encoding neuraminidase genes present in each hybrid (Table tively by mobility differences between human and mouse 2). Hybrid cells with activity close to that of mouse parental enzymes following cellulose acetate electrophoresis (unpub- cells (within one standard deviation of the mean: 4.1 ± 3.1 lished data). nmol/hr per mg for RAG hybrids and 4.9 ± 1.5 nmol/hr per In each hybrid the human chromosome content was determined karyotypically and by scoring for the presence of previously mapped human enzyme markers. The correlation Table 1. Neuraminidase activity of human fibroblasts and between the presence of human neuraminidase and each mouse cells human chromosome is shown in Tables 2 and 3, expressed as 4-Methylumbelliferyl- the discordance frequency. Instances where chromosomes neuraminidase activity were present in <15% of the 30 metaphase spreads analyzed Cell cultures n Mean SD Range for each hybrid were not considered in calculating discord- ance percentages in Table 3. The findings from a panel of 28 Human fibroblasts RAG-human hybrids indicated that the expression of Normal 60 66.3 45.5 17.6-189 neuraminidase was completely concordant only with chro- Sialidosis 9 1.6 1.7 0.0- 5.3 mosome 10 and the enzyme markers glutamic-oxaloacetic Galactosialidosis 15 7.7 4.5 0.8- 17.0 transaminase 1 (GOTI) and adenosine kinase (ADK), which Mouse cells had been mapped previously to chromosome 10 (32, 33). For LM/TK- 22 4.9 1.5 2.6- 8.7 all other chromosomes, the expression of neuraminidase and RAG 22 4.1 3.1 0.0- 10.3 the chromosome was discordant in four or more hybrids Downloaded by guest on September 30, 2021 Genetics: MueUer et al. Proc. Natl. Acad. Sci. USA 83 (1986) 1819

Table 3. Segregation of human neuraminidase gene in human-mouse somatic cell hybrids Chromosome 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X RAG hybrids scored Neur (+)/chrom (+) 8 5 13 9 9 13 10 7 3 16 10 10 4 11 7 8 12 12 9 11 12 4 10 Neur(+)/chrom(-) 4 9 3 8 8 4 3 11 13 0 4 7 14 6 8 10 5 4 8 6 4 13 2 Neur (-)/chrom(+) 0 1 2 3 1 1 4 3 0 0 3 1 5 4 0 2 3 1 2 0 4 3 5 Neur(-)/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 scored Neur (+)/chrom(+) 3 2 4 5 5 5 5 4 0 8 6 3 3 5 3 0 8 5 3 9 6 0 5 Neur (+)/chrom(-) 6 5 4 4 4 2 4 5 9 0 3 6 6 4 6 9 0 3 5 0 3 8 3 Neur(-)/chrom(+) 1 3 6 4 7 3 3 10 0 6 5 8 4 5 8 3 15 5 4 0 9 2 4 Neur(-)/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 27 The 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 the activity thresholds as discussed in the text.

