J Biol Chem, Vol. 275, Issue 15, 11432-11439, April 14, 2000

Transcriptional Analysis of the Glutamate Dehydrogenase Gene in the Primitive Eukaryote, Giardia lamblia
IDENTIFICATION OF A PRIMORDIAL GENE PROMOTER^*

Janet Yee, Michael R. Mowatt, Patrick P. Dennis^§, and Theodore E. Nash

From the Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 and the ^§ Department of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, Canada

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We studied gene expression in the ancient eukaryote, Giardia lamblia, by taking advantage of assays developed recently in our laboratory, which allow new genetic analyses of this organism. We examined the transcription of a 2.2-kilobase segment of the Giardia genome that contains the glutamate dehydrogenase (GDH) gene and a portion of a second open reading frame encoding an uncharacterized gene. Nuclear run-on analyses showed that the genes are transcribed as two separate units spaced less than 200 base pairs apart, and transcription of the GDH gene initiates just 3-6 nucleotides upstream of its translation start codon. We characterized the GDH promoter by transfecting Giardia with DNA constructs that used the GDH upstream sequence to drive the expression of a luciferase reporter gene. By deletion and mutational analyses, we localized promoter function to three motifs within a 50-base pair region of the GDH upstream sequence. Using band shift assays and UV cross-linking, we demonstrated specific binding of a 68-kDa protein from Giardia nuclear extracts to short poly(T) tracts contained within two of the sequence motifs on single-stranded DNA from the promoter region. This report describes one of the first functional gene promoter and its cognate DNA-binding protein in this primitive eukaryote.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The protozoan Giardia lamblia is a representative of the most primitive lineage within the eukaryotic domain (1-3). This protozoan is found in freshwater lakes and streams throughout the world, and is an intestinal parasite of humans and other mammals. Giardia has just two stages in its simple life cycle: a swimming trophozoite form, and an infective cyst form. The transformation of the trophozoite into the cyst is accompanied by the activation of gene transcription of cyst-specific genes (4). Giardia trophozoites have the unusual ability to switch the make-up of the major protein on its cell surface by varying the expression of a repertoire of genes encoding these variant-specific surface proteins (VSPs).¹ Although antigenic variation in Giardia is commonly thought to be a mechanism to evade the immune response of the host, the biological significance of VSPs in Giardia has not been determined (5, 6).

The initiation of transcription is a key step in regulating gene expression and has been studied extensively in organisms ranging from yeast to man. Although results from these studies were presumed to be representative of all eukaryotes, little is known about this process in organisms lower than yeast on the evolutionary tree. In higher eukaryotes, the basal level of transcription by RNA polymerase II is dependent on the assembly of a set of general transcription factors on the core promoter which consists of either a TATA box, an initiator, or both (reviewed in Ref. 7). While the binding of the transcription factor TBP (TATA box-binding protein) to the TATA box has been well characterized, less is known about proteins which interact with the initiator. Since TBP is required for transcription from TATA-less promoters, it is possible that TBP can bind directly to the initiator, or alternatively, it is recruited indirectly to this region by its interaction with another protein which binds this motif (8). In some genes, additional factors and their specific recognition sites are used to modulate transcription regulation. For example, the mammalian -globin and the histone H2B genes have an additional promoter element, called the CAAT box, located between 40 and 100 bp upstream of the transcription initiation site (reviewed in Ref. 9). This element, with the consensus sequence GG(C/T)CAATCT, is bound by a family of transcription factors which up-regulate transcription of these genes.

Since little information is available about promoter structure and function in lower eukaryotes, we were interested in determining the requirements of a "minimal" gene promoter in one of the lowest eukaryotes, G. lamblia. Furthermore, knowledge about the constitutive transcription of a housekeeping gene, such as GDH, would be a starting point for future studies on regulated gene expression, such as during the processes of encystation and antigenic variation in this medically important parasite.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Construction-- All transfection plasmids used were based on the constructs P11 and 3'2 described previously (10, 11). The constructs used in the nuclear run-on experiments are shown in Fig. 1. Plasmid 1 represents the P11 construct, a pGEM-3Zf() vector (Promega) carrying a 2.29-kb genomic DNA fragment from G. lamblia (Portland-1 isolate) that contains the entire coding region of the GDH gene (1350 bp) and its accompanying 5' (820 bp) and 3' (120 bp) flanking sequences. Plasmids 2-7 contain subcloned DNA fragments generated from P11, by either restriction enzyme digests or PCR amplification. Plasmids CWP-1 and -2 are pCR II-based vectors (Invitrogen) that contain the first 400 bp of coding region subcloned from constructs containing the full-length cyst wall protein genes (12, 13). Plasmid vectors, pGEM-3zf() and pCR II, which contain no inserts, were used as negative controls. The rRNA construct, carrying a 5.6-kb fragment of the Giardia ribosomal RNA operon, was a gift from T. Edlind (14).

In the deletion analysis, the plasmids used are shown in Fig. 2. To make 3'2N, the MscI/PvuII cloning junction between the GDH and the luciferase genes within the 3'2 construct was replaced by a NcoI site. Plasmid 3'2N/H3 resulted from the removal of the 600-bp HindIII DNA fragment containing a portion of GDH 5' noncoding region from 3'2N. Further deletions of the GDH 5' noncoding region from 3'2N/H3 are contained within plasmids 5'2, 3, 4, 5, 6, and 7. In N2N, the DNA fragment containing the remaining 18 codons of the GDH gene in 3'2N/H3 was removed. Deletion of the 5' noncoding region in N2N to 44 bp resulted in the plasmid 5'5N.

