Vol.
170,
July
31,
No.
2,
1990
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
1990
Pages
A COMMON GOLGI
PEPTIDE
APPARATUS:
STRETCH
AMONG
STRUCTURAL
ENZYMES
SIMILARITY
LOCALIZED
OF
879-882
TO
THE
GOLGI-ASSOCIATED
GLYCOSYLTRANSFERASES Brad Bendiak The Biomembrane
Institute
and Department
of Pathobiology,
201 Elliott Avenue West, Seattle, Washington Received
.June 22,
University of Washington, 98119
1990
Summary: A common peptide motif has been discovered among a series of Golgi-localized glycosyltran.sferases. The peptide stretch, (Ser/Thr)-X-(Glu/Gln)-(ArgLys), always occurs near a hydrophobic domain close to the N-terminus of these enzymes which is believed to anchor them to the membrane lipid bilayer (Paulson and Colley, J. Biol. Chem., 264, 1761517618, 1989). The finding that this similar peptide motif is not associated with catalytic activity of these enzymes, and its presence near the hydrophobic domain suggest that the stretch may be involved in localization of these enzymes to the Golgi apparatus. Q1990 Academic
Pres:;,
Inc.
In eukaryotes, different
cellular membranes
proteins which enable each membrane not haphazardly intermingled,
are largely defined by characteristic
to carry out specific functions.
These proteins are
but come to be localized with high probability
to specific
organelles or suborganellar regions. A problem that remains unresolved is the mechanism by which many proteins come to reside in their membranes on a relatively long term basis, often referred to as “sorting” (1). As all proteins are initially translated from mRNA which can only define the linear sequence of a number compartmentalization
of amino acids, the information
for
of a particular protein to a specific organelle must reside, in essence,
in that protein’s sequence. The Golgi apparatus plays a central role in the traffic of integral membrane proteins through the cell (2). Even though many of these proteins are temporary residents in the Golgi apparatus en route to other destinations, there is a set of proteins which may be considered
permanent
residents of Golgi membranes,
and are known
to be marker
enzymes. Among these proteins are a number of the glycosyltransferases (3,4). The active sites of these enzymes are located in the lumen of the Golgi compartment, participate
in the assembly of the carbohydrate
where they
portions of glycoproteins and glycolipids. 0006-291x/90
879
$1.50
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol.
170,
No.
2,
1990
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
The structures of several Golgi-localized glycosyltransferases have been investigated; eight full-length sequences of five enzymes have been reported from different mammalian sources (5-13).
Extensive analysis of their sequences has now revealed a common stretch
which, it is postulated,
may be involved in localization
of these enzymes to the Golgi
apparatus. RESULTS AND DISCUSSION A
short
sequence
glycosyltransferases
reported
motif
was
discovered,
common
to
to date, which is shown in Fig. 1.
the
Golgi-localized
Although
exhaustive
computer searches have been performed by us (10) and others (5-9, 11-13) in the past, the common stretch was only found by consideration of the similar structures of Gln and Glu and the similar positive charge of Arg and Lys residues. The conserved stretch, (Ser/Thr)X-(Gln/Glu)-(Arg/Lys),
occurs adjacent to a nearby, more N-terminal
in four of the enzymes. Most important this common peptide hydrophobic domain.
for an understanding
Met-Pro sequence
of the possible function of
stretch is its position on the various glycosyltransferases
near a
In this regard, the overall domain structure of these enzymes is very
relevant, and Paulson and Colley (4) have presented a general structural hypothesis very similar to that shown in Fig. 2a. Essential common features of Golgi-localized
glycosyltransferases
are a catalytic
domain comprising much of the molecule toward the C-terminus, a “stem” region which is very susceptible to proteolytic cleavage and is probably solvent-exposed, and a hydrophobic domain near the N-terminus, which is probably embedded in the Golgi lipid bilayer. been postulated
(4) that the hydrophobic
It has
domain spans the membrane.
The common peptide stretch, in the enzymes examined to date, can occur either more toward the N-terminal but always occurs near it.
or C-terminal
of the protein than the hydrophobic
If this domain
actually
spans the lipid
domain,
bilayer
in all
glycosyltransferases, this would place the common peptide motif either in the cytoplasm, or in the Golgi lumen.
Met Met
Pro Pro
Lys Gly
1 Ile va1 GUY '=Y Ala Ala GUY
Murine 011-3 Galtransferase (5) Rat CY2-6 Sialyltransferase (6) Human 1X1-3 GalNActransferase "A" (7) "8" (8) Human al-3 Galtransferase Human, bovine 01-4 Galtransferase (9,101 Murine 01-4 Galtransferase (11,12) Bovine 011-3 Galtransferase (13)
Figure 1. Alignment of glycosyltransferase peptide motif. Numbers in brackets indicate references.
880
sequences having the common
structural
Vol.
170, No. 2, 1990
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Figure 2. Schematic representations of glycosyltransferaseswith the common structural motif (boxed) either C-terminal (a) or N-terminal (b) to the hydrophobic domain. It would be reasonable to surmise that if the common peptide retention
stretch were a
signal, that it should be located on the same side of the membrane
glycosyltransferases.
Such a situation could arise if both the N- and C-terminal
for all portions
of the proteins were located in the Golgi lumen, as shown in Fig. 2b. If this were the case, the hydrophobic
domain would comprise a hairpin loop, embedded in the membrane.
presence of hairpin precedented
loops in proteins
having one or more hydrophobic
The
domains
is
(14).
Also relevant is the fact that the similar peptide stretch occurs in enzymes which differ
in
their
catalytic
acetylgalactosamine,
properties
(one
transfers
one
transfers
N-terminal
Of the glycosyltransferases which have
sequences have established that proteolysis
of the initial
translation products occurs. This appears to be an artefact of purification. the polypeptide
N-
and three transfer galactose). Moreover, the common peptide stretch
is not associated in any way with catalytic activity. been purified,
sialic acid,
Cleavage of
occurs in the stem region, which releases the C-terminal
enzyme from the hydrophobic
domain.
portion of the
The freed C-terminal domain is a soluble enzyme
which retains full catalytic activity and substrate specificity. To date, N-terminal of the soluble enzyme forms (6,10,15-17) indicate that the tetrapeptide remain with the catalytic portion
of these enzymes after proteolysis
remains with the hydrophobic domain).
motif does not (but presumably
Therefore, the common peptide stretch cannot be
required for catalysis, and its location near the hydrophobic be involved in retention
sequences
domain suggests that it could
of these glycosyltansferases to the Golgi apparatus.
While it is also reasonable to postulate
many other functions for the common
peptide motif, its possible role as a Golgi retention signal cannot be ignored.
Of course,
other localization mechanisms may also exist, and this stretch may not be common to all proteins
resident
in the Golgi
apparatus,
or even to all glycosyltransferases.
881
The
Vol.
170,
No.
2,
1990
transferases reported
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
up to now synthesize the terminal structures of the carbohydrate
moieties of glycoproteins
and/or glycohpids; those Golgi-localized
enzymes which transfer
sugars nearer to the core of the carbohydrate units may be located in different cisternae of the Golgi stack (1). ACKNOWLEDGMENTS I thank Drs. S. Hakomori, comments in preparation
F. Yamamoto,
and S. Pawar for critical evaluation and helpful
of the manuscript.
Studies were supported by the Biomembrane
Institute. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
Pfeffer, S.R. and Rothman, J.E. (1987) Ann. Rev. Biochem. 56, 829-852. Farquhar, M.G. and Palade, G.E. (1981) J. Cell Biol. 91, 77s-103s. MorrC, D.J., Merlin, L.M., and Keenan, T.W. (1969) Biochem. Biophys Rex Commun. 37, 813-819. Paulson, J.C. and Colley, K.J. (1989) J. Biol. Chem. 264, 1761517618. Larsen, R.D., Rajan, V.P., Ruff, M.M., Kukowska-Latallo, J., Cummings, R.D., and Lowe, J.B. (1989) Proc. Natl. Acad. Sci. lJ.SX 86, 8227-8231. Weinstein, J., Lee, E.U., McEntee, K., Lai, P.-H., and Paulson, J.C. (1987) J. Biol. Chem. 262, 17735-17743. Yamamoto, F., Marken, J., Tsuji, T., White, T., Clausen, H., and Hakomori, S. (1990) J. Biol. Chem. 265, 1146-1151. Yamamoto, F., Clausen, H., White, T., Marken, J., and Hakomori, S. (1990) Nature 345, 229-233. Masri, K.A., Appert, H.E., and Fukuda, M.N. (1988) Biochem. Biophys. Res. Commun. 157, 657-663. D’Agostaro, G., Bendiak, B., and Tropak, M. (1989) Eur. J. Biochem. 183, 211-217. Shaper, N.L., Hollis, G.F., Douglas, J.G., Kirsch, I.R., and Shaper, J.H. (1988) J. Biol. Chem. 263, 10420-10428. Nakazawa, K., Ando, T., Kimura, T., and Narimatsu, H. (1988) J. Biochem. (Tokyo) 104, 165-168. Joziasse, D.H., Shaper, J.H., Van den Eijnden, D.H., Van Tunen, A.J., and Shaper, N.L. (1989) J. Biol. Chem. 264, 14290-14297. Jennings, M.L. (1989) Ann. Rev. Biochem. 58, 999-1027. Appert, H.E., Rutherford, T.J., Tarr, G.E., Thomford, N.R., and McCorquodale, D.J. (1986) Biochem. Biophys. Res. Commun. 138, 224-229. Navaratnam, N., Ward, S., Fisher, C., Kuhn, N.J., Keen, J.N., and Findlay, J.B.C. (1988) Eur. J. Biochem. 171, 623-629. Clausen, H., White, T., Takio, K., Titani, K., Stroud, M., Holmes, E., Karkov, J., Thim, L., and Hakomori, S. (1990) J. Biol. Chem. 265, 1139-1145.
882