(Table 3). This indicates that a gene necessary for neur- demonstrate any increases in neuraminidase activity (Table aminidase expression is encoded on human chromosome 10. 5), indicating that LM/TK- cells, unlike RAG cells, lack the Further, the presence of neuraminidase activity in hybrid gene or protein that complements the galactosialidosis de- XTR-3-BsAg H with a deletion in chromosome 10 in the fect. Since LM/TK- cells apparently complement neither the q23-*qter region indicated that this gene is located in the sialidosis nor the galactosialidosis mutations, there is a (pter-*q23) region of chromosome 10. A second hybrid panel requirement for both of these genes before human consisting of 27 human-LM/TK- somatic cell hybrids indi- neuraminidase can be expressed in LM/TK- hybrid cells cated that human neuraminidase expression was apparently constructed from normal human cells (Table 3). All concordant with the presence of human chromosome 20 LM/TK--human hybrid cells in Table 3 that are scored (Table 3). For all other chromosomes, the discordance positive for neuraminidase have both chromosomes 10 and frequency was 21% or more, including six hybrids that 20. All six hybrid clones in which the expression of demonstrated discordance with chromosome 10(24% discord- neuraminidase and chromosome 10 was discordant contained ance, explained below). This suggests that another gene chromosome 10 but had no detectable human neuraminidase controlling the expression of neuraminidase is encoded on activity. All of these hybrid clones lack human chromosome human chromosome 20. 20, explaining the absence of human neuraminidase activity. Expression of Neuraminidase Deficiencies in Somatic Cell Chromosomal Assignment ofSialidosis and Galactosialidosis Hybrids. To explain this apparent discrepancy in the mapping Mutations. Complementation analysis also was used to de- results from these two hybrid panels, we constructed hybrid termine whether either of the genes on chromosomes 10 and cells using sialidosis and galactosialidosis fibroblasts as the 20 was mutated in the sialidosis and galactosialidosis disor- human parental cells and tested for human neuraminidase ders. Two hybrid clones from the LM/TK--human hybrid expression. Sialidosis and galactosialidosis fibroblasts are panel were identified that have the chromosomal phenotypes known to be caused by mutations in different genes, since neuraminidase is complemented in polykaryocytes formed 10+,20- and 10-,20+. These hybrids were used in comple- from a sialidosis-galactosialidosis fibroblast fusion experi- ment (17-19). Hybrid cells formed from fusion of mouse RAG Table 4. Neuraminidase expression in human-mouse hybrid cells with galactosialidosis fibroblasts (GAR hybrids) have signif- made with human neuraminidase-deficient cells icant levels of neuraminidase activity in those hybrids re- taining chromosome 10 (Table 4). The activity in each GAR Hybrid Neuraminidase hybrid varied somewhat according to the numbers of chro- clone Activity Score Chromosome 10 Chromosome 20 mosome 10 present, but chromosome 20 was not necessary Galactosialidosis fibroblast-RAG for expression. This indicates that the neuraminidase gene on GAR-2 29.8 (+) + chromosome 10 is functionally normal in this galactosialido- GAR-3 45.0 (+) + + sis subject and also that RAG cells have a gene or protein GAR-9 27.1 (+) + + similar to that defective in this disorder. Therefore, in GAR-12 15.0 (w) + human-RAG hybrid cells made with normal human cells, GAR-15 17.5 (+/w) + + human neuraminidase can be expressed in the absence ofthis GAR-16 18.7 (+/w) + gene, and mapping analysis detected only a single require- ment for chromosome 10 (Table 3). Sialidosis fibroblast-LM/TK- Hybrid cells formed from the fusion of sialidosis fibro- SIL-1 0.9 (-) + blasts-LM/TK- cells have no detectable human neuramin- SIL-2 1.2 (-) - idase activity, whether human chromosomes 10, 20, or both SIL-3 1.0 (-) + 10 and 20 were present (Table 4). Further, no increases in SIL-5 2.6 (-) + + neuraminidase activity could be detected in enriched SIL-6 3.6 (-) + + polykaryocytes formed from fusion of these cells (Table 5). Neuraminidase activity is the average of three separate harvests Hybrid cells from the fusion of LM/TK--galactosialidosis and is expressed in nmol/hr per mg of homogenate protein. Scoring fibroblasts were not successfully constructed. However, for human neuraminidase and chromosomes (+, w for weak, and -) complementation analysis using enriched polykaryocytes was determined according to the thresholds established for Tables 2 from fusion of LM/TK--galactosialidosis fibroblasts did not and 3. Downloaded by guest on September 30, 2021 1820 Genetics: MueUer et al. 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 cells Sialidosis FB-mouse LM/TK- 7.5 5.8 Galactosialidosis FB-mouse LM/TK- 4.5 4.5 Human-hybrid cells 10+,20- hybrid only (n = 7) 2.3- 6.1 10-,20+ hybrid only (n = 5) 7.3-11.7 Sialidosis FB x 10+,20- hybrid 0.1 10.9 + Sialidosis FB x 10-,20+ hybrid 3.1 Galactosialidosis FB x 10+,20- hybrid 2.8 4.5 Galactosialidosis 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 of the fused cell cultures compared with the range of appropriate parental cell activities and of cocultivated mixtures using Student's t test. FB, fibroblasts.