In Fig. 3, the unaltered sequence of the GDH 5' noncoding region remaining in 5'5N is shown on the top line. Constructs 2 to 10 were made by replacing this region (a HindIII/NcoI fragment) in 5'5N with duplex oligonucleotides containing nucleotide substitutions within this sequence. These duplexes were generated by annealing different sets of complementary oligonucleotides that left overhanging HindIII and NcoI sites at the 5' and 3' ends, respectively. All constructs were checked by DNA sequencing.

Nuclear Run-on Analysis of the GDH Gene-- The following procedure was adapted from one used for Trypanosoma (15). Giardia cultures were washed and concentrated to 3 × 10⁸ cells/ml in buffer A (150 mM sucrose, 20 mM KCl, 20 mM HEPES pH 7.2, 3 mM MgCl₂, 1 mM DTT, and 1 µg/ml leupeptin). Each 400-µl sample was incubated for 5 min on ice followed by permeabilization with lysolecithicin (500 µg/ml, Sigma) for 1 min. The cells were washed twice and resuspended into 100 µl of buffer A at room temperature. An equal volume of transcription buffer (2× stock: 20 mM HEPES pH 7.2, 180 mM KCl, 7 mM MgCl₂, 50 mM phosphocreatine, 1.2 mg/ml creatine kinase, 4 mM ATP, 2 mM GTP, 2 mM CTP, 1 mCi/ml [-³²P]UTP, 1 mM DTT, 10 µg/ml leupeptin) was added and transcription was allowed to proceed for 10 min at room temperature. After transcription, 600 µl of RNA STAT-60 (Tel-Test "B," Inc.) was introduced to each sample, and RNA was extracted according to manufacturer's directions. The recovered RNA was used immediately to probe slot blots containing fragments of the GDH gene and its 5'- and 3'-flanking sequences.

Giardia Transfections and Luciferase Assays-- Giardia (P-1 isolate) were grown and prepared for electroporation as described previously (11). Plasmid DNA (50 µg) was added and the cells were immediately electroporated at 350 V, 1000 microfarads, and 720  (low voltage setting on a BTX Electro Cell Manipulator 600). The cuvette was placed back on ice for 15 min before the cells were added to 15 ml of culture medium in a capped glass culture tube.

Six hours after electroporation, lysates of the transfected cells were prepared as described previously (11). For each assay recorded (Turner TD-20e luminometer), 20 µl of the cell lysate supernatant was used along with 100 µl of luciferase assay reagent (Promega). The relative luciferase activities listed under "Results" (as percentages of the values obtained in control transfections) are averages of at least six independent experiments and have standard deviations within 10% of the presented value. In every experiment, electroporation of each sample was performed in duplicate and each duplicate was assayed twice for luciferase activities.

Primer Extension Analysis-- RNA was extracted from Giardia cultures 6 h after transfection with plasmid 3'2. Electroporations and RNA purification were performed as described above. A sample of the RNA (10 µg) was treated with 5 units of DNase I (amplification grade, Life Technologies, Inc.) for 30 min at room temperature. The digestion was stopped by the addition of EDTA (2 mM) followed by incubation at 65 °C for 10 min. A 5' end-labeled oligonucleotide, oJY-Luc/PE (sequence complementary to codons 6-13 of the luciferase gene), was hybridized to the DNase I-treated RNA and extended with avian myeloblastosis virus reverse transcriptase and dNTPs (16). The resulting extension products were electrophoresed on a denaturing 8% polyacrylamide gel alongside a DNA sequence ladder generated from the plasmid 3'2 using the same oligonucleotide as a primer.

Preparation of Nuclear Extracts-- Nuclear extracts were prepared from freshly harvested Giardia cultures using a modification of the procedure of Andrews and Faller (17). Giardia was lysed in hypotonic buffer (10 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride, 0.75 µg/ml leupeptin) and by the addition of 0.2% Nonidet P-40. After incubation on ice for 10 min, a nuclear pellet was recovered by centrifugation of the cell lysate at 4 °C (12,000 rpm for 15 s in an Eppendorf centrifuge). The nuclear pellet was resuspended in extraction buffer (20 mM HEPES-KOH, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride, 0.75 µg/ml leupeptin), and incubated on ice for 20 min with occasional inversion of the sample tube. The extraction mixture was centrifuged for 4 min at 4 °C (12,000 rpm in an Eppendorf centrifuge) and the extracted protein was recovered in the supernatant. The protein concentration of the extract was determined by the Bradford assay using bovine serum albumin as the standard. The extract was divided into small aliquots, frozen in a dry ice/ethanol bath, and stored at 70 °C. The nuclear protein preparation procedure was monitored by fluorescent dye staining (Hoechst 33258 dye diluted 1:1000 from a 1 mg/ml stock solution) of the Giardia nuclei at various stages during its isolation and extraction.

Preparation of Probes and Competitors for Band Shift Analysis-- The oligonucleotides used are shown in Fig. 4. Complementary oligonucleotides are labeled "a" for the plus strand and "b" for the minus strand. For duplex oligonucleotides, the two complementary oligonucleotides were added in equimolar amounts to a 200-µl volume of an annealing buffer (50 mM Tris-HCl, pH 8.0, 10 mM MgCl₂, 50 mM NaCl), heated in a 90 °C water bath for 5 min, and cooled slowly to room temperature over 1 h. Both single-stranded and double-stranded oligonucleotides were radiolabeled with T4 polynucleotide kinase and [-³²P]ATP. The oligonucleotide duplex containing the yeast GAL4-binding site has the sequence: plus strand, TCGACGGAGTACTGTCCTCCG; minus strand, TCGACGGAGGACAGTACTCCG. The oligonucleotide duplex containing the adenovirus E1B TATA box has the sequence: plus strand, CTAGAGGGTATATAATGCGCCAGCT; minus strand, GGCGCATTATATACCCT. Both these oligonucleotide duplexes were gifts from I. Sadowski.

A 245-bp HindIII-SacI fragment containing the GDH sequence from 223 to +21 was cloned into a pGEM3Zf() vector (Promega). A 51-bp fragment containing the minimal promoter region of the GDH gene was prepared from the pGEM clone by using oligonucleotides 1b and 2a (see Fig. 4) as primers in a PCR amplification reaction. The primers were annealed at 50 °C and extended at 72 °C for 30 cycles in a 100-µl reaction. The 51-bp PCR product was gel purified and its sequence was checked by direct sequencing using the CircumVent thermal cycle dideoxy DNA sequencing kit (New England Biolabs). The fragment was radiolabeled by the substitution of half the total concentration of either dATP or dCTP with [-³²P]dATP or [-³²P]dCTP, respectively, in the PCR amplification. For radiolabeling purposes, amplification was performed for 15 cycles in a 25-µl reaction. The radiolabeled GDH fragment was gel purified prior to use.

Transcription of RNA in vitro was from the 245-bp HindIII/SacI fragment cloned into the pGEM vector described above. The pGEM clone was linearized at the EcoRI site in the polylinker and transcription was initiated at the SP6 promoter by the addition of nucleotide triphosphates and SP6 RNA polymerase. The RNA was radiolabeled by the incorporation of [-³²P]CTP. The DNA template was removed after transcription by treatment with RQ1-RNase-free DNase (Promega) and the RNA was gel purified prior to use. Poly(dI-dC), poly(dA), and poly(dT) were obtained from Amersham Pharmacia Biotech.

Band Shift Assays-- Binding reactions contained 2.5 µg of poly(dI-dC), 3-5 µg of nuclear extract protein, and 0.02-0.05 pmol of radiolabeled probe in a 20-µl volume of binding buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10% glycerol, 1 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride, 0.75 µg/ml leupeptin, 0.05% Nonidet P-40). In competition experiments, all components, including probe and competitor DNA, were pre-mixed before the final addition of the protein extract. In the studies on single-stranded DNA binding, the poly(dI-dC) was first heated at 100 °C and cooled quickly on ice prior to its addition to the binding reaction. The reactions were incubated on ice for 20 min. A 5% polyacrylamide gel was pre-electrophoresed for 1-2 h at 200 V in 0.5 × TBE, 1% glycerol with a replacement of fresh running buffer prior to sample loading. Samples were loaded onto the gel and subjected to electrophoresis at 150 V for 1.5-2 h at room temperature. After electrophoresis, the gel was fixed (10% acetic acid, 10% methanol for 30 min), dried onto filter paper, and subjected to autoradiography.

UV Cross-linking-- The binding reactions using oligonucleotides 1a/b and 2a/b as probes were scaled up 5-fold and performed in duplicates. After incubation on ice, samples were subjected to UV irradiation by a pulsed Nd:YAG laser (5-ns pulse at 266 nm, 60 mJ/pulse; Quanta-Ray, GCR14). One set of each duplicate reaction was subjected to electrophoresis through a standard band shift gel. After electrophoresis, the gel was subjected to autoradiography and gel slices corresponding to the shifted bands on the autoradiograph were removed. The gel slices were placed in the sample wells of a discontinuous 10% SDS-polyacrylamide gel along with the remaining set of UV-treated binding reactions and a molecular weight marker. After electrophoresis, the gel was fixed, dried, and subjected to autoradiography as described above.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To examine transcription of the Giardia GDH gene contained within a 2.2-kb genomic DNA fragment, we developed a nuclear run-on assay in this organism by adapting a procedure used for Trypanosomes (15). In this procedure, Giardia cells were permeabilized and newly synthesized RNA was pulsed-labeled by the addition of [-³²P]UTP. The RNA was extracted and used to probe slot blots containing different segments of the 2.2-kb genomic fragment (Fig. 1), all of which detect only the GDH locus in Southern blot hybridizations of Giardia genomic DNA. Strong hybridizations were obtained to DNA fragments containing the GDH coding region (fragments 1, 2, and 6) and a portion of its associated 3'-noncoding sequence (fragment 7). In contrast, no hybridizations were observed for noncoding sequences beyond 6 bp upstream of the GDH translation start codon (fragments 4 and 5) except for weak hybridization to a DNA fragment located 200 bp further upstream (fragment 3) which corresponds to a portion of an open reading frame encoding an uncharacterized gene. The negative controls consisting of genes for two Giardia cyst wall proteins, CWP1 and CWP2, and two empty cloning vectors, pGEM and pCRII, showed no hybridizations while the positive control rRNA showed the strongest hybridization.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Localization of transcription initiation in the GDH gene by nuclear run-on analysis. The 2.2-kb genomic DNA fragment containing the GDH gene is represented by: open box, gene coding region; hatched bar, 5'-flanking region; open bar, 3'-flanking region; arrow (+1), putative transcription initiation site. Restriction enzyme sites indicated are EcoRI (RI) and HindIII (H3). Portions of the GDH gene contained within each sample slot are depicted by heavy lines. Results from the hybridization of these samples to radiolabeled RNA extracted from cells after pulse labeling with [32P]UTP are shown on the right.

