(Received for publication, October 29, 1996)
From the Laboratory of Cell Biology, Department of Biochemistry, Imperial College of Science, Technology, and Medicine, Exhibition Road, London SW7 2AY, United Kingdom
The Ras superfamily of small G proteins governs unidirectional cellular processes by virtue of GTP hydrolysis and concomitant conformational changes, which are in turn regulated by a number of accessory factors. Members of the Rab subfamily are important for correct targeting and fusion of intra-organellar vesicles loaded with trafficking proteins and lipids. During evolution from a prototype gene, novel functions may be acquired by duplicated daughter genes; for Rab proteins, this can be tested by location, which is specifically related to the function of each Rab. We have found an example of two rab genes in Trypanosoma brucei (trab genes) that clearly arose by tandem duplication, being highly related to each other and remaining juxtaposed in the genome, whose products have dramatically different subcellular locations, indicative of discrete functions. These two trab genes, isolated on a single genomic clone, are separated by a short intervening sequence and are in a head-to-tail orientation. The nucleotide sequences of the open reading frames and intervening sequence were determined and show that the genes are paralogues, probably arising from an ancient tandem duplication. Both genes are most homologous to ypt1 and sec4 in the Saccharomyces cerevisiae genome, while phylogenetic reconstruction indicates that although they have clearly diverged, the proteins are more closely related to each other than to other Rab protein sequences available in the data base. Immunofluorescence microscopy, using antibodies raised against the recombinant Trab proteins, clearly demonstrates that the native Trab proteins have completely distinct subcellular locations in the trypanosome. Trab1p is present in a widespread reticular location similar to BiP, suggesting an endoplasmic reticulum location, while Trab7p is observed in a discrete structure adjacent to the kinetoplast. Most interestingly, the Trab7p-positive compartment also appears to divide at the same time, or just prior to, the kinetoplast, i.e. early in mitosis, suggestive of association with structures in the flagellar pocket region. An estimate of the divergence time indicates that the trab1/trab7 duplication occurred ~100 million years ago, and therefore, the persistence of this pair suggests an essential role in the survival of T. brucei.
Trypanosoma brucei spp., the causative agent of nagana in ungulates and sleeping sickness in humans, is a parasitic protozoan of economic and public health importance. Studies of this organism have revealed many unusual biological phenomena later observed in other systems, perhaps most remarkably trans-splicing (1) and glycosylphosphatidylinositol protein anchors (2). The paradigm glycosylphosphatidylinositol-anchored protein, the variant surface glycoprotein (VSG),1 accounts for ~10% of total protein biosynthesis in the mammalian bloodstream form (BSF) of the parasite and provides an essential immunological barrier. The entire VSG coat, comprising 107 identical molecules, is replaced during antigenic switching in the BSF by a process of continued biosynthesis of new VSG and internalization/shedding of the old coat protein. Much VSG recycles: internalization is concomitant with surface delivery of newly synthesized VSG in the flagellar pocket (FP). In the insect procyclic form, VSG is replaced by the abundant glycosylphosphatidylinositol-anchored procyclic acidic repetitive protein (PARP) (3).
Rates of surface protein internalization are comparable with that of a mammalian cell, but in a trypanosome, only the FP, a small area of the cell surface, is available for exo- and endocytic activity due to a subplasmalemmal microtubular network over the rest of the cell body (4). The FP has four specialized adjacent microtubules running parallel down one side of the structure, coincident with a specialized area, the cytosome. This may be the site of exo- and endocytosis since it is associated with vesicles carrying VSG arising from the Golgi complex and with vesicular tubules and flattened cisternae structurally similar to the trans-Golgi network. Coated vesicles budding from the FP are observed only in BSF parasites (5). The FP is of interest for several reasons; first, for membrane trafficking of glycosylphosphatidylinositol-anchored glycoproteins; second, for sorting mechanisms, as endo- and exocytosis are presumably simultaneously active within a small membrane area (6, 7); and finally, as a potential site for therapeutic attack.
Detailed exploration of intracellular trafficking requires suitable markers, but the wide evolutionary distance between trypanosomes and most other eukaryotes is unfortunately reflected in a lack of immunological cross-reactivity between homologues. Marker proteins that are available include trypanosome BiP, an endoplasmic reticulum marker (8, 9), and CRAM, a cysteine-rich acidic transmembrane protein found in the FP and associated endocytic vesicles (10). Also available is an antibody to the CB1 epitope of a developmentally regulated 44-kDa protein of unknown function in the endocytic pathway, found in the FP, endosomes, and lysosomes (7).
The Rab family consists of GTPases with ~30% identity to Ras and with most homology confined to four GTP-binding motifs (11). Rab proteins are essential for protein trafficking in all eukaryotic organisms. Family members have low intrinsic GTPase activity, requiring activating proteins (GTPase-activating proteins) to increase turnover and act as effectors (12). GTP- and GDP-bound Ras have distinct conformations (13), and it can be inferred that Rab proteins will similarly undergo a conformational change upon GTP hydrolysis since their structure is highly homologous. Each Rab protein binds a subset of organelles; the C terminus is responsible for intracellular localization, while the N terminus recognizes the target organelle and is required for vesicle fusion (14). The paucity of markers suitable for trafficking studies led us to initiate the cloning of rab genes from T. brucei (15) since each Rab protein has a discrete subcellular location within the secretory pathway (16, 17). The study of T. brucei Rab proteins (Trab proteins) will yield new sequences important for targeting proteins to trypanosome organelles, provide information on the compartmentalization of trypanosomes, and generate subcellular markers.
A degenerate PCR cloning approach has yielded eight expressed sequence tags (EST) of small GTP-binding proteins from T. brucei, including Ran/TC4 (18) and seven Trab homologues (15), while separate studies have identified three additional trab genes (19, 20). We isolated two trab genes on a single genomic clone (15), and here we report the genomic organization of these genes (previously rtb1 and rtb7) and the subcellular location of the encoded proteins.
Molecular
biology-grade water was from an Elgastat Option 4 system (Elga)
sterile-filtered through a 0.2-µm filter (Sartorius Corp.).
Restriction and DNA-modifying enzymes were from Bioline, Promega, MBI
Fermentas, New England Biolabs Inc., and Stratagene. Nitrocellulose was
from Schleicher & Schuell (obtained from Anderman & Co. Ltd., London).
Radioisotopes were from DuPont NEN and Amersham International. A P1
T. brucei library filter was a gift from Dr. S. Melville
(Department of Pathology, Cambridge University, Cambridge, UK). Vector
pGEX2tk and glutathione-Sepharose 4B were from Pharmacia Biotech Inc.,
and the ZAPII kit was from Stratagene. Molecular biology methods and
antibody preparations followed standard procedures (21, 22). Plasmids
were grown in Escherichia coli XL1-Blue (Stratagene)
following transformation by electroporation with a BTX 600 ECM
electroporator. PCR products and gel-embedded DNA were purified using
PCR cleanup kits (Promega), and plasmid and
DNAs were purified
using QIAGEN kits following the manufacturers' instructions.
Small-scale plasmid preparations were performed using the Promega
Wizard system. Rabbit antibodies to T. brucei BiP were from
J. Bangs (University of Wisconsin, Madison, WI). Secondary antibodies
were from Sigma.
PCR was performed in 50-µl reactions with 25 pmol of
each primer in ammonium buffer (Bioline), 2.5 mM
MgCl2, and 1 unit of Taq polymerase in a Model
480 thermal cycler (Perkin-Elmer) as follows: one cycle at 98 °C for
5 min; one cycle at 95 °C for 5 min; and 35 cycles at 94 °C for 1 min, 56 °C for 1 min, and 72 °C for 3.5 min. Templates were 1 µl of a 1:1000 dilution of R111 phage lysate or 1 µl of a
ZAPII plaque equilibrated in 0.4 ml of SM phage dilution medium (21)
or 10 ng of plasmid; phages were boiled for 10 min prior to PCR. The
primers for determining orientation of trab1 and
trab7 were RAT1FUS, RAT7PCR, and two reverse primers
designed from the EST (15). The primers for subcloning the open reading
frames (ORFs) into expression vectors were as follows: trab7
5
, GTGAAGTTTGGATGGATGAAAGAGGAACCC; trab7 3
,
CCACCATAAGGAATTCAGCACCTCCACTT; trab1 5
,
GAGTCCACTGGATCCATGATCACAGCAGC; and trab1 3
,
CGACGAGAGGAATTCTTAACAGCAACCACCC. For expression of recombinant Trab
proteins, each ORF was amplified from pBS.R7X1 (see below), digested
with EcoRI/BamHI for 16 h at room
temperature, and ligated into EcoRI/BamHI-cleaved
pGEX2tk.