mentation experiments with sialidosis and galactosialidosis ment located on chromosome 20, this chromosome is not fibroblasts (Table 5). The 10-,20+ hybrid has weak levels of sufficient for neuraminidase expression, as demonstrated human neuraminidase activity and was not included in the with hybrid cells made by fusion of sialidosis and LM/TK- initial assignment studies (Tables 2 and 3), according to the cells (Table 4). Complementation analysis findings indicated established criteria discussed above. This hybrid clone may that sialidosis fibroblasts have a mutation in the gene on contain a low amount of chromosome 10, which was not chromosome 10 and that the gene on chromosome 20 is detected by enzyme marker measurement. It was neverthe- apparently normal (Table 5). Sialidosis-LM/TK- hybrid less useful for these complementation studies because scor- cells retaining chromosome 20 do not have any detectable ing was based on activity increases over that measured in human neuraminidase activity, indicating that a functional cocultivated mixtures of parental cells. Both of these hybrid chromosome 10 gene is also required for neuraminidase clones were made with normal human cells, therefore activity. The different contributions of the two parental complementation in these fusions is expected if the hybrid mouse cell lines are most likely a result of the accumulation cell supplies the chromosome containing the mutated gene in of different mutations during mutagenesis and selection for the neuraminidase-deficient (sialidosis or galactosialidosis) hypoxanthine/guanine phosphoribosyl transferase-deficient fibroblasts. The 10+,20- hybrid-sialidosis fibroblast fusion or TK- phenotypes. The two-chromosome requirement ob- resulted in a significant increase in neuraminidase activity, served in human-LM/TK- cell hybrids also suggests a suggesting that sialidosis fibroblasts have a functionally reason for the relatively low neuraminidase activity observed normal chromosome 20. The 10-,20+ hybrid-sialidosis fibro- in LM/TK- hybrid cells scored as positive, relative to that of blast fusion resulted in no change in neuraminidase activity RAG-human hybrid cells scored positive (Table 2). Human over control cultures, confirming that the gene on chromo- chromosomes in hybrid cells are rarely present at their some 10 is mutated in sialidosis. In the two experiments diploid number (60 homologues out of the 30 metaphase involving galactosialidosis fibroblasts, the fusion with the spreads that were analyzed), and hybrid cells with as few as (10-,20+) hybrid resulted in a significant increase in activity, 15% of possible homologues (9/60) were scored as positive. indicating that galactosialidosis fibroblasts have a normal Therefore, the low levels of human neuraminidase activity in gene on chromosome 10. The fusion between the galacto- LM/TK- hybrid cells with both chromosomes scored posi- sialidosis fibroblast and the 10+,20- hybrid resulted in no tive reflect the low numbers of hybrid cells containing complementation, confirming that chromosome 20 is mutated sufficient numbers of both chromosomes. in this disorder. Therefore, the inferred chromosomal phe- We also have shown that the deficiency of neuraminidase notype for sialidosis is 10m,20+ and for galactosialidosis is activity in two subjects with sialidosis and galactosialidosis is 10+,20m. associated with a loss in function of the genes on chromo- somes 10 and 20, respectively. It has not been determined whether this can be generalized for all subjects with these DISCUSSION diseases, although no genetic heterogeneity has thus far been Genetic complementation analysis of human diseases asso- identified within sialidosis or galactosialidosis (17, 19). The ciated with a deficiency of glycoprotein-specific