In order to use Giardia transfections to identify the GDH promoter, we needed to ensure that transcription initiation from plasmids transfected into Giardia is the same as that from its endogenous genes. We examined the 5' ends of RNA transcribed from the transfected plasmid 3'2 (11) which uses the GDH 5'-flanking sequence to drive the expression of a luciferase reporter. RNA was extracted from the cells after transfection and an oligonucleotide specific to the luciferase portion of the transcript was annealed and extended using reverse transcriptase (data not shown). The sizes of extension products obtained indicate that transcription from the transfected construct initiates at the same major site as that determined previously for the endogenous GDH gene (10).

To map promoter activity within the GDH 5'-flanking sequence, effects on luciferase activity from deletions of this region contained on expression plasmids were measured in transfected cells (Fig. 2). Only minor effects on luciferase activity were observed when the 5'-flanking sequence was deleted to within 44 bp of the translation start codon (5'5). However, luciferase activity decreased dramatically if deletions extended beyond this point (5'6 and 5'7), and dropped further to background levels if the 5'-flanking sequence was completely eliminated (5'2). Removal of the small amount of GDH coding region remaining (18 codons) in the transfection constructs (N2N and 5'5N) did not affect luciferase activity significantly.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Identification of a minimal promoter region within the GDH gene. Giardia cultures were transfected with luciferase reporter constructs containing incremental deletions of the GDH gene and assayed for luciferase activity. The composition of these constructs are represented by: filled box, GDH coding region; stippled bar, 5' noncoding region of the GDH gene; open box, luciferase coding region; H3, HindIII restriction enzyme site. The numbers proximal to the stippled bars indicate the amount (bp) of 5' noncoding region remaining within each plasmid. Luciferase activity obtained from transfected Giardia are shown as relative percentages of the activity obtained upon transfection of the control plasmid (3'2). These percentages are the averages of at least six independent transfection experiments and have standard deviations within 10% of each value.

The upstream sequence of the GDH gene, containing the 44-bp region identified in the deletion analysis, was aligned against similar regions from other Giardia genes reported in the GenBank data base (data not shown). An A/T-rich region, varying from 10 to 16 nucleotides in length, was found in all sequences in the alignment with the exception of the variant specific antigen genes. The A/T-rich element contains the transcription initiation site in genes which have been transcript mapped and overlaps frequently with the putative translation start codon in all genes examined. A second motif, present in single or multiple copies on either DNA strand, is located upstream of, and occasionally overlaps with, the AT-rich element. We referred to this motif as the g-CAB element, for the Giardia CAT box, and its consensus sequence is CA_nT, where n  3.

Nucleotide substitutions were made on the 5'5 plasmid (Fig. 3, line 1) to determine whether the g-CAB and the AT-rich elements have promoter function in vivo and to identify other promoter elements within the minimal 44-bp sequence. Results obtained from Giardia transfected with these mutant constructs showed that substitutions within a TATA box-like sequence, located between the AT-rich and the g-CAB element (line 3) have no significant effect on luciferase activity; while alterations to the g-CAB element (line 2) have a slight effect on luciferase activity. In contrast, alterations within the AT-rich element (line 9) resulted in a marked reduction of luciferase activity. This reduction was observed even when alterations to this element were limited to nucleotides that are not sites for transcription initiation (line 10). Surprisingly, luciferase activity also decreased when the order of nucleotides within this element was changed (line 8) although the A/T content remained the same as the wild type sequence. Mutations to an AG-dinucleotide, located between the g-CAB and the AT-rich element, also caused a decrease in luciferase activity (line 4), but mutations to another AG-dinucleotide further downstream (line 5) did not. When alterations to the g-CAB element and the upstream AG-dinucleotide were combined in the same mutant construct (line 7), the decrease in luciferase activity was greater than that obtained with alterations to either motifs alone (line 2 and 4).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Characterization of sequence elements required for the function of the GDH promoter. Giardia cultures were transfected with plasmid constructs containing mutations within the minimal 5' noncoding region (44 bp) of the GDH gene requisite for gene expression. In the wild-type DNA sequence presented on the top line, the CAAAT and the AT-rich element are boxed; the AG dinucleotide is overlined; inverted DNA repeats are denoted by open arrows; and the major transcription initiation site is indicated by the (+1) arrow. Shown below are the specific nucleotide substitutions within each mutant construct. Luciferase activity obtained from Giardia transfected with each plasmid are calculated as relative percentages of that obtained upon transfection of the wild-type plasmid. The percentages shown are the averages of at least eight independent transfection experiments and have standard deviations within 10% of each value.