0.5-µg aliquots of R111 DNA were
cut with the AvaII, BglII, ClaI,
HindIII, HincII, PvuII,
SacII, SalI, XbaI, or XhoI
restriction enzyme, and duplicate filters were prepared by high salt
transfer Southern blotting (21). Filters were probed at high stringency (0.05 × SSC and 0.1% SDS at 75 °C) with trab1 or
trab7 ESTs cut from pBluescript with
EcoRI/BamHI, gel-purified, and labeled using a
random hexamer priming kit (Boehringer Mannheim). To map R111, 0.5-µg
aliquots were cut with one or two restriction enzymes
(AvaII, ClaI, SalI, XbaI,
and XhoI), blotted, and probed with the WD box primer (15)
labeled with [
-32P]ATP using polynucleotide kinase at
low and high stringency (6 × SSC and 0.1% SDS at room
temperature, and 1 × SSC and 0.1% SDS at 65 °C,
respectively). Single and double restriction digests of pBS.R7X1 were
blotted and probed with trab1 and trab7 ESTs. To
estimate the total number of trab genes in the T. brucei genome, a bacteriophage P1 library filter (SM5 No. 12) of
T. brucei strain TREU 927/4 genomic DNA in grid format, with
~3 × haploid genome coverage, was probed with the WD primer at low
stringency. Autoradiography was with X-Omat 5 film (Eastman Kodak Co.)
at
85 °C with an intensifying screen; exposure times were 16 h to 7 days.
A 6-kb XbaI fragment
from R111 (15) was subcloned into pBluescript using a ZAPII kit;
recombinant
plaques (from the whole R111 XbaI digest)
were screened on duplicate filters with trab1 and
trab7 EST probes. Secondary screening was by PCR (15), and one clone was rescued to produce plasmid pBS.R7X1.
Sanger sequencing of pBS.R7X1 through the ORFs and intervening region (IVS) was performed using a deaza-GTP kit (Pharmacia) and primer walking in both directions. Dye terminator cycle sequencing of untranslated regions and constructs was performed using a Taq thermal cycle sequencing kit and a Model 377 DNA sequencer (Perkin-Elmer) with polyethylene glycol-precipitated DNA (6.5% polyethylene glycol 6000 and 0.4 M NaCl on ice 20 min; centrifuged at 13,000 × g for 30 min at 4 °C; washed with 70% EtOH).
AntibodiesExpression plasmids pGEX2tk.Trab1 and
pGEX2tk.Trab7 in E. coli were grown in L-broth and induced
with 1.0 mM
isopropyl--D-thiogalactopyranoside. Fusion proteins were
affinity-purified on glutathione-Sepharose 4B and thrombin-cleaved to
produce recombinant Trab1p and Trab7p, as described (23), with the
following modifications. Protease inhibitors (pepstatin, leupeptin,
antipain, and chymostatin, all at 2 µg/ml (protease inhibitor
mixture)) and TLCK (dry) were added to the initial lysate. Cleared
lysate was equilibrated with glutathione-Sepharose for 16 h at
4 °C. Thrombin treatment was at room temperature for 2 h, with
more thrombin added after 1 h. Protein was quantitated by SDS-PAGE
and Coomassie Blue staining. Typical yields were 0.4-4 mg of rTrab
protein/liter of E. coli, with a purity of 95%. Rabbits were immunized three times with a total of 3 mg of rTrab1p or 0.9 mg of rTrab7p using the MPL® + TDM + CWS Adjuvant System (Sigma). Animals were exsanguinated, and serum was stored at 4 °C with azide.
For affinity purification, 0.2 g of CNBr-Sepharose (Pharmacia) was
activated and resuspended with an equal volume of coupling buffer (100 mM NaHCO3 and 500 mM NaCl, pH 8.5)
containing 0.5 mg of rTrab protein in phosphate-buffered saline and
coupled overnight at 4 °C. Efficiency was ~45%. Purified
antibodies were prepared from 50% ammonium sulfate-precipitated serum
and stored at
85 °C.
Culture-adapted BSF T. brucei strain 427, a gift from S. Croft (London School of Hygiene and Tropical Medicine), was grown in HMI-18 (24) at 37 °C in a 5% CO2 incubator. Procyclic form T. brucei strain 427 was grown in SDM79 (25); media and supplements were from Sigma. For boiling SDS lysates, freshly harvested trypanosomes were added to SDS-PAGE sample buffer (26) at 95 °C, heated for 5 min, and reduced with dithiothreitol. For hypotonic lysates, trypanosomes were resuspended in 10 mM Tris, pH 7.5, protease inhibitor mixture, phenylmethylsulfonyl fluoride, and TLCK, followed by two cycles of freeze/thaw and centrifugation at 13,000 × g for 15 min at 4 °C.
Western BlottingWestern blotting was performed on proteins electrophoresed on 17% SDS-polyacrylamide minigels and wet-blotted at 100 V for 3 h in transfer buffer (Hoefer Pharmacia Biotech Inc.) onto 0.45-µm nitrocellulose. Filters were blocked in 3 or 5% bovine serum albumin, phosphate-buffered saline, and 0.1% Tween 20 at 4 °C for 16 h. All remaining manipulations were at room temperature. Blots were probed with primary antibodies (in a 2-4-ml block) for 1 h, washed four times with phosphate-buffered saline/Tween 20, and then incubated with secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit IgG diluted 1:2000) in blocking buffer for 20 min and washed. Detection was with either NiCl2/diaminobenzidine (21) or ECL (Amersham Corp.) exposed to x-ray film. Purified antibodies at concentrations of 0.43 mg/ml (anti-Trab1p) and 0.17 mg/ml (anti-Trab7p) were typically used diluted by 1:1000 and 1:200, respectively.
Immunofluorescence AnalysisImmunofluorescence analysis was as described (8) with the following modification. Cells were fixed for 80 min with 1.5% formaldehyde and 0.32% glutaraldehyde (v/v). Purified anti-Trab antibodies were used at a 1:200 dilution. Fluorescence microscopy was performed on a Nikon Microphot-FX microscope using a 100× oil immersion objective and Ilford HP5 Plus ISO400 film.
Computer AnalysisThe GenBankTM Data Bank was searched via the Internet using Netscape Version 2.0 (Netscape Communications Corp.) and FASTA or TXBLASTP against the non-redundant data base "nr" (National Center for Biological Information). The Saccharomyces cerevisiae genome was searched at the Stanford Medical School Web site.2 Turbogopher was used for sequence retrieval from GenBankTM. The University of Wisconsin Genetics Computer Group package was used at the Daresbury site (United Kingdom) via MacIP and Versaterm Pro for the following algorithms: SEQED, BESTFIT, GAP, PILEUP, MFOLD, SQUIGGLES (PLOTFOLD), COMPARE (DOTPLOT), and PEPTIDESTRUCTURE. Navigator and Autoassembler (Perkin-Elmer) and NIH Image were used on a Macintosh LCIII computer. For phylogenetic analysis, PAUP 3.1.1 (27) was used on a Macintosh Quadra 650; initial tree searches were done in branched and bound or exhaustive mode, with Ha-Ras assigned as outgroup. Preliminary results were bootstrapped (1000 replicates) in exhaustive mode. Divergence time was calculated using the program of Li (28). Three-dimensional computer modeling of Trab1p was performed automatically via SwissProt at EXPASY, and that of Trab7p was performed using INSIGHT II (Homology module) on an Indigo computer (Silicon Graphics Inc.) against the Ras pdb file (221P); four non-aligned loops were modeled de novo in free space and energy-minimized.
Limited restriction
endonuclease mapping of R111, a genomic clone containing both
trab1 and trab7 (15), showed the insert was at
least 26 kb long. The genomic organization of the trab genes
was further explored by PCR using forward (F) and reverse (R) primers
to each EST. We tested all four possible orientations: trab1
and trab7 oriented head-to-tail with trab1 first
(1F/7R), head-to-tail with trab7 first (7F/1R), head-to-head
(1F/7F), or tail-to-tail (1R/7R). A ~2.4-kb product was obtained
using only the second primer combination, 7F/1R (Fig.
1A, second lane), indicating that
trab7 preceded trab1 in a head-to-tail
orientation, with an IVS of ~1.8 kb between ORFs. Identity of the PCR
product was confirmed by digestion with XhoI and
HincII (Fig. 1B), previously demonstrated to
separate the ESTs in a Southern blot (Fig. 1C). The
locations of the restriction sites were confirmed by sequence analysis
(see below). The size of the IVS was initially overestimated due to
anomalous migration of the PCR product; the gene organization was
confirmed by sequence analysis (see below).
trab Genes in T. brucei
We analyzed clone R111 for additional trab genes to determine if trab7 and trab1 represent a simple duplication event or form part of a more extensive array. A degenerate oligonucleotide annealing to the conserved sequence WDTAGQE from the second GTP-binding site motif of Rab proteins (WD primer; see "Materials and Methods") was used in a low stringency Southern blot of genomic clone R111. No fragments other than those hybridizing to trab1 and trab7 probes were identified (data not shown), suggesting that these two trab genes do not form part of an extensive array. A 6-kb XbaI fragment from R111, which hybridized to trab7 and trab1 probes, was subcloned into pBluescript for sequence analysis (pBS.R7X1; see "Materials and Methods"). Mapping the position of the trab7 gene demonstrated that R111 contains only ~2 kb of DNA upstream of the trab7 ORF, and so the possibility of another trab gene lying upstream has not formally been excluded, although we believe it highly unlikely (see below and data not shown).
We also examined the total number of trab sequences in T. brucei by probing a ~3-fold degenerate (for haploid genome) T. brucei genomic library array with the WD box oligonucleotide at low and high stringency. Thirty-three positives were identified, corresponding to ~20 distinct trab loci. As previous Southern analysis and the studies of R111 have failed to identify extensive trab gene arrays (15), this number also provides an estimate of the total number of individual trab genes present in the T. brucei genome.
Sequence Data for trab1 and trab7 Genes and Their Intervening SequenceWe sequenced ~3.3 kb of subclone pBS.R7X1 including
the trab1 and trab7 ORFs, the IVS, and the
surrounding regions (Fig. 2). The trab1 and
trab7 ORFs were identified by hypothetical translation of
the nucleotide sequences. The orientation of the genes was confirmed,
and the length of the IVS was determined to be 1165 base pairs. The ORF
nucleotide sequences are 48.8% identical between trab1 and
trab7.
Both genes are expressed as assessed by Northern blotting and reverse transcription-PCR (15), and we identified consensus features for mRNA production (Fig. 2). Mini-exon splice sites and ATG start codons were located by comparing the sequence of the EST, derived from reverse transcription-PCR, with genomic sequences. A potential polyadenylation sequence for trab7 was inferred from poly(T) tracts, with an appropriate GA addition site conforming to the consensus sequence (29). The predicted trab7 message is 1.8 kb, consistent with Northern blotting data. The polyadenylation site of trab7 is placed in very close juxtaposition to the mini-exon splice site of trab1: these two processes are known to be highly coordinated in trypanosomes (30). There is no polypyrimidine tract upstream of trab7, the first gene, suggesting that there is no gene upstream of trab7 that is polyadenylated prior to the trab7 splice site, further evidence that there are no other genes in this array.
trab1 and trab7 messages are expressed at low
levels compared with T. brucei Ran/TC4 mRNA (18) and
other well characterized mRNAs such as PARP and VSG, all of which
are high abundance messages. Also, these trab genes are
constitutively expressed (15). As expected, none of the sequence motifs
demonstrated to be involved in the maintenance of high expression
levels (31) or stage-specific expression of PARP (32) or VSG (33) were
present in the trab1 and trab7 genes. Although
the structure of the trab7 and trab1 locus
suggested a tandem duplication event, no strong candidates for possible
recombination sites were found within the sequenced region of pBS.R7X1,
i.e. related sequences upstream, downstream, and between
trab7 and trab1, nor were repetitive sequence
motifs implicated in recombination between VSG genes observed (34). However, homology was found between the IVS (between genes) and a
repetitive element in the T. brucei mini-exon donor gene,
which in turn had homology to the reverse complement of a region in the
3-untranslated region of trab1
(double-underlined in Fig. 2), a potential Z-DNA site, and
three other regions of internal homology were identified by the program
COMPARE (boxed in Fig. 2). The presence of a limited stretch
of repetitive DNA observed elsewhere in the T. brucei
genome, may be significant for the evolution of this locus.
A divergence time of 100 million years ago was calculated from the
nucleotide ORF sequences of trab1 and trab7 using
the method of Li (28). This suggests that these genes are paralogues
arising from an ancient duplication event. The homology of the amino
acid sequence (64% similar) is greater than that of the nucleotide sequence (49%), so although synonymous nucleotide substitutions were
saturated (28), considerable constraints on the divergence of the
protein sequences are evident, indicating that the two gene products
are important for T. brucei survival. The ORF sequences of
trab7 and trab1 were used to search the data
base, and both are clearly Rab homologues at the nucleotide and amino
acid levels. However, Trab1p and Trab7p could not be unambiguously
assigned as particular Rab homologues, which suggests a
trypanosome-specific function for Trab1p and Trab7p. Because of species
redundancy in the GenBankTM Data Bank, we also determined which genes
in another unicellular eukaryote, S. cerevisiae, were most
similar to trab1 and trab7. trab1 was most
homologous to sec4, while trab7 was most similar
to ypt1, which are themselves closely related. We analyzed
the relationship of Trab1p and Trab7p to one another and to other
homologous Rab sequences by phylogenetic reconstruction. Trab1p and
Trab7p always cosegregated as a monophyletic group whether assessed
against their closest Rab homologues (Fig. 3A and data not shown) or compared with kinetoplastid Rab sequences (Fig.
3B), including the putative Rab5 homologues Trab5Ap
(Tbrab5 (19)) and Trab5Bp.3
Protein Sequences of Trab1p and Trab7p
Hypothetical
translations of the Trab1p and Trab7p ORFs indicate that the protein
sequences are 43% identical and 64% similar. Both sequences contain
all of the major homology boxes typical of a Rab protein, including
most of the residues involved in GTP/GDP binding and GTPase-activating
protein binding (Fig. 4A). Trab1p also has a
typical C-terminal amino acid motif for double geranylgeranylation, CCGG.
Interestingly, Trab7p contains several atypical features. Most
significantly, the GTPase motif characterized by the WD box, normally
WDTAGQE for Rab proteins (PM3 in Fig. 4A), is altered in
Trab7p to WDTGQ. The Glu Asp change is
not uncommon, but the Ala
Ser change (corresponding to a G
T
nucleotide transversion in the first position of the codon) is highly
unusual and possibly functionally significant; human RhoE also has an
Ala
Ser substitution at this position and binds GTP but not GDP.
Additionally, both Trab7p and RhoE have Ser at codon 12 (Ras
numbering); combined with the Ala
Ser substitution, this results in
lack of GTPase activity in RhoE (35). Also, constitutively active viral
Ki-Ras has an Ala
Thr mutation in the WD box (36). Overall, these data suggest that Trab7p may possess abnormal GTPase activity.
Trab7p has a 12-amino acid insertion when compared with Ha-Ras, Ki-Ras, Ran, and at least 10 other Rab proteins (Fig. 4A and data not shown). Alignments place this insertion as an extension of loop 8 (nomenclature of Wittinghofer and Pai (13)). Residues important in the adjoining secondary structures are conserved in both Trab1p and Trab7p (Fig. 4B). The 12-residue loop of Trab7p was hypothetically large enough, by three-dimensional modeling, to interact with the adjacent nucleotide-binding site (data not shown).
Trab1p and Trab7p in TrypanosomesTo examine the expression
and location of Trab1p and Trab7p, we raised polyclonal antisera to
recombinant proteins. Affinity-purified anti-rTrab1p and anti-rTrab7p
antibodies were used in Western analysis of trypanosome lysates.
Specific signals were obtained from material migrating at 27 and 29 kDa, respectively, from both life stages (Fig.
5A). The theoretical molecular masses are 24 and 21.5 kDa for Trab1p and Trab7p, respectively, suggesting aberrant migration on SDS-PAGE. However, this was not due to post-translational modifications because the migration positions of rTrab1p and rTrab7p produced in E. coli, which does not isoprenylate proteins,
were identical to those of the native trypanosome proteins (data not shown). Trab7p was also found to be highly labile and was seen only in
fresh extracts.
Membrane Association of Trab Proteins
The unusual C-terminal sequence of Trab7p suggested that this protein may not be isoprenylated efficiently, and therefore, we wished to test if the protein was membrane-associated. Trab1p, by contrast, with its canonical C-terminal isoprenylation motif, is predicted to be isoprenylated and therefore membrane-associated. Procyclic cells were hypotonically lysed and centrifuged, and the soluble and particulate material was resolved by SDS-PAGE. Trab1p and Trab7p were located by Western blotting, and proteins were quantitated in identical samples (see "Materials and Methods"). Trypanosome proteins fractionated equally into soluble and particulate material, indicating efficient cell lysis, while both Trab1p and Trab7p appeared solely in particulate fractions (Fig. 5B). These data indicate that Trab7p is membrane-associated. In additional, our data are consistent with previous observations that some Rab proteins are exclusively membrane-bound and have no cytoplasmic pool (37).
Subcellular Locations of Trab1p and Trab7pTo begin to
ascribe a function to Trab1p and Trab7p, we next chose to investigate
the subcellular locations of the two proteins. We used
affinity-purified antibodies to immunolocalize the native proteins by
immunofluorescence analysis. Anti-Trab1p staining revealed a reticular
pattern in both procyclic and BSF cells (Fig. 6,
A and C) very similar to the appearance of cells
stained for BiP, an endoplasmic reticulum luminal marker (Fig. 6,
B and D). 10 µg of rTrab1p premixed with
primary antibody for 10 min was sufficient to abolish the signal in BSF
cells, demonstrating specificity (data not shown). Therefore, an
endoplasmic reticulum location is tentatively assigned to Trab1p.
In contrast, cells stained with anti-Trab7p antibody revealed one or two discrete and highly compact structures in each trypanosome cell. The Trab7p-positive compartment was located close to the kinetoplast in both procyclic and BSF trypanosomes (Fig. 6, E-J). Prior incubation of the antibodies with 10 µg of rTrab7p abolished the signal completely (data not shown). Most significantly, in cells containing two kinetoplast DNA networks (as revealed by Hoescht stain), two Trab7p-positive structures were observed positioned at either end of the kinetoplast (Fig. 6, G-J). The location of Trab7p is similar to that described for the basal bodies of trypanosomes (38), but clearly, further analysis is required to refine this assignment. Taken together, the immunolocalization data demonstrate radically different subcellular locations for Trab1p and Trab7p. Therefore, a tandem duplication resulted in acquisition of functional differences in these highly related Rab homologues.
We have identified a pair of small GTPase genes juxtaposed in a head-to-tail orientation in the genome of T. brucei. We were unable to find evidence for a reverse transcriptase-mediated duplication mechanism, e.g. mini-exon or poly(A) tail vestiges. While there are two genes present in the locus (trab1 and trab7), five other trab genes and the gene for Ran/TC4 are dispersed and exhibit no linkage with each other (15, 18), suggesting that the trab1/trab7 pair is unusual. In addition, all our data suggest that trab1/trab7 are not part of a more extensive trab gene array. Phylogenetic reconstruction indicates that Trab1p and Trab7p are more similar to each other than to any other Rab sequence, evidence that they are paralogues. Taken together, these data strongly argue that these genes are the result of an ancient tandem duplication event, estimated to have occurred ~100 million years ago.
Duplication, probably by homologous recombination, has resulted in many gene arrays in T. brucei, e.g. the PARP locus with arrays of two or three genes (31) and the extensive tubulin locus (39). Homologous recombination is highly active (40-42) and is probably responsible for producing and manipulating the reservoir of up to 1000 different VSG genes providing diverse antigenicity (34, 43). Tandem arrangement of genes allowing polycistronic expression is a common strategy in T. brucei and may facilitate coordination of gene expression. Indeed, trab1 and trab7 are transcribed at low levels and in equal amounts (15).