neuramini- indirect evidence from the literature (i.e., a gene-dosage dase activity indicated that there are apparently two distinct effect in obligate heterozygotes and kinetic alterations in the genetic variants (17-19). We have determined that there are enzyme ofhomozygotes) suggests that the sialidosis disorder two genes required for the expression of this enzyme in is caused by a mutation in the neuraminidase structural gene. somatic cell hybrids located on chromosomes 10 and 20. We presume, therefore, that the gene on chromosome 10 Although one of the two hybrid cell panels used demonstrat- encodes the structural enzyme protein. Oohira et al. (34) ed a single-gene requirement for chromosome 10, it was reported that there was a suggestion of linkage between a shown by complementation analysis that the mouse RAG locus causing a case of sialidosis II (single neuraminidase genome substitutes or complements one ofthe two necessary deficiency) with the HLA complex located on chromosome 6. genes (the gene on chromosome 20). In a second hybrid clone Their analysis, however, was based on only seven informa- panel made with mouse LM/TK- cells, the data indicate that tive family members, and this may not be sufficient data to there was no apparent contribution by the mouse genome, establish linkage. and neuraminidase expression was shown to require the Our data also indicate that the galactosialidosis disorder is presence of both chromosomes 10 and 20. Although these caused by a loss in function of the gene located on chromo- data also can be interpreted to indicate a single-gene require- some 20. Evidence from complementation analysis and from Downloaded by guest on September 30, 2021 Genetics: MueUer et al. Proc. Nati. Acad. Sci. USA 83 (1986) 1821

the lack of a demonstrated gene dosage effect in tissues of 10. Miyatake, T., Yamada, T., Suzuki, M., Pallmann, B., obligate heterozygotes suggests that the defect is not in the Sandhoff, K., Ariga, T. & Atsumi, T. (1979) FEBS Lett. 97, structural gene of either neuraminidase or 13-galactosidase. 257-259. 11. Thomas, G. H., Tipton, R. E., Ch'ien, L. T., Reynolds, L. W. d'Azzo et al. (23) have proposed that the primary defect in & Miller, C. S. (1978) Clin. Genet. 13, 369-379. galactosialidosis is an absence of a 32-kDa "protective 12. Mueller, 0. T. & Wenger, D. A. (1981) Clin. Chim. Acta 109, protein," which forms part of the P-galactosidase and 313-324. neuraminidase enzyme complex. This protein was found to 13. Thomas, G. H., Goldberg, M. F., Miller, C. S. & Reynolds, be encoded on chromosome 22 (35). The function of the L. W. (1979) Clin. Genet. 16, 323-330. chromosome 20 gene and its relationship to the protective 14. Suzuki, Y., Sakuraba, H., Potier, M., Akagi, M., Sakai, M. & protein remains to be determined. Beppu, H. (1981) Hum. Genet. 58, 387-389. The galactosialidosis mutation is complemented or cor- 15. Maire, I. & Nivelon-Chevallier, A. (1981) J. Inherited Metab. rected by one of the two mouse cell lines (RAG) that were Dis. 4, 221-223. 16. Loonen, M. C. B., Reuser, A. J. J., Visser, P. & Arts, W. F. used to make cell hybrids. The cocultivated mixture of RAG M. (1984) Clin. Genet. 26, 139-149. and galactosialidosis cells also had a slight increase in activity 17. Hoogeveen, A. T., Verheijen, F. W., d'Azzo, A. & Galjaard, over that expected by averaging the parental cells' activities. H. (1980) Nature (London) 285, 500-502. This could be due to diffusion of the postulated corrective 18. Swallow, D. M., Hoogeveen, A. T., Verheijen, F. W. & factor or protective protein from the hybrid cells, resulting in Galjaard, H. (1981) Ann. Hum. Genet. 45, 105-112. partial correction of the activity of galactosialidosis fibro- 19. Mueller, 0. T. & Shows, T. B. (1982) Hum. Genet. 