Jun 18, 1993 - mark membranes in transit at this site. For the present, the .... diluted 1/100 in 1% NGS, 30 min at room temperature; (iii) mouse or rab- ..... Kantrowitz,E.R., Pastra-Landis,S.C. and Lipscomb,W.N. (1980) Trends. Biochem.
The EMBO Journal vol.12 no.10 pp.3753-3761, 1993
SCAMP 37, a new marker within the general cell surface recycling system
Susan H.Brand and J.David Castle1 Department of Anatomy and Cell Biology and the Molecular Biology Institute, University of Virginia Health Sciences Center, Charlottesville, VA 22908, USA 1Corresponding author Communicated by P.DeCamilli
Secretory carrier membrane proteins (SCAMPs) are widely distributed as components of post-Golgi membranes that function as recycling carriers to the cell surface. In fibroblasts, SCAMPs are concentrated in compartments involved in the endocytosis and recycling of cell surface receptors while in neurons and other cell types having regulated transport pathways, SCAMPs are also components of regulated carriers (synaptic vesicles, secretion granules and transporter vesicles). Their presence in multiple pathways distinguishes them from proteins (e.g. recycling cell surface receptors and synaptic vesicle proteins) which are concentrated in selected pathways. The SCAMPs also do not appear to reside beyond the boundaries of these pathways. This distribution suggests that SCAMPs are general markers of membranes that function in cell surface recycling. The primary sequence of SCAMP 37 reveals a novel polypeptide containing a series of structural motifs, including a calcium binding domain, a leucine zipper and two zinc fmgers. The very broad tissue distribution, subcellular localization and sequence analysis all predict that SCAMPs play a fundamental role in cell surface recycling. Key words: endocytosis/recycling/SCAMPs/secretion/ synaptic vesicles
Introduction The operations of endocytosis and secretion in eukaryotic cells are mediated by membrane-bound carriers that recycle between the cell surface and internal compartments. Progress in the study of these operations has been driven largely by different experimental approaches in different cell types. It is only recently that the considerable similarity of these pathways and the extent of their interaction are becoming clear (Cameron et al., 1991; Green and Kelly, 1992; Hopkins, 1992). Of all the carriers that are involved in these operations, by far the most is known about the composition of synaptic vesicles. The primary structures of most of their major resident proteins have been determined (reviewed in Sudhof and Jahn, 1991) and several have been implicated to function in vesicular transport (Petrenko et al., 1991; Alder et al., 1992a,b; Schiavo et al., 1992; Elferink et al., 1993; Sollner et al., 1993) in ways that remain to be precisely defined. Synaptic vesicles have been regarded primarily as unique ( Oxford University Press
structures in composition, location and function. They are distinguished by their ability to regenerate rapidly following exocytosis, and until recently, this local recycling function within axon termini was regarded as having little relationship to other recycling pathways. However, an important conceptual turning point in this view arose with the discovery of synaptic-like microvesicles in neuroendocrine cells (Navone et al., 1986). Further, the realization that a synaptic vesicle protein expressed in a fibroblast is targeted to recycling early endosomes (Johnston et al., 1989) and the subsequent characterization of the interrelationship of synaptic-like microvesicles to constitutive secretory and recycling endocytic pathways (Clift O'Grady et al., 1990; Cameron et al., 1991; Regnier-Vigouroux et al., 1991) have also helped to promote a new view in which synaptic vesicles may represent a specific adaptation of a general recycling system that encompasses all membranes that communicate with the cell surface (Cameron et al., 1991). This view has gained further support from the identification of proteins that are related (if not identical) to synaptic vesicle components in secretion granule membranes and other regulated vesicular carriers (Abdelhaleem et al., 1991; Brand et al., 1991; Perin et al., 1991; Cain et al., 1992; Laurie et al., 1993). Thus all post-Golgi pathways may prove to be related as derivatives of a general recycling system between intracellular compartments and the cell surface (Cameron et al., 1991; Sudhof and Jahn, 1991). Much effort has been focused on defining the efficient sorting processes that underlie the formation and restitution of specific carriers involved in recycling (Hopkins et al., 1990; Regnier-Vigouroux et al., 1991; Bennett et al., 1992; Brose et al., 1992; Disdier et al., 1992; Dunn and Maxfield, 1992; Hopkins et al., 1992; Koedam et al., 1992). Less attention has been directed toward determining whether components that are specific to one pathway might have analogs in other branches of the recycling system. Thus an important goal remains to identify polypeptides or families of polypeptides that are general recycling markers and thus may subserve the functions that are shared by all vesicular shuttles involved in cell surface recycling. We have previously identified a family of secretory carrier membrane proteins (SCAMPs 35-40) that have the distinction of being present in all types of membranes that function in regulated transport pathways-secretion granules (exocrine and endocrine), synaptic vesicles (Brand et al., 1991) and insulin-regulated glucose transporter vesicles (Laurie et al., 1993). The SCAMPs are absent in purified rough microsomal and lysosomal fractions and are barely detected, if at all, in plasma membranes from exocrine acinar tissue (Brand et al., 1991). SCAMPs are also found in cell types (hepatocytes, fibroblasts) that have no apparent regulated transport pathway. We now provide evidence that they are general markers of cell surface recycling pathways. This distribution is demonstrated by colocalization with an 3753
S.H.Brand and J.D.Castle
endocytic tracer in fibroblasts and is supported by immunocytochemistry and cell fractionation of rat brain. We also present the primary structure of SCAMP 37 as deduced by cloning and sequencing from a rat brain library. This polypeptide has a novel sequence and a unique combination of structural motifs that have been implicated elsewhere in formation and regulation of macromolecular complexes. These motifs include a potential calcium binding site, a leucine zipper and two zinc fingers. Also we identify a novel alanine repeat motif that may be an amphipathic or mostly nonpolar (membrane-inserted?) helix and is just upstream from a C-terminal tripeptide identical to that of synaptophysin.
Results Tissue distribution of SCAMPs as shown by Western blotting of total membranes In extending our initial observations of SCAMPs in cells that exhibit regulated secretion, we now show that they may be distributed ubiquitously among cell types. SCAMPs 37 and 39 have been identified in all tissues examined so far, although SCAMP 39 is present at such low levels in brain that it is not detected at the protein loads shown in Figure 1. It should be noted that SCAMPs were originally stated to have apparent Mr in the range 31 000-35 000 (Brand et al., 1991). Our present results, including a prediction from
Fig. 1. Distribution of SCAMPs in various rat tissues. Immunoblot showing SCAMPs (Mr 37 000 and 39 000) following separation of total membrane samples (70 isg protein each) on an 11 % SDS-PAGE.
molecular cloning, suggest that the Mr estimates should be revised to 35 000-40 000. Association of SCAMPs with cell surface recycling pathways The broad distribution of SCAMPs among cell types, irrespective of their function in regulated secretion, caused us to examine the presence of SCAMPs in more widely distributed recycling pathways. Accordingly, we compared the immunocytochemical distribution of SCAMPs with the distribution of fluorescein transferrin (FITC-TF) that had been internalized by endocytosis in NRK fibroblasts. After extended FITC-TF uptake (1 h at 37°C), there is a striking, almost complete, colocalization with the SCAMPs (Figure 2). This colocalization is observed both at the cell periphery [likely within the 'rapid recycling pathway' (Hopkins and Trowbridge, 1983; Hopkins et al., 1990)] and in the perinuclear region [likely within later endocytic compartments (Hopkins and Trowbridge, 1983; Yamashiro et al., 1984; Hopkins et al., 1990)]. Thus the ubiquitous presence of SCAMPs in cells may relate to its presence in cell surface recycling pathways. We also examined the distribution of SCAMPs in neuronal tissue where SCAMP 37 is unusually enriched. Our previous studies showed by direct labeling that SCAMPs are synaptic vesicle proteins but did not address their presence in other compartments (Brand et al., 1991). In addition, synaptic vesicle proteins characteristically are concentrated in axon termini where they are mostly localized to synaptic vesicles (DeCamilli and Jahn, 1990). Using both immunocytochemistry (Figure 3a and b) and cell fractionation (Figure 3c), it can be seen clearly that neuronal SCAMPs are much more broadly distributed than other synaptic vesicle proteins. In cerebellar sections, synaptophysin is concentrated in synaptic termini that are most evident in the granule layer and is largely absent in the soma and dendrites of Purkinje cells (Figure 3b). In contrast, SCAMPs are found in synaptic termini as well as in Purkinje cell bodies and dendrites (Figure 3a). Further, the more diffuse staining of SCAMPs than synaptophysin within synaptic termini may reflect a wider distribution among membranes involved in local recycling. Velocity sedimentation on glycerol gradients also demonstrated only a partial overlap of distributions of SCAMP 37 and synaptophysin among membranes compris-
Fig. 2. Colocalization of SCAMPs with FITC-transferrin in NRK fibroblasts. Paired immunofluorescent images showing extensive colocalization of SCAMPs (SC) and FITC-transferrin (TF) following a 60 min incubation in FITC-TF. Colocalization is evident in both the peripheral cytoplasm and in the perinuclear region. A few FITC-TF positive/SCAMP negative spots (arrowheads) may represent binding of FITC-TF at the cell surface.
3754
SCAMP distribution and structure
ing a crude synaptic vesicle fraction from rat cortices (Figure 3c). While synaptophysin is almost completely concentrated in the slowly sedimenting synaptic vesicle peak, only about half of the SCAMP 37 cosediments with synaptic vesicles while the other half sediments with larger or more dense vesicles at the bottom of the gradient. We also note that the antigenic peaks attributed to synaptic vesicle associated synaptophysin and SCAMP 37 do not exactly coincide. This could reflect heterogeneity in relative antigen concentrations
;4 1, '4;.,
;f
i! RWR, t)iK }i'.tr''4,
b.
-A4>% 4
-4
I
It
among synaptic vesicle subpopulations or the presence of other SCAMP-containing membranes with slightly different sedimentation properties than synaptic vesicles. Purification of SCAMP 37 from rat brain SCAMP 37, which is unusually enriched in rat brain membranes, has been purified on an immunoaffinity column containing antibody SG7C 12. As can be seen in the silverstained SDS -PAGE profile of the eluate from the affinity column (Figure 4a), SCAMP 37 is sufficiendy pure to enable easy dissection for digestion and microsequencing. Notably, both protein staining and blotting of the immunopurified antigen showed that SCAMP 37 has a strong tendency to form SDS-resistant aggregates (Figure 4a) with molecular weights that are consistent with formation of homooligomers. Tryptic digests of purified SCAMP 37 were separated by HPLC and the structures of four microsequenced peptides are shown in Figure 4b.
Cloning and sequencing of SCAMP 37 Two overlapping oligonucleotides, made to the ends of the longest tryptic peptide fragment of SCAMP 37 (Figure 4b and Materials and methods), were used to screen a rat brain cDNA library. After screening one million recombinants, two clones (2.0 and 3.5 kb) were identified, each having a distinct restriction map. The 2.0 kb clone was sequenced in entirety in both directions and revealed an open reading frame of 338 amino acids from its initiator methionine, giving a calculated Mr of 37 900 (Figure 5a). The 3.5 kb clone encodes a highly related but incomplete cDNA and is presently being analyzed further. All four microsequenced peptides were identified in the deduced amino acid sequence of the 2.0 kb SCAMP cDNA (underlined in Figure Sa). An N-terminal domain having the general structural features of a signal sequence (von Heijne, 1985) was not identified. Hydropathy analysis (Figure Sb) suggests that SCAMP 37 consists of three domains: a hydrophilic N-terminal
.N
a 9 \.
. -'J.'
C
4,
re
.4b 2
:-5.-
H
3 No "I :
.
Z-'
"
-
,
.,
-r I " "D .t I. .. .7'Dl -,Z'
;.; :.
,
-
.0
Fig. 3. Comparison of SCAMP and synaptophysin distribution in rat brain. Immunocytochemical localization of SCAMPs (a) and synaptophysin (b) in rat brain cerebellar sections. SCAMPs are identified in both axon termini (small arrows) of afferent neurons synapsing in the granule layer, as well as in the cell bodies and dendrites (arrowheads) of Purkinje cells, whereas synaptophysin is concentrated in the axon termini. Bar 30 /sm. (c) Immunoblots showing differing relative amounts of SCAMP and synaptophysin (synphsn) in fractions collected following glycerol gradient sedimentation of a partially purified cortical synaptic vesicle. Fraction numbers are in order of increasing density.
Fig. 4. Purification and microsequencing of SCAMP 37 from rat brain. (a) Silver-stained 11% SDS-PAGE (Stain) and Western blot (Blot) of immunoaffinity-purified SCAMP. Silver staining shows that >50% of this fraction is SCAMP 37. Lower mobility bands on the stained profile and especially on the Western blot have molecular weights that are consistent with formation of SCAMP 37 multimers. (b) Amino acid sequences of the tryptic peptide fragments of SCAMP 37 purified by HPLC. *indicates tryptic peptide with incomplete sequence. The positions of the two overlapping oligonucleotides that were used to generate probes for cDNA library screening are also shown.
3755
S.H.Brand and J.D.Castle
a
(TCCCGGATCTGTGCGGCGGCGGCGGCGACGTGAGCGCGCAGGGGGGCGGCG( CGTCTGTGTCTCGGCGCCGGGC)TCGGGTGGGTGACGCGCAGTGCCGGGGC,
120 ,AATGTCGGA N S D 3 TTTTGACAGCAACCCCTTTGCGGATCCGGATCTCMCMCCCGTTCMGGA ,CCCATCTGT 180 F D S N PF A D P D L N N P F K D F S 23
CACACAGGTGACAAGAAATGTGCCACCGGGACTTGATGAATATAATCCTTT(CTCTGATTC T a v T R N V P P G L D E Y N P F D S
43
TAGAACGCCTCCACCAGGTGGTGTGAAAATGCCTAATGTACCCAACACCCACGCCAGCMT
300
0
240
P A 1 63 AATGAAGCCGACCGAGGAGCATCCGGCTTACACACAGATCACAAAGGAGCAlTGCCTTGGC 360 N K P T E E N P A Y T0 I T K E N A L A 83 TCAAGCCGAGCTTCTCAAGCGCCAAGAGGAGCTGGAGAGAAAAGCTGCAGAI GTTACATCG 420 Q A E L L K R a E E L E R K A A E L D R 103 TCGAGAGCGAGAGATGCAGAACCTGAGTCAGCACGGGAGAAAAAATAATTGI GCCACCTCT 480 R E R E N O N L S Q N G R K N NW P P L 123 R T P P P G G V K N P N V P N T
o
540 .ATTCCCGT I 143
TCCTAGCAATTTCCCTGTGGGGCCTTGTTTCTATCAAGACTTTTCTGTAGAC P S N F P V G P C F Y Q D F
S V D
AGAGTTCCAGAAGACAGTGAAGCTTATGTACTACCTATGGATGTTCCA GCtGGTAACCCT J E F 0 K T V KIL M Y Y L W f H A V T L
600 163
660 183 CTTTGGACTCAGCATCCTGTGGTTCTTGCTCTTTACCCCCTGTTCATTTGTCCTGTTGGTA 720 IF G L S I L W F L L F T P C S F V GTTCCTGAACATCTTCGGATGCCTGGCTTGGTTTTGTGTTGATTCCTCAAGAGCAGTTGA F L N I F G C L A W F C VID S S R A V D
CAGGCCACTCTATGGGGCCTTCAGGAGTGACAGTTCCTTCAGGTTCTTTGTGGTTCTTCTT R P L Y G A F R S D S S F R F F V
780
FEi 223 CGTCTATATTTGTCAGTTTGCTGTACATGTGCTCCAGGCTGCAGGATTTCATrACTGGGG a4O V Y I c Q f A V H V L 0 A A G FIH N W G 243 AAACTGTGGTTGGATCTCATCCCTCACTGGTCTCAACAAGAA MTTrCCGGT CCGGUTCAT 900 N C G W I S S L T G L N K P NlI
V
GATGATCATCATCGCAGCGCTCTTTACAGCGTCCGCTGTCATCTCGCTCGTC N I I I A A L F T A S A V I S L V
NTK
283
GCMAGTT 1020 K V N G L Y R T T G A S F E K A Q Q E F 303 TGCAACGGGAGTGATGTCCAATAAAACTGTTCAGACAGCAGCTGCAAATGCC :GCGTCCAC 1080 T A A A N A A T G V N S N K T V
AAAAGTGCATGGACTGTATCGCACAACAGGTGCTAGCTTTGAGAAGGCCCAG
C
A S T
323
1140 338
TATTTTTATGTTGGAAACACACAGTACATAAAGCATGGGTATTTCCTATCCACATGTGCA TGGTCTAAAAACCAGAGAAACTTCCTAGTCTTATTACTTTGCCTAATAGTTTCTTACTAT TTCCATGCCCCTCACAGAAGAAAACTATCATCTGCAAAACAAATATTCTCCATATTTTTG GGGGATGGTGTTCCTCAGATTATTICAGTGGTGACACACTGAAATGAGTATGTTACTGAA GATTTAAAAGGCACTCGATGCTAGCCACAGAAGCTCTTCCAAGACGCAGTAGAGACACAG ATTATACATTGCAAATGAAACAACCTTTGTCCCCCCCCCCCCATATTAGTTTGGAAGGAA CAAGGTCAGCCAAGGCTTGCGGTGTTTTGAAATTGGCTTGCTTTATATGCGGTTCAGTCA GCAGAGAGGAGCCTGCGTGCTCTGTAGTAGATAATGTATATTTTACCTTCTTTTArTTTA TTTTTAGACTAGTTGATATTTTGATACAATCTCTGATCTTGCATGGGCACCACATCTCT
TAAGTGAATGAATTAGTATTTTGGGTTCTGTACTGCTATAGTTGTAGGACCTGGGGCTC
ATAArWE; T TAATATATiU"_lra (T#-
l
TATCAAA(AGbAGGAClAGGAATTATG
GCTAGTATATAGTGATAATTTCTTTTCCCCTCCTAATTTCCAGTATTCATAGAGTAAAAT GTTACTGTAT
A
AA AA
1200 1260 1320
domain, a mostly nonpolar central domain and a predominantly amphipathic C-terminal domain. Four hydrophobic segments, 19-26 residues each (boxedstrongly in Figure 5a), present in the nonpolar central domain and may be are ae i oplr dmi a
transmembrane helices. This organization would place the N- and C-termini on the same side of the membrane (cytoplasmic, see below) with the putative transmembrane helices being separated by reasonably short (7-19 amino acids) connecting segments. As for other polytopic sget o oyoi membrane proteins (Wessels and Spiess, 1988), the first transmembrane domain of SCAMP 37 may serve as an uncleaved signal sequence. The deduced amino acid sequence contains two potential N-glycosylation sites (Asni 0 and 3 10); however, we have shown previously that 1 n ae peiul the SCAMPs are not glycosylated (Brand et al., 1991). In vitro transcription and translation of the 2.0 kb clone generated a polypeptide with an apparent Mr of 37 000 (Figure 6a), in excellent agreement with the calculated Mr. Much more polypeptide of the same size was synthesized in the presence of canine pancreatic rough microsomes, a result consistent with the translation product being an integral membrane protein with an internal signal sequence. Lower molecular weight bands (34, 31 and 27 kDa) were also generated by the in vitro translation and their sizes are consistent with translation initiation from downstream methionines at positions 53, 64 and 108 in the primary sequence. The 37 000 product of the in vitro transcription/ translation can also be immunoprecipitated with SG7C12 (Figure 6a) thus confirming that the isolated cDNA encodes ta noe
a
b
1380 1"0 1500 1560 1620 1680 1740 1800 1860 1920 1980 2015
......
......
b
Fig. 5. Sequence and hydropathy analysis of SCAMP 37. (a) Nucleotide and deduced amino acid sequence of the 2.0 kb rat brain SCAMP cDNA. The four microsequenced peptides are underlined and four nonpolar, potential transmembrane, segments are boxed. The 82 nucleotides at the 5' end (shown in parentheses) come from a related 2.1 kb SCAMP cDNA that has an identical restriction map to the 2.0 kb cDNA that was sequenced. An upstream stop codon (bold letters) is included in the 5' untranslated domain. (b) Hydropathy plot of the deduced amino acid sequence of SCAMP 37 showing hydrophilic N- and C-termini and four closely spaced hydrophobic segments, 19-26 residues each. Window = 9 for the Kyte-Doolittle algorithm (Kyte and Doolittle, 1982). 3756
Fig. 6. The 2.0 kb cDNA encodes SCAMP 37 from rat brain. (a) Autoradiograph showing the major 37 kDa product of the in vitro transcription and translation of the 2.0 kb SCAMP cDNA in the presence (+ As) and absence (- As) of canine pancreatic microsomes. Also shown are other smaller peptides (34, 31 and 27 kDa) that probably reflect translation initiation at downstream methionines (residues 53, 64 and 108, respectively). Immunoprecipitation (Ippt) of the 37, 34 and 31 kDa translation products, but not the 27 kDa product, with the SCAMP monoclonal antibody is also shown. (b) Comigration of the 35S-labeled 37 000 in vitro translation product (in vitro) and SCAMP 37 from rat brain membranes (brain). Samples were run side by side on an 11 % SDS gel, transferred onto nitrocellulose and either autoradiographed (in vitro) or immunoblotted (brain).
SCAMP distribution and structure
the SCAMP 37 originally identified with this antibody. Notably, the shortest (27 000) translation product synthesized in vitro was not efficiently immunoprecipitated. Thus the cytoplasmically oriented epitope for SG7C 12 (Brand et al., 1991) probably includes residues between methionines at positions 64 and 108. Expression of residues 1-149 using
a B3Cala
E 'J
II
SCAMP 37
D
S
M L
a, Integrin
SG
A1
0, i
D S
F P E F L T M
D
y
P D L N N P F
D
Q D G
L
x
r
N P F A
F
D
T
z
Y
N D
-Y
-x
V
AI
G A P F
-Z
d
b
E jvJj
' !
E
&t
+
.
'R (-
EA
I
RK R
R+ l
C,
EARQE hI K
C
G LS F
D
I
L
V
W
A R
F
S
S
A
I
If7
D
T
L
V
P
V
@)G
XWY R PL 2-
Fa
*e
V
OI
224
G
F
GWIg 245
Fig. 7. Structural motifs of SCAMP 37. (a) Diagram depicting prospective N-terminal calcium binding motif (amino acids 1-18 of SCAMP 37) as compared with the Ca2+ binding loop of bovine brain calmodulin and the prospective Ca2' binding domain of the human a1-integrin. The highly conserved amino acids that coordinate calcium ions are boxed and the corresponding coordinate positions are indicated. (b) Longitudinal projection of the a-helical leucine zipper motif (residues 84-109) drawn according to Landschulz et al. (1988) at 3.5 residues per turn. The Q at position 84 in the predicted sequence is placed in position number 1. The aligned leucines (and methionine) and the opposing ordered arrays of positively and negatively charged residues are evident. (c) The two putative zinc finger motifs located between residues 170 and 201, and 227 and 245. Boxes identify the Zn-coordinating C and H residues and circles identify conserved nonpolar residues that interact with one another in other zinc fingers. (d) Diagram depicting prospective a-helical segment (residues 312-338) and adjoining C-terminus highlighting the alaninerich non-polar surface and surrounding S/T-rich surface that endows this domain with either amphipathic or largely nonpolar character. The tripeptide terminus, identical to that of synaptophysin, is boxed.
the pGEX system confirms that the epitope is present in this N-terminal domain (unpublished observation). Finally, the 37 000 in vitro translation product encoded in the 2.0 kb clone exactly comigrates with SCAMP 37 from brain membranes when run side by side on SDS -PAGE and transferred to nitrocellulose (Figure 6b). Although a Kozak consensus sequence for translation initiation (Kozak, 1984) is not present upstream from the presumed start site, the comigration of translated polypeptide and antigen strongly suggests that the 2.0 kb cDNA encodes the full-length polypeptide. Indeed, a recently isolated 2.1 kb cDNA having an identical restriction map to the SCAMP 37 clone has an in-frame stop codon 5' to the presumed initiator methionine (bases -84 to -82), indicating that this methionine is responsible for translation initiation. Comparisons to other polypeptides and domain analysis Searches of the PIR, SwissProt and GenBank data bases indicated that the 2.0 kb cDNA does not encode any previously sequenced proteins. However, several interesting structural features, including similarities to other synaptic vesicle membrane proteins, were revealed from examination of the deduced amino acid sequence. As for many synaptic vesicle proteins, SCAMP 37 seems to lack an N-terminal signal sequence, has a relatively long 3'-untranslated domain (Sudhof and Jahn, 1991) and has a proline-rich cytoplasmic domain (Trimble and Scheller, 1988). Further, SCAMP 37 resembles synaptic vesicle proteins synaptophysin and synaptoporin in size, C-terminal tripeptide NQX (X either M or I) (Knaus et al., 1990) and tendency to self-aggregate (Figure 4a; Thomas et al., 1988; Johnston and Sudhof, 1990). Although the four potential transmembrane helical segments mentioned above would add to the similarity, the topology of SCAMP 37 remains to be determined. Beyond these similarities, however, we have not found other obvious relationships between the proteins in either primary structure or organization within polar and nonpolar domains. Thus SCAMP's hydrophilic N-terminal domain is extended whereas synaptophysin's and synaptoporin's are quite short; SCAMP 37's C-terminal domain is amphiphilic and contains none of the proline repeat motif found in synaptophysin and synaptoporin; and three of SCAMP 37's prospective transmembrane helices would have polar faces whereas only one of synaptophysin's and synaptoporin's would have a polar face. SCAMP 37 has a number of subdomains that may prove to be of functional interest. At the N-terminus, a single consensus sequence for metal ion binding was identified. It resembles the Ca2+ binding sites of the al-chain of human integrin (Fitzgerald et al., 1987) and is less like those found in bovine brain calmodulin II which have helix forming amino acids in the -Z coordinate position and a full EFhand motif (Gariepy and Hodges, 1983) (Figure 7a). Also within the N-terminal cytoplasmic domain is an interesting and highly charged leucine zipper motif between residues 84 and 109 (Figure 7b). This motif is found within a longer stretch (residues 72- 112) of helix-forming amino acids (Chou and Fasman, 1978) and resembles leucine zippers of many DNA binding proteins (Landschulz et al., 1988). However, in contrast to other reported leucine zippers where the charged amino acids surrounding the leucine face of the helix are not noticeably ordered, most of the acidic and basic
3757
S.H.Brand and J.D.Castle
residues in the SCAMP 37 motif are aligned by charge in alternating rows along the helical axis (Figure 7b) and can interact as an electrostatic network (Landschulz et al., 1988). Another motif that is present in various nucleic acid binding proteins (Berg, 1986; Schwabe and Rhodes, 1991) and certain yeast vacuole-associated proteins (Preston et al., 1991; Robinson et al., 1991; Weisman and Wickner, 1992) appears in two forms in SCAMP 37. There are two prospective zinc fingers extending between residues 170 and 201 and residues 227 and 245 (Figure 7c). The first of these has an extended finger similar to that found in aspartate carbamoyl transferase while the second resembles those found in retroviral nucleic acid binding proteins (Berg, 1986). Notably, these motifs span part of the nonpolar segments observed in the primary structure (Figure 5a). If they prove to be bonafide metal coordination sites, then they may serve a structure stabilizing role within the protein interior [as for aspartate carbamoyl transferase (Kantrowitz et al., 1980)]. Their configurations would alter the view that SCAMP 37 contains four transmembrane helices as derived from the hydropathy analysis. A novel repeating sequence A-A-X-X in which X corresponds to polar amino acids, particularly S and T, occurs near the C-terminus of SCAMP 37, spanning amino acids 316-332. It is predicted to be helical (Chou and Fasman, 1978; Garnier et al., 1978) and when projected laterally (Figure 7d), it appears amphipathic with one face (shown centrally) consisting almost entirely of alanine and an opposing face (shown peripherally) consisting of polar residues. Because the numerous S and T residues can hydrogen bond with carbonyl oxygens in the preceding helical turn (Gray and Matthews, 1984), this helix may assume a nonpolar character. Thus it might interact with the nonpolar domains on other proteins or insert into the bilayer interior as has been predicted for the fusion domains of viral glycoproteins (White, 1992). SCAMP mRNA isoforms We used the 2.0 kb SCAMP cDNA to probe a Northern blot of total cellular RNA from several rat tissues (Figure 8). Under very stringent conditions (see Materials and methods) we identified three sizes of SCAMP-related mRNA (2.3, 2.5 and 3.7 kb). All three mRNAs are present in each of the tissues examined, although the relative levels may vary. Northern analysis using poly(A) + RNA from a subset of tissues confirmed these findings (not shown). Thus there are at least three highly related SCAMP messages in these rat tissues. Current progress in cloning and sequencing a cDNA corresponding to the 3.7 kb message indicates that differences within both the coding and non-coding domains may account for the different sizes of mRNA and that alternative splicing is likely to contribute to the diversity of SCAMP isoforms. The conserved profile of SCAMP messages in different tissues clearly reiterates the broad polypeptide distribution (Figure 1) and supports the idea that SCAMPs may play a universal role in cell surface membrane
trafficking.
Discussion SCAMPs as markers of the general cell surface recycling systemOufidnssoigcociztoofSAPadedOur findings showing colocalization of SCAMPs and endocytosed transferrin in fibroblasts (Figure 2) illustrate that
3758
Fig. 8. Northern analysis of total cellular RNA from various rat tissues. Autoradiograph showing three common messages (2.3, 2.5 and 3.7 kb) from the indicated rat tissues that were identified by probing total cellular RNA (10 ,ug/lane) with the 2.0 kb SCAMP cDNA. Lighter exposure of the autoradiograph clearly indicates two distinct messages between 2.3 and 2.5 kb. Darker exposure of this same autoradiograph confirns the presence of the same-sized messages in all the tissues.
SCAMPs are distributed throughout intracellular pathways that function in constitutive recycling of cell surface receptors. Based on the findings of others (e.g. Stoorvogel et al., 1991; Dunn and Maxfield, 1992; Green and Kelly, 1992), these pathways mainly comprise early endocytic compartments but may also include late endocytic compartments and the trans-Golgi network. Much of the SCAMP immunostaining observed in the soma and dendrites of highly differentiated neurons in brain tissue (Figure 3a) may be associated with analogous pathways as the network of endocytic pathways in cultured hippocampal neurons shows the same sort of distribution (Parton et al., 1992). We strongly suspect that this distribution is shared by all cell types, irrespective of their specialization for regulated secretion. In cell types that have regulated recycling pathways, it is now clear in several instances that the distribution of SCAMPs also includes the regulated pathways. Indeed, we have demonstrated the presence of SCAMPs not only in neuronal synaptic vesicles (Brand et al., 1991 and Figure 3 herein) but also in exocrine and endocrine secretion granules (Brand et al., 1991 and unpublished observations) and adipocyte glucose transporter vesicles (Laurie et al., 1993). Thus a distinguishing feature of the SCAMPs in comparison with other proteins that have been identified in various recycling pathways is that they are not restricted to subsets of recycling carriers. In this respect, they contrast with the transferrin and mannose-6-phosphate/insulin-like growth factor II receptors, which are both excluded from regulated transport pathways. Also they contrast with other synaptic vesicle proteins, which are almost exclusively concentrated in the regulated pathway (Figure 3; DeCamilli and Jahn, 1990). As a consequence of these multiple associations, we hypothesize that the SCAMPs are markers throughout the general cell surface recycling system. This system encompasses all post-Golgi pathways functioning in macromolecular export and import and in the recirculation of cell surface components. In considering the extent to which the SCAMPs collectively are restricted in their distribution, we have already noted that they either are not present or are barely detected in rough microsomal, lysosomal and plasma membrane fractions (Brand et al., 1991). We take these observations
SCAMP distribution and structure
indicate that SCAMPs probably do not reside at or beyond the boundaries of the general surface recycling system. The apparent absence of SCAMPs in lysosomes could reflect exclusion from or efficient turnover in this compartment. It is notable in that it distinguishes the SCAMPs from the vacuolar proton pump which is distributed in both recycling membranes and lysosomes. Our deduction that SCAMPs do not reside in the plasma membrane is based primarily on the analysis of fractions deriving from exocrine acinar cells where most post-Golgi membrane trafficking is regulated and constitutive cell surface trafficking is limited. The immuofluorescent labeling in fibroblasts (Figure 2) would also argue that SCAMPs are concentrated at intracellular sites and not at the cell surface. Neither finding, however, excludes the possibility that there is trafficking of SCAMPs to the cell surface. If this is the case, as we suspect, then the low steady state levels at the surface may reflect efficient reinternalization during recycling and detection at the cell surface may be contingent on identifying a labelable ectodomain or enhancing export relative to import in a regulated transport pathway. Recent studies in adipocytes suggest that SCAMPs may indeed be transported to the cell surface where they are efficiently reinternalized (Laurie et al., 1993). The prospect that SCAMPs are efficiently sequestered and internalized from the cell surface has stimulated our interest in internalization signals. Although structural determinants of internalization on polytopic membrane proteins have not been examined in much detail, several tyrosines present in SCAMP 37 (e.g. residues 37, 73, 134, 207 and 289) may be part of exposed tight-turn motifs that fit the characteristics of clustering/internalization signals in type I and type II membrane proteins (Hopkins, 1992; Vaux, 1992). In view of the close correlation of SCAMP and transferrin distributions in fibroblasts (Figure 2), we suspect that SCAMPs are not localized to the stacked cisternae of the Golgi complex. However, as the trans-Golgi network is an important station for cell surface recycling (Green and Kelly, 1992; Hopkins, 1992) and is the origin of the regulated secretory pathway, we suggest that SCAMPs are likely to mark membranes in transit at this site. For the present, the SCAMPs appear to be distinguished by being distributed throughout, yet confined to, the collection of pathways that communicate with the cell surface. to
SCAMP isoforms and subsorting We have only limited insight regarding the structural basis for the multiple SCAMP isoforrns identified by Western blot and Northern analyses. The sequences of two distinct cDNAs already suggest that alternative splicing may distinguish some of the different forms. Post-translational modifications other
glycosylation (ruled out previously by Brand et al., 1991) may also contribute to the diversity of SCAMPs. It is possible that the unusually broad distribution of SCAMPs reflects the aggregate distribution of several isoforms, each containing the epitope of SG7C 12 and marking a subset of recycling pathways. Notably, however, our findings to date using cell fractionation and Western blotting techniques indicate that the different-sized SCAMPs (37 000, 39 000) are largely codistributed (Brand et al., 1991; Laurie et al., than
1993) and thus may function together. Nevertheless, the issues of isoform diversity and subsorting are high priorities for ongoing investigation. Further, we acknowledge the
possibility that proteins related to the SCAMPs but lacking the SG7C 12 epitope could be present in ER/Golgi-related transport pathways that are not part of the cell surface recycling system. SCAMP 37 structure and function As a common denominator in various branches of the cell surface recycling system in all cells, the SCAMPs are likely to perform functions that are shared by each of the pathways and these roles could be manifest in both import and export directions. The modular organization of SCAMP 37 with motifs potentially involved in calcium binding, intermolecular interactions and even intermembrane interactions being distributed along its entire length makes this interesting new family of polypeptides a candidate for operating at many different levels of recycling processes. Thus SCAMPs may function in forming, targeting or translocating vesicles, assembling membrane fusion machinery or in expediting the fusion process itself. Moreover, it will be important to explore ways in which SCAMPs might interact with other proteins or macromolecular assemblies that have been implicated in these processes.
Materials and methods Subcellular fractionation Total cellular membranes for Western blotting were isolated by homogenizing rat tissues (5-10% w/v) in 0.27 M sucrose containing 0.25 mM phenylmethylsulphonyl fluoride (PMSF, freshly added from an ethanol stock) and 0.2 sg/ml each of pepstatin, antipain and leupeptin, centrifuging the homogenates at 600 g for 10 min and spinning the resulting post-nuclear supernatants (PNS) for 1 h at 190 000 g to pellet all membranes. Membranes were rehomogenized in 50 mM NaHCO3 and repelleted using the same conditions. Total protein was assayed using a modified Lowry procedure (Markwell et al., 1978). Crude synaptic vesicle fractions from rat brain were used for velocity sedimentation on glycerol gradients. The P3 pellet prepared according to Hell et al. (1988) was cleared of large particulates by centrifugation (10 min at 26 000 g) and the supernatant (200 Ag protein) was layered on a 5-25 % glycerol gradient containing 150 mM NaCl, 10 mM HEPES, 1 mM EGTA and 0.1 mM MgCl2, and spun for 110 min at 175 000 g (Clift-O'Grady et al., 1990; Cameron et al., 1991). Twelve 1 ml fractions were collected manually and subjected to SDS -PAGE and immunoblotting.
Western blotting and immunoprecipitation Western blotting on nitrocellulose and immunoprecipitation using the antiSCAMP monoclonal antibody (SG7C12) were performed as described previously by Brand et al. (1991). Synaptophysin was immunoblotted on nitrocellulose (Blotto procedure) with a rabbit polyclonal antibody (gift of Dr R.Jahn) and a secondary antibody (goat anti-rabbit IG) conjugated to horseradish peroxidase. The Enhanced Chemiluminescence (ECL) method was used to develop both the SCAMP and synaptophysin immunoblots.
Cell culture and immunocytochemistry NRK cells were cultured in DMEM with 5 % fetal calf serum. For FITCTF uptake experiments, cells were washed rapidly three times and incubated (37°C) for 30 min in serum-free medium followed by a 60 min incubation (37°C) in 15 Ag/ml FITC-TF (Molecular Probes, Eugene, OR). Cells were washed rapidly in ice cold PBS, fixed in 2% formaldehyde in 0.1 M potassium phosphate, permeabilized in 0.2% Triton X-100 and blocked in 16% normal goat serum (NGS) in 10 mM Tris and 150 mM NaCl, pH 7.4 (TBS). Coverslips were then incubated successively in 125 Agg/ml affinity purified SG7C12 and 7.5 jig/ml affinity purified goat anti-mouse IG conjugated to Cy3 (Jackson Immunoresearch, West Grove, PA). Coverslips were washed five times in TBS after each antibody incubation. Immunofluorescent images were collected using a Bio-Rad MRC600 Laser Sharp confocal box adapted to a Nikon Diaphot microscope equipped with a 40 x UV apochromatic objective (NA 1.3) and a microcomputer outfitted with COMOS software (Bio-Rad). Specimens were illuminated with an argon/krypton laser using excitation/emission combinations of 488 nm/ 515 nm (fluorescein) and 568 nm/585 nm (Cy3) provided by a standard
3759
S.H.Brand and J.D.Castle set of Bio-Rad filters. Appropriate control samples were viewed with these optics to show that the various fluorophores were detected only in the appropriate channels. The confocal detector aperture was set at 3/4 maximum for low intensity signals and images were Kalman averaged and enhanced using a standard adaptive histogram procedure. Illustrations were prepared using Adobe Photoprint software. Rat cerebellar tissue was fixed in situ by transcardial perfusion with chilled 4% formaldehyde, 0.1 M potassium phosphate, pH 7.4 followed by immersion of dissected tissue (2 h, 4°C) in the same fixative. Tissue pieces (1 mm thick) were dehydrated and infiltrated with plastic at 4°C for 1 h in 95% ethanol, for 1 h in 1:1 95% ethanol:methacrylate mix and for 1 h in 100% methacrylate mix (methacrylate mix is methyl methacrylate:butyl methacrylate, 2:8 with 5% dibutylphthalate and 0.25% w/v benzoin methyl ether). Samples were polymerized by UV illumination overnight at 4°C. Sections (8ym) were mounted on gelatin coated slides, deplasticized in xylene (1.5 min) and rehydrated through an ethanol series to water. The sections were then rinsed in 10 mM Tris and 150 mM NaCl, pH 7.4 (TBS) and blocked by incubation with 3% NGS in TBS for 30 min at room temperature. Antibody staining was carried out by successive incubation in: (i) either purified anti-SCAMP monoclonal antibody SG7C12 (260 ,tg/ml) or synaptophysin polyclonal antiserum diluted (1/100) in 3% NGS or no primary antibody, overnight at 4°C; (ii) goat anti-mouse IgG or goat anti-rabbit IgG (as appropriate; both from Jackson Immunoresearch, West Grove, PA) diluted 1/100 in 1% NGS, 30 min at room temperature; (iii) mouse or rabbit peroxidase anti-peroxidase (Stemberger Monoclonals, Baltimore, MD) diluted 1/50 in 1% NGS, 30 min room temperature. Sections were washed three times for 2 min with TBS after each antibody incubation. The peroxidase reaction was carried out by incubation in 0.05% DAB, 0.01 % H202 and 0.1 M citric acid/ammonium acetate buffer, pH 5.8 for 10 min at room temperature. Immunopurification of SCAMP SCAMP 37 was purified from detergent-solubilized rat brain membranes by affinity chromatography on immobilized anti-SCAMP antibody that had been coupled to Affigel-Hz (Bio-Rad Inc., Richmond, CA). Rat brains (frozen upon dissection in liquid N2 and stored at -80°C) were used as a source of antigen. They were homogenized (17% w/v) in 0.25 M sucrose, 1 mM EDTA, = 0.2 /ig/ml (each) of leupeptin, antipain and pepstatin and 0.25 mM PMSF using a Tekmar Tissumizer (15 s, half speed) followed by a Teflon/glass homogenizer (5 strokes, = 300 r.p.m.). The homogenate was filtered through four layers of cheesecloth and centrifuged for 15 min at 2000 gav The resulting supernatant was spun 4 h at 87 000 gav to pellet total membranes. The membranes were resuspended in homogenization medium (20-30 mg/ml protein) and stored in aliquots at -80°C until use. Brain membranes (180 mg protein) were solubilized at 0.9 mg/ml in 1% Triton X-114, 10 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA and proteinase inhibitors (as above). After stirring for 40 min at 4°C, insoluble material was pelleted (30 min at 46 000 gay). The supernatant was warmed to 25°C for 8 min to induce phase separation and then spun at the same temperature for 10 min at 2000 gaw The lower Triton micelle phase was restored to the original volume using chilled 10 mM Tris, 150 mM NaCl and 1 mM EDTA (pH 7.4) and loaded at 4°C onto the anti-SCAMP imnmunoaffinity column (4 ml gel bed; flow rate 40 ml/h). The solution was passed twice over the column. The column was then washed in succession with 80 ml (20 vol) 1% Triton X-1 14 in 10 mM Tris, 150 mM NaCl and 1 mM EDTA (pH 7.4), 40 ml 1% CHAPS in 10 mM Tris, 500 mM NaCl and 1 mM EDTA (pH 7.4), and 20 ml 1% octyl-O-Dglucopyranoside in 5 mM Tris, 150 mM NaCl, 0.5 mM EDTA (pH 7.4). Antigen was then eluted using 0.2% sodium deoxycholate (DOC) in 25 mM triethylamine, pH 11.5 (Mellman and Unkeless, 1980). Eluted fractions (1 ml each) were immediately adjusted to pH 8 by adding 50 1l of 2.3 M Tris pH 7.3. Pooled fractions of purified antigen were concentrated (Centricon-10) at 15°C and exchanged using 0.2% DOC, 5 mM Tris (pH 8.2) to a final composition of 0.2% DOC, 10 mM triethylamine and 30 mM Tris and a final volume of 0.6-0.8 mi per preparation. The samples were concentrated 10-fold further by lyophilization (but not taken to dryness) and diluted by adding 0.5 vol of solubilization buffer containing 10% SDS. Samples were reduced with 15 mM dithiothreitol (3 min at 80°C) and carboxamidomethylated with 40 mM iodoacetamide (30 min, room temperature in the dark). Subsequently, they were electrophoresed (Laemmii, 1970) on 0.75 mm thick polyacrylamide gels (11% resolving gel) and processed further for either microsequencing (see below). Coomassie blue and silver staining or Western blotting depending on the experiment. -