By probing of a genomic library of known complexity, the total number of Trab-related genes was estimated to be ~20. Since 11 trab clones and ESTs have been reported to date, it is likely that a representative sampling of trab sequences, covering most of the trypanosome intracellular trafficking pathways, is already in hand. This number is similar to that obtained for S. cerevisiae, also a unicellular eukaryote, and is in agreement with our earlier estimate of total trab numbers (15). Therefore, these data suggest that, as a first approximation, T. brucei possesses a secretory pathway of equivalent complexity to yeast.
We propose that the continued linkage of trab1 and trab7 is due to functional selection. Conservation of protein sequence contrasted with saturated synonymous nucleotide replacement strongly suggests a functional selection on the gene products. While both Trab sequences are homologous to a selection of Rab sequences in the data base, neither exhibited strong homology to a particular Rab sequence, which suggests instead that they are trypanosome-specific. The immunolocalization data for Trab7p, indicating association with the kinetoplast and FP, are clearly consistent with this proposal. In addition, we were unable to detect Trab1p or Trab7p in COS-7 cells, either by Western blotting or by immunofluorescence analysis, or in Leishmania major, by Western blotting.3 While the former result was not unexpected, the lack of immunological cross-reactivity between Trab proteins and Leishmania Rab proteins is also consistent with a species-specific function.
Trab1p has canonical Rab sequences, including a C-terminal
isoprenylation signal, and shows diffuse reticular staining consistent with a Ypt1p/Rab1p function. The homology to sec4 (and the
closely related ypt1) detected by searching the S. cerevisiae genome may reflect this. In contrast, Trab7p has
several unusual features. An important alteration in Trab7p is an Ala
Ser substitution in the WD box, potentially resulting in a
GTP-bound, constitutively active protein like RhoE (35) or viral
Ki-Ras, which has an Ala
Thr mutation (36). A second change in
Trab7p is a 12-amino acid loop adjacent to the nucleotide-binding
pocket. We believe it is likely that Trab7p binds only GTP and does not
hydrolyze it, but clearly, this needs to be shown empirically. The
subcellular location of Trab7p is also highly unusual, consisting of a
single spot in nondividing cells located close to the kinetoplast. This structure divides at a similar time to the kinetoplast, clearly preceding nuclear division, suggesting association of the Trab7p compartment with basal body/FP structures. Like the Trab7p compartment, the basal body and FP are located close to the kinetoplast and divide
with it, prior to nuclear division (38).
Carboxyl-terminal isoprenylation is essential for Rab function and occurs in T. brucei (44), and in all Rab homologues reported in T. brucei with the exception of Trab7p (KWRC), canonical isoprenylation sequences are found, i.e. SCAC (Trab4Bp), GCAC (Trab5Ap), GCCG (Trab5Bp), CCGG (Trab1p), and DSCC (Ypt1p from kinetoplastid L. major). Despite its unusual signal sequence, native Trab7p is membrane-associated, which suggests that it is a substrate for the trypanosome isoprene transferases. It remains to be determined which isoprene is present on Trab7p.
The divergence of Trab7p from the canonical Rab sequence and the association with a kinetoplast-specific organelle are suggestive of a novel role for Trab7p in the trypanosome cell. We propose that trab7 arose by duplication of a progenitor trab gene, with Trab1p now fulfilling a more basic endoplasmic reticulum-associated function and Trab7p co-opted into a specialized role. Trab1p and Trab7p are 43% identical and 64% similar, reminiscent of Arf1 and Arf6, which are 66% homologous and also have different locations (45). Because of the absence of clear homologues for Trab1p and Trab7p, the precise functions of these proteins must be determined directly, which is being actively pursued. Efforts to further localize the individual proteins and to gain more insight into their function by overexpression and direct study of the GTPase activities of these potentially important trypanosome proteins are currently under way. In summary, we report the presence of a tandem pair of rab genes in the protozoan parasite T. brucei. Immunolocalization and sequence data suggest a species-specific role for at least one (Trab7p). The trab1/trab7 pair has remained together for a considerable period of time, indicative of functional importance.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) X99951[GenBank].
We thank Anneke Loos for sequencing IVS and mapping studies, Jay Bangs for antibodies to trypanosome BiP, Simon Croft for culture-adapted BSF trypanosomes, Richard Thomas (Museum of Natural History, London) for analysis of the divergence time, and Ann Ridley for the thrombin cleavage protocol.