60, 158-162. blasts as noted (23-25). Most human structural gene muta- 20. Van Diggelen, 0. P., Schram, A. W., Sinnott, M. L., Smith, tions are not corrected by the mouse genome, although P. L., Robinson, D. & Galjaard, H. (1981) Biochem. J. 200, correction of the lysosomal enzyme deficiencies in I-cell 143-151. 21. Van Diggelen, 0. P., Hoogeveen, A. T., Smith, P. J., Reuser, disease in man-mouse hybrid cells was demonstrated (36). A. J. J. & Galjaard, H. (1982) Biochim. Biophys. Acta 703, This disease is caused by a deficiency of a Golgi enzyme, 69-76. N-acetylglucosaminyl phosphotransferase, involved in intra- 22. Suzuki, Y., Sakuraba, H., Hayashi, K., Suzuki, K. & Imahori, cellular lysosomal enzyme transport. This suggests that the K. (1981) J. Biochem. (Tokyo) 90, 271-273. defective gene in galactosialidosis also may be involved in 23. d'Azzo, A., Hoogeveen, A., Reuser, A. J. J., Robinson, D. & posttranslational processing or regulation of the affected Galjaard, H. (1982) Proc. Natl. Acad. Sci. USA 79, 4535-4539. enzymes. This postulated regulatory gene would function to 24. Strisciuglio, P., Creek, K. E. & Sly, W. S. (1984) Pediatr. Res. modulate or coordinate the activity of P-galactosidase and 18, 167-171. neuraminidase and would have a key role in the initial steps 25. Hoogeveen, A., d'Azzo, A., Brossmer, R. & Galjaard, H. (1981) Biochem. Biophys. Res. Commun. 103, 292-300. of glycoprotein catabolism. 26. Mueller, 0. T., Henry, W. M., Haley, L. L., Byers, M. G., This work was supported in part by grants HD 051% and GM 20454 Eddy, R. E. & Shows, T. B. (1984) Am. J. Hum. Genet. 36, from the National Institutes of Health and 1-935 from the March of 205S (abstr.). Dimes. 27. Shows, T. B., Eddy, R., Haley, L., Byers, M., Henry, M., Fujita, T., Matsui, H. & Taniguchi, T. (1984) Somatic Cell 1. Okada, S., Kato, T., Miura, S., Yabuuchi, H., Nishigaki, M., Mol. Gen. 10, 315-318. Kobata, A., Chiyo, H. & Furuyama, J. (1978) Clin. Chim. Acta 28. Shows, T. B., Brown, J. A., Haley, L. L., Byers, M. G., 86, 559-562. Eddy, R. L., Cooper, E. S. & Goggin, A. P. (1978) Cytogenet. 2. Wenger, D. A., Tarby, T. J. & Wharton, C. (1978) Biochem. Cell Genet. 21, 99-104. Biophys. Res. Commun. 82, 589-595. 29. Warner, T. G. & O'Brien, J. S. (1979) Biochemistry 18, 3. Spranger, J., Wiedemann, H. R., Tolksdorf, M., Graucob, E. 2783-2787. & Caesar, R. (1%8) Z. Kinderheilkd. 103, 285-306. 30. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, 4. Cantz, M., Gehler, J. & Spranger, J. (1977) Biochem. Biophys. R. J. (1951) J. Biol. Chem. 193, 265-275. Res. Commun. 74, 732-738. 31. Norwood, T. H., Ziegler, C. J. & Martin, G. M. (1976) So- 5. Durand, P., Gatti, R., Cavalieri, S., Borrone, C., Tondeur, M., matic Genet. 263-270. Michalski, J.-C. & Strecker, G. (1977) Helv. Paediatr. Acta 32, Cell 2, 391-400. 32. Shows, T. B. (1974) Cytogenet. Cell Genet. 13, 143-145. 6. Goldberg, M. F., Cotlier, E., Fichenscher, L. G., Kenyon, 33. Klobutcher, L. A., Nichols, E. A., Kucherlapati, R. S. & K., Enat, R. & Borowsky, S. A. (1971) Arch. Intern. Med. Ruddle, F. H. (1976) in Human Gene Mapping 3: Birth De- 128, 387-398. fects, Original Article Series, ed. Bergsma, D. (Karger, Basel, 7. Galjaard, H., Hoogeveen, A., Keijzer, W., de Wit-Verbeek, Switzerland), Vol. 12, pp. 171-174. H. A., Reuser, A. J. J., Ho, M. W. & Robinson, D. (1975) 34. Oohira, T., Nagata, N., Akaboshi, I., Matsuda, I. & Naito, S. Nature (London) 257, 60-62. (1985) Hum. Genet. 70, 341-343. 8. Lowden, J. A. & O'Brien, J. S. (1979) Am. J. Hum. Genet. 31, 35. Sips, H. J., de Wit-Verbeek, H. A., Westerveld, A. & 1-18. Galjaard, H. (1985) Hum. Genet. 69, 340-344. 9. Andria, G., Del Giudice, E. & Reuser, A. J. J. (1978) Clin. 36. Champion, M. J. & Shows, T. B. (1977) Nature (London) 270, Genet. 14, 16-23. 64-66. Downloaded by guest on September 30, 2021