To investigate whether the minimal promoter region contains binding sites for transcription factors, a 51-bp PCR fragment encompassing this region was amplified from a larger cloned fragment of the GDH upstream sequence to use as a probe in band shift assays (Fig. 4). When the radiolabeled PCR fragment was incubated with proteins from a Giardia nuclear extract, two shifted complexes, called C1 and C2, were observed (Fig. 5A). The C1 complex is initially observed at low protein concentrations and C2 appears as the protein concentration increases.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Alignment of the probes used in the band shift assays to the upstream region of the GDH gene. The top line represents the upstream region from the GDH gene with the AT-rich and the CAAAT boxed. The AG dinucleotide in the middle of the intervening sequence between the AT-rich and the CAAAT motifs is overlined. The +1 position in the nucleotide sequence is assigned to start of transcription initiation as indicated by the arrow. The duplex probes, represented by the solid lines, are made from annealing complementary oligonucleotides: labeled "a" for the plus strand and "b" for the minus strand. The DNA sequence of each plus strand corresponds to the upstream sequence of the GDH gene indicated and the minus strand corresponds to the complementary sequence. Specific substitutions within each mutant probe are indicated on the plus strand.

View larger version (59K): [in this window] [in a new window]   Fig. 5.   Binding assays with double-stranded DNA probes representing the GDH promoter. A, increasing amounts of a nuclear protein extract was added to binding reactions with the radiolabeled 51-bp DNA fragment containing the GDH promoter region. The two shifted complexes, C1 and C2, are indicated by the arrows. B, competitors used against protein binding to the 51 bp probe are indicated at the top. Competition was at 20-, 40-, and 80-fold molar excess of the unlabeled DNA relative to the probe. C, band shift assays with either the duplex oligonucleotide 1a/b containing the AT-rich element or with the duplex oligonucleotide 2a/b containing the g-CAB element as probes are indicated by the brackets at the bottom. Increasing amounts of the unlabeled duplex oligonucleotides (20-, 40-, and 80-fold molar excess relative to the probe) used as competitors are indicated at the top. D, UV cross-linking of proteins bound to duplex oligonucleotides 1a/b and 2a/b was done as described under "Experimental Procedures." The autoradiograph from the SDS-polyacrylamide gel is presented. The two outer lanes are the total UV cross-linked binding reactions and the middle lanes are the shifted bands in each complex isolated after UV cross-linking and electrophoresis through a standard band shift gel. The sizes from a molecular weight standard are indicated on the left. The arrow indicates the position of the 68-kDa protein.

The protein-binding sites on the 51-bp fragment were localized by using short oligonucleotide duplexes that span this region in band shift assays (Fig. 4). The oligonucleotide duplex 1a/b represents the region of the GDH sequence between 7 and +10, and contains the AT-rich element. The oligonucleotide duplex 2a/b represents the region of the GDH sequence between 25 and 41, and contains the g-CAB element. Lastly, the oligonucleotide duplex 3a/b represents the region (8 to 24) between the two motifs and contains the first AG-dinucleotide. The oligonucleotide duplexes 1a/b and 2a/b, containing the AT-rich and the g-CAB elements, respectively, were able to compete for protein binding to the 51-bp fragment while the oligonucleotide duplex containing the region between the two motifs did not compete (Fig. 5B). Furthermore, addition of duplex oligonucleotides containing either the binding site for the yeast GAL4-binding site (G4-a/b) or the adenovirus EIB TATA-box (E1B-a/b) did not compete for protein binding to the 51-bp fragment (data not shown).

To investigate further the sequence specificity of protein binding to the 51-bp fragment, oligonucleotides were constructed to contain mutations to the g-CAB and the AT-rich element which were found to cause a reduction in luciferase activity in Giardia transfections (Fig. 3). When an oligonucleotide duplex was made to contain alterations to the AT-rich element (as in line 8 of Fig. 3), this altered duplex (1*a/b) could not compete for protein binding (Fig. 5B). Similarly, when the g-CAB element was changed (as in line 2 of Fig. 3), the altered oligonucleotide duplex (2*a/b) could not compete for protein binding (Fig. 5B).

To examine directly the protein binding ability of the AT-rich and the g-CAB elements, the oligonucleotide duplexes 1a/b and 2a/b containing each of these motifs were radiolabeled and incubated with nuclear protein extracts (Fig. 5C). On a nondenaturing polyacrylamide gel, both probes in the presence of nuclear extract exhibited a complex with retarded mobilities relative to the mobility of the free probes. Surprisingly, both oligonucleotide duplexes can also cross-compete with each other with approximately equal affinities while oligonucleotide duplex 3a/b which contains the region between the two motifs could not compete for protein binding. This observation suggested that the two duplex probes, containing the AT-rich and the CAAAT element, respectively, bind to the same protein or proteins in the nuclear extract.

