(Ile)~~~-Ile(Leu, Val, Ala) - The Journal of Biological Chemistry

Tm JOURNALOF B I O ~ I C CHEMISTRY AL 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 9, Issue of March 4, pp. 6622-6631, 1994 Printed in U.S.A.

The Residues Le~(Ile)~~~-Ile(Leu, Val, Ala)476,Contained in the for Targeting of Extended Carboxyl Cytoplasmic Tail, Are Critical the Resident Lysosomal MembraneProtein LIMP I1 to Lysosomes* (Received for publication, August 9, 1993, and in revised form, October 7, 1993)

Ignacio V. Sandoval+& JuanJ. ArredondoS, Jose Alcalde+$ Alfonso Gonzalez Noriegall, Joel Vandekerckhove**,Maria A. Jimenezll, and Manuel RicoSS From the SCentro de Biologia Molecular “Severo Ochoa,” Facultad de Ciencias, Consejo Superior de Znuestigaciones Cientificas, Universidad Autonoma de Madrid, Cantoblanco, 28049 Madrid, Spain, the **Laboratory of Physiological Chemistry, Rijksuniversiteit Gent, Ledeganckstraat 35, B-9000 Gent, Belgium, the S$Znstituto de Estructura de la Materia, Consejo Superior de Znvestigaciones Cientificas, Serrano 119, 28006 Madrid, Spain, and the lhstituto de Znvestigaciones BiomPdicas, Unidad de Gene‘tica de la Nutricion, Universidad Nacional Autonoma de Me‘xico, MPxico, D.F: 04530, MPxico

LIMP11, a type I1 lysosomal integral membrane pro- Kornfeld, 1987; Kornfeld and Mellman, 1989; Fukuda, 1991). tein, and the CD36hIMP I1 construct are targeted to Soluble hydrolases are transported to lysosomes by a mechalysosomes by means of a signal expressed in the tyro- nism dependenton mannose 6-phosphate (Kaplanet al., 19771, sine-lackingcarboxyl cytoplasmictail of LIMP I1 (Vega, a signal which, acquired in the early Golgi (Pohlman et al., M. A, Rodriguez, F., Segui, B., Cales, C., Alcalde, J., and 1982; Deutscher et al., 1983; Minnifield et al., 19861, is recogSandoval, I. V. (1991) J. Biol. Chem. 266,16269-16272; nized by specific receptors (Sahagianand Neufeld,1983; Vega, M. A, Segui-Real,B., Garcia, J. A., Cales, C., Rod- Hoflack and Kornfeld, 1985a, 1985b)that transport thehydroriguez, F., Vandekerckhove, J., and Sandoval,I. V. (1991) lases from the trans-Golgi to lateendosomes (i.e. prelysosomal J. Biol. Chen. 266, 16818-16824). Substitution of Leu475 compartment; Griffiths et al., 1988; Vladutiu, 1983; Fedde and with Ile resulted in a decreased efficiencyof targeting. Sly, 1985; Campbell and Rome, 1983; Geuze et al., 1985, 1987; Mutant forms produced substituting by Leu476by hydroBrown et al., 1986; Lemansky et al., 1987; Schmid et al., 1989; phobic residues with either large (Val) or small (Ala, Gly) of lysosomes side chains, or by a charged residue (Asp), showed in- Jadot etal., 1992), theimmatureprecursors 1991). (Murphy, hibited targeting. In contrast, the contiguous Ile476 resiIn contrast, targetingof membrane proteins tolysosomes is due could be replaced by either Leu, without loss in the independent of the mannose 6-phosphate signal (Barriocanal et efficiency of targeting, or byVal or Ala, with some impediment. Substitutionof Ile476by either Gly or Aspin- al., 1986) and appears to involve signals expressed in their hibited completely the targeting. The addition of the cytoplasmic tails (for a review see Fukuda, 1991). Transport to sequence Ser-“rp-Asp to the carboxyl end of the con- lysosomes of the membrane proteinsH-lamp1-1 ( i e . LIMP 111) struct did not interfere with targeting. Data from ‘H (Williams and Fukuda, 19901, the precursor of lysosomal acid NMR analysis of the icosapeptide correspondingto the phosphatase (Peters et al., 1990; Lehmann et al., 1992; Eberle carboxyl cytoplasmic tail of LIMP I1 indicated the pre- et al., 19911, antigen CD36 (Metzelaar et al., 1991), and lamp-2 dominance of structures with extended random coil con- (i.e. LIMP IVI2 appear toinvolve a tyrosine-based signal. Howformations, suggesting that the targeting signal is con- ever, this signal does not appear to be universal among lysotained in a domain with an extended configuration. somal membraneproteins. Recent cloning, sequence, and transport studies of LIMP I1 have revealed that this protein is targeted to lysosomes by virtue of a signal expressed in its Lysosomes are digestive organelles that resultfrom matura- tyrosine-lacking cytoplasmic tail (Vega et al., 1991a, 1991b). tion of late endosomes produced by the fusion of Golgi-derived Therefore, at least two targeting signals could operate in the primary lysosomes and plasma membrane-derived earlyendo- transport of proteins to thelysosomal membrane. somes (de Duve, 1963; Straus, 1963; Novikoff, 1963). Here, we reportthatthe 2 consecutive residuesand Newly synthesized soluble hydrolases and membrane pro- Ile476expressed in thecarboxyl cytoplasmic tail of the protein teins are transported to lysosomes through the endoplasmic are critical for targeting the CD36LIMP I1 construct to lysoreticuludGolgi pathway. In mammalian cells, we know much somes. The sequences formedby the combination of L e ~ ( I l e ) ~ ~ ~ of transport of Ile(Leu, Val, Ala)476expressed various targeting capacities. more of the targeting mechanisms and pathway soluble hydrolases to lysosomes than of those involved in the EXPERIMENTAL PROCEDURES targeting of lysosomal membrane proteins (for references see Materials

* This work was supported by grants from the Comision Interminis-

Enzymes used in molecular cloning were obtained from Boehringer terial para Ciencia y Tecnologia del Ministerio Espaiiol de Educacion y Ciencia (to I. V. S.), the EECC (to I. V. S.), the Concerted Research Mannheim, Bio-Rad or New England BioLabs. The sources of the antiActions of the Flemish Community (to J. V.), and by the Belgium Uni- bodies to CD36 (Kieffer et al., 1989),H-lamp-1 (Hughes and August, versity Research Fund (to J. V.). The costs of publication of this article were defrayedin part by the payment of page charges. This article must The abbreviationsused are: lamp, lysosome-associated protein; therefore be herebymarked “advertisement”in accordance with 18 LIMP, lysosomal integral membrane protein; NOESY, nuclear OverU.S.C.Section 1734 solelyto indicate this fact. total correlation spectroscopy; This paper is dedicated to the memory of Prof. Severo Ochoa (1905- hauser effectspectroscopy;TOCSY, 1993). 1Supported by a grant from the EECC (to I. V. S.). 5 To whom correspondence should be addressed. Tel.: 91-397-84-55; Fax: 91-397-47-99.

ROESY, rotating frame Overhauser effect spectroscopy; MPR, mannose phosphate receptor. I. V. Sandoval, J. J. Arredondo, J. Alcalde, A. G . Noriega, J. Vandekerckhove, M. A. Jimenez, and M. Rico, unpublished results.

6622

Targeting of Lysosomal Membrane Protein

6623

19821, CD8 (Malissen et al., 19821, cis-Golgi 15C8 (Yuan et al., 1987) and Mutant 27-P474D/oligo:GAAAGGGCAGACCTCATACGG/temprotein disulfide isomerase (Nietoet al., 1990) have been reported. The plate: wild. expression vector pcEXV-3 (Miller and Germain, 1986) was the generMutant 28-StopS479: adds Ser-Trp-Asp to the carboxyl terminus of ous gift of Carmela Cales (Autonoma Universityof Madrid). All of the the proteidoligo:ACGGACCTCATGGGACTAACC?T/template:wild. chemicals were of the highest quality commercially available. Mutant 29-T302L in the carboxyl cytoplasmic tail of CD36/oligo: AGATCGAAATTAATAAAATMtemplate:wild. Mutagenesis of the cytoplasmic tailof CD8 was performed with the cDNA Constructs following oligos and templates. Mutant 30-An internal Leu-Ile sequence was introduced in thecyThe construction of cDNAs encoding the extracellular and transtoplasmictail of CD8 by the S231Isubstitution,usingthe oligo/ membrane domains of human CD36 (Oquendo et al., 1990) and the carboxyl cytoplasmic tailof either LIMPI1 or CD8 (Littman et al., 1985) CCCAGCCTTATCGCGAGATAC/template:wild. Mutant 31-Three rounds of mutagenesis were performed for the called CD36LIMP I1 and CD36/CD8, respectively, has been reported introduction of the carboxyl-terminal Pro-Leu-Ile-Arg-Thr sequence in (Vega et al., 1991b). the tail ofthe CD8 protein. Thefirst round (mutant 31.1) was performed by substituting Valz2' + Leu and Ile, using the oligo: Mutagenesis of CD36ILIMP 11, CD36, and CD8 cDNAs CCCCGGCCTCTGATCAAATCGGGMtemplate: wild;inthe second Restrictionfragments encoding the completesequences of CD36/ round (mutant 31.2) the substitutions S224T and G225Stop were introduced using the oligo:CTGATCAAAACGTGAGACAAGCCC/template: LIMP 11, CD36, and CD8, and the transmembrane and luminal domains mutant 31.1; finally, mutant 31 was developed in a third round by of CD36 and the cytoplasmic tail of LIMP I1 (CD36LIMP 11) were substituting L y P 3 + Arg, using the o1igo:CGGCCTCTGATCAsubclonedinto the phage M13mp19 in a 3' to 5' direction.SingleGAACGTGAGACAAG/template: mutant 31.2. stranded uracilated templates were generated from Escherichia coli Mutant 32-E466Qloligo:GTCTACGGATCAGGGAACTGC/template: CJ236 infected with phage particles enclosing the subcloned cDNA rewild. striction fragments and were used as templates for synthesis of the Mutant 33-E47lQ/oligo:ACTGCAGATCAAAGGGCAC/template: mutagenic strands. Site-directed mutagenesis of the LIMP I1 carboxyl wild. cytoplasmic tail was performed using the following oligonucleotides Mutant 34-S463A/oligo:TCGAGGACAGGGGGCTACGGATGAG/ (oligos) and templates. template: wild. Mutant I"E471Stop deletes the 8 carboxyl-terminal amino acids/ Mutant 35-T468A/oligo:ACGGATGAGGGAGCTGCAGATGAAAGt oligo:GGGAACTGCAGA'MAAAGGGCACCCC/template: wild. Mutant 2-L475T,I476H,R477T,T478D/oligo:AGGGCACCCACTC- template: wild. After primer extension by T4 DNA polymerase the resulting DNAs ATACGGACTAATGGGACT/template: wild. E. coli strain MV1109. Phage plaques were Mutant 3-R477G/oligo:CCCCTCATAGGAACCTAATGG/template: were used to transform isolated and single-strandedDNA prepared for sequence of the DNA by mutant 6. 2.0 (U. S. Biochemical Corp.). Mutant 4-R477A/oligo:CCCCTCATAGCAACCTAATGG/template: the dideoxy methodusingSequenase Inserts containing the desired mutations were subcloned into the exmutant 6. Mutant 5-R477E/oligo:CCCCTCATAGAAACCTAATGG/template: pression vector pcEXV-3 and used for transient transfection experiments. wild. Mutant 6-R477Q/oligo:CCCCTCATACAAACCTAATGG/template: Dansfection Experiments wild. Mutant 7-R477Woligo:CCCCTCATAAAAACCTAATGG/template: COS cells were transfected by a DEAE-dextran procedure that inwild. Mutant 8-T478G/oligo:CTCATACGGGGCTAATGGGAG/template: cluded a 4-h incubation with a mixture of 20 pg of DNA and 200 pg of DEAE-dextran, prepared in Dulbecco's modified Eagle's medium, folwild. sulfoxide, preparedin Mutant 9-T478S/oligo:CTCATACGGAGCTAATGGGAG/template: lowed by a2-min shock with10%dimethyl Hanks' balanced solution (Lopata et al., 1984). wild. Mutant lO-T478V/oligo:CTCATACGGGTCTAATGGGAG/tempIate: Immunofluorescence Microscopy Experiments wild.

-

Mutant ll-T478Woligo:CTCATACGGATCTAATGGGAG/tempIate: Studies of the cellular distribution of newly synthesized CD36LIMP wild. 11, CD8, and CD36 wild and mutant species were performed by immunMutant 12-L475G/oligo:AGGGCACCCGGCATACGGACC/tem- ofluorescence microscopy as described (Barriocanalet al., 1986; Yuan et plate: wild. al., 19871, using specific antibodies to CD36 and CD8 (see "Materials"). Mutant 13-L475A/oligo:AGGGCACCCGCCATACGGACC/template: Nonpermeabilized cells were incubated for 1 h at 4 "C with the correwild. sponding antibody, diluted 1 : l O O in phosphate-buffered saline. After Mutant 14-L475V/oligo:AGGGCACCCGTCATACGGACC/template: extensive washingthe cells werefixed with 2% fresh paraformaldehyde wild. (Alcalde et al., 1992) and, following incubation for 15 min at room Mutant 15-L475Woligo:AAGGGCACCCATCATACGGAC/template: temperature with 0.1M glycine in phosphate-buffered saline, incubated wild. with fluorescein-conjugatedspecific antibodies. Cells were fixed-permeMutant 16-L475D/oligo:AGGGCACCCGACATACGGACC/template: abilized by incubation for 2 min with cold (-20 "C) methanol and prowild. cessed for single or double immunofluorescence as described before Mutant 17-1476G/oligo:GCACCCCTCGGACGGACCTAA/tempIate: (Barriocanal et al., 1986; Yuan et al., 1987). Localization of the exmutant 19. pressed proteins to the endoplasmic reticulum, Golgi complex, and lyMutant 18-1476A/oIigo:GCACCCCTCGCACGGACCTAA/template: sosomes was performed by double immunofluorescence microscopy exwild. periments usingspecific antibodies to protein disulfide isomerase, Golgi Mutant 19-1476V/oligo:GCACCCCTCGTACGGACCTAA/template: membrane protein (GMP,~,), and H-lamp-], respectively Vega (see et al., wild. 1991a, 1991b). Mutant 20-1476Woligo:GCACCCCTCCTACGGACCTAA/template: wild. ' H NMR Spectroscopy Mutant 21-1476D/oIigo:GCACCCCTCGACCGGACCTANtemplate: mutant 24. NMR samples were 10 mM peptide in H,O/D,O 9:1, pH 5.0. NMR Mutant 22-1476Eloligo:GCACCCCTCGAACGGACCTAA/template: experiments were performed on a Bruker "600 spectrometer. pH wild. measurements were not corrected for isotope effects. Sodium 3-trimethMutant 23-1476R,R477L/oligo:GCACCCCTCAGACTGACCTAA/ ylsilyl (2,2,3,3-2H,)-propionatewas used as an internal reference. All of template: wild. the two-dimensional spectra wererecorded in the phase-sensitivemode Mutant 24-P474G/oligo:GAAAGGGCAGGCCTCATACGG/tem- using the time proportional phase incrementation technique (Marion plate: wild. and Wuthrich, 1983) with presaturation of the water signal. COSY (Ane Mutant 25-P474Aioligo:GAAAGGGCAGCCCTCATACGG/template: et al., 1976) and NOESY (Kumar et al., 1980) experiments were rewild. corded using standard phase-cycling sequences. Short mixing times Mutant 26-P474V/oligo:GAAAGGGCAGTCCTCATACGG/template: (200 ms) were used in the NOESY experiments to minimize spin diffumutant 25. sion effects. TOCSY (Bax and Davies, 1985) and ROESY (Bothner-By et

6624


A Wild

B

LIMP I1

lo0l 80

="

60

0

s

40

20

0

15

18

13

15

18

13

hours

hours

C Wild

D

CD36/LIMP II

'7

loo 80

1

CD36/LIMP II ( i4'6-> L)

-

I"

cn

5

-

60-

0

z

13

40 -

18

15 hours

13

18

15 hours

E CD36/LIMPll

(I

F 475

''%V)

CD36/LIMPll (L

loo

18

13

15 hours

18

-> V)

1

13

15 hours

FIG.1. Timedependent changes in the cellular distributionof wild LIMP II, wild CD36, wild CD361LIMp II, and mutant CD3W LIMP I1 protein species expressed inCOS cells. COS cells were transfected withthe wild and mutant species indicatedin the panels. The species shown wererepresentativeof three distinctgroups showing normal distributions (wild LIMP 11; wild CD36; wild CD36LIMPII; mutant 20,

Targeting of Lysosomal Membrane Protein al., 1984) experiments were performed using the standard MLEVl7 spinlock sequence and 80- and 200-ms mixing time, respectively. The size of the data matrix was 2,048 x 512 words i n f . and f l , respectively, and prior to Fourier transformation the two-dimensional data matrix was multiplied by a phase-shifted square-sine bell window function in both dimensions. The phase shift was optimized for every spectrum. Complete assignment of the ‘H NMR spectra of the peptide was performed by using standard two-dimensional sequence-specific methods. Temperature dependence of the amide signals was followed through a series of one-dimensional and TOCSY experiments recorded at different temperatures within the 5 4 5 “C range. The corresponding temperature coefficients were calculated by the least squares method. RESULTS

Use of the CD36ILIMP II Construct Helps to Discriminate between Mutants of the Carboxyl Cytoplasmic Tail of LIMP II Displaying Correct or Inhibited Targeting to Lysosomes-The carboxyl cytoplasmic tail of LIMP I1 has been shown to express a signal necessary and sufficient to targetboth LIMP I1 and its homologous plasma membrane protein CD36 to lysosomes (Vega et al., 1991a, 1991b). Pulse-chase experiments revealed that LIMP I1 was not proteolitically processed upon arrival to lysosomes in normal rat kidney cells (Barriocanal etal., 1986). The lack of this modification, frequently associated with transport of proteins to lysosomes (Braulke et al., 1987), prevented the study of its kinetics of transport by biochemical means. Furthermore, the remarkable survivalof LIMP I1 in thelysosome (half-life of 20 h) (Barriocanal et al., 1986), added an extra difficulty to the study of the targetingof the protein in transient transfectants, since mutants of LIMP I1 missorted to the cell surface could undergo passiveinternalization and,by surviving inlysosomes, produce the false impression that theywere correctly sorted to lysosomes. To circumvent those problems COS cells were transfected with the wild and mutant forms of CD36LIMP 11, a construct made by substituting thecarboxyl cytoplasmic tail of LIMP I1 for the carboxyl tail of CD36 (Vega et al., 1991a). Since the extracellular domain of CD36 interacts with the extracellular matrix protein collagen IV (Oquendo et al., 1990; Vega et al., 1991b) it was expected that CD36LIMP I1 mutants missorted to the plasma membrane would be retained or delayed in the latter, thus, preventing their early passive transport to lysosomes and allowing their discrimination from wild type and mutants sorted efficiently to lysosomes. As shown in Fig. 1, C and D, the wild and mutant species of CD36LIMP I1 which were efficiently targeted to lysosomes were not detected on the cell surface at times shorter than 18 h after transfection. In contrast, mutants of CD36LIMP I1 showing either partial (for mutant 19 see Fig. 1 E ) or complete (for mutant 14see Fig. W ) inhibition of the targeting to lysosomes, accumulated rapidly on the cell surface, and in the latter case no staining of lysosomes was observed 18 h after transfection.Therefore, the studies of the cellular distribution of CD36LIMP I1 mutants in transiently transfected COS cells were always performed between 13 and 18 h after transfection. Deletion of the Last 8 Amino Acids of the Carboxyl Cytoplasmic Tail of LIMP II Impairs the Sorting of CD36ILIMP II to Lysosomes-A mutant of CD36LIMP I1 (mutant 1)lacking the last 8 residues of the carboxyl cytoplasmic tail (mutant 1; E471Stop) when expressed in COS cells accumulated on the was plasma membrane (Fig. 2and Fig. 3, C and D). This result

6625

consistent with the destructionof the signal(s)involved in the targeting of the protein to lysosomes. However, it could not distinguish whether the targeting signal was partly or totally included in the deleted segment or if the truncation caused a change in the tailconformation and the inactivation of an upstream signal. Substitution of the Carboxyl-terminal 4 Amino Acids Leu-IleArg-Thr by the Sequence Thr-His-Thr-Asp Impairs the Sorting of CD36ILIMP II toLysosomes-To investigate further the cause of the missortingof mutant 1, thefour carboxyl-terminal amino acids, Leu-Ile-Arg-Thr, were replaced by the 4-amino acid sequence Thr-His-Thr-Asp (mutant 2: L475T,I476H, R477T,T478D). As shown in Fig. 2 the pattern of cellular distribution of mutant 2 was indistinguishable from that of mutant 1, suggesting that an undetermined number of the 4 carboxyl-terminalaminoacids were involved intargetingthe protein to lysosomes. The Contiguous Leu475 and Ile476Residues Are Critical for Targeting CD36ILIMP II to Lysosomes-Characterization of the residues of the Leu-Ile-Arg-Thr sequence involved in targeting CD36LIMPI1 to lysosomes was achieved by performing separate substitutions of each of the 4 amino acidsby residues showing similar or different physicochemical properties. Replacement ofby either Gly, Ala, Val, or Asp (mutants 12, 13, 14, and 16) prevented the targeting of the protein to lysosomes, as shown by their massive accumulation in the plasma membrane (seeFig. 2) (for mutant 14see also Figs.lF and 3,G and H). Leu475could only be replaced by Ile (mutant 15) but with some loss in the efficiency of targeting, as shown by the early accumulation of significant amounts of the protein on both plasma membrane and lysosomes 13 h after transfection (see Fig. 2 and Fig. 3, E and F ) . Similar studies in which Ile476was replaced by either Gly, Asp, or Glu (mutants 17, 21, and 22) also resulted in impeded targeting of the protein to lysosomes (Fig. 2) (for mutant 17see also Figs.3,O andPI. However, replacement of Ile by either Ala (mutant 18, Figs. 3, M and N ) or Val (mutant 19, Fig. lE and Fig. 3, K and L ) did not block the targeting completely, as shown by theirearly accumulationin both lysosomes and plasma membrane. Finally, transport was normal whenthe Ile residue was replaced by Leu (mutant 20) (Fig. 2 and Fig. 3, Z and J ) . As recorded in Fig. 2 the replacement ofArg477by either Gly, Ala, Glu, Gln,or Lys (mutants 3-7) did not imped the targeting of CD36LIMP I1 to the lysosomal membrane. Similar results were observed when Thr478 was replaced by either Gly, Ser, Val, or Ile (mutants 8-11), disregarding that the 2 carboxyl-terminal amino acids were part of the sorting signal (see “Discussion”). Changing the Ile-Leu-Arg Sequence to Ile-Arg-Leu Blocks the Bansport of CD36ILIMP II and LIMP II to Lysosomes-To investigate if targeting to lysosomes required the consecutive presentation of the Ile and Leu residues, thepositions of Ile476 7 exchanged (mutant 23). As recorded in Fig. 2, and A r ~ 7 were the mutant was targeted to the plasma membrane. The result indicates that separationof Leu and Ile by an interposed residue caused the inactivation of the targeting signal. Extensive Substitution of Pro474Does Not Interfere with the Targeting to Lysosomes-Since proline residues have littleconformational freedom even in unfolded structures (see below),

I476L) and partly (mutant 19,I476V) or complete block (mutant 14, L475V) of CD36LIMP I1 targeting to lysosomes. The cellular distributions of the newly synthesized construct species were studied by immunofluorescence microscopy at different times after transfection. Fifty transfected cells were scanned for each of the cDNA species transfected. Bars indicate construct localization restricted to the secretory pathway (i.e. endoplasmic reticulum and Golgi complex) (black bars), lysosomes (white bars),plasma membrane (lightly hatched bars), and plasma membrane plus lysosomes ( h e a d y hatched bars).It should be noted that cells showing staining of lysosomes or plasma membrane frequently showed staining of the Golgi complex.

6626

Targeting of Lysosomal Membrane Protein CD36AIMPII mutans

Localization

LYS

'WA

Wild type

RGCGSTDEGTAD

Mutant 1

++

Mutant 2

++

m d RGQGSTDEGTADERAPTHTD 'WARGQGSTDEGTADERAPLIGT A E Q

K

'TA

Mutant 3 4 5 6 7

RGQGSTDEGTADERAPLIRG S V I

Mutant 8 9 10 11

RGQGSTDEGTADERAPGIRT A V I D

Mutant 12 13 14 15 16

'VA 'PA

RGQGSTDEGTADERAPLGRT A V L D

E

y/A RGQGSTDEGTADERAPLRIT

'XA

Mutant 17 18 19 20 21 22

++ ++ ++ ++

++ ++

++

++

++

+

+ + ++

++ ++ ++ ++

RGCGSTDEGTADERAPLIRTSWD

Mutant 28

++

Mutant32

++

RGQGSTDEGTADQRAPLIRT

Mutant 33

++

RGQGATDEGTADERAPLIRT

Mutant 34

++

Mutant 35

++

'FA

'rd

RGQGSTDEGAADEFWPLIRT

++ + +

++

++

Mutant 23 Mutant 24 25 26 27

RGQGSTDQGTADERAPLIRT

++ ++ ++ + ++

++

RGQGSTDEGTADERAGLIRT A V D

'WA

vTA

++

RGQGSTDEGTADERAPLIRT

w '4

rA

PM

CD36 mutant Wild type

++

Mutant 29

++

NHRNRRRVCKCPRPVVKSGDKPSLSARYV

Wild type

++

NHRNRRRVCKCPRPWKSGDKPSLIARYV

Mutant 30

++

NHRNRRRVCKCPRPLIRT

Mutant 31

++

CD8 mutants

'm

FIG.2. Cellular distribution of wild CD36nIMF' I1 and mutantsof CDWLIMP 11, CD36, and CD8. The bows represent the carboxyl transmembrane domains of the proteins. The sequences of amino acids correspond to their carboxyl cytoplasmictails. Substitute amino acids are

6627


TABLEI ‘Hchemical shifts of the icosapeptide correspondingto the carboxyl cytoplasmictail of LIMP I1 Experimental conditions were: peptide10 a, pH 5.0,5 “C. Valuesfor the cis Pro peptide are shownin italics. (6,ppm from sodium 3-trimethvlsilvl (2.2.3.3-’&)-~ro~ionate).

8.98 2.04 8.76 8.82 8.40 8.47 2.66 8.48 8.55 8.62 4.30 8.18 8.58 8.44 8.43 8.39

2.16,

2.77,

1.79 8.32

1.95,

1.71, 1.71

C& 3.24, 3.24; N.H 7.35

2.42, 2.42

N,H 7.71, 7.01

3.95, 3.89 4.32

1.22

2.15, 1.98

2.37, 2.35 1.23

1.41 2.77, 2.64 2.12, 2.00 1.88, 1.79

2.36, 2.34 1.67, 1.67

4.43

1.39 8.36

1.67,

C,H 3.23, 3.21; N.H 7.40

1.67

4.58

8.21 1.36 1.89

4.11 1.95 4.10,4.04 4.36 4.06, 4.01 4.56 4.41 4.65 4.32 4.03,4.03 4.38 4.29 4.58 4.28 4.32

4.35 2.30,

2.03, 2.03

4.42

2.15,

1.992.35,4.63 8.47

4.32

1.64, 1.57

8.63

4.33

1.67, 1.58

8.35

4.18

1.85

2.15

C,H 3.82, 3.66

3.61,

3.51

1.64

Cd-IH, 0.95, 0.91

1.51, 1.20

CaH3 0.86

CJ-I, 0.91 8.66 1.78

1.92,

4.49

8.67

4.48

4.29 8.03

4.17

the possibility that Pro474could be necessary for the targeting activity of the Leu-Ile sequence was also studied. In mutants 24-27, Pro474 was replaced by either Gly, Ala, Val, or Asp, respectively. Gly was introduced because, unlike Pro, it has many different conformations in different unfolded structures of the same protein and thereby contributes to the diversity of unfolded conformations. Ala, Val, and Asp completed the spectrum of amino acids with preferences for a-helix, P-sheet, and P-hairpin bend, respectively. As recorded in Fig. 2, none of the four mutants showed any alteration of the targeting to lysosomes, disregarding that the activity of the targeting signal was conditioned by Pro474. Addition of 3 Extra, Ser-Dp-Asp, Amino Acids to the Carboxyl End of CD36ILIMP11 Does Not Interfere with Targeting of the Protein to Lysosomes-The proximity of the Leu-Ile sequence to the carboxyl end of the protein and the frequency with which the activity of a targeting signal depends on its exact positioningwith respect to thecarboxyl end of the protein (Munro and Pelham, 1987; a u l d et al., 1987) led us to study the importance of the Ile-Leu position within the protein tail. For that purpose three arbitrary residues, Ser-Trp-Asp (mutant 28) were added to the carboxyl end of the protein. As shown in Fig. 2, the mutant was targeted to lysosomes, indicating that the position of the Ile-Leu signal withrespect to the carboxyl end was notcritical for targeting. Mutants Expressing the Leu-Ile Sequence in the Cytoplasmic Tails of CD36 and CD361CD8 Are Not flzrgeted toLysosomesSince patching of the carboxyl tail of LIMP I1 to the plasma membrane proteinsCD36 and CD8 resulted ineffective targeting of both reporter proteins tolysosomes (Vega et al., 1991a), we sought to study if the Leu-Ile pair wassufficient for efficient targeting. For that purpose we developed mutants of CD36 (mutant 29) and CD36/CD8 (mutants 30 and 31) with the LeuIle pair indifferent positionsof their carboxyl cytoplasmic tails. Particularly interestingwas mutant 31 since the insertedLeuIlesequence waswithinthe sequence Pro-Leu-Ile-kg-Thr

1.67, 1.63

C,H 3.22, 3.22; N.H 7.30

1.18

found in LIMP I1 (see “Discussion”). However, as shown in Table I none of the three mutants was targeted to lysosomes (see Fig. 2 and Fig. 3, Q-TI. Mutations Substituting the Potentially Phosphorylated Ser463 and Thr4- Residues or Disrupting the n o Casein 11 Kinase Cassettes Ser(Thr)Xaa-Asp-Glu,Expressed in the Cytoplasmic Tail of LIMP 11, Do Not Alterthe Targeting of LIMP 11 to Lysosomes-The carboxyl cytoplasmic tail of LIMP I1 has 2 residues, Ser463 and Thr468,that could be potentially phosphorylated by casein kinase I1 (Meggio et al., 1984; Kuenzel et al., 1987). To investigate if their phosphorylation by casein kinase I1 was involved in targeting CD36LIMP I1 to lysosomes the residuesG ~ andu G ~ ~ were ~ u separately ~ ~ ~ replaced ~ by Gln (mutants 32 and 33, respectively)and Sef163and Tre468by Ala (mutants 34 and 35, respectively). As shown in Fig. 2 all four mutants were targeted to lysosomes (Fig. 2). Analysis of the Carboxyl Cytoplasmic Tail of LIMP 11 by ‘H NMR Indicates the Predominance Random of Coil Conformations-”H NMR analysis of the peptide representing the cytoplasmic tail of lysosomal acid phosphatase has shown that the tyrosine critical for transport of the protein to lysosomes stabilizes a type I p-turn, which might be required for correct targeting (Eberleet al., 1991). To gain further insights into the signal that targets LIMP I1 to lysosomes the peptide corresponding to the carboxyl cytoplasmic tail of LIMP I1 was studied by “H NMR spectroscopy. Table I lists the ‘H chemical shift values for the peptide in aqueous solution. There were no significant differences between the recorded 6 and those accepted for a random coil conformation. Furthermore, thedifferences between the chemical shifts measured under nondenaturing and denaturing peptide conditions were within experimental error, and allof the amide shift temperature coefficients were within the range expected for random coiled peptides(Jim6nez et al., 1986). Briefly, the results indicated the scarcity of peptide stable structures. Although some nonsequential NOE connectivities

printed in boldface letters. Wild and mutants proteins were expressed in COS cells, and their cellular distributions were studied 13 h after transfectionby immunofluorescence microscopy.Cells showing exclusive staining of endoplasmic reticulum and Golgi complex were not scored (see “Experimental Procedures” andFig. 1).

6628


FIG.i!. Immunofluorescence microscopy of COS cells transfected with wild CD36/LIMP I1 and mutantsof CD36/LIMP 11, CD36, and CD36/CDF(. I:< h aftrr transfection COS cells were elther inculxltcd for 1 h at 4 (; with thr corresponding antihody and fixed with 2% paraformaldehyde to study the surface expression of the corresponding foreign protein (labeled n p ) or fixed-pcrmeahilized with cold methanol

Targeting of Lysosomal Membrane Protein 465 459

470

6629

475 478

RGQGSTDEGTADERAPLIRT dNN(i.i+l) daN(i.i) daN(i.i+l) daN(i,i+P)

-.--.

.. ._.I+.I. .I. *

*.

-+I.-+

-*

-

,

"

daN(i.1+3)

-

dsch(i,i+Z)

-"

dsch(i.1+3)

476

-

475

I

6. ppm

'

,

8 . 08 . 28 . 4

'

,

'

,

I

'

ppm

t

4.4

'

V

'

l

4.2

FIG. 5 . lbo-dimensionalROE spectrum of the Pro-Leu-ne-ArgThr region of the icosapeptide corresponding to the carboxyl cytoplasmic tailof LIMP 11. Experimental conditions wereas in Fig. 3.Nonsequential NOE connectivities are boxed.

-

FIG.4. ROE connectivitiesrecorded in the'H N M R analysis of the icosapeptide corresponding to the LIMP I1 carboxyl cytoplasmic tail. Experimental conditions were 10 m~ peptide, pH 5.0, 5 "C, 200-ms mixing time. Correlations either close to the diagonal or overlapping with the solvent signal are marked with an asterisk. The abbreviation dsch indicates ROE connectivities involving side chains.

Ile-Arg-Thr; (iii) Gly, Ala, Val, or Asp residues substituted for Leu475., (iv) Gly or Asp instead of Ile476. The extensive mutagenesis of the L e ~ ~ ~ ~sequence - I l ein~ ~ ~ dicates that targetingto lysosomes was normal only when the first position of the dipeptide was occupied by Leu. The only viable, although less efficient, alternative to Leu475was Ile,the wereidentified in a NOESY experiment, performed with a aliphatic amino acid closest to Leu. It also revealed that the short mixing time (200 ms) to avoid the generation of nonse- second position allowed a broader range of substitutions, since quential NOES attributable to spin diffusion effects (Marion, mutants with Ile, Leu, Val, and Ala in position 476 were tar1985), the results of a ROESY experiment, less sensitive to geted to lysosomes to different extents. Another important reqthose artifacts (Bax e t al., 19861, confirmed most of the nonse- uisite is that the Leu and Ile residues must occupy adjacent quential NOE connectivities recorded in the NOESY experi- positions, as shown by the inhibition of the transportwhen the ment. The ROESY correlations between protons in different positions of Ile476 and Ar&77 were exchanged. Therefore, 2 residues are illustrated schematically in Fig. 4. Some of the consecutive hydrophobic amino acids in positions 475 and 476 observed spectral correlations involving resonance inthe appear to be critical for targeting the protein to lysosomes. Leu475and Ile476side chains areshown in Fig. 5. Note that they Targeting is most effective when residues 475 and 476 display were spread throughout thewhole length of the peptide, indi- long branched aliphatic side chains with two terminal methyl cating the presence of small populations of the peptide with groups ( i e . Leu, Ile). transient nonrandom conformations (involving protons Less critical appears to be the position of the pair Leu-Ile brought together at distances in the range of 2-5 A) in a dy- with respect to the carboxyl end of the protein since extension namic equilibrium with the more abundant extended struc- of the cytoplasmic tail with 3 randomlychosen amino acidshad tures (characterized by intense sequential C,Hi-NHi+l cross- no effect on the targeting. signals). Moreover, among the sequences showing transient The Leu-Leu sequence appears tobe the essential element in conformations, those correspondingto the segment Leu-Ile-Arg the 6-amino acid sequence Asp-Lys-Gln-Thr-Leu-Leu of Tac for required theadoption by Ile476of main chain 4 and torsions targeting the truncated antigento lysosomes (Letourneur and close to the values of an a-helix turn (see"Discussion"). Finally, Klausner, 1992). Furthermore, deletion of the carboxyl-termithe possibility that the nonsequential NOE connectivities were nal Leu-Leu-His-Val of the cation-independent mannose produced by peptide aggregates (Dyson and Wright, 1991) was 6-phosphate receptors (MPRs)has been shown to interfere parexamined in a 10-fold dilution experiment, going down from 10 tially with the sortingof soluble lysosomal hydrolases (Johnson to 1mM peptide. The results did not reveal any signal narrow- and Kornfeld, 1992a). Similarly, alanine scanning experiments ing, thus confirming that the observed NOE effects were pro- have revealed that thecarboxyl-terminal His-Leu-Leu residues duced by monomer peptide molecules and not by aggregates. of the cation-dependent MPR are critical for efficient sorting of lysosomal hydrolases (Johnson and Kornfeld, 1992b). MoreDISCUSSION over, the carboxyl-terminal Leu-Leu-Arg-Ile(Thr) sequenceexOur studies have sought the characterization of the signal(s) pressed in the cytoplasmic tail of the mouse hepatitis coronathat targets LIMP I1 to lysosomes (Vega et al., 1991a, 1991b). virus E l could explain why the protein is targetedto lysosomes Targeting of LIMP I1 to lysosomes occurs by a mechanism in- when the transmembrane domains that retain it in theGolgi dependent of mannose 6-phosphate (Barriocanal et al.,1986) complex are deleted (Armstrongand Patel,1991). Similarly, the and ismediated by a tyrosine-lacking signal(s)expressed in the Leu-Ile pair found in thecytoplasmic tail of the invariantchain carboxyl cytoplasmic tail of LIMP I1 (Vega et al.,1991b). This of the major histocompatibility complex class I1 molecules feature distinguishes LIMP I1 from the rest of resident lyso- might also be involved in targeting the class I1 molecules to somal membrane proteins(Williams and Fukuda,1990; Harter either endosomes or lysosomes (Bakke andDobberstein, 1990). and Mellman, 1992) and the lysosomal acid phosphatase preThe failureof Leu-Ile sequencesto target theCD36 and CD8 cursor (Peterset al.,1990; Lehmann et al.,1992) whose target- reporter proteins to lysosomes could result from neutralization ing tolysosomes is mediatedby a tyrosine-based signal located of the single Leu-Ile signal placed in a wrong context. With in their carboxyl cytoplasmic tails. respect to this, it could be interesting to compare the NMR The experimentsperformed here establish that the sequence profiles of the carboxyl cytoplasmic tails of wild CD36LIMP 11, Leu-Ile, found in the 20-amino acid long carboxyl cytoplasmic and the CD8 and CD36 mutants with Leu-Ile signals. Moretail, is critical for targeting the CD36LIMP I1 construct to over, the precedent that the sequence Leu-Leu acts in tandem lysosomes. This is indicated by the missorting to the plasma with tyrosine-based motifs to sort Tac-TCR constructs to lysomembrane of CD36LIMP I1 mutants displaying: (i) a truncated somes (Letourneur and Klausner, 1992) and MPRs to late encarboxyl tail (E471Stop) of 12 amino acids; (ii) a carboxyl tail dosomes (Johnson andKornfeld, 1992a, 1992b) alsoraises the with the terminal sequence Thr-His-Thr-Asp instead of Leu- possibility that a second signal mightbe required for targeting

+

before incubation with the appropriate antibody to study both their intracellular and surface distributions (labeled p ) . The proteins studied are indicated in the lower right corners of the pictures. Bars, 15 pm.

6630


LIMP 11 to lysosomes (see below). It is important that the peptide representing the carboxyl cytoplasmic tail of LIMPI1 displays predominant extended structures with no turn conformations. This suggests that, unlike the tyrosine-based signals (Eberle et al., 1991; Bansal and Gierasch, 1991),the Leu-Ile signal is presented in a structure with an extended conformation. More work is required to gain further insights into the meaning of the nonsequential NOE connectivities detected in the NMR analysis of the peptide. In view of the adoption by Ile476of main chain 4 and $ torsions close to the values of an a-helix turn and of its involvement in the process of targeting, it will be interesting to compare the NMR spectra of the wild peptide and Ile476mutants showing normal or impeded targeting to lysosomes. The carboxyl-cytoplasmictail of LIMP I1contains two potential casein I1 kinase phosphorylation sites, yet their neutralization has no detectable effect on the targetingof CD365IMP I1 t o lysosomes. However,these negative results do not exclude the possibility that phosphorylation plays an important role in the targeting of LIMP 11 to lysosomes. It should be noted that LIMP I1 is phosphorylated in transfected cells' and that the negative results discussed here were performed in cells expressing abnormal high levels of the protein, which may have resulted in the loss of the fine regulation of the mechanisms of targeting. With regard to this, it isinteresting that Meresse et al. (1990)have recently described a casein kinase 11,associated with the 47-kDa subunit of the Golgi adaptor complex, which phosphorylates 2 serine residues in the cytoplasmic tail of the MPR. Experiments in cell lines expressing constitutive levels of the wild and mutant proteins might help to study this possibility. Since the MPRs (Griffiths et al., 1988; Denget al., 1991)and LIMP I1 (Barriocanal et al., 1986) colocalize in clathrin-coated Golgi-derived vesiclesand in late endosomes (Le. prelysosomal compartment) but only LIMP I1 is found in lysosomes, it is likely that theLeu-Ile(Leu1sequences play an important role in their sorting at the trans-Golgi network. The nature of the pathway that transports proteins bearing the tyrosine motif to the membrane of lysosomes has been the subject of an intense discussion (Nabi et al., 1991; Mathews et al., 1992; Harter and Mellman, 1992; Carlsson and Fukuda, 1992). In this respect, the Gly-Tyr-Gln-Thr-Ile and Gly-TyrGlu-Gln-Phe sequences might serve as a plasma membrane internalization signal, when acting alone, or as aGolgi sorting signal, when acting in tandem with a Leu-Leu signal (Johnson and Kornfeld, 1992a, 199213: Letourner and Klausner, 1992: Harter andMellman, 1992).With respect to the involvement of the tyrosine residue in surface internalization, it isnoteworthy that tyrosine-based signals frequently play a role in the internalization of surface receptors (Lerhman et al., 1985; Mostov and Deitcher, 1986; Davies et al., 1987; Iacopetta et al., 1988; Jing et al., 1990; Lobel et al., 1989; Miettinen et al., 1989; Breitfeld, 1990; Collawn et al., 1990), some of which are degraded in lysosomes. Our results do not say whether LIMP I1 is targeted t o lysosomes directly from the Golgi or through the plasma membrane. With respect to this, surface Scatchard binding analysis and internalization studies using 1251-monoclonalantibody to LIMP 11, performed at 4 and 37"C, respectively, show that 0.02% of the total of LIMP I1 is expressed onto the cell surface and that the protein is rapidly internalized and partially recycled to the cell surface.2 The small amount of LIMP I1 detected on the cell surface might be a result of either inefficient sorting in theGolgi or the transport of the protein to lysosomes via the plasma membrane. Our results with the CD36LIMP I1 construct, which helps to discriminate between mutants efficiently sorted to lysosomes and mutants missorted to the

plasma membrane, do not discard the possibility that theconstruct could be transported to lysosomes through the plasma membrane since its rapid internalization could pass unnoticed in immunostained cells. If the Leu-Ile dipeptide acts with a second signal to target LIMP I1directly from the Golgi to lysosomes, the nature of that second signal must be different from those describedpreviously (Johnson and Kornfeld, 1992a, 199213; Letourner and Klausner, 1992; Harter and Mellman, 1992) since the tail of LIMP I1 contains no Tyr residues. Furthermore, inactivation of the Leu-Ile signal blocks completely the targeting of the protein to lysosomes, a result that is in contrast with the partial loss in the Golgi sorting of the TacTCR constructs (Letourneur and Klausner, 1992) and MPRs (Johnson and Kornfeld, 1992a, 1992b) when their Leu-Leu pairs are inactivated. Finally, transport of LIMP I1 from endosomes to lysosomes might not require additional targeting and could result from maturation of the former into lysosomes. Whereas it is clear that a signal in the carboxyl tail of the cation-independent MPR interaots with an undetermined component of the Golgi adaptor complex (Glickman et al., 1989),it remains to be determined if LIMP 11, and by extension any resident lysosomal membrane protein with a Leu-Ile, or equivalent signal, interacts with the Golgi adaptor complex. Acknowledgments-We are gratefulto M. Goethals for help with the synthesis of the peptide. We thank Dr. T. August, J. Castafio, and N. Kieffer for the gift of the antibodies to lamp-1, protein disulfide isomerase, and CD36 and to Drs. J. M. Cuezva, J. Mata, and B. Sed-Real for critical readingof the manuscript. We acknowledgethe institutional support of the Fundacion Areces. REFERENCES Alcalde,J., Bonay, P.,Roa,A., Vilar6, S., and Sandoval, 1. V.(1992)J . Cell B i d . 116, 69-83 Ane, W. P., Bertholdi, E., and Emst, R. R. (1976) J . Chem. Phys. 64,222-224 Armstrong, J., and Patel, S. (1991) J . Cell Sci. 98, 5 6 7 5 7 5 Bakke, O., and Dobberstein, B. (1990)Cell 63,707-716 Bansal, A., and Gierasch, L. M. (1991) Cell 67, 1195-1201 Bamocanal, J. G., Bonifacino, J. S., Yuan, L., and Sandoval, I. V. (1986) J . Biol. Chem. 261,16755-16763 Bax, A., and Davies, D. G . (1985) J . Mugn. Reson. 66,355-360 Bax, A., Sklener, V., and Summers, M. F. (1986) J. Magn. Reson. 7 0 , 3 2 7 4 3 1 Bothner-By,A. A,, Stephens, R. L., and Lee, J. (1984) J. Am. Chem. Soc. 106, 811-813 Braulke, T., Geuze, H. J., Slot, J. W., Hasilik, A., and von Fimra, K. (1987) Eur. J. Cell Biol. 43, 316-321 Breitfeld, P. P., Casanova, J. E., McKinnon, W. C., and Mostov, K E. (1990)J . Biol. Chem. 266, 13750-13757 Brawn, W. J., Goodhouse, J., and Farquhar, M. G . (1986) J. Cell Biol. 103,12351247 Campbell, C. H., and Rome, L. H. (1983) J . Biol. Chem. 268, 13347-13452 Carlsson, S. R., and Fukuda, M. (1992)Arch. Biochem. Biophys. 296,630-639 Collawn, J. F., Stangel, M., Kuhn, L. A., Esekogwu, V., Jing, S. Q., Tmwbridge, I. S., and Tainer, J. A. (1990) Cell 63, 1061-1072 Davies, C. G . , van Driel, I. R., Russell, D. W., Brown, M. S., and Goldstein, J. L. (1987) J. Biol. Chem. 262,4075-4082 de Duve, C. (1963) inLysosomes. Ciba Foundation Symposium (de Reuck, A. V. S., and Cameron, M. P., eds) pp. 1-35, Churchill Livingstone, London Deng, Y. P., Griffiths, G., and Stome, B. (1991) J. Cell Sei. 99,571-582 Deutscher, S. L., Creek, K. E., Merion, M.,and Hirschberg, C. B. (1983)Roc.Natl. Acad. Sci. Li. S. A. 80,393a-3942 Dyson, H. J., and Wright, P. E. (1991) Annn. Reu. Biophys. Biophys. Chem. 20, 519-538 Eberle, W., Sander, C., Klaus, W., Schmidt, B., von Figura, K., and Peters, C. (1991) Cell 67,1203-1209 Fedde, K. N., and Sly,W. S. (1985)Biochem. Biophys. Res. Commun. 133,614620 Fukuda, M. (1991) J . Biol. Chem. 266,21327-21330 Geuze. H. J.. Slot. J. W.. Strous.. G. J... Hasilik A.. and von Fieura. - . K. (1985)J. Cell Bioi. 101; 2253-2262 Geuze, H. J., Slot, J. W., and Schwartz, A. L. (1987) J. Cell Biol. 104, 1715-1723 Glickman, J. N.,Conibear, E., and Pearse, B. M. (1989) EMBO J. 8, 1041-1047 Gould, S. G., Keller, G . A., and Subramani, S. (1987)J. Cell Biol. 106,2923-2931 Griffiths, G., Hoflack, B., Simons, K., Mellman, I., and Kornfeld,S. (1988) Cell 62, 329-341 Harter, C., and Mellman, I. (1992) J. Cell Biol. 117, 311425 Hoflack, B., and Kornfeld, S. (1985a) Proc. Natl. Acud. Sci. U.S. A. 8 2 , 4 4 2 U 3 2 Hoflack, B., and Kornfeld, S. (1985b) J . Biol. Chem. 260,12008-12014 Hughes, E. N., and August, T. (1982) J . Biol. Chem. 267,397C-3977 Iacopetta, B. J., Rothenberger, S., and Kuhn, L. C. (1988) Cell 1 2 , 4 8 5 4 8 9 Jadot, M., Canfield,W. M., Gregory,W., and Kornfeld,S . (1992)J . B i d . Chem. 267, 11069-11077


6631

Jimenez, M. A,, Nieto, J. L., Rico., M., Santoro, J., Herranz, J., and Bermejo, F. J. Meggio, F., Marchiori, F., Bonn G., Chessa, G., and Pinna, L. A. (1984)J. Biol. Chem. 269, 14576-14579 (1986)J . Mol. Struct. 143, 435-438 Jing, S., Spencer, T., Miller, K., Hopkins, C., and Trowbridge, I. S. (1990)J. Cell Meresse, S., Ludwig, T., Frank, R., and Hoflack, B. (1990)J. B i d . Chem. 286, Biol. 110, 283-294 18833-18842 Johnson, K. F., and Kornfeld, S. (1992a)J. Cell Biol. 119,249-257 Metzelaar, M. J., Wijngaard, P. L., Peters, P. J., Sixma, J. J., Nieuwenhuis, H., and Johnson, K. F., and Kornfeld, S. (1992b)J . Biol. Chem. 267, 1711C-17115 Clevers, H. C. (1991)J. Biol. Chem. 266,32393245 Kaplan, A,, Achord, D. T., and Sly, W. S. (1977)Proc. Natl. Acad. Sci. U. S. A. 74, Miettinen, H. M., Rose, J. K., and Mellman, I. (1989)Cell 58, 317-327 Milller, J., and Germain, R. N. (1986)J. Exp. Med. 164, 1478-1489 2026-2030 Kieffer, N., Bettaieb,A., Legrand,C., Coulombel, L., Vainchenker, W., Edelman, L., Minnifield, N., Creek, K. E., Navas, P., and Morre, D. J. (1986)Euc J. Cell Biol. 42, Breton-Gnus, J. (1989)Biochen. J. 262,835-842 92-100 Mostov, K. E., and Deiteher, D. L. (1986)Cell 46,613421 Kornfeld, S. (1987)F U E B J. 1, 462-468 Munro, S., and Pelham, H. R. (1987)Cell 48,899-907 Komfeld, S., and Mellman, I. (1989)Annu. Reu. Cell B i d . 5,483-525 Kuenzel, E. A,, Mulligan, J. A,, Sommercorn, J., and Krebs, E. G . (1987)J. Biol. Murphy, R. F. (1991)%rids Cell Biol. 1, 77-82 Nabi, I. R., Le Bivic, A,, Fambrough, D., and Rodriguez-Boulan, E. (1991)J. Cell Chem. 262,91369140 Biol. 115, 1573-1584 Kumar,A., Ernst, R. R., and Wiithrich, K. (1980jBiochem.Biophys. Res. Commun. Nieto, A., Mira, E., and Castafio, J. G. (1990)Biochem. J. 267,317423 95, 1 4 Novikoff, A. (1963)in Lysosomes. Ciba Foundation Symposium(de Reuck,A. V S., Lehmann, L. E., Eberle, W., Krull, S., Prill, V.,Schmidt, B., Sander, C., von Figura, and Cameron, M. P., eds) pp. 3677,Churchill Livingstone, London K.,and Peters, C. (1992)EMBO J . 11,43914399 Oquendo, P., Hundt, E., Lawler, J., and Seed, B. (1990)Cell 68,9%101 Lemansky, P., Hasilik, A., von Figura, K., Helmy, S., Fishman, J., Fine, R. E., Peters, C., Braun, M., Weber, B., Wendland, M., Schmidt, B., Pohlmann, R., WaKedersha, N. L., and Rome, L. H. (1987)J. Cell Biol. 104, 1743-1748 heed, A,, and von Figura, K. (1990)EMBO J. 9,3497-3506 Lerhman, M. A,, Goldstein, J. L., Brown, M. S., Russell, D. W., and Schneider, W. Pohlmann, R., Waheed, A,, Hasilik, A., and von Figura, K. (1982)J. Biol. Chem. J. (1985)Cell 41,735743 Letourneur, F., and Klausner, R. D. (1992)Cell 69, 1143-1157 257,53234325 Littman, D. R., Thomas, Y., Maddon, P. J., Chess, L., andAxel, R. (1985)Cell 40, Sahagian, G . G., and Neufeld, E. F. (1983)J. Biol. Chem. 258, 7121-7128 Schmid, S., Fuchs, R., Kielian, M., Helenius, A,, and Mellman, I. (1989)J. Cell Biol. 237-246 108. 1291-1300 Lobel, P., Fujimoto, K.,Ye, R. D., Griffiths, G . , and Kornfeld, S. (1989)Cell 57, Straus; W. (1963)in Lysosomes. Ciba Foundation Symposium (de Reuck, A. V. S., 787-796 Lopata, M. A,, Cleveland, D. W., and Sollner-Webb, B. (1984)Nucleic Acids Res. 12, and Cameron, M. P., eds) pp. 151-175,Churchill Livingstone, London Vega, M. A,, Rodriguez, F., Segui, B., Calks, C., Alcalde, J., and Sandoval, I. V. 5707-5717 (1991a)J . Biol. Chem. 266, 16269-16272 Malissen, B., R e b i , N., Liapeuf, B. A., and Mowas, C. (1982)Euc J. Immunol. 12, Vega, M. A,, Segui-Real,B., Garcia, J. A,, Calks, C., Rodriguez,F., Vandekerckhove, 734-745 J., and Sandoval, I. V. (1991b)J. Biol. Chem. 266,16818-16824 Marion, D. (1985)FEES Lett. 192,99-103 Marion, D., and Wiithrich, K. (1983)Biochem. Biophys. Res. Commun. 113, 967- Vladutiu, G. D. (1983)Biochim. Biophys. Acta 760,3634370 Williams, M. A., and Fukuda, M. (1990)J . Cell B i d . 111, 955-966 974 Mathews, P. M., Martinie, J. B., and Fambrough, D. M., (1992)J. Cell Biol. 118, Yuan, L., Bamocanal, J. G., Bonifacino, J. S., and Sandoval, I. V. (1987)J . CeZl Biol. 105, 215-227 1027-1040