Purification of yolk protein 2 of Drosophila melanogaster and identification of its site of tyrosine sulfation.

We have identified the site of tyrosine sulfation in an insect secretory protein, yolk protein 2 of Drosophila melanogaster. Yolk proteins were purified from [35S]sulfate-labeled flies, and yolk protein 2 was separated from yolk protein 1 and yolk protein 3 by preparative two-dimensional polyacrylamide gel electrophoresis. After digestion of yolk protein 2 with trypsin and reversed-phase high performance liquid chromatography, the sulfate label was recovered in two distinct sulfopeptides which, however, had identical NH2-terminal sequences and contained 3 tyrosine residues each. After chymotryptic digestion of the two tryptic sulfopeptides, the sulfate label was recovered in one sulfopeptide which contained a single tyrosine residue. NH2-terminal sequencing showed that this tyrosine residue corresponded to tyrosine 172 of the yolk protein 2 precursor (Hung, M.-C., and Wensink, P. C. (1983) J. Mol. Biol. 164, 487-492) in the sequence Glu-Thr-Thr-Asp-Tyr(S)-Ser-Asn-Glu-Glu. This insect tyrosine sulfation site is very similar to the known vertebrate tyrosine sulfation sites in terms of amino acid composition and secondary structure. In the accompanying paper (Friederich, E., Baeuerle, P. A., Garoff, H., Hovemann, B., and Huttner, W. B. (1988) J. Biol. Chem. 263, 14930-14938), we report on the expression of Drosophila yolk protein 2 in mouse fibroblasts and show the in vivo sulfation of tyrosine 172 by the vertebrate tyrosylprotein sulfotransferase.

Sulfation is the most abundant modification of tyrosine residues known. In multicellular eukaryotic organisms many secretory proteins and a few membrane proteins have been shown to undergo tyrosine sulfation (for reviews, see Refs. 1-3). Sulfate transfer is catalyzed by the enzyme tyrosylprotein sulfotransferase, an integral membrane protein of the trans Golgi (4, 5). Common structural features of the sequences surrounding sulfated tyrosine residues have been noted (1-3, 6). However, with the exception of hirudin, these sequences are all from vertebrate proteins, and little is known about the evolutionary conservation of tyrosine sulfation sites.
The function of protein tyrosine sulfation has so far only been elucidated for some of the small tyrosine-sulfated peptides, but not for any of the identified larger tyrosine-sulfated * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Present address: Whitehead Institute for Biomedical Research, Nine Cambridge Center, Cambridge, MA 02142.
Recipient of Grants Hu 27513-2 and Hu 27513-3 from the Deutsche Forschungsgemeinschaft. To whom correspondence and reprint requests should be addressed.
polypeptides. In the case of small tyrosine-sulfated peptides, studies on the role of tyrosine sulfation were facilitated by the fact that peptides can be chemically synthesized in sulfated and unsulfated form for comparative analyses. Using synthetic peptides it has been shown that tyrosine sulfation is required for the hormonal activity of cholecystokinin (for review, see Ref. 7). However, in the case of larger polypeptides, chemical synthesis of the sulfated and unsulfated form of a ~l~p t i d e is not feasible. Moreover, even if such synthesis were feasible, possible roles of tyrosine sulfation of proteins prior to their secretion could not be studied using synthetic polypeptides. A highly specific approach to investigating functional aspects of tyrosine sulfation of a protein before its secretion is to express the protein in cells after rendering it unsulfatable by site-directed mutagenesis. Prerequisites for this approach are the cloning of DNA coding for the protein and the identification of its tyrosine sulfation site(s).
We have previously shown that the three yolk proteins of Drosophila melanogaster are the major tyrosine-sulfated proteins in female flies (8). Yolk protein 2 (YP2)l contains about 1 mol of tyrosine sulfate/mol of polypeptide, in contrast to YP1 and YP3 which have a higher content of tyrosine sulfate (8). Thus, YP2 is likely to contain only a single site of tyrosine sulfation. The sequence of YP2 is known from the sequence of the cloned DNA (9), which facilitates the identification of its site of tyrosine sulfation. In order to compare the structure of a tyrosine sulfation site of an insect secretory protein with that of the known tyrosine sulfation sites of vertebrate secretory proteins and as the first step towards functional studies involving site-directed mutagenesis of a tyrosine sulfation site, we report here the purification of YP2 from Drosophila flies and the identification of the site of tyrosine sulfation.

DISCUSSION
We have previously shown that YP2 of D. ~l a n o g~t e~ contains 1 mol of tyrosine sulfate/mol of polypeptide and that all isoelectric variants of this protein are sulfated to the same extent (8). This means either that all YP2 molecules are sulfated on the same tyrosine residue or that the YP2 mole-' The abbreviations used are: YP, yolk protein; PMSF, phenylmethanesulfonyl fluoride; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; TPCK, tosyl-L-phenylalanine chloromethyl ketone; HPLC, high performance liquid chromatography; TLCK, tosyl-L-lysine chloromethyl ketone.
Portions of this paper (including "Experimental Procedures," "Results," Figs. S1-S3, Tables SI-,9111, and Footnotes 3-5) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

Tyrosine Sulfation Site
in Drosophila Yolk Protein 2 cules are sulfated on various tyrosine residues in a mutually exclusive manner, each molecule containing only 1 sulfated residue. The present study shows that the former is the case since only 1 tyrosine residue of YP2, tyrosine 172, was found to be sulfated (see Fig. 1). Our results are consistent with tyrosine 172 being the exclusive site of sulfation in YP2, although they do not strictly exclude the presence of a very minor second site of sulfation. However, as will be reported elsewhere (27), the complete lack of sulfation of YP2, mutated to phenylalanine in position 172, confirms that tyrosine 172 is indeed the exclusive site of sulfation in YP2. It has been noted (4,22,23) that the sequences surrounding sulfated tyrosine residues contain multiple acidic amino acids. More recently, the sequences surrounding identified tyrosine sulfate residues in several, mostly vertebrate, proteins have been compared and analyzed in detail (1, 6). As a result of these studies, the role of acidic residues has been characterized, and additional consensus features of tyrosine sulfation sites have been outlined (I, 2, 6). It is of interest to note that the insect tyrosine sulfation site identified in the present study is very similar to the known vertebrate tyrosine sulfation sites in terms of amino acid composition and secondary structure. The hallmark of stoichiometrically sulfated tyrosines in vertebrate proteins is the presence of an acidic amino acid residue in position -1 and a total of at least 3 acidic amino acid residues between positions -5 and +5 of the sulfated tyrosine (1, 6). Tyrosine 172 is the only tyrosine residue in YP2 with this feature (see Table I). Three other tyrosine residues in YP2 (tyrosines 92, 377, 401; see Table I) are also preceded by an acidic amino acid residue in position -1 but lack the total of three acidic amino acid residues between positions -5 and +5. One tyrosine residue in YP2 (tyrosine 220; see Table I) is characterized by 3 acidic residues between positions -5 and +5, with one in position -2, but lacks an acidic residue in position -1. Since none of these other 4 tyrosine residues in YP2 are sulfated to any detectable extent, it appears that the re~uirements for the presence of acidic amino acids in tyrosine sulfation sites are very similar, if not identical, in vertebrates and insects. The lack of sulfation of tyrosine 220 in YP2 also supports the notion (see Fig.  7 in Ref. 1) that in the absence of an acidic residue in position -1, a total of 3 acidic residues between positions -5 and +5 may not be sufficient for sulfation and that 4 or more acidic residues are required, as is the case in factor X and gastrins (see Fig. 6 in Ref . 1).

Q R Y N L Q P Y E T T~~S N E E O S~R~~~E E~~T~R~R K~N G E~~~T K
An additional explanation as to why tyrosine 172 but not tyrosine 220 (despite the presence of 3 acidic residues from positions -5 to +5) is sulfated, concerns the secondary structure of tyrosine sulfation sites. The presence of turn-inducing amino acids has been proposed to be another consensus feature of tyrosine sulfation sites (1, 2). The sequence surrounding tyrosine 172 contains a total of 5 residues with significant turn-conformational preference (>1.3; Pro, Gly, Asp, Ser, Asn; Ref. 24) between positions -7 and +7 (with 2 such residues at positions -1 and +l), whereas that surrounding tyrosine 220 contains only 2 residues with turn-conformational preference between positions -7 and +7 (at positions -4 and +4). Recently, the fJ loop has been described as a novel category of protein secondary structure that occurs almost always at the protein surface (25). It is worth noting that in the sequence surrounding tyrosine 172 in YP2, all residues between positions -6 and +4 belong to the category of amino acids found preferentially in loops ( f > 1.0; see Table   3 in Ref. 25). Since such amino acids are also frequently found in the sequences surrounding identified tyrosine sulfate residues in vertebrate proteins (see Fig. 6 in Ref. l), we propose that loop structures are a typical feature of tyrosine sulfation sites that is conserved in vertebrates and insects.
The similarity in structure between the insect tyrosine sulfation site identified here and the known vertebrate tyrosine sulfation sites suggests a high evolutionary conservation of the interaction of protein substrates with tyrosylprotein sulfotransferase. In the accompanying paper (26), it is shown that tyrosine 172 of Drosophila YP2 is specifically and stoichiometrically sulfated in uiuo by a mammalian tyrosylprotein sulfotransferase after expression of the protein in mouse fibroblasts.

TABLE I
Comparison of sequences surrounding "acidic" tyrosine residues in YP2 The sequences surrounding tyrosine residues (vertical lines, position 0) with acidic amino acid residues at positions -1 or -2 are shown from positions -7 to +7. The sequence positions of tyrosine residues refer to the YP2 precursor (9). Bold letters, acidic amino acids; underlined letters, amino acids with a turn-conformational preference >1.3 (Pro, Gly, Asp, Ser, Asn; Ref. 24). A t least 3 residues Sequence with turnposition of "Acidic" tyrosine residues in YP2 Z 2 e residues conformational Sulfated  Trvntic rulfnneolidcn of YP2 -After preparative two-dimensional PAGE, sulfopeptides were gcncnttcd hy trypsiill trcatment ofgel picccs containing YP2. This resulted in an essentially quantitative elution (-95%) of the 35S-radioactivity. In order to remove gel debris and Coomassle blue and to concentrate sulfopeptides. thc tryptic eluate was subjected to chromatography on a CIS SEP-PAK canridge. A small ponlon of the 3hadioactivity did not bind to the resin and was recovered in the flow-through (Table II of  radioactwity (but a11 of the Coomassie blue) were elured in a subsequent step using 100% methanol (Table 11).
YP? revc;llcd the prcwce of<evcral sulfnpeptides (not shown). There sulfopeplider were apparently '1IPI.C of the flow-through fraction of the SEP-PAK CIX chromatography of the tryptic dngert of d e r w d fronl cont:min;ltin$ YPI rathcr than from YP2 smcc they were also rcen. in lager amounts. in thc flow-lhmurh Itlction of a trypt~c digert of YPI (data not shown). Since YPI contains -2.5 times a , much wllntc a\ YPZ (XI. the -15% of 35S-radloaclwNy found in the flow-through fraction of tryptrc digest, of YP? (Table 11) cnrrerponded to -6% of contaminating YPI.
absorbance of which only two contained significant amounts of 3%-radioactivity ( Fig. 2 and Table II Analysis When ihe 35S-r;!dioactivity recovered in the sum of the HPLC fractions was compared to that whlruted t o IIPLC. significant losses (10.30%) of 35S-radioactivity were noted. There losses were !not the r c~u l t ofdcrulfation of rulfopeptider priort0,orduring. HPLC since no 3hadioactivitywas iwnd in the first H P I S fractions (where sulfate would be recovered) and since tymsine sulfate was lnund to k ruble 11) the IIPLC huffcra used (data not shown). Rather. there losses appeared to be due chrnms1ogr;tphy of rulfopeptider A and R.
10 umpcctlic atlsorpmn of rullopeptidcs durmg HPLC slnce simdar losses were observed upon re-SThc rccnvery of 3sS-radionctivity in the various fractions obtalned after SEP-PAK CIS chmm;un~raphy and IIPLC. and the profile of 3%-radioactivity after HPLC. were very similar to thwc oh\erved when 1~~SI*ulf~~tc-labeled YP2 obrsined from a two-dimensional gel of the total cttrrier YP? bhich had k e n purified from unlabeled flier by the method descnbed in Fig. I (see Table   prolrin of 1~~S l~u l f~~t c -l a k l e d female flies was digested wlth trypsin in the presence of unlabeled II: dibti~ not shownl.ll~i\ !ndr:tted th:u the wlfation of YP2 dtd not change dunng the punficauon.  Values are expressed for each entire fraction. The first value in brackets gives the recovery of 35S-radioactivity with respect to the material subjected to the respective chromatography performed; the second value gives the recovery with respect to the total 35S-radioactivity recovered after the respective chromatography. The starting material for experiment 1 was YP2 purified from a mixture of unlabeled and [35S]sulfate-labeled flies as described in Fig. 1. The starting material for experiment 2 was a mixture of [35S]sulfate-labeled YP2 obtained from t w~d i m~n s i o n~ gels of whole fly homogenate and YP2 purified from unlabeled flies as described in Fig. 1. MeOH: methanol; total peak: sum of %radioactivity in the fractions containing sulfopeptides A and B plus the adjacent fractions (see Fig.2 Fig. 1 o i type-\el pan of paper). Since the ebter bond of tyravne sulfate is hvdrolyced ~n the digestion of rulfopeptider 4 and B (Fig. 3 of miniprint section). Chymotryptic cleavage was not generation of a rulfopeptlde contalntng a single tymslne residue. This was achieved by chymotryptic observed In the presence of 0.1 m M CaC12 (not shown), but occurred almost quantitatively in the premIce of 5 mM EDTA (Fig. 3). Both sulfopeptides A and B were convened into single new acetnnitnle concentranon (Fig. 3). After one mund ofchymotryptic digestion and HPIC. sulfopeptlder sulfvprptides. referred to as rulfopeptide5 CA and CB, respectively, which eluted at a lower C A and CB conlained more than 80% of the 3X+ddloactwity originally contained in sulfopeptides A and R. respectwely (Table Ill. The 15-207i of the 35S-radioactivity remaining in sulfopeptides A and B rhlfted to the yoailions of sulfupeptidea CA and CB, respectively, upon re-digestion with chymot~pcin(datanot shown).
N-terminal sequencing ofsulfopeptiiks CA and Cy through 3 cycles revealed identical N-temrini with the sequence Glu-Thr-Thr (Fig. 3). corresponding to sequence positions 168.170 of the YP2 precursor (ref. 9 Fee Fig. I of typr-sel pan of paper). Thar, chymotryptic cleavage of the tryptic rulfopeptides A and B occurred at lyrosme 167, but no1 at tyrosine 172 (see Fig. I of type-set pan of $?spec'). It hdb ken observed that ~h y~t r y p s i~ doer not cleave at rulfated tyrosine residues 121). ' R e arnlnn acld malyrn of wlfapeptlde CAB agreed k s t with the calculated amino acid composlt~on ofa peptide correrponding to sequence positions 16X-180 of the YP2 precursor (Table 111 of rningprint rectlon; see Flg 1 of type-set pan of paper). This means that during incubation with chymotrypsin and ED'I'A. arginine I80 had become accessible to cleavage. probably by a trypsin contamination of the chymotrypsin used (non TLCK-treated). After some chymotryptic digestions of sulfopeptidcs A and with the calculated amino acid composition of a peptide corresponding to sequence positions 168.
13. amino acld analysis oithe rerullirig rulfopeptidea CA and CB (no1 shown) was also compatible 19011Y I of the YP2 precursor (see Fig. 1 of type-set pan of paper). raising the possibility that cleavage of arginine 180 did not always occur. At any rate. any sulfopptide generated by = h y m~t~~t i~ dlgcstlon of sulfnpeptidea A and B contained only one tyrosine residue, tyrosine 172. Together with t h e data on h e recovery of the s s S -~~~i~~~l , " i t y contained in YP2 hrouph h e isolation of s " l f~p [ i d~ CAJB lsee Table II of