The size of the protein bound to each probe was analyzed by UV cross-linking (Fig. 5D). Duplicates of the binding reactions with probes 1a/b and 2a/b were subjected to UV irradiation and one set of each reaction was subjected to electrophoresis through a standard band shift gel. The major shifted band that was observed with each probe were cut out separately from the gel and loaded onto a SDS-polyacrylamide gel along with the remaining set of UV-treated binding reactions. The autoradiograph of this SDS-polyacrylamide gel showed that the major shifted band from both oligonucleotide probes migrated at the same molecular mass of approximately 68 kDa (Fig. 5D). A band of this molecular weight was also apparent in the unfractionated binding reactions. In a control experiment, gel slices of the complexes were cut from band shift gels of identical binding reactions that were not cross-linked. The DNA recovered from these gel slices was subjected to electrophoresis on a SDS-polyacrylamide gel. Only bands representing the free probes were observed in the autoradiograph of this gel (data not shown).

To determine whether the 68-kDa protein can also bind single-stranded DNA, a series of experiments was performed using single-stranded oligonucleotides as either probes or competitors in band shift assays (Fig. 6). When the single-stranded oligonucleotides were used as probes in band shift assays, both plus and minus strands containing the AT-rich element (1a and 1b) and the minus strand of the g-CAB element (2b) could bind proteins (Fig. 6A).

View larger version (44K): [in this window] [in a new window]   Fig. 6.   Band shift assays with single-stranded oligonucleotides as either probes or competitors. Lanes: () probe alone; (+) probe plus proteins. A, the different single-stranded probes used in the binding assays are listed at the top. B, 2a·/b represents the binding reaction with the duplex probe consisting of radiolabeled oligonucleotide 2a annealed to unlabeled oligonucleotide 2b Similarly, 2a/b· represents the binding reaction with unlabeled 2a annealed to radiolabeled 2b. C, protein binding to the duplex 51-bp probe. The single-stranded oligonucleotides competitors, used at 40-fold molar excess over the probe, are listed at the top.

Although we assumed that the complexes observed previously with the duplex oligonucleotide probes represented binding of the 68-kDa protein to the double-stranded DNA, it is possible that the protein was bound to single-stranded DNA melted from the duplexes. To investigate this possibility, oligonucleotide 2a, which could not bind protein was radiolabeled and annealed to the unlabeled complementary oligonucleotide 2b which could bind protein (Fig. 6A). After annealing, this mixture was subjected to electrophoresis through a 20% nondenaturing polyacrylamide gel beside a sample of radiolabeled 2a. Examination of the autoradiograph obtained from this gel revealed that >90% of the radiolabeled oligonucleotide 2a when mixed with 2b was in the duplex form (data not shown). When this mixture was used in the band shift assay, no radiolabeled complex was observed (Fig. 6B, 2a·/b). In the complimentary experiment with the duplex probe containing unlabeled oligonucleotide 2a and radiolabeled oligonucleotide 2b, a radiolabeled complex was observed (Fig. 6B, 2a/b·). Thus, the 68-kDa protein appears to bind single-stranded oligonucleotide 2b and 2b derived from the 2a/b duplex but not with the 2a/b duplex itself.

When single-stranded oligonucleotides were used as competitors to protein binding to the 51-bp PCR fragment, the oligonucleotides that were shown to bind protein (1a, 1b, and 2b) were also able to compete for binding, whereas oligonucleotide 2a that did not bind protein could not compete for binding (Fig. 6C). Since the sequences of both the AT-rich and the g-CAB elements are composed almost entirely of As and Ts, we tested the effect of using an excess of unlabeled poly(A) and poly(T) DNA homopolymers as competitors in the binding reaction. While poly(T) was an effective competitor and abolished all protein binding to the 51-bp fragment, poly(A) could not compete for binding (Fig. 6C).

To examine whether the 68-kDa protein can bind to RNA, RNA were transcribed in vitro using an SP6 promoter system from a fragment of the GDH upstream sequence which contains the 51-bp region used in the band shift assays. An excess of this RNA could not inhibit protein binding to the 51-bp fragment (Fig. 6C). When the RNA was radiolabeled and incubated with proteins from the nuclear extract, complexes were obtained but these could not be inhibited by an excess of unlabeled oligonucleotides from the GDH promoter region (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To study the initiation of transcription in Giardia more precisely, we modified a nuclear run-on assay (15) to examine newly transcribed RNAs originating from a 2.2-kb genomic DNA fragment in permeabilized Giardia cells. Previous studies suggest that Giardia RNAs have unusually short 5'-UTRs that are only 1-6 nucleotides in length (18). However, these analyses are comprised mainly of primer extension and S1 protection assays performed exclusively on steady-state RNAs. Therefore, it is unclear whether the 5' ends of these transcripts correspond specifically to sites of transcription initiation or to ends resulting from the cleavage of longer 5'-UTRs. In the kinetoplastid protozoa such as trypanosomes, analyses of promoter sequences have been confounded by the phenomenon of transsplicing in which a leader sequence is added to the 5' end of every RNA as it is being transcribed (19). Using nuclear run-on assays on a 2.2-kb genomic DNA fragment containing the Giardia GDH gene, we localized transcription initiation of the GDH gene to a region that coincide with the sites that were determined previously by primer extension and S1 protection analyses. This finding confirms that the GDH transcripts, and probably those from other Giardia genes, are not transspliced and therefore, have the shortest 5'-UTRs among all eukaryotic RNAs.

Although immunoprecipitation experiments with 5'-methyl-cap antibodies have been performed with total RNA from Giardia, the results are unclear as to whether the mRNAs have modified 5' ends (20). Consequently, it is not known how translation initiates in these transcripts since other eukaryotic mRNAs required longer 5'-UTRs to allow "ribosome scanning" and eubacterial transcripts require a Shine-Delgarno sequence to permit ribosome binding before translation can begin. Interestingly, transcripts from Archaebacteria, a group of non-nucleated, unicellular organisms that is believed to be the ancestor of all eukaryotic cells, also have short 5'-UTRs that are uncapped and lack Shine-Delgarno sequences (21-23). Since phylogenetic analyses indicate Giardia as the eukaryote most closely related to archaebacteria (1), it is conceivable that these organisms have similar mechanisms to initiate translation of their respective transcripts.

In our nuclear run-on assays discussed above, we also detected a small amount of RNA from an uncharacterized open reading frame, located less than 200 bp upstream from the major GDH transcription start site (Fig. 1, sample 3). This upstream ORF is probably a part of a separate transcription unit since no transcripts were observed in the region between this ORF and the one encoding GDH (Fig. 1, samples 4 and 5). Although this is the first demonstration that two such closely spaced ORFs are transcribed separately in Giardia, it also suggests that previous reports of multiple open reading frames within similarly sized genomic DNA fragments (24, 26)² may be other examples of closely spaced, independent transcriptional units within the Giardia genome. The possibility that the Giardia genome is densely packed with expressed genes is supported by C₀T analysis which detected an abundance of single copy DNA versus small amounts of repetitive DNA present within this parasite's genome (27); and by results from the Giardia genome sequencing project (26). We speculate that this minimization may also occur within Giardia genes since no introns have been found so far in any of its genes examined. Furthermore, the ribosomal rRNAs of Giardia are the among the shortest observed in both eukaryotes and eubacteria (14, 27).

In a previous study, we developed a method to transfect Giardia with DNA constructs containing luciferase as a marker for gene expression. We found that the detection of luciferase activity in transfected cells requires the inclusion of the GDH upstream sequence in front of the luciferase gene on the plasmid constructs (11) suggesting that regulatory elements are contained in the upstream sequence of the GDH gene. In this study, we determined that the major site of transcription initiation within the GDH upstream region contained on the transfected plasmid (3'2) is the same as that from the endogenous gene (data not shown) which validates the use of Giardia transfections to study gene expression.

Giardia transfected with reporter plasmids containing incremental deletions of the GDH upstream sequence showed that only 44 bp of DNA sequence immediately upstream of the ATG codon for the GDH gene is required to maintain full luciferase activity (Fig. 2). Such short promoters in Giardia would allow minimal "spacer" DNA between genes, and would help in further reducing the size of its genome.

An AT-rich region and a g-CAB motif were observed in an alignment of the upstream sequence of the GDH gene against similar regions from other Giardia genes (data not shown). The sequence conservation and localization of these elements suggest that they may be components of a Giardia gene promoter. Similar elements have been identified recently in the promoter of the Giardia ran gene (28). Surprisingly, these elements are missing from the DNA sequences upstream of the Giardia variant-specific surface antigen genes (VSP 1267 and TSA 417). This absence, together with their highly regulated expression during the process of antigenic variation, suggest that the VSP genes have a promoter that is different from that of other protein-encoding genes. In the parasite Trypanosoma brucei, in which antigenic variation also occurs, the antigen genes are transcribed by a RNA polymerase that has different drug sensitivities and nuclear localization than the one used to transcribe other protein encoding genes (29-32). By analogy, the VSP genes in Giardia may also be transcribed by a different polymerase and therefore, have a promoter that is different than the one for the GDH gene and other protein-encoding genes.

To identify the specific sequences which contribute to promoter activity within the minimal (44 bp) upstream sequence, we examined effects of selected nucleotide substitutions to this region contained on constructs transfected into Giardia (Fig. 3). Three sequence motifs were identified in this analysis. Two of these elements correspond to the g-CAB (lines 2 and 7) and the AT-rich (lines 8-10) motifs that were identified previously by their sequence conservation in the alignment of upstream sequences (data not shown). The third element is defined by an AG-dinucleotide affected in the mutation on line 4 (Fig. 3). This dinucleotide is likely a part of larger element at this location since identical nucleotide substitutions to a second AG-dinucleotide 7 bp downstream (line 5) did not affect luciferase activity. However, the consensus sequence of this third element is not as well conserved as the AT-rich and g-CAB elements in the upstream sequences of other Giardia genes (data not shown).

Among the three elements identified in the minimal GDH promoter, the AT-rich sequence has the greatest effect on gene expression. This observation along with its location suggest that it is a Giardia equivalent of the C/T-rich initiator element found in higher eukaryotes (33). Since most Giardia genes are GC-rich (up to 85% G + C in the third codon position (18)) one advantage of an A/T-rich initiator is that it may facilitate the melting of the promoter region prior to transcription initiation. However, this initiator element may also contain the binding site for a protein involved in directing the formation of the transcription preinitiation complex. This proposal is supported by the demonstration that the specific order of A and T nucleotides within this element is important for promoter activity in transient transfection assays (Fig. 3, line 8) which suggest a binding site for a protein is contained on this element. In the transcriptional analyses of another ancient eukaryote, Trichomonas vaginalis, the gene promoters examined contain initiator-like elements but not TATA boxes (34). The absence of TATA boxes in the promoters of both Giardia and Trichomonas suggest that the appearance of this motif is preceded by the initiator motif during the evolution of the eukaryotic promoter.

Band shift assays were used to identify possible sites for protein binding on the Giardia GDH promoter. When a 51-bp fragment containing the GDH promoter region was radiolabeled and incubated with proteins from a Giardia nuclear extract, two shifted complexes, C1 and C2, were formed (Fig. 5A). Since the C1 complex forms at low protein concentrations and the C2 complex forms at higher concentration, our interpretation is that there are two levels of protein binding to the 51-bp fragment. Using duplex oligonucleotides representing different regions of the promoter as competitors in band shift assays showed that one of the protein-binding sites is contained within the AT-rich motif and a second binding site is located on the g-CAB motif within the 51-bp probe (Fig. 5B). Therefore, the C1 complex observed with the 51-bp fragment probe in the band shift assays may represent protein binding at one of these motifs, and the C2 complex represents protein binding at both motifs.

When the AT-rich and the g-CAB oligonucleotide duplexes were used as probes in band shift assays, we found that they can cross-compete with each other for protein binding (Fig. 5C). The possibility that the duplexes contain binding sites for the same protein is supported by the UV cross-linking experiment which showed that both the AT-rich and the g-CAB oligonucleotide duplexes bind a 68-kDa protein from the nuclear extract (Fig. 5D). Since this weight is larger than those of the TBPs identified in yeast, Drosophila, and human (35-37), the relation of this binding protein from Giardia to known transcription factors in other eukaryotes cannot be clarified until it is purified and further characterized.

Additional band shift experiments showed that the 68-kDa protein can bind single-stranded DNA from the GDH promoter in a sequence-specific manner (Fig. 6). Although the protein appear to be able bind duplex oligonucleotides in previous experiments (Fig. 5), we demonstrated that the protein is binding exclusively single-stranded DNA which has melted within the duplex (Fig. 6B). Similarly, the observed protein binding to the 51-bp fragment could also be to single-stranded regions melted in this small DNA fragment. Specifically, the 68-kDa protein can bind both DNA strands containing the AT-rich element and the minus DNA strand containing the g-CAB element (Fig. 6, A and C). The binding specificity of the 68-kDa protein appears to involve the short poly(T) tracts contained within the AT-rich and the g-CAB elements on these single-stranded DNA as the poly(T) DNA homopolymer is an effective competitor in the binding assay (Fig. 6C). However, the 68-kDa protein cannot bind RNA corresponding to the same region used in the DNA binding assays. Proteins which can bind specific sequences on single-stranded DNA but not on RNA have been characterized and shown to have roles in the regulation of transcription in other eukaryotes (25, 38, 39).

Our results suggest that we have characterized one of the first promoters in a lower-branching eukaryote. Since GDH genes in other eukaryotes, and protein encoding genes in general, are transcribed by RNA pol II, it is reasonable to presume that we have identified a promoter in Giardia that is associated with the transcription by a similar polymerase. In our nuclear run-on assays in this organism, we observed that the transcription of protein-encoding genes, in addition to those encoding rRNA, is insensitive to concentrations of -amanitin as high as 1 mg/ml,³ in sharp contrast to the high drug sensitivity of RNA pol II transcription in higher eukaryotes. However, the number of different RNA polymerases active in Giardia are not known as none have been purified.

Finally, our results suggest a genetic downsizing in Giardia: this organism appears to have a genome filled with closely spaced transcriptional units, a short and simple gene promoter, as well as having transcription initiation occurring close to the start of the protein-coding region. This streamlined organization may provide Giardia with greater efficiency in its replication, growth, and gene regulation. It is tempting to speculate that these characteristics are relics of early evolution that are retained by this organism, but genomic analysis need to be performed in other primitive eukaryotes, such as Microsporidia and Trichomonads, before this hypothesis can be confirmed.

	Acknowledgments

We thank I. Sadowski for advice on the bandshift assays and the gifts of the oligonucleotides containing the yeast GAL4 binding site and the adenovirus E1B TATA-box. We thank T. D. Edlind for the gift of the Giardia rRNA operon contruct. We are grateful to M. Roberge for suggesting the UV cross-linking experiment and help with the use of the laser. We also thank S. Rafferty for the critical reading of this manuscript.

	FOOTNOTES

^* A portion of this work was supported by Grant MT6340 from the Medical Research Council of Canada (to P. P. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Present address: Dept. of Biology, Trent University, 1600 West Bank Dr., Peterborough, Ontario K9J 7B8, Canada. To whom correspondence should be addressed: Tel.: 705-748-1048; Fax: 705-748-1625; E-mail: jyee@trentu.ca.

² P. Upcroft, A. Healey, J. Upcroft, and S. Townson, GenBank^TM accession number L27221.

³ M. R. Mowatt and T. E. Nash, unpublished data.

	ABBREVIATIONS

The abbreviations used are: VSP, variant-specific surface protein(s); GDH, glutamate dehydrogenase; bp, base pair(s); TBP, TATA box-binding protein; kb, kilobase(s); PCR, polymerase chain reaction; DTT, dithiothreitol; UTR, untranslated region; pol, polymerase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES