Characterization of Two Protein Activities That Interact at the Promoter of the Trypanosomatid Spliced Leader RNA*

All trypanosome mRNAs have a spliced leader (SL). The SL RNA gene in Leptomonas seymouri is a member of the small nuclear RNA gene family. However, the SL RNA is required in stoichiometric amounts for trans-splicing during mRNA formation. Expression of the SL RNA gene requires sequence elements at bp −60 to −70 and bp −30 to −40 upstream from the transcription initiation site. Using conventional and affinity chromatography, we have identified and characterized an ∼122-kDa protein, promoter-binding protein (PBP) −1, that binds to double-strand DNA. The PBP-1-binding site is within the bp −60 to −70 element determined by DNase I footprinting. Therefore, PBP-1 is the first characterized double-strand DNA binding activity that interacts with a trypanosome gene promoter. A second protein, PBP-2, interacts with the PBP-1:DNA complex and its DNase I footprint extends to include the second promoter element (bp −30 to −40). An alteration of the spacing between the two promoter elements or mutation of the second element decreases PBP-2/PBP-1:DNA stability. Taken together, these data suggest that PBP-1 and PBP-2 are components of a transcription initiation complex that assembles within the SL RNA gene promoter.

The SL RNA genes are members of the small nuclear (sn) RNA class of eukaryotic genes (3)(4)(5). These genes encode short, nonpolyadenylated RNAs that participate in mRNA-processing reactions. In higher eukaryotes, in which snRNA genes have been extensively studied, snRNAs are synthesized from independent transcription units that contain promoter elements specific to this gene class (6,. The presence of common promoter elements was unexpected since some snRNA genes are transcribed by RNA polymerase II and others by RNA polymerase III. In trypanosomatids, the SL RNA genes appear to be transcribed by RNA polymerase II, whereas other snRNA genes, including the U2 snRNA, U6 snRNA, and U3 snRNA homolog (U-snRNA B) genes, are transcribed by RNA polymerase III (5, 10 -12).
To characterize transcription of the SL RNA, we began by assessing the effects of mutations on a marked copy of an SL RNA gene. The marked gene was reintroduced into Leptomonas seymouri, an easily manipulated trypanosomatid, using a stably maintained extrachromosomal vector (13,14). Three upstream cis-acting elements, positioned between bp Ϫ1 to Ϫ10, Ϫ30 to Ϫ40, and Ϫ60 to Ϫ70 upstream from the transcription initiation site (ϩ1) were identified (4) (see Fig. 1). Analysis of the SL RNA gene promoter from two other trypanosomatids, Leishmania tarentolae and L. amazonensis, revealed a similar SL RNA gene promoter structure (3,5). It has been suggested that the bp Ϫ60 to Ϫ70 region of the SL RNA gene promoter in trypanosomes is analogous to the proximal sequence element (PSE) found within the snRNA genes of other eukaryotes (3)(4)(5). PSEs, identified in vertebrates and sea urchins, are located ϳ55 bp upstream of the snRNA gene transcription initiation site in these eukaryotes (8,9,15). The trypanosomatid SL RNA gene PSE sequence contains a trypanosome-specific 5Ј-GAC-3Ј core region (5) (and see below) but is divergent from higher eukaryotic PSE sequences, as expected.
In preliminary biochemical experiments, we detected two specific protein-DNA complexes that required wild type (WT) DNA sequences in two promoter elements defined in vivo (4). In this study, we have characterized these complexes (now called I and II) and identified two proteins that assemble at the SL RNA gene promoter.
EXPERIMENTAL PROCEDURES Parasites L. seymouri (ATCC 30220) was maintained in logarithmic phase at room temperature with gentle stirring in Crithidia medium described previously (16) and supplemented with 13 mM sodium phosphate adjusted to pH 7.4.
Co-purification of Complex I and II Forming Activities-Protein from 1.2 ϫ 10 12 cells, extracted as described above, was applied to a 70-ml Trisacryl-M DEAE column which was developed with a discontinuous gradient of increasing KCl concentrations (100 mM, 200 mM, 300 mM) in Buffer A. The peak activity fractions were pooled and diluted 2-fold with Buffer A to a final concentration of 100 mM KCl and applied to a 30-ml heparin-Sepharose CL-6B column (Pharmacia Biotech Inc.) that was developed with a linear 100 -500 mM KCl gradient in Buffer A. Peak activity fractions were pooled, dialyzed against Buffer A containing 100 mM KCl, and applied to a 1.5-ml double strand DNA-cellulose column (Sigma). Protein was eluted using a linear 0.1-1 M KCl gradient in Buffer A lacking Nonidet P-40. Active fractions were combined and dialyzed against affinity buffer (20 mM Hepes-KOH, pH 7.6, 10% glyc-erol, 0.05% Nonidet P-40, 1 mM EDTA, 2.5 mM dithiothreitol, and 90 mM KCl).
Specific DNA Affinity Chromatography-The activity peak (0.96 mg of protein) obtained from the nonspecific double-strand DNA-cellulose chromatographic step was applied to a specific DNA affinity column containing the basal SL RNA gene promoter. To obtain only complex I, 25 g/ml poly(dI-dC)⅐poly(dI-dC) was added to the dialysate and the 1.5-ml column was loaded and run at 0.1 ml/min. To obtain complexes I and II, 4 mM MgCl 2 was also added to the dialysate and the column was loaded and run at 0.05 ml/min. In addition, loaded protein was retained on the affinity column for 10 min before the wash step. The column was washed extensively with affinity buffer and developed using a linear 0.09 -0.5 M KCl gradient.
Preparation of a Specific DNA Affinity Column-Plasmid pHL7, which contains two head-to-head copies of the SL RNA gene promoter sequence, from bp Ϫ10 to Ϫ86 (4) (see Fig. 1), 2 cloned into the polylinker region of pBluescript (pBS/SK II; Stratagene Co.), was digested using SacI and BamHI. The BamHI 3Ј ends were selectively biotinylated using dATP, dGTP, and biotinylated dUTP (Boehringer Mannheim) and the Klenow fragment of DNA polymerase I. The promoter-containing DNA fragment was separated from the vector sequences after A-15 chromatography and digestion with HindIII. The 0.19-kilobase HindIII-BamHI fragment that contained the SL promoter DNA was bound to 4 ml of immobilized streptavidin (Boehringer Mannheim) and used to make a 1.5-ml specific DNA affinity column (containing 108 pmol of DNA) (17). 2 H. Luo and V. Bellofatto, unpublished data. Gel Filtration Chromatography-A Superdex 200 HR 10/30 column (S-200, Pharmacia Biotech.) was calibrated at a flow rate of 0.1 ml/min, using affinity buffer containing 0.3 M KCl and the following standards (total protein, 150 g in 0.1 ml): ferritin, 61 Å; rabbit muscle aldolase, 48.1 Å; bovine serum albumin (BSA), 35.5 Å; and carbonic anhydrase, 20.1 Å. Affinity-purified protein, containing either complex I or both complex I and II forming activity, was concentrated in a Centricon TM 10 (Amicon) to 0.1 ml (ϳ35 g) and run under the conditions used for the standards.
Glycerol Gradient Sedimentation-Triplicate 5-ml 15-30% (v/v) linear glycerol gradients in affinity buffer containing 0.3 M KCl were set overnight at 4°C. The gradients were overlaid with 10 -30 g/0.1-ml S-200-purified proteins or 30 g/0.1-ml standards (bovine liver catalase (11.30 S), rabbit muscle aldolase (7.35 S), BSA (4.30 S), hen egg ovalbumin (3.66 S) (Pharmacia Biotech). Protein eluted from the S-200 column was concentrated in a Centricon TM 10 to 0.1 ml and loaded onto the gradient. The gradients were centrifuged at 42,000 rpm in a Beckman SW 55 rotor, for 28 h at 4°C. The gradients were fractionated from top to bottom and 0.125-ml fractions were collected.

EMSA Analysis
Substrates-The 66-bp WT promoter (bp Ϫ17 to Ϫ83 region) was made by annealing two 66-nt complementary oligonucleotides (see Fig.  1). The 95-bp WT substrate (bp Ϫ1 to Ϫ95 region) was made by polymerase chain reaction amplification of a cloned copy of the SL RNA gene and appropriate primers (4).
The "37-bp" and "29-bp" probes were made by digesting the 66-bp WT promoter with Tru 9 I. The 37-bp probe contained the bp Ϫ46 to Ϫ83 region, the 29-bp region contained the bp Ϫ45 to Ϫ17 region.
The "␦-43/46" probe was made by annealing an oligonucleotide that contained the sequence from nt Ϫ83 to Ϫ65 of the coding strand (5Ј-GGCTACTATATATACATAGA-3Ј) to a 64-nt oligonucletide that contained the noncoding strand from nt Ϫ17 to Ϫ88, but without nt Ϫ43 to Ϫ46. The coding strand was extended using the Klenow fragment of DNA polymerase I to produce a double-strand DNA.
The "⌬-31/40" probe was made by annealing two complementary 66 nt oligomers, from nt Ϫ17 to Ϫ83, with a 10-bp mutation of nt Ϫ31 to Ϫ40 (5Ј-(C) 8 CT-3Ј on the coding strand) to 5Ј-GAGGTTAACG-3Ј.  c Specific activity is total activity per mg of protein. d Peak activity fractions were pooled and diluted 2-fold prior to loading onto heparin.
The "␦-60/74" substrate was made by annealing an oligonucleotide (5Ј-CTATATACACGCTGGTACCAGCTGGAGCGGGTGCATTAACTC-CCCCCCCTCATTTCGTCATGGGCA-3Ј) to a primer complementary to nt Ϫ17 to Ϫ36. The Klenow fragment of DNA polymerase I was used to complete the double-strand molecule. The product was cloned into pBS/SK II and a standard polymerase chain reaction, using SK and KS primers (Stratagene), produced a 133-bp fragment that contained the mutated promoter.
Nonspecific substrates were made by amplifying a 70-bp DNA fragment from pBS/SK II in a standard polymerase chain reaction using the SK and KS primers or by annealing two 69-nt oligonucleotides that contained the polylinker region of pBS/SK II. Each substrate DNA was polyacrylamide gel-purified before use.
Assay Conditions-Reaction mixtures (20 l), containing affinity buffer, 0.5 g of poly(dI-dC)⅐poly(dI-dC), 100 mM KCl, 4 mM MgCl 2 , and 20 fmol of the ␥-32 P-labeled substrates were incubated with titrated amounts of protein for 30 min at room temperature (4). The reaction products were applied to a neutral 4% polyacrylamide gel in 0.5 ϫ Tris borate-EDTA buffer and electrophoresed at 13 mA for 2 h at room temperature. Protein-DNA complexes were quantitated by Phosphor-Imager TM analysis of dried gels.

DNase I Protection Assays
A 130-bp HindIII/BamHI fragment from pLL 95-6 (a derivative of pBS/SK II containing the bp Ϫ1 to Ϫ95 SL RNA gene promoter region) was gel purified and uniquely 3Ј-end labeled at the recessed 3Ј HindIII site using the Klenow fragment of DNA polymerase I and [␣-32 P]dATP. 10 ng of probe (10 5 cpm) was incubated with 8 l (ϳ1 g) of affinitypurified protein in a 20-l reaction under standard EMSA conditions. DNase I (25 ng) and CaCl 2 (1 mM) were added, and the reaction was terminated with 5 mM EDTA after 1 min at room temperature (18). DNase I-nicked free and bound substrate was separated by EMSA and visualized by autoradiography. The DNA fragments were excised from the EMSA gel and electrophoresed on a 6% polyacrylamide, 7 M urea gel in 1 ϫ Tris borate-EDTA buffer. The corresponding dideoxy-sequenced SL RNA gene promoter region was run on the same gel.

Photocross-linking
Uniformly-labeled DNA was prepared by annealing an oligonucleotide (nt Ϫ83 to Ϫ63, top strand) to the bottom strand (66 nt oligonucleotide) WT SL promoter template and filling in with dATP, dCTP, 5Ј-bromouracil, and [␣-32 P]dGTP. The product was digested with Tru 9 FIG. 5. Density sedimentation fractionation of the PBP-1 fraction. EMSA analysis of fractions obtained from density centrifugation. A, protein from S-200 chromatography (load) was applied to a 15-30% glycerol gradient. The probe was the WT 95-bp DNA. B, the binding specificity of fractionated PBP-1 was assessed using competitor DNAs in an EMSA. Fractions 20 -23 were pooled, incubated with WT 95-bp probe in the presence of a 50-fold molar excess of either unlabeled WT 95-bp DNA ("S") or a pBS DNA fragment ("N"). Complex I and free DNA are indicated by arrows.
FIG. 6. Complex I and II bind within essential promoter elements. PhosphorImager analyses of a 6% denaturing gels are shown. A, DNase I protection of promoter DNA using affinity-purified PBP-1 (see fraction 12, Fig. 10, A). Lane "I" is DNA extracted from complex I after EMSA. "F" and "F * " are DNA extracted from the unbound substrate; DNA in F * migrated ahead of the majority of free substrate (F) and was the more extensively digested substrate. The vertical bars show the protected DNA region. A, DNase I-hypersensitive site lies near the middle of the binding site, possibly due to localized DNA distortion by PBP-1. Numbers on the right indicate distance (in bp) from the transcription initiation site (see Fig. 1). B, DNase I protection of promoter DNA using affinity-purified complex II forming activity (fraction 22, see I, and the 37-bp DNA fragment (bp Ϫ83 to Ϫ46) was gel purified. Protein (ϳ1.5 g) was incubated with DNA in a standard EMSA prior to irradiation at 2 cm from the UV source (302 nm, relative intensity at 3 in. is 1500 microwatts/cm 2 ) for 40 min on ice. CaCl 2 concentration was adjusted to 2 mM and the DNA was digested for 10 min using DNase I (10 units, Boehringer Mannheim). 5 ϫ loading buffer (0.25 M Tris-HCl, pH 6.8, 15% SDS, 50% glycerol, 0.02% bromphenol blue) was added, heated at 65°C for 3 min, and resolved on a SDS-PAGE gel.

Identification of Nuclear Proteins
Interacting Specifically with the SL Promoter-Nuclear extracts contained proteins that generated two specific protein-promoter complexes. The more slowly migrating complex, complex II, was in greater abundance. Fractionation of extract on phosphocellulose resulted in the loss of complex II forming activity and a substan-tial increase in complex I forming activity. Fig. 2 shows that complex II could be restored by mixing the complex I forming fraction (fraction 42) with a fraction (fraction 64) that alone did not produce a significant gel shift. When a fixed amount of fraction 42 was mixed with increasing amounts of fraction 64, more complex II was produced. As was the case for unfractionated extract, both complexes could be competed only with WT promoter-containing unlabeled DNA (data not shown). These data suggest that complex I and II share a common specific DNA-binding protein. This protein is named promoter binding protein (PBP)-1. The protein that supershifts complex I to complex II is referred to as PBP-2.
Partial Purification of PBP-1-We fractionated nuclear extract through several chromatographic steps and obtained a 320-fold purification ( Table I). The essential features of the fractionation included a high salt extract of protein from a crude preparation of nuclei (19); the use of an ion-exchange chromatography step followed by heparin-Sepharose, a nonspecific double-strand DNA-cellulose column, and a specific DNA affinity column. EMSA of fractions obtained from the final chromatography step showed that the peak of PBP-1 activity eluted at 0.26 M KCl (Fig. 3, panel A). Although the recovery of PBP-1 from this final column was low (10%) this step was key in removing a majority of the protein that co-purified with PBP-1 through the nonspecific DNA column. We found that three polypeptides co-fractionated with the peak of complex I forming activity (Fig. 3, panels A and B). These proteins have apparent molecular masses of 57, 46, and 36 kDa. The affinitypurified PBP-1 was used to analyze the activity of this protein on DNA and for its physical characterization.
Physical Properties of PBP-1-PBP-1 was subjected to two analytical procedures. Gel filtration chromatography on Superdex-200 showed that PBP-1 had a molecular mass of approximately 140 kDa and a Stokes' radius of 41 Ϯ 3.3 Å (Fig. 4, panel  A). After concentration of the PBP-1 peak from gel filtration, PBP-1 was sedimented through a 15-30% glycerol gradient. EMSA analysis of protein fractionated in the glycerol gradient is shown in Fig. 5, panel A. The activity was present in fraction 17-28; the peak of activity was in fraction 17-23. These fractions contained an approximately quantitative recovery of PBP-1 activity. Fig. 5, panel B, shows that the PBP-1 recovered from the sedimentation procedure retained the same DNA binding specificity as did the input material. This analysis of PBP-1 gave an s 20,w of 7.25 Ϯ 0.8 (Fig. 4, panel B).
Since these data were obtained using nondenaturing conditions, a reasonable estimate for both native molecular weight and frictional ratio could be made. The native molecular weight was calculated by incorporating the Stokes' radius and sedimentation coefficient values into the standard hydrodynamics equation of Siegel and Monty (20). A native molecular mass of 122 Ϯ 16 kDa (x Ϯ 1 S.D.) was calculated for PBP-1. Using this value, we calculated a shape factor of f/f o of 1.2, which indicated that PBP-1 was a relatively symmetric molecule. This predicted symmetry was qualitatively consistent with the independent molecular weight calculations made by comparing either the peak of PBP-1 activity that eluted from the gel filtration column or the relative sedimentation of PBP-1 with molecular weight standards. These data indicate that PBP-1 is either a single polypeptide that was proteolytically cleaved into the 57-, 46-, and 36-kDa polypeptides during purification or that PBP-1 is a multisubunit protein.
PBP-1 Binds to an Essential Region of the SL RNA Gene Promoter-DNase I footprint analysis was used to localize the binding site of PBP-1 on the SL promoter. Fig. 6, panel A, shows that the DNase I-resistant area was between nt Ϫ55 to Ϫ74 on the bottom strand of the promoter (the top strand represents the SL RNA coding region that is downstream from the promoter). The breadth of the footprint was 18 nt, indicating that PBP-1 was a large protein, consistent with a molecular mass of 122 kDa.
To assess the sequence specificity of the PBP-1:DNA interaction, a mutated DNA probe (⌬-60/74; Fig. 1) was tested for PBP-1 binding. Fig. 7, panel A, shows that complex I did not assemble on this mutated DNA. In the converse experiment, panel B shows that the 37-bp region (bp Ϫ46 to Ϫ83) was sufficient for PBP-1 binding.
To test if PBP-1 was specific for double-or single-strand DNA we performed the EMSA analysis shown in Fig. 8. These data show that neither strand of the promoter nor two irrelevant DNA strands could compete complex I (Fig. 8, panel A,  lanes 1-8). Complex I was destabilized only when the oligonucleotides that correspond to the SL promoter were annealed prior to being used as competitor (lanes 9 -12). To determine if PBP-1 bound weakly to single-strand promoter DNA, we used radiolabeled single-strand DNA probes in EMSA. Fig. 8, panel  B, lanes 4 and 5, shows that PBP-1 did not bind to these probes. These data demonstrate that PBP-1 is a sequence-specific double-strand DNA-binding protein.
Photocross-linking was used to determine which polypeptide present in the affinity-purified PBP-1 preparation interacted with the SL promoter. Fig. 9, panel A, shows that the 37-bp promoter region (bp Ϫ46 to Ϫ83) cross-linked to the 46-kDa species (see Fig. 3, panel B). The 46-kDa protein binding was specific for the promoter sequence since this interaction could be competed by the WT SL promoter and not by the mutated promoter (⌬-60/74; Fig. 1). Panel B shows the complexes re-  2-7). Lane 1 shows the binding of protein to DNA without any competitors. Complex I, II, and free DNA are indicated by arrows. These results show that detection of the 46-kDa protein by SDS-PAGE was UV-dependent and observed only under conditions that produced complex I. These data show that the 46-kDa protein that co-purifies with complex I activity is likely responsible for the specific PBP-1 interaction with the promoter.
Analysis of PBP-2-Complex II forming activity co-eluted with complex I forming activity through DEAE, heparin-Sepharose, and double-strand DNA cellulose. The specific DNA affinity step was used to co-purify complex II with complex I by altering the loading and running conditions of this column (see "Experimental Procedures"). Fig. 10, panel A, shows the elution of complexes I and II. Complex II activity eluted at a slightly higher salt concentration (0.3 M KCl) than did complex I (0.26 M KCl). Competition experiments (panel B) show that complex II forms specifically on the SL promoter.
PBP-2 was separated from PBP-1 by gel filtration chromatography of the 0.3 M KCl fraction from the affinity column. Fig.  11 shows that PBP-2 eluted from the S-200 column regenerated complex II when mixed with affinity-purified PBP-1. Fig. 4 shows that the elution of PBP-2 in gel filtration analysis indicated a Stokes' radius of 62 Ϯ 1.4 Å. S-200-fractionated PBP-2, sedimented in a glycerol gradient, indicated an s 20, w of 6.2 Ϯ 0.6. The native molecular weight of PBP-2, determined using the Siegel and Monty (20) hydrodynamic equation, was 157 Ϯ 17-kDa (x Ϯ 1 S.D.). This molecular weight and Stokes' radius indicated a shape factor for PBP-2 of 1.6, which suggested an asymmetric molecule. Such asymmetry was consistent with the finding that PBP-2 eluted from gel filtration before PBP-1 but sedimented more slowly than did PBP-1 during density centrifugation.
PBP-2 Interaction with the SL Promoter-The 0.3 M KCl affinity-purified fraction, which contained complex I and II forming activities, was used in DNase I footprint analysis. Fig.  6, panel B, shows the DNase I protection pattern obtained when enzyme-nicked complex II was subject to denaturing gel electrophoresis. The protected region was from bp Ϫ28 to Ϫ82. This region includes both the bp Ϫ60 to Ϫ70 and bp Ϫ30 to Ϫ40 promoter elements (see Fig. 1). These footprint data suggest that both elements are required for complex II assembly.
To test this prediction, we assayed complex II formation on several promoters with mutations. A deletion in the spacing between the bp Ϫ30 to Ϫ40 and bp Ϫ60 to Ϫ70 regions (␦-43/46, Fig. 1) markedly reduced complex II assembly (Fig. 12, panel  A). Sequence substitution of the bp Ϫ30 to Ϫ40 region (Fig. 12,  panel B) showed similar results. These alterations did not disrupt the PBP-1-DNA interactions required to assemble complex I (Fig. 12, panel A, lanes 4 and 8; panel B, lane 2). In addition, PBP-2 did not form a complex with the 29-bp probe (bp Ϫ17 to Ϫ45 Tru 9 I fragment, data not shown). The 37-bp probe (bp Ϫ46 to Ϫ83 Tru 9 I fragment) could only form complex I when incubated with PBP-1 and PBP-2 (Fig. 7, panel B). Mutation of the PBP-1-binding region not only blocked complex I formation but also destroyed complex II assembly (Fig. 12, panel C). Taken together, these data show that PBP-2 supershifts complex I into complex II and this larger complex requires the WT sequence and spatial arrangement of two SL promoter elements defined in vivo.
Characterization of Complex I and II Assembly on the SL Promoter- Fig. 13 shows the effect of varying KCl and MgCl 2 concentrations on complex I and II formation in EMSA. Complex II formation was increased with increasing concentrations of KCl whereas complex I was decreased (panel A). Complex II was markedly affected by MgCl 2 concentrations; optimal binding was observed at 2-4 mM MgCl 2 (panel B). Complex I was insensitive to varying MgCl 2 concentrations. Addition of ATP did not appear to have an effect on formation of either complex (data not shown). Panel C shows that BSA mixed with either PBP-1 (lanes 4 -6) or PBP-2 (lanes 7-9) did not yield complex II. Therefore, PBP-2 cannot be replaced with a nonspecific protein to assemble complex II.
We examined the association and dissociation rates of complexes I and II. PBP-1 was incubated alone or with PBP-2 and DNA under standard EMSA conditions for increasing amounts of time. Complexes were detected by EMSA (Fig. 14). PBP-1 alone bound DNA very rapidly; the rate was so rapid that by "0.1" min significant amounts of complex I were formed. Complex II formed much more slowly; initially rates were low and the amount of complex formed increased during the 32-min assay (panel A).
The dissociation rate of both complexes I and II were compared by mixing either PBP-1 alone or with PBP-2 and DNA for 10 min, to establish binding, and then challenging the complexes with unlabeled 50-fold molar excess WT promoter DNA (Fig. 14, panel B). PBP-1 alone readily dissociated from DNA; this reaction was very rapid and most of complex I was desta-bilized by the first time point (0.1 min). In contrast, complex II was stable and did not significantly dissociate after 32 min.

DISCUSSION
Studies of SL RNA gene expression in three organisms of the Trypanosomatidae family of protozoa have shown that the basal SL RNA gene promoter lies upstream from the transcription initiation site (ϩ1) between bp Ϫ1 and Ϫ100. In L. seymouri, base substitution mutagenesis data indicated that sequences between bp Ϫ1 to Ϫ10, Ϫ20 to Ϫ40, and Ϫ50 to Ϫ70 were essential for gene expression in vivo (summarized in Fig.  1). PBP-1 binds in a sequence specific fashion to double-strand DNA in the bp Ϫ60 to Ϫ70 promoter element. This promoter element contains a 5Ј-GAC-3Ј sequence that is also present in the SL promoters of two Leishmania species (3,5). These data suggest that PBP-1 is a transcription factor within the SL initiation complex. Three polypeptides, 57, 46, and 36 kDa, co-fractionated with PBP-1 binding activity through multiple chromatographic steps and finally through sedimentation density centrifugation and gel filtration chromatography at 0.3 M KCl. The 46-kDa species was photocross-linked to the SL gene promoter. These data suggest that PBP-1 may be a tightly assembled trimeric complex which has an apparent molecular mass of 122 Ϯ 16 kDa. Proof of these contentions awaits in vitro transcription experiments in which the role of PBP-1 can be assessed directly. We have recently developed an L. seymouri in vitro transcription system that accurately initiates SL RNA gene transcription (21). This system will be instrumental in defining PBP-1's role in transcription.
In vivo experiments showed that both the bp Ϫ30 to Ϫ40 and the bp Ϫ60 to Ϫ70 sequences were necessary for SL transcription. Both regions were necessary in their WT configuration for complex II formation. Kinetic experiments showed that complex II had a slower dissociation rate from DNA than did complex I. In addition, complex II elutes from the specific DNA affinity column at a higher salt concentration than did complex FIG. 14. Kinetics of complex I and II association and dissociation. EMSAs are shown. A, association rate of complex I and II. An EMSA was performed under standard conditions. Aliquots were loaded onto the gel after incubation for different amounts of time, from 0.1 to 32 min. B, dissociation rate of complex I and II. An EMSA was performed under standard conditions for 10 min before a 50-fold molar excess of unlabeled WT 66-bp DNA was added as competitor. Aliquots were loaded onto the gel after incubation for different amounts of time, from 0.1 to 32 min. Lane "ϩ" is no competitor; Lane "Ϫ" is competitor added before the 10-min preincubation.
I. Therefore, complex II is a more stable protein-promoter interaction than is complex I. These experiments argue that complex II may facilitate the formation of the initiation complex that assembles at the SL promoter.
The data presented here support two models for complex II formation. Complex II is PBP-2 bound to complex I and to the bp Ϫ30 to Ϫ40 promoter element (model 1). Alternatively, complex II is PBP-2 bound independently or in the presence of a third protein (or a modified PBP-1) to the two promoter elements (model 2). Both models are consistent with the finding that complexes I and II have different KCl and MgCl 2 optima, different dissociation and association rates, and overlapping DNase I footprint patterns. Model 2 is supported by the findings that complex II could not be completely lost during S-200 fractionation and that complex I could not be completely supershifted to complex II with increasing amounts of S-200 purified PBP-2. However, the majority of our findings support model 1. First, PBP-2, isolated from P-11 chromatography, did not bind to DNA in an EMSA unless it was mixed with PBP-1. Second, limiting amounts of PBP-1 incubated with PBP-2 generated increased amounts of complex II (Fig. 2, lanes 2-4). S-200-purified PBP-2 only produced complex II when incubated with affinity-purified PBP-1 and not with an irrelevant protein (Fig. 11, lane 6, and Fig. 13, panel C, lanes [7][8][9]. In addition, PBP-2 is required specifically for the supershift with PBP-1 since it could not be replaced with BSA (Fig. 13, panel C, lanes 4 -6). Third, the difference in migration of complex II as compared with that of complex I in EMSA is consistent with the addition of a protein (PBP-2) added to complex I. Finally, increased stability by the addition of proteins to a promoter: DNA-binding protein complex is reminiscent of transcription factor assembly at other eukaryotic promoters (22).
Furthermore, the findings that support model 2 do not exclude model 1. S-200-fractionated PBP-2 appeared to shift a very small amount of DNA into complex II (see Fig. 13, panel C, lanes 2 and 7-9). This is likely to be due to contamination by small amounts of PBP-1 since PBP-1 and PBP-2 activity peaks were only 1.5 ml apart. Addition of PBP-1 to the S-200 PBP-2 fraction produced significant amounts of complex II, but not at the obvious expense of complex I. Although we would expect that if complex II is a supershifted form of complex I, PBP-2 should titrate all of complex I into complex II. However, this need not occur for several reasons. First, the overall amount of total probe shifted in these experiments was small, and the amount of PBP-2 activity in this final PBP-2 fraction was limiting. Second, the association and dissociation rates of PBP-1 binding to DNA are very rapid whereas the binding of PBP-2 to complex I is relatively slow (see Fig. 14). The result is that complex I cannot be easily captured by PBP-2 to form complex II. Finally, complex II may require an additional protein, or a modified form of PBP-1, that is limiting in the affinity-purified PBP-1 fraction. A detailed analysis of the PBP-2 interaction with PBP-1 awaits the production of purified recombinant proteins.
Studies of snRNA gene transcription in mammalian cells may provide useful paradigms for PBP-1 structure and function in SL RNA gene expression in trypanosomes. In mammalian cells, snRNA genes are expressed from promoters that contain an essential element, the PSE, located ϳ55 bp upstream from the transcriptional start site. The PSE is bound by SNAPc (snRNA activating protein complex), also called PTF (PSE binding transcription factor), which is a stable 4-subunit complex with an apparent molecular mass of 200 kDa (23,24). Our data suggest the Leptomonas PBP-1 protein may be analogous to SNAPc/PTF since it interacts within the PSE of the SL RNA gene promoter.
The only promoter-binding protein in trypanosomes that has been characterized is a 40-kDa single-strand DNA-binding protein from Trypanosoma brucei (25). We have now identified a sequence-specific double-strand DNA-binding protein that interacts at the SL RNA gene promoter. Further analysis of these proteins will identify their structure as well as their specific role in trypanosome gene expression. As ancient eukaryotes, trypanosomes most likely possess primordial components of the eukaryotic transcription machinery. As parasitic organisms, trypanosomes may demonstrate a flexibility in gene expression that is key to their survival. Detailed studies of the protein-DNA complexes that assemble at trypanosome gene promoters will provide insights into molecular aspects of trypanosome biology.