-
Microsequencing of SCAMP tryptic peptides Affinity-purified SCAMP was electrophoresed (as above) on polyacrylamide gels prepared from ultrapure reagents: polyacrylamide, bis-acrylamide and
3760
glycine from Bio-Rad (Richmond, CA), Tris base from ICN Biomedical (Cleveland, OH), SDS (99% pure) from BDH Chemicals (Poole, UK) and deionized glass-distilled water. Following electrophoresis, the gels were fixed, Coomassie-stained and processed for in situ proteolytic digestion using a procedure (Ward et al., 1990) modified by the William Keck Biotechnology Research Laboratory, Yale School of Medicine (K.Stone and K.Williams, personal communication). Bands containing 14-15 ,g SCAMP (estimated by comparison to Coomassie-stained standards) were excised along with blanks of the same electrophoretic mobility and were incubated for 24 h at 37°C with trypsin-TPCK (Worthington Biochemicals, Freehold, NJ) at 1:25 w/w ratio of enzyme to protein (final volume = 200 Al). After digestion, peptides that had not eluted were extracted by successive washes with 0.1 M NH4HCO3 (8 h) and 2 M urea in 0.1 M NH4HCO3 (24 h). Dried extracts were dissolved in H20, filtered (0.22 Am) and used for amino acid analysis (to determine amount of protein) and HPLC. Ninety-five picomoles of protein were recovered in the peptide digest corresponding to a 25% yield of the SCAMP polypeptide estimated to be present in the gel. Narrow bore HPLC of the digests [SCAMP, blank and carbonic anhydrase (positive control for digestion)] was performed on a Vydac C18 reverse phase column (2.1 m x 25 cm) in the William Keck Biotechnology Research Laboratory (Stone et al., 1992). Four peptide peaks were selected for sequencing. cDNA cloning and sequencing Two overlapping oligonucleotides were made (Operon, Alameda, CA) to the opposite ends of the longest tryptic peptide fragment of SCAMP 37 (Figure 3c) using most frequent codon usage predictions (Lathe, 1985). The sequences of these oligonucleotides are: (5'-ATG-CCC-AAT-GTG-CCCAAC-ACC-CAG-CCT-GCC-ATC-ATG-AAG-CCC-ACA-3' and 5'-TTGGTG-ATC-TGG-GTG-TAG-GCA-GGA-TGC-TCC-TCT-GTG-GGCTT-3'). The two oligonucleotides were mixed, allowed to hybridize at their 3'-ends and opposite strands were filled in using the Klenow fragment of DNA polymerase in a reaction mixture containing 60 nM of each oligonucleotide, 1 x nick translation buffer (BRL, Grand Island, NY), 0.2 mM dATP, 0.2 mM dGTP, 0.2 mM dTTP, 2 /tCi/Al [a-32P]dCTP and 0.1 units/4l Klenow. Following incubation for 2 h at room temperature, 150 Al of TE (10 mM Tris-HCl, pH 7.4 and 1 mM EDTA, pH 8.0) were added and the mix was spun through a G25 Sephadex column to remove unincorporated [ca-32P]dCTP from the labeled double stranded probe. Probes containing > 1 x 108 c.p.m./,ug were used to screen one million recombinant phage from a rat brain Xgtl 1 cDNA library (XZAPII, Stratagene, La Jolla, CA). The probe was boiled for 10 min and cooled on ice prior to hybridization. Prehybridization and hybridization were carried out at 42°C in a solution containing 5 x SSPE, 0.2% SDS, 10 x Denhardts, 0.05% sodium pyrophosphate and 100 ,ug/mi boiled salmon testis DNA. Positive clones were purified by several rounds of screening and the plasmids containing the cDNA were excised from the XZAP phage according to the manufacturer's instructions. The clones were sized and mapped using restriction digestion. Sequencing using the dideoxy method (Sanger et al., 1977; Sequenase Version 2, US Biochemicals) was performed on nested deletions that were prepared using the Erase-a-base kit, (Promega, Madison, WI). In vitro transcription and translation SCAMP 37 cDNA was transcribed and translated in vitro using rabbit reticulocyte lysate (Promega, Madison, WI) in the presence or absence of canine pancreatic microsomes (Anderson and Blobel, 1983). Samples were either used directly for SDS-PAGE or immunoprecipitated with antiSCAMP (SG7C12). Labeled bands were detected by fluorography or electrotransferred to nitrocellulose for direct comparison with Western blots of SCAMP antigen in rat brain membranes.
Northern analysis Total RNA was isolated from rat tissues by acid guanidinium thiocyanate-phenol-chloroform extraction (Chomczynski and Sacchi, 1987). RNA (10 jig each) was electrophoresed on a 1.5% agarose formaldehyde gel and transferred onto Magnagraph nylon (Micron Separations Inc. Fisher, Pittsburgh, PA) in 10 x SSC using standard procedures. Following transfer, the blot was soaked for 5 min in 2 x SSC and baked for 2 h at 80°C. Prehybridization was carried out at 65°C for 10 min in 250 mM NaHPO4 pH 7.2, 250 mM NaCl, 5% SDS, 10% PEG 8000, 1 mM EDTA and 100 jig/ml salmon testis DNA. Hybridization was in the same buffer at 65°C for 18 h. The 1.9 kb SCAMP cDNA clone was nick-translated (Rigby et al., 1977) and used to probe the Northern blot (Amisino, 1986). The blot was washed at high stringency: 1 x 10 min with 50 mM NaHPO4, 0.5% SDS at room temperature, 2 x 20 min with 25 mM NaHPO4, 0.5% SDS at 65°C and exposed to x-ray film.
SCAMP distribution and structure
Acknowledgements We are grateful to Dr Kevin Lynch and members of his laboratory and Dr Anna Castle for advice and numerous valuable discussions, Dr Jim Lechleiter for use of his confocal microscope and help in preparation of the fluorescent images, Dr Charles Little for use of his video processing facilities, Dr Marilyn Fisher for advice on immunocytochemistry, Dr Reinhard Jahn for the gift of anti-synaptophysin antibody, Michael Palladino for advice on Northern blotting and Yun Shim for assistance in preparing the illustrations. These studies were supported by an NIH grant (DE09655) and by the University of Virginia DERC Pilot and Feasibility
Program.
References Abdelhaleem,M.M., Hatskelzon,L., Dala,B.I., Gerrard,J.M. and Greenberg,A.H. (1991) J. Immunol., 147, 3053-3059. Alder,J., Lu,B., Valtorta,F., Greengard,P. and Pu,M.-M. (1992a) Science, 257, 657-661. Alder,J., Xie,Z.-P., Valtorta,F., Greengard,P. and Poo,M.-M. (1992b) Neuron, 9, 759-768. Amisino,R.M. (1986) Anal. Biochem., 152, 304-307. Anderson,D.J. and Blobel,G. (1983) Methods Enzymol., 96, 111-120. Bennett,M.K., Calakos,N., Kreiner,T. and Scheller,R.H. (1992) J. Cell Biol., 116, 761-775. Berg,J.M. (1986) Science, 232, 485-487. Brand,S.H., Laurie,S.M., Mixon,M.B. and Castle,J.D. (1991) J. Biol. Chem., 266, 18949-18957. Brose,N., Petrenko,A.G., Sudhof,T.C. and Jahn,R. (1992) Science, 256, 1021-1026.
Cain,C.C., Trimble,W.S. and Leinhard,G.E. (1992) J. Biol. Chem., 267, 11681 -11684.
Cameron,P.L., Sudhof,T.C., Jahn,R. and De Camilli,P. (1991) J. Cell Biol., 115, 151-164. Chomczynski,P. and Sacchi,N. (1987) Anal. Biochem., 162, 156-159. Chou,P.Y. and Fasman,G.D. (1978) Annu. Rev. Biochem., 47, 251-276. Clift O'Grady,L., Linstedt,A.D., Lowe,A.W., Grote,E. and Kelly,R.B.
(1990) J. Cell Biol., 110, 1693 -1703. and Jahn,R. (1990) Annu. Rev. Physiol., 52, 625-645. Disdier,M., Morrissey,J.H., Fugate,R.D., Bainton,D.F. and McEver,R.P. (1992) Mol. Biol. Cell, 3, 309-321. Dunn,K.W. and Maxfield,F.R. (1992) J. Cell Biol., 117, 301-310. Elferink,L.A., Peterson,M.R. and Scheller,R.H. (1993) Cell, 72, 153-159. Fitzgerald,L.A., Poncz,M., Steiner,B., Rall,S.C.,Jr, Bennet,J.S. and Phillips,D.R. (1987) Biochemistry, 26, 8158-8165. Gariepy,J. and Hodges,R.S. (1983) FEBS Len., 160, 1-6. Gamier,J., Osguthorpe,D.J. and Robson,B. (1978) J. Mol. Biol., 120,
DeCamilli,P.
Navone,F., Jahn,R., Di Gioia,G., Stukenbrok,H., Greengard,P. and DeCamilli,P. (1986) J. Cell Biol., 103, 2511-2527. Parton,R.G., Simons,K. and Dotti,C.G. (1992) J. Cell Biol., 119, 123-137. Perin,M.S., Brose,N., Jahn,R. and Sudhof,T.C. (1991) J. Bio. Chem., 266, 623-629. Petrenko,A.G., Perin,M.S., Davletov,B.A., Ushkaryov,Y.A., Geppert,M. and Sudhof,T.C. (1991) Nature, 353, 65-68. Preston,R.A., Manolson,M.F., Becherer,K., Weidenhammer,E., Kirkpatrick,D., Wright,R. and Jones,E.W. (1991) Mol. Cell Biol., 11, 5801 -5812.
Regnier-Vigouroux,A., Tooze,S.A. and Huttner,W.B. (1991) EMBO J., 10, 3589-3601. Rigby,P.W.J., Diekmann,M., Rhodes,C. and Berg,P. (1977) J. Moi. Biol., 113, 237-251. Robinson,J.S., Graham,T.R. and Emr,S.D. (1991) Mol. Cell Bio., 12, 5813-5824.
Sanger,F., Nicklen,S. and Coulsen,A.R. (1977) Proc. NatlAcad. Sci. USA, 74, 5463-5467. Schiavo,G., Benfenati,F., Poulain,B., Rossetto,O., Polverino de Laureto,P., DasGupta,B.R. and Montecucco,C. (1992) Nature, 359, 832-835. Schwabe,J.W.R. and Rhodes,D. (1991) Trends Biochem. Sci., 16, 291-296. Sollner,T., Whiteheart,S.W., Brunner,M., Erdjument-Bromage,H., Geromanos,S., Tempst,P. and Rothman,J.E. (1993) Nature, 362, 318-324.
Stone,K., McNulty,D., LoPresti,M., Crawford,J., DeAngelis,R. and Williams,K. (1992) Tech. Prot. Chem., 3, 23-34. Stoorvogel,W., Strous,G.J., Geuze,H.J., Ooshot,V. and Schwartz,A.L. (1991) Cell, 65, 417-427. Sudhof,T.C. and Jahn,R. (1991) Neuron, 6, 665-677. Thomas,L., Hartung,L., Langosch,D., Rehm,H., Bamberg,E., Franke,W.W. and Betz,H. (1988) Science, 242, 1050-1052. Trimble,W.S. and Scheller,R.H. (1988) Trends Neurosci., 11, 241-242. Vaux,D. (1992) Trends Cell Biol., 2, 189-192. von Heijne,G. (1985) J. Mol. Biol., 184, 99-105. Ward,L., Reid,G., Moritz,R. and Simpson,R. (1990) In Villafranca,J. (ed.), Current Research in Protein Chemistry: Techniques, Structure and Function. Academic Press, New York, pp. 179-190. Weisman,L.S. and Wickner,W. (1992) J. Biol. Chem., 267, 618-623. Wessels,H.P. and Spiess,M. (1988) Cell, 55, 61-70. White,J.M. (1992) Science, 258, 917-924. Yamashiro,D.J., Tycko,B., Fluss,S.R. and Maxfield,F.R. (1984) Cell, 37, 789-800.
Received on 29 March, 1993; revised on 18 June, 1993
97-120.
Gray,T.M. and Matthews,B.W. (1984) J. Mol. Biol., 175, 75-81. Green,S.A. and Kelly,R.B. (1992) J. Cell Biol., 117, 47-55. Hell,J.W., Maycox,P.R., Stadler,H. and Jahn,R. (1988) EMBO J., 7, 3023-3029.
Hopkins,C.R. (1992) Trends Biochem. Sci., 17, 27-32. Hopkins,C.R. and Trowbridge,I.S. (1983) J. Cell Biol., 97, 508-521. Hopkins,C.R., Gibson,A., Shipman,M. and Miller,K. (1990) Nature, 346, 335-339.
Johnston,P.A. and Sudhof,T.C. (1990) J. Biol. Chem., 265, 8869-8873. Johnston,P.A., Cameron,P.L., Stukenbrok,H., Jahn,R., DeCamilli,P. and Sudhof,T.C. (1989) EMBO J., 8, 2863-2872. Kantrowitz,E.R., Pastra-Landis,S.C. and Lipscomb,W.N. (1980) Trends Biochem. Sci., 5, 150-153. Knaus,P., Marqueze-Pouey,B., Schaer,H. and Betz,H. (1990) Neuron, 5, 453-462.
Koedam,J.A., Cramer,E.M., Briend,E., Furie,B., Furie,B.C. and Wagner,D.D. (1992) J. Cell Biol., 116, 617-626. Kozak,M. (1984) Nucleic Acids Res., 12, 857-873. Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105-132. Laemmli,U.K. (1970) Nature, 227, 680-685. Landschulz,W.H., Johnson,P.F. and McKnight,S.L. (1988) Science, 240, 1759-1764.
Lathe,R. (1985) J. Mol. Biol., 183, 1-12. Laurie,S.M., Cain,C.C., Leinhard,G.E. and Castle,J.D. (1993) J. Biol. Chem., in press. Markwell,M.A.K., Haas,S.M., Beiber,L.L. and Tolbert,N.E. (1978) Anal. Biochem., 37, 206-210. Mellman,I.S. and Unkeless,J.C. (1980) J. Exp. Med., 152, 1048-1069.
3761
