Download Directed mutagenesis of the Trypanosoma cruzi trans

Glycobiology vol. 7 no. 3 pp. 445-451, 1997
Directed mutagenesis of the Trypanosoma cruzi trans-sialidase enzyme identifies two
domains involved in its sialyltransferase activity
Lynne E.Smith and Daniel Eichinger1*2
Departments of Pathology and 'Medical and Molecular Parasitology, New
York University School of Medicine, 341 East 25th Street, New York, NY
10010, USA
^To whom correspondence should be addressed
Of the increasing number of sialidases found to be made by
microorganisms, the trypanosome trans-sialidase is unique
hi its added ability to efficiently carry out a sialyltransferase reaction using preformed glycoconjugates. The enzyme
is predicted to have a multidomain structure, with one domain containing sequence and expected structural features
found in bacterial sialidases. The trans-sialidase is very
similar in overall sequence to another trypanosome enzyme
that has only sialidase activity. Hybrid expression constructs containing pieces of these trypanosome transsialidase and sialidase genes were used to determine which
regions of trans-sialidase are required for sialyltransferase
activity. Two domains were found to influence the enzymatic activity: the N-terminal catalytic domain, and a
downstream domain that resembles an Fn3-like module.
Key words: mutagenesis/enzyme/sialyltransferase/trypanosome
Introduction
Several protozoan parasites of vertebrates (including Trypansoma cruzi, T.brucei, and Endotrypanum sp.) have been shown
to express a unique type of sialyltransferase called transsialidase (TS; Schenkman et al., 1991; Engstler et al, 1992;
Medina-Acosta et al., 1995). This enzyme is capable of removing terminal a2,3-linked sialic acid from sialyl-3-galactose donor molecules and reattaching it to other 3-galactoseterminated glycoconjugates (Vandekerckhove et al, 1992;
Scudder et al., 1993). In this reaction the enzyme thus functions both as a sialidase, similar to viral neuraminidase and
bacterial sialidases, and as a glycosyltransferase. It differs,
however, from the sialyltransferases found in the Golgi apparatus which use only nucleotide-monophosphate-sugar donor
substrates (Komfeld and Kornfeld, 1985), although both enzyme types can be considered to use similar acceptor substrates. Neither the reaction mechanism nor the structural features of the TS protein that support its efficient sugar transfer
activity are known. The trypanosome parasite, unable to synthesize sialic acid (Previato et al., 1985), uses the enzyme to
scavenge the sugar from host glycoconjugates and to sialylate
mucin-like acceptor molecules found on its plasma membrane.
These molecules, and/or the enzyme itself, are involved in the
obligate entry of the trypomastigote form of the parasite into
host cells where it differentiates and multiplies (Schenkman et
al, 1991; Hall et al, 1992; Burleigh and Andrews, 1995).
© Oxford University Press
The sequences of several genes encoding T. cruzi TS and
TS-like proteins have been reported (Pereira et al., 1991; Pollevick et al., 1991; Uemura et al., 1992), and the predicted
amino acid sequences of the subclass of enzymatically active
gene products indicate that the parasite proteins have structural
features previously found in influenza neuraminidase and several bacterial sialidases (Vimr, 1994). In particular, the trypanosome proteins contain all the amino acid residues (or
classes) which have been described to interact with sialic acid
in the substrate binding pockets of Salmonella typhimurium,
Vibrio cholorae and Micromonospora viridifaciens sialidases
(Crennell et al., 1993, 1994; Gaskell et al., 1995). The parasite
enzymes also contain multiple copies of a conserved consensus
sequence (aspartic acid box (ASP box)) found in viral, bacterial, and mammalian sialidases (Roggentin et al., 1989; Miyagi
et al., 1993). Crystal structure analyses of the three bacterial
sialidases indicate that their protein chains are folded into sixsided 3-sheet barrels similar to that of influenza neuraminidase. The length of the open reading frames (ORFs) encoding
the catalytic domain of the trypanosome enzymes, as well as
die positions of the conserved sialic acid binding site residues
and ASP boxes in primary structure alignments and hydropathicity plots, predicts that the tertiary structure of the catalytic
domain of trypanosome TS will be found to be similar to those
of die viral and bacterial enzymes. The N-terminal catalytic
domain in TS is followed, in order, by a domain of unknown
function, a fibronectin type HI domain (Fn3 domain), a large
segment composed of 12 amino acid tandem repeats, and finally by a GPI anchor addition sequence (Pereira et al., 1991).
The last two regions are not required for enzymatic activity
(Schenkman et al., 1994a).
The low overall level of amino acid similarity (35%) between the bacterial and trypanosome enzymes renders uninformative any cross-genera sequence comparisons designed to
reveal unique amino acids or subdomains involved in the sialic
acid transfer function of TS. Another trypanosome species,
T.rangeli, was previously shown to express a sialidase type
enzyme that hydrolyzes the same sialylated donor substrates as
TS (Pereira and Moss, 1985), but lacks sialic acid transfer
activity (Pontes de Carvalho et al., 1993). The relatedness of
the T.cruzi and rangeli parasites suggested that sequence comparisons between me two parasite enzymes might reveal structural differences relevant to the sialic acid transfer function of
TS. We recently described the cloning and expression of a gene
fragment encoding T.rangeli sialidase (Smith et al, 1996). The
T.rangeli protein contains the same predicted structural features as TS, including ASP boxes and a C-terminal Fn3 domain. Most importantly, the same amino acids within the presumed sialic acid binding pocket of TS are found in the T.rangeli sialidase, suggesting that amino acids outside of this
pocket are involved in the TS sialic acid transfer reaction.
Here we describe the construction, expression, and functional activity analysis of T.cruzi trans-sialidase/T.rangeli si445
LJLSmith and D.Ekhinger
alidase hybrid proteins which served to roughly define a region
in the TS catalytic domain required for transfer function, and
the site specific mutagenesis of this region to further delimit
the involved segment. The amino acids in this segment are
predicted to lie in a loop just outside of the sialic acid binding
pocket, and in this position are likely to interact with the galactose portion of donor/acceptor molecules. By exchanging a
single amino acid within this segment, we added sialyltransferase activity to a hybrid protein that originally expressed only
sialidase activity, thereby creating a novel trans-sialidase enzyme. The C-terminal Fn3-like domain of TS, although well
separated from the catalytic domain in the primary structure,
was also found to influence the relative levels of the sialidase
and sialyltransferase activities of TS.
Results
The approach taken to dissect TS for the region required for
sialyltransferase activity was based on the similarity in the
positions of the sialic acid binding pocket residues and Asp
boxes in primary structure alignments of the catalytic domains
of T.cruzi TS and T.rangeli sialidase with the S.typhimurium
sialidase. Figure 1 shows a schematic of the p-sheet and strand
136
131
4
123 4
structure of the bacterial sialidase with the approximate positions of the sialic acid binding pocket residues indicated. The
amino acids that accommodate the sialic acid N-acetyl group
on C5 (M99, W 121 , W 128 , L175) and the glycerol chain on C6
(W128) are all within P-sheets 2 and 3 of the enzyme. On the
other hand, the amino acids (E 231 , R246, R309, Y 342 , E 361 , with
the exception of R37 in (3-sheet 1) that interact with Cl and C2
at the opposite "end" of the sialic acid molecule are located in
P-sheets 4—6. Since hydrolysis and subsequent transfer of sialic
acid to galactose occurs here, the gene segments of TS and
T.rangeli sialidase corresponding to P-sheets 4—6 were exchanged with one another, using the pQE-based expression
constructs pA4 (expressing TS) and pG5EX (expressing sialidase) previously described (Schenkman et al., 1994a; Smith et
ai, 1996).
The region of T.rangeli sialidase containing the three presumed P-sheets 4—6, plus the flanking region between the sixth
P-sheet and the Fn3 domain, was inserted behind P-sheets 1-3
of TS. The resulting construct, pS7-l, encoded a product with
sialidase, but no trans-sialidase activity (Figure 2). This glycolytic activity was greatest at the lowest pH tested, (4.0), but
dropped off abruptly with any increase in pH (Figure 3). The
fact that this hybrid expressed any activity at all was signifi-
I35
123
66
4
12
3
S. typhimurium
LT2 sialidase
T. cruzi trans-sialidase
T. rangeli G5EX sialidase
Fig. 1. Aligranent of the P-sheet/strand structure of the S.typhimurium sialidase with the ORFs of trypanosome enzymes. The six p-sheets (P1-6) of the
bacterial sialidase are each made up of four strands (1-4) that are connected by loops (not shown to scale). The approximate positions of the amino acids that
comprise the sialic acid binding pocket are indicated as bold letters (R, arginine; M, methionine; W, tryptophan; L, leucine; E, ghitamic acid; Y, tyrosine). N
and C indicate the termini of the bacterial sialidase ORF. Schematics of the ORFs of T.cruzi trans-sialidase and T.rangeli sialidase catalytic domains are
aligned below to indicate the conserved positions of the Asp boxes in primary sequence alignments. Vertical dotted lines indicate the positions within the
presumed secondary structure (based on that of S.typhimurium sialidase) of the junction points of the hybrid proteins shown in Figure 2 (E, EcoNI).
446
Trans-sialidase domains
T. cruzlTS154A4A
Activity
S
T3
5122
5188
9786
0
3S52
o
3336
o
3081
6630
8508
17S9
7403
1200
8801
1762
T.rangollG5EX
Hybrid Constructs
•7-1
Fig. 2. Hybrid expression constructs and enzymatic activities. The top two
schematics represent the ORFs of the full-length T.cruzi trans-sialidase
(TS154A4) and T.rangeli sialidase (G5EX) expression constructs (Smith et
al, 1996). These and all other expression constructs contain the intervening
and Fn3-like domains downstream of the catalytic domain from TS154A4
(to the right of the Bs restriction site), plus a small number of C-terminal
tandem repeats (not shown). Hybrid constructs made with these two genes
are indicated below, and shaded and unshaded segments represent T.cruzi
TS- and T.rangeli sialidase-derived sequences, respectively. When a
mutagenic oligonucleotide was used, the template and amino acids
exchanged by the oligo are indicated (PGS, P). Enzyme activity encoded by
each construct is indicated to the right, in fluorescence units for sialidase
(S), which was assayed at pH 5.0, and in c.p.m. foT trans-sialidase (TS),
which was assayed at pH 7.2 (E, EcoNl; Bs, BsiWT).
cant, since it indicated that the functionally important amino
acids in the two "halves" of the sialic acid binding pocket, one
from TS, and the other from T.rangeli sialidase, were properly
positioned in the hybrid protein. However, the hybrid gene
resulting from the reciprocal exchange, placing the TS P-sheets
4—6 and the flanking C-terminal region after the T.rangeli sialidase ^-sheets 1-3 of pG5EX, encoded an enzymatically inactive product (data not shown).
The next hybrid construct, pS3—2, inserted all of P-sheets 4
and 5 and the first strand of sheet 6 from T. rangeli sialidase in
place of the TS sequence. The product of this hybrid expressed
only sialidase activity (Figure 2). Again, the reciprocal hybrid
construct protein product was inactive (data not shown).
The third hybrid, pS8-l, replaced the fourth strand of the
fourth p-sheet (P4S4), the entire fifth P-sheet and the first
strand of the sixth p-sheet (P6S1) of the pS3-2 gene product
with the same region from TS. Now the encoded product expressed both sialidase and sialyltransferase activities. The pS81-encoded protein had reduced sialidase activity and the broad
pH optimum range similar to the wild type TS enzyme. The
sequence and encoded activity differences between the pS3-2
and pS8-l constructs served to define a segment of 88 amino
acids that contained residues required for TS sialic acid transfer
activity.
Next, using pS3—2 as template, oligo-directed mutagenesis
was used to systematically exchange small portions of the
T.rangeli sialidase-derived 88 amino acid-encoding segment
with those found in TS. Of the eight oligos designed to span
this region, the changes introduced by only one were informative. Oligonucleotide Mon-2 was designed to replace the 3
contiguous amino acids, Q, D, and C (284—286) of sialidase
with the P, G, and S, respectively (315-317), found in the
corresponding positions of TS. Exchange of these three amino
acids resulted in a hybrid gene (s3-2mon2) that encoded a
protein with sialidase and sialyltransferase activities. The sialidase activity of this mutant exhibited the same pH sensitivity
as the pS3-2 encoded product (not shown). Its level of sialyltransferase activity ranged from 12 to 25% that of the wild type
T.cruzi TS enzyme.
To determine if any of the TS-derived sequence at the Cterminal end of the catalytic domain of pS3—2 was required to
reveal the sialyltransferase activity resulting from addition of
the PGS triplet, the same Mon-2 oligo was used to mutagenize
plasmid S7-1, which had the entire C-terminal half of the
catalytic domain derived from G5. This mutant (s7-lmon-2)
also acquired a similar level of sialyltransferase activity as the
mon2 mutant of S3-2 (Figure 2).
Each of these three amino acids from TS was next introduced individually into the pS3-2 and pS7-l sialidase expression constructs. Inserting proline for glutamine added sialyltransferase activity to both expression constructs, whereas the
other two amino acid substitutions (D to G, and C to S) resulted
only in retention of sialidase activity (Figure 4). The level of
sialyltransferase activity of the P to Q mutant was approximately 25% that of wild-type TS (Figure 2).
We had previously described the inability to express enzyme
activity from constructs expressing only the catalytic domains
(i.e., without any attached C-terminal domains) of either the TS
or T.rangeli sialidase genes (Smith et al., 1996). To express a
protein product with any enzyme activity, these genes required
the full length ORFs encoding the N-terminal catalytic domain
and the C-terminal Fn3-like domain (which was noted by others to contain a conserved sequence VTVxNVxLYNR; Cross
and Takle, 1993), plus the intervening domain of unknown
function. A recent description of the crystal structure of sialidase from M.viridifaciens (Gaskell et al., 1995) revealed the
presence of two additional domains attached to its P-barrel
Fluorescence units (Thousands)
4.0
4.5
5.0
5.5
6.0
6.6
7.0
7.5
8.0
8.5
9.0
Fig. 3. Relative sialidase activities of wild type and hybrid expression
constructs. The pH of the sialidase reaction buffer was adjusted to those
indicated and the levels of released MU from MU-NANA were measured.
The expression constructs assayed are indicated to the right, and correspond
to those in Figure 2.
447
L.ELSmlth and D.Eichinger
Enzyme Activity
S
TS
T.c. TS154
T.r. G5
G5/Mon 2
G5/Mon 10
G5/Mon 11
G5/Mon 12
VGTLSRVWGPSPKSNQPGSQS
LGTLSHVWTNSPTSNQQDCQS
LGTLSHVWTNSPTSNQPGSQS
PDC..
QGC..
QD8..
Fig. 4. Amino acid sequence and enzyme activities of oligo-directed
mutants. Amino acids from the presumed loop between strand 4 of |J-sheet
4 and strand 1 of sheet 5 of the parasite enzymes are shown for T.cruzi
trans-sialidase (TS154) and T.rangeli sialidase (G5). The three amino acids
(P,G,S), where substituted into the G5 sequence, are indicated in bold, and
unchanged residues are indicated with dots. Sialidase (S) and trans-sialidase
(TS) activities are indicated to the right.
catalytic domain. In this case a C-terminal jelly roll domain
containing a galactose binding pocket was proposed to position, via an intervening Ig-like domain, a bound galactose molecule to within 30 A of the sialic acid binding pocket of the
catalytic domain. The gross similarity between the domain
structure of this bacterial sialidase and those of the parasite
enzymes, and the apparent requirement of the C-terminal domain of the parasite enzymes for activity, prompted us to align
the amino acid sequences of the bacterial and parasite protein
C-terminal domains (Figure 5). The R 572 and E 578 residues of
M.viridifaciens sialidase that interact with 0 3 and O4 of the
bound galactose were in fact conserved in the T.cruzi and
T.rangeli proteins. The aligned arginine in the parasite proteins
was the same one found at the end of the conserved VTVxNVxLYNR sequence that defines the trypanosome TS/sialidase
superfamily of cell surface proteins (Campetella et al., 1992;
Cross and Takle, 1993; Schenkman et al., 1994b). Oligodirected mutagenesis of the corresponding R (to A) in each
parasite enzyme greatly reduced their sialidase activities when
using MU-NANA as substrate, and significantly increased the
sialyltransferase activity of TS (Figure 5). Changing the conserved E (to V) in each parasite enzyme had the same effects
on the two enzymatic activities.
Discussion
Sequence alignments of T.cruzi TS, T.rangeli sialidase, and
S.typhimurium and M.viridifaciens sialidases predict that the
three amino acids exchanged by the Mon-2 oligo fall within a
loop connecting the last strand of (3-sheet 4 and the first strand
of sheet 5 in the parasite enzymes. This loop in the bacterial
enzymes has amino acid side chains that extend into the space
just outside of the sialic acid binding pocket (Crennell et al,
1993; Gaskell et al., 1995). Therefore, these residues in the
parasite proteins might be capable of interacting with the penultimate (galactose) residue of a donor substrate, or with the
terminal galactose of an acceptor substrate. This predicted loop
is 4—5 amino acids longer in both trypanosome proteins than in
any of the bacterial enzymes, and varies in sequence between
the T.cruzi and T.rangeli enzymes in places other than the three
amino acids exchanged by oligo Mon-2. These other differences (when exchanged independently of the Mon-2 residues)
were not found to be significant with respect to sialyltransferase function (data not shown).
The finding that a single proline substitution into the sialidase-expressing construct added sialyltransferase function was
surprising. The praline is not likely to be directly involved in
Micromonospora viridifaciens sialidase
14
1 1
TFWHTEWSRADAPGYPHRISLDLGGTHTISGLQYTRRQNSANEQVADY
PDGRTPDISHFYVGGYGRSDMPTISHVTVNNVLLYNRQLNAEEIRTLF
(A)
(V)
Trypano3oma cruzi trans-aialida3e
h
T. cruzi trans-sialidase
T. cruzi TS-mutAEV
T. cruzi TS-mutLYNA
Sialidase
5122
962
369
Trans-sialidase
5188
10904
10310
Fig. 5. Selected mutagenesis of the third domain of T.cruzi trans-sialidase. The ORF domains of the M.viridifaciens sialidase and T.cruzi TS are shown in
schematic form. Domains I-in of M.viridifaciens sialidase are the catalytic, Ig-like, and jelly-roll, respectively (Gaskell et at, 1995), and domains 1-TV of
T.cruzi TS are the catalytic, intervening, Fn3-like and tandem repeat, respectively (Pereira et at, 1991). Amino acid alignment of a portion of each domain III
is shown. Arrows above the amino acids indicate those of the bacterial sialidase which comprise the galactose binding site, the positions of sequence identity
are shown with dots, and the conserved sequence found in members of the Trypanosoma TS/sialidase superfamily is underlined. Enzyme activity levels of the
two single amino acid mutants of T.cruzi trans-sialidase are shown at the bottom. TS-mutLYNA had R MJ changed to A, and TS-mut AEV had E 648 changed
to V (shown in parentheses). Enzyme activities are indicated in fluorescence units for sialidase and in c.p.m. for trans-sialidase.
448
Trans-sialklase domains
protein—sugar interactions, as amides are the type of amino
acid side chain in lectins most often found hydrogen bonded to
sugar OH groups (Weiss and Drickamer, 1996). The gain of
function resulting from insertion of a second proline into this
segment of the T.rangeli protein indicates that all the other
amino acids essential for transfer function are already present,
and that only a conformational alteration of the loop is necessary to add some level of transfer function. We do not know at
this time how the proline described here changes the local
structure of the T.rangeli sialidase, and whether the significant
repositioned residues are in fact found within the same or
adjacent loops. There are three prolines in this presumed loop
of the T.cruzi TS protein, suggesting that further conformational adjustments may improve on the low (25% of wild type)
level of sialyltransferase function of the Mon-10 mutant of
hybrid s7-l.
A proline in a different position of TS was previously shown
to be required for high levels of TS sialyltransferase activity in
proteins that already expressed some level of transferase activity (Cremona et al., 1995). Both the T.cruzi clone 154 TS
and the T.rangeli G5 sialidase contain this particular proline,
which is predicted to be part of strand 1 of (3-sheet 4. It, most
likely, is critical for positioning the adjacent upstream glutamic
acid, which is one of the conserved amino acids of the substrate
binding pocket, and thereby directly effects the enzyme's sialidase activity.
The implication that the amino acids which are essential for
sialyltransferase activity are present in the T.rangeli enzyme is
consistent with the notion that the T.rangeli enzyme expressed
in the insect stage of the parasite originally had and then lost
sialyltransferase activity (Briones et al., 1995), gained mutations that increased the sialidase activity at low pH levels, and
drifted in regions of the protein not required for sialidase function. It is worth noting, therefore, that a reported sequence
(Buschiazzo et al., 1993) of a trans-sialidase-uke gene from an
unspecified strain of T.rangeli is predicted to encode a protein
with the PGS sequence in the critical loop, rather than the QDC
sequence of the three T.rangeli sequences described previously
(Smith et al., 1996). Although no evidence was presented for
enzymatic activities expressed by this unspecified strain or the
protein product of the described gene, it leaves open the possibility that a different stage of the T.rangeli parasite does
express TS activity, as postulated by those authors, and that the
T.rangeli genome contains two types of sialidase-related genes.
The sequences of trans-sialidases from earlier diverging trypanosomatids, such as T.brucei, will be interesting for comparison in this regard.
The role of the Fn3-like domain in the parasite enzymes
remains unclear based on the limited analysis described here. It
contains conserved amino acids that may be involved in galactose binding, by comparison to the M.viridifaciens sialidase.
M.viridifaciens makes two forms of sialidase from the same
gene, one which contains only the N-terminal catalytic domain,
and another which has the attached Ig-like and jelly roll domains (Sakurada et al., 1992). Both of these forms express
siahdase activity, and the growth medium of a casein-induced
culture of M.viridifaciens, which should express the full length
gene product, contained readily detectable sialidase but no
trans-sialidase activity when assayed under the same conditions as the parasite enzymes, (D.E., unpublished observations). Previous deletion constructs indicated that an intact
Fn3-like domain is required for expression of enzymatically
active products from both T.cruzi TS (Schenkman et al, 1994)
and T.rangeli sialidase (Smith et al, 1996) genes. On the other
hand, the Fn3-like domain of TS is apparently not sufficient to
confer sialyltransferase activity to a protein with sialidase activity, since it is found in the wild type T.rangeli sialidase
protein, and since exchanging the Fn3-like domain from TS
with that of the T.rangeli sialidase did not add sialyltransferase
function to the sialidase. Individual elimination of the conserved R and E residues of this domain of TS actually served
to increase its apparent sialyltransferase activity. The fact that
mutations in this region did have an effect on enzyme activity
indicates that the parasite proteins are folded so as to bring the
substrate-binding region of the catalytic domain and (at least)
the R and E of the Fn3-like domain close together, as is proposed for the sialic acid and galactose binding sites of the
M.viridifaciens siahdase (Gaskell et al., 1995). Whether or not
the Fn3-like domain of TS forms an essential component of a
binding pocket for donor and/or acceptor substrates will best be
resolved with additional mutagenesis and kinetic studies, in
combination with protein crystal structure data obtained from
enzymatically active forms of these proteins. Such detailed
information about the roles of the catalytic and Fn3-like domains of TS should also help to establish the feasibility of
eventually adding glycosyltransferase activity to other glycosidases.
Materials and methods
Materials
Oligonucleotides were purchased from Operon Technologies, Inc. (Alameda,
CA). Methyl umbelliferyl-N-acetylneuraminic acid (MU-NANA), sialyUactose (SL), IPTG, and buffers were purchased from Sigma.
Bacterial strains
For plasrrud preparations E.coli strain TGI (F' traD36 laqlq A(lacZ)M15
proA+B+/supE A(hsdm-mcrB)5 (rk- n v mcrB") thi A(lac-proAB)) was used.
All expression plasmids were maintained in strain CC114 (sup° lacZ"™1 recA
hsdR hsdM+). Micromonospora viridifaciens was obtained from ATCC (catalog number 31146) and grown as described previously (Sakurada et al, 1992).
Construction of hybrid genes
Overlap extension PCR was performed using a modification of the method of
Yon and Fried, 1989. Oligo DE17 (5'-GCC CAT GGC ACC CGG ATC GAG
CCG AGT T), which established an initiation codon as part of an Ncol restriction site at the point of leader sequence cleavage (Pollevick et at, 1993),
and the linking oligo DE37 (5'-CGA GGA CTG CCG GTT CAG AGC AGC
CAA AAT CAC T) were used in the PCR with T.cruzi pTS154A4A (pA4)
(Schenkman et al, 1994a) template to amplify the N-terminal half of the TS
catalytic domain containing the presumed B-sheets 1-3, (see Figure 1). This
reaction product, plus oligo RT154 (5'-CGG GAT CCG GGC GTA CTT CTT
TCA CTG CTG CCG CT) and T.rangeli sialidase expression construct pG5EX
(Smim et al, 1996) as template were combined and amplified for 10 cycles,
and then oligo DEI7 was added and amplification was continued for 20 more
cycles. The resulting product was digested with Ncol and BsiWI and inserted
in place of the same restriction fragment of the pA4 expression vector, yielding
plasmid pS7-l (see Figure 2). Two other constructs were made with additional
sequence from the TS gene. The EcoNl-BsiWl fragment of pS7-l was inserted in place of the same fragment of pA4, yielding pS3-2. An overlap
extension PCR, containing oligo DE17, linking oligo DE47 (5'-AAT TCC
CCA TGT CGC TCG ATT CGT AGA CCA AGC GA) and pS3-2 in the first
round, and the first round product plus DE17, RT154 and pA4 template in the
second round, generated an Ncol/BsiWI insert which was inserted in place of
the same fragment of pA4, yielding pS8-l.
Oligonucleotide-directed mutagenesis
Mutagenesis was carried out using the QuickChange Site-Directed Mutagenesis Kit (Stratagene) following the manufacturer's instructions. In order to
distinguish mutated plasmids from wild type plasmids that carried through the
449
L.E-Smith and D.Eichinger
reaction, each oligo pair was designed to add or destroy a restriction site
without changing the encoded amino acids except for those being changed by
mutagenesis. Tbe pair of oligos (Mon 2: 5'-TCA CCA ACA TCG AAC CAG
CCA GGA TCC CAG AGC AGC TTC GTT GCT; Mon 2 IC: 5'-AGC AAC
GAA GCT GCT CTG GGA TCC TGG CTG GTT CGA TGT TGG TGA)
inserted the amino acids P, G, and S (residues 315—317, numbering based on
T.cruzi trans-sialidase clone 154; Smith et al., 1996) in place of Q, D, and C
(residues 284-286 of T.rangeli sialidase), respectively, into the hybrid constructs pS7-l and pS3-2. Single amino acid changes to pS7-l and pS3-2 were
made using the following oligo pairs: Q2** to P (Mon 10: 5'-AACTCA CCA
ACT TCG AAC CAG CCG GAC TGT CAG AGC AGC; Mon 10 ic: 5'-GCT
GCT CTG ACA GTC CGG CTG GTT CGA ACT TGG TGA GTT); D 285 to
G (Mon 11: 5'-AAC TCA CCA ACT TCG AAC CAG CAG GGC TGT CAG
AGC AGC; Mon 11 ic 5'-GCT GCT CTG ACA GCC CTG CTG GTT CGA
AGT TGG TGA GTT ); C28* to S (Mon 12: 5'-AAC TCA CCA ACT TCG
AAC CAG CAG GAC AGT CAG AGC AGC; Mon 12 ic 5'-GCT GCT CTG
ACT GTC CTG CTG GTT CGA AGT TGG TGA GTT). In addition, the pA4
plasmid was used as the template to change the sequence P 313 to Q (Mon 13
5'-AAA TCG AAC CAG CAG GGC AGT CAG AGC AGC TTC ACT GCC
GTG; Mon 13 ic 5'-CAC GGC AGT GAA GCT GCT CTG ACT GCC CTG
CTG GTT CGA TTT). Two single amino acid mutations were made in the Fn3
domain of pA4 as foUows: R 642 to A (AEV 5'-CAG CTG AAT GCC GAG
GTG ATC AGG ACC TTG TTC and AEV ic 5'-GAA CAA GGT CCT GAT
CAC CTC GGC ATT CAG CTG) and E 648 to V (LYNA 5'-AAT GTT CTT
CTT TAC AAC GCT CAG CTG AAT GCC GAG GAG and LYNA ic 5'-CTC
CTC GGC ATT CAG CTG AGC GTT GTA AAG AAG AAC ATT).
Sialidase and trans-sialidase assays
E.coli containing expression plasmids were grown overnight in L broth plus 50
(ig/ml ampicillin (LB/amp) at 37°C, diluted 1:10 in LB/amp containing 2 mM
IPTG, and grown for 2.5 h. Bacterial lysates were prepared using the method
described by Schupp et al, 1995. The total protein concentration in the lysates
was determined by the method of Udenfriend et al. (1972), and 100 p,g total
protein was used in all assays. Sialidase and trans-sialidase activities were
measured as described (Schenlanan el al., 1991). The pH curves for sialidase
activity were generated using 100 (ig total lysate protein and 50 mM each of
the following buffers: sodium acetate, pH 4.0-5.5; MES, pH 6.0; HEPES, pH
6.6-7.5; and Tris, pH 8.0-9.0.
Sequencing
All TS/sialidase gene junction sites and site-directed mutations were confirmed
by sequencing across the modified regions of double-stranded templates using
Sequenase 2.0 (USB).
Acknowledgments
We thank M. Briones and S. Schenkman for helpful discussions, and Laura
Pologe and Stephen Tomlinson for critical reading of the manuscript. This
work was supported by The Mizutani Foundation for Glycoscience (D.E.),
NSF Grant MCB-9418190 (D.E.), and NIH Training Grant 5T32CA09161
(L.E.S.).
Abbreviations
TS, trans-sialidase; Fn3, fibronectin type HI domain; IPTG, isopropyl-P-r>
thiogalactopyransoside; MES, 2-(N-morpholino)etbanesulfonic acid; HEPES,
(N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid]).
References
Briones,M.R.S., Egima.CE., Eichinger.D. and Schenkman^. (1995) Transsialidase genes expressed in mammalian forms of Trypanosoma cruzi
evolved from ancestor genes expressed in insect forms of tbe parasite. J.
MoL EvoL, 41, 120-131.
BurleighJJ-A. and AndrewsJJ.W. (1995) The mechanisms of Trypanosoma
cruzi invasion of mammalian cells. Annu. Rev. Microbiol., 49, 175-200.
Buschiazzo,A., Crcmona,M.L., Campetella,O., Frasch,A.CC. and Sanchez,
D.O. (1993) Sequence of a Trypanosoma rangeli gene closely related to
Trypanosoma cruzi trans-sialidase. MoL Biochem. ParasitoL, 62, 115-116.
CampeteUa,O., SanchezJ}., CazzuloJJ., and Frasch,A.CC (1992) A superfamily of Trypanosoma cruzi surface antigens. ParasitoL Today, 8, 378—
381.
450
CremonaJd.L., SanchezJD.O., Frasch,A.C.C. and Campetalla,O. (1995) A
single tyrosine differentiates active and inactive Trypanosoma cruzi transsialidases. Gene, 160, 123-128.
CreimeU.SJ., GarmanJE.F., Laver.W.G., Vimr,E.R. and Taylor.G.L. (1993)
Crystal structure of a bacterial sialidase (from Salmonella typhimurium LT2)
shows the same fold as an influenza neuraminidase. Proc. NatL Acad ScL
USA, 90, 9852-9856.
CrenneU^J., Garman,E., Laver.G., Vimr,E. and Taylor.G. (1994) Crystal
structure of a Vibrio cholorae neuraminidase reveals dual lectin-like domains in addition to the catalytic domain. Structure, 2, 535-544.
Cross.G.A.M., and Takle.G.B. (1993) The surface trans-sialidase family of
Trypanosoma cruzi. Annu. Rev. Microbiol., 47, 385—411.
EngstleT,M., Reuter,G. and Schauerjt. (1992) Purification and characterization
of a novel sialidase found in procyclic culture forms of Trypanosoma bruceL
Mol. Biochem. ParasitoL, 54, 21-30.
GaskellA, Crennell.S. and Taylor.G. (1995) The three domains of a bacterial
sialidase: a f$- propeller, an immunoglobulin module and a galactosebinding jelly-roll. Structure, 15, 1197-1205.
HaU.B.F., WebsterJ>., Ma,AJC., JoinerJCJV. and Andrews,N.W. (1992) Desialylation of lysosomal membrane glycoproteins by Trypanosoma cruzi: a
role for the surface neuraminidase facilitating parasite entry into the host cell
cytoplasm. /. Exp. Med, 176, 313-325.
Komfekyt. and Komfeld.S. (1985) Assembly of asparagirte-linked oligosaccharides. Annu. Rev. Biochem., 54, 631-664.
Medina-Acosta,E., Pereira,A.F., Jansen^A.M., SampolJvL, Neves,N., Pontes
de Carvalho,L., Grimaldi.G. and Nussenzweig.V. (1994) Sialidase and
trans-sialidase activities discriminate between morphologically indistinguishable trypanosomatids. Eur. J. Biochem., 225, 333-339.
Miyagi.T., Konno.K., Emori.Y., Kawasaki.H., Suzuki.K., Yasui,A. and
Tsuiki.S. (1993) Molecular cloning and expression of cDNA encoding rat
skeletal muscle cytosolic sialidase. J. BioL Chem., 268, 26435-26440.
Pereira,M.E J \., MejiaJS., Ortcga-Barria,E., MatsilevitchJ). and Prioli,RP.
(1991) The Trypanosoma cruzi neuraminidase contains sequences similar to
bacteria] neruaminidases, YWTD repeats of the low density lipoprotein
receptor, and type El modules of fibronectin. J.Exp. Med, 174, 179-191.
Pereira,M.E.A. and MossJ). (1985) Neuraminidase activity in Trypanosoma
rangeli MoL Biochem. Parasitol, 15, 95-103.
Pollevick,G.D., AfrranchinoJ.L., FraschAC.C. and Sanchez,D.O. (1991) The
complete sequence of a shed acute-phase antigen of Trypanosoma cruzi
MoL Bioechm. ParasitoL, 47, 247-250.
Pollcvick,G.D., Sanchez.D.O., Campetella,O., Trombetta,S., Souza>I., Henrikssonj., Hellman.U., Petersson.U., CazzuloJJ. and Frasch,A.C.C. (1993)
Members of the SAPA/trans-sialidase protein family have identical Nterminal sequences and a putative signal peptide. Mol. Biochem. ParasitoL,
59, 171-174.
Pontes de CarvalhoJL.C, Tomlinson.S. and Nussenzweig.V. (1993) Trypanosoma rangeli sialidase lacks trans-sialidase activity. MoL Biochem. Parasitol, 62, 19-26.
PreviatoJ.O., AndradeAF., Pessolani,M.C. and Mendonca-Previato,L. (1985)
Incorporation of sialic acid into Trypanosoma cruzi macromolecules: a proposal foT a new metabolic route. MoL Biochem. Parasitol, 16, 8 5 - % .
RoggentinJ3., Rothe.B., KaperJ.B., GalenJ., Lawrisuk.L., Vimr.E., and
Schauer,R. (1989) Conserved sequences in bacterial and viral sialidases.
Glycoconjugate J., 6, 349-353.
SakuradaJC., Ohta.T. and Hasegawajv). (1992) Qoning, expression, and characterization of the Micromonospora viridifaciens neuraminidase gene in
Streptomyces lividans. J. Bacterial, 174, 6896-6903.
Schenkman,S., Jiang,M., Hart,G. and Nusenzweig.V. (1991) A novel cell surface trans-sialidase of Trypanosoma cruzi generates a stage-specific epitope
required for invasion of mammalian cells. Cell 65, 1117-1125.
Schenkman.S., ChavesJ_.B., Pontes de Carvalho.L.C. and Eichinger.D.
(1994a) A proteolytic fragment of Trypanosoma cruzi trans-sialidase lacking the carboxy-terminal domain is active, monomeric, and generates antibodies that inhibit enzymatic activity. J. BioL Chem., 269, 7970-7975.
Schenkman^., Eichinger.D., Pereira,M. and Nussenzweig.V. (1994b) Structural and functional properties of Trypanosoma cruzi trans-sialidase. Annu.
Rev. Microbiol, 48, 499-523.
SchuppJJ^., Travis.S.E., Price,L.B., Shana\R.F. and Keim,P. (1995) Rapid
bacterial permeabilization reagent useful for enzyme assays. BioTechniques,
19, 18-20.
Scudder.P., DoomJ.P., Chuenkova.M., Manger.I.D. and Pereira.M.E.A.
(1993) Enzymatic characterization of P-D-galactoside a2,3-trans-sialidase
from Trypanosoma cruzi J. BioL Chem., 268, 9886-9891.
Smith,L.E., Uemura,H. and Eichinger J). (1996) Isolation and expression of an
open reading frame encoding sialidase from Trypanosoma rangeli. MoL
Biochem. ParasitoL, 79, 21-33.
Trans-sialidase domains
Udenfriend.S., Stein,S., Bohlen.P., Dairman.W., Leimgruber.W. and
Weigele,M. (1972) Huorescamine: a reagent for assay of amino acids, peptides, proteins, and primary amines in the picomole range. Science, 178,
871-872.
Uemura,H., Schenkman,S., Nussenzweig.V. and EichingerJD. (1992) Only
some members of a gene family in Trypanosoma cruzi encode proteins
which express both trans-sialidase and neuraminidase activities. EMBO J.,
11, 3837-3844.
Vandekerclchove,F., Schenkman.S., Pontes de Carvalho,L.C., Tomlinson.S.,
Kiso,M., Vrujhirfa M Hasegaw^A. and Nussenzweig.V. (1992) Substrate
specificity of the Trypanosoma cruzi trans-sialidase. Gfycobiology, 2, 541548.
Vimr,E-R. (1994) Microbial sialidases: Does bigger always mean better?
Trends MicrobioL, 2, 271-277.
Weis.W.I. and Drickamer^C (1996) Structural basis of lectin-carbohydrate
recognition. Annu. Rev. Biochem., 65, 441-473.
YonJ. and Fried>f. (1989) Precise gene fusion by PCR. Nucleic Acids Res.,
17, 4895.
Received on October 11, 1996; revised on November 12, 1996; accepted on
November 21, 1996
451