INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Auxins are phytohormones that regulate various aspects of plant growth and development including elongation growth, photo- and gravitropism, apical dominance, lateral root initiation, the differentiation of vascular tissues, embryogenesis, and fruit ripening (Macdonald, 1997). Although the mechanism by which auxins are perceived by the cell is still unclear, several auxin-binding proteins (ABPs) have been identified and are thought to play a role in auxin perception (Jones and Prasard, 1992). Of these, ABP1 has been implicated as an auxin receptor because it binds the most active auxins in vitro (Ray et al., 1977; Lobler and Klambt, 1985). However, studies that showed that ABP1 is localized predominantly to the endoplasmic reticulum and to a much lesser extent to the cell membrane were puzzling since it is expected that ABP1 binds auxins at the cell membrane (Lazarus et al., 1991; Napier, 1997). Two main lines of evidence subsequently established ABP1 as an auxin receptor. First, ABP1 was found to bind auxins at the cell membrane and not the endoplasmic reticulum despite the latter being its predominant location (Barbier-Brygoo et al., 1989, 1991; Tian et al., 1995). Second, both transgenic tobacco plants and maize (Zea mays subsp. mays) cell lines overexpressing Abp1 displayed an increased capacity for auxin-induced cell expansion (Jones et al., 1998). A model was suggested in which ABP1 is secreted to the outer surface of the cell membrane through its association with a membrane-spanning docking protein, possibly a G-protein-coupled receptor (Macdonald, 1997). Auxin binding at the cell membrane would induce a conformational change in ABP1 that activates the auxin signal transduction pathway.

Comparison of the maize (Z. mays subsp. mays) Abp1 genomic and cDNA clones failed to reveal any TATA box motifs in the genomic sequences immediately upstream of the cDNAs considered to be full-length (Lazarus et al., 1991). Initial attempts to determine the transcriptional start site (+1) yielded inconsistent results (Lazarus et al., 1991). However, the Abp1 +1 was mapped to the CC A CT at 320 bp upstream of the start of translation (ATG) by consensus sequence analysis and primer extension (Schwob et al., 1993). Although this +1 is located 45 bp downstream from a consensus TATA motif, the predicted transcript is much longer than the mRNA detected by northern analysis (Inohara et al., 1989) and the longest cDNA sequenced (Hesse et al., 1989). The TATA and CAAT box motifs as well as the +1 were reported to be located within a transposable element (TE), Tourist-Zm11, inserted 299 bp upstream of the ATG (Bureau and Wessler, 1992; Bureau et al., 1996). Tourist-Zm11 was, thus, suggested to contribute the Abp1 core promoter sequences.

Tourist-Zm11 belongs to a novel superfamily of TEs called miniature inverted-repeat TEs (MITEs). MITEs are characterized by their small size, presence of conserved terminal inverted repeats (TIRs), and a target site preference (Bureau and Wessler, 1992; Bureau et al., 1996). MITEs and MITE-like sequences are frequently associated with the non-coding regions of normal (wild-type) plant genes (Bureau and Wessler, 1992, 1994a, 1994b; Bureau et al., 1996; Casacuberta et al., 1998; Charrier et al., 1999; Surzycki and Belknap, 1999) and are also present in non-plant systems including the mosquito (Aedes aegypti; Tu, 1997), human (Smit and Riggs, 1996), and teleost fish (Izsvak et al., 1999). In addition to MITEs, remnants of copia-like retrotransposons have been shown to flank many normal plant genes (White et al., 1994). In contrast, short and long interspersed nuclear elements are the predominant TEs in association with mammalian genes (Eickbush, 1992). Although some TE sequences have been demonstrated to contribute to the control of expression of nearby genes (McDonald, 1993), their role in the evolution of normal gene expression is not well understood.

To determine the impact of Tourist-Zm11, we characterized the Abp1 5'-flanking region in maize and its wild relatives, the teosintes. We show that this region is highly polymorphic due to the insertion of several TEs, and we discuss their significance in the regulation of Abp1 gene expression.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The Abp1 5'-Flanking Region Contains Multiple TE Insertions

Tourist-Zm11 was first identified in the 5'-flanking region of maize Abp1 by sequence similarity searches (Bureau and Wessler, 1992). Database searches using the published maize Abp1 sequence (GenBank accession no. L08425) as a query also revealed that the 5' upstream-most region (870-1240 bp upstream of the ATG) shares similarity with the Ds1 element insertion of the maize Bz1 gene (Schiefelbein et al., 1988; EMBL accession no. X14155) and with a Ds1 insertion in Zea perennis (MacRae and Clegg, 1992; EMBL accession no. X54711). Although the 3' TIR could be recognized (5'-ATCCATCCCTA-3'), the L08425 sequence did not extend far enough upstream to include the 5' TIR (Fig. 1).

View larger version (23K): [in this window] [in a new window]   Figure 1.   Genomic (A), 5'-flanking region (B), and transcript (C) organization of maize (Z. mays subsp. mays) Abp1. Exons and TEs are represented by shaded and empty rectangles, respectively. Arrows represent primers. Primer names are derived from their positions on the genomic sequence (accession no. L08425). The thick black lines represent the probes used in this study and the thin lines starting with arrowheads represent the two in vitro transcripts used in Figure 5. The flags labeled 320 and 110 represent the transcriptional start sites determined by Schwob et al. (1993) and in this report and located 320 and 110 nt upstream of the ATG, respectively.

To search for insertion polymorphism and to further characterize the Abp1 upstream sequences, the Abp1 5'-flanking regions in maize (cv W22) and the teosintes (Z. mays subsp. parviglumis, Z. mays subsp. heuheutenangenesis, Z. mays subsp. mexicana, Zea diploperennis, and Zea luxurians) were amplified using primers ZMAUX563S or ZMAUX574S and ZMAUX1316A shown in Figure 1. Nucleotide sequence (GenBank accession nos. AF292696-AF292709) comparison revealed the presence of several TE insertions. Based on their TE content, the Abp1 5'-flanking sequences can be grouped into four main types as represented in Figure 2 and Table I. The first group (type A) is represented by the only maize sequence amplified (ZmW22) and one sequence amplified from Z. mays subsp. huehuetenangenesis (ZmH3). Both sequences contain only one TE insertion, Tourist-Zm11. The second group (type B) of Abp1 sequences lacks Tourist-Zm11 and instead contains a different insertion located 21 bp downstream of the position where Tourist-Zm11 has inserted in ZmW22 and ZmH3. This new element was called Batuta (after Ibn Batuta, a traveler and author), is flanked with a 5-bp target site duplication (TSD, consensus; 5'-TTCTT-3'), and is 221 to 222 bp in length. Members of this group of sequences include ZmP1, ZmH1, ZmM7, ZL1, ZL5, ZD1, and ZD2. ZL1 and ZL5, and ZD1 and ZD2 are allelic sequences that share 91.8% and 96.7% nucleotide identity, respectively. The third group (type C) of Abp1 sequences contains both Tourist-Zm11 and Batuta. This latter Batuta insertion is also located 21 bp downstream of the Tourist-Zm11 target site, has a 5-bp TSD (5'-TTCTT-3') but is 12 to 13 bp shorter than its counterpart of the second group. Abp1 sequences of the third group include ZmP3, ZmH8, ZmM6, and ZL4. The fourth group (type D) of Abp1 sequences is represented by a sequence obtained from Z. mays subsp. mexicana, ZmM1. This sequence contains three TE insertions; Tourist-Zm11, Batuta, and a third insertion designated Pilgrim. Pilgrim is located 109 bp downstream of Batuta, has an 11-bp TSD (5'-CTGCGTGGTGC-3'), and is 165 bp in length.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Schematic depiction of the four types of the Abp1 5'-flanking sequences. The first, second, third, and fourth types (as described in the text) are represented by A, B, C, and D, respectively. The shaded rectangles represent the 5'-untranslated region (UTR), and the TE insertions are represented by triangles. The transcriptional start site is represented by a flag. All sequences were amplified using primer 574S or 563S (represented by the circle) and 1316A.

The Maize and Teosinte Abp1 Transcripts Are Approximately the Same Size

Northern-blot hybridization was performed using a probe spanning the translated region of exon I and a part of the 5'-UTR (EX1; Fig. 1) and RNA extracted from maize (cv W22 and cv Seneca Horizon) and the teosintes. Figure 3 shows that only one Abp1 mRNA species of about 1 kb in size was observed in all samples tested. This suggested that the Abp1 transcripts of maize and the teosintes may have the same 5'-UTRs regardless of the TE insertions at their 5'-flanking regions. They may alternatively have different 5'-UTRs and the differences in sizes are compensated for by differences at their 3'-UTRs. The maize Abp1 transcript was reported to start at 320 bp upstream of the ATG (Schwob et al., 1993) and to contain an open reading frame of 603 nucleotides (nt; Hesse et al., 1989; Inohara et al., 1989; Tillmann et al., 1989). In addition, our 3' RACE results (data not shown), and primer-directed RNaseH digestion of maize and teosinte Abp1 mRNA (Fig. 4) revealed that the size of the 3'-UTR including the poly(A) tail is approximately 300 nt. Taken together, the predicted transcript size should be approximately 1,225 nt. Although the transcript size reported here (approximately 1.0 kb) is in agreement with (approximately 1.0 kb in Inohara et al., 1989) or larger than (approximately 0.8 kb in Hesse et al., 1989; approximately 850 nt in Tillmann et al., 1989) that previously reported, it does not correspond to the predicted size of approximately 1,225 nt.

View larger version (141K): [in this window] [in a new window]   Figure 3.   Northern-blot analysis of Abp1. Total RNA from Z. mays subsp. mays cv Seneca Horizon (lane 1), Z. mays subsp. mays cv W22 (lane 2), Z. mays subsp. parviglumis (lane 3), Z. mays subsp. mexicana (lane 4), and Z. diploperennis (lane 5) was separated on formaldehyde agarose gels, blotted to nylon membranes, and probed with EX1. RNA size standards (in kb) are given on the left-hand side.

View larger version (113K): [in this window] [in a new window]   Figure 4.   Primer-directed RNaseH digestion of Abp1 mRNA. Total RNA was incubated with RNaseH and ZMAUX3435A (lanes 1-5), ZMAUX3435A only (lane 8), or RNaseH only (lane 9), or was untreated (lanes 6 and 7), followed by northern analysis. The membrane was probed with EX1 and subjected to autoradiography (A), and then probed with 3PE and re-exposed (B). Lane 1, Z. mays subsp. mays cv Seneca Horizon; lanes 2 and 6, Z. mays subsp. parviglumis; lanes 3 and 7, Z. mays subsp. mexicana; lanes 4 and 8, Z. diploperennis; lanes 5 and 9, Z. luxurians. The asterisks indicate the positions of the intact Abp1 transcript (*), the 5' fragment (**), and the 3' fragment (***).

The Abp1 Transcript 5' End Is Conserved in Maize and the Teosintes

To resolve the structure of the 5'-UTR and the discrepancy in mRNA size, we adopted a number of approaches to map the 5' end of Abp1 mRNAs. 5' RACE was performed with maize and teosinte mRNA as templates and using primers that spanned the junction of exons I and II (ZMAUX2993A) and the junction of exons II and III (ZMAUX3435A) for first-strand cDNA synthesis (Fig. 1). The longest 5' RACE products terminated at 110 nt upstream of the ATG. Since Schwob et al. (1993) mapped the maize Abp1 mRNA 5' end to 320 nt upstream of the ATG, we attempted to amplify maize sequences upstream of nt number 110 using reverse transcriptase (RT)-PCR and different combinations of primers ZMAUX920S, ZMAUX949S, and ZMAUX1040S and primers ZMAUX1078A, ZMAUX1242A, ZMAUX1316A, ZMAUX2993A, and ZMAUX3435A (see Fig. 1 and "Materials and Methods"). Despite using several different RTs (AMV, M-MLV, and rTth), no amplification products were obtained (data not shown).

The same primers were used successfully to amplify the corresponding maize genomic sequences. This suggested that the transcript 5' end for maize and probably the teosintes may actually be located further downstream from that previously reported for maize Abp1 (Schwob et al., 1993). RT-PCR using primer ZMAUX1197S, which is located 44 nt upstream of the ATG, and any of primers ZMAUX1316A, ZMAUX2993A, or ZMAUX3435A yielded the expected amplification products (data not shown). In addition, RT-PCR was performed using mRNA purified from seven different tissues and developmental stages (young seedling leaf, 10-week-old leaf, immature tassel, post-pollen tassel, immature ear, mature ear, and fertilized ear). Whereas the expected amplification products were obtained when primers ZMAUX1197S and ZMAUX2993A were used, amplification using primers ZMAUX949S and ZMAUX2993A yielded no products.

To get a precise measure of the size of the Abp1 mRNA in maize and the teosintes, we performed primer-directed RNaseH digestion of total RNA followed by northern-hybridization analysis. Figure 4 shows that the size of the 5' fragment of the mRNA in both maize and the teosintes is approximately 330 nt, which puts the transcript 5' end at approximately 110 nt upstream of the ATG. The sum of the sizes of the 5' (approximately 330 nt) and 3' (approximately 650 nt) fragments, and primer ZMAUX3435A (25 nt) yields an Abp1 mRNA of approximately 1 kb, which is consistent with our (Fig. 3) and previous (Inohara et al., 1989) northern results.

Northern hybridization was performed on maize total RNA using three different probes; EX1, 3' end (3PE, which spans exons III-V), and 5'-non-transcribed region (5NT, which covers sequences upstream of nucleotide number 110 from the ATG) (Fig. 1 and "Materials and Methods"). According to Schwob et al. (1993) all three probes should hybridize to Abp1 mRNA. As seen in Figure 5A, although both the EX1 and 3PE probes gave the expected 1-kb signal, the 5NT probe failed to recognize any Abp1 RNA. The 5NT probe hybridizes specifically, however, to an in vitro transcript covering the ZmW22 sequence but not to that covering full-length cDNA (Figs. 1 and 5B).

View larger version (103K): [in this window] [in a new window]   Figure 5.   Northern-blot analysis of maize total RNA (A), and in vitro-transcribed maize Abp1 sequences (B, also see Fig. 1). A, Ten micrograms of Z. mays subsp. mays total RNA was separated on three replica lanes, blotted onto a nylon membrane, and the membrane was cut into three strips. Each strip was then hybridized with the probe indicated. B, Ten nanograms of in vitro transcripts of ZmW22 (lanes 1) and full-length Abp1 cDNA (lanes 2) was separated and probed with the same probes used in A.

Primer extension was performed to confirm our 5' RACE results. As seen in Figure 6, primer ZMAUX1242A gave rise to two extension products that put the 5' end of Abp1 mRNA at the A and G residues located at positions 110 and 109 upstream of the ATG. Primers ZMAUX1078A and ZMAUX969A, on the other hand, did not yield extension products (data not shown). Primer extension with cv Seneca Horizon RNA also yielded two extension products terminating at the T and C residues located at positions 108 and 107 upstream of the ATG, respectively (data not shown).

View larger version (91K): [in this window] [in a new window]   Figure 6.   Primer extension mapping of the Abp1 transcript 5' end. 32P-Labeled primer ZMAUX1242A was used in extension reactions with 20 µg of total RNA extracted from Z. diploperennis (lane 1), Z. mays subsp. mexicana (lane 2), and Z. mays subsp. mays cv W22 (lane 4). Lane 3 is blank. Extension products were separated on 8% (w/v) polyacrylamide/urea gels alongside with 35S-labeled sequencing reactions primed with ZMAUX1242A. The sequence of the region flanking the +1 is shown and residues 107, 108, 109, and 110 upstream of the ATG are underlined.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

TE Insertions within the Abp1 5'-Flanking Region of Maize and the Teosintes

Sequence comparison between the 5'-flanking region of maize and teosinte Abp1 genes reveals the presence of two novel TE insertions, Batuta and Pilgrim. Computer-assisted database searches (in September 1999) using either the Batuta or Pilgrim sequences as queries did not reveal similarity to known TEs. However, several lines of evidence suggest that both insertions are actually TEs. Insertion polymorphisms were identified for both Batuta and Pilgrim in the different teosintes. Furthermore, Pilgrim contains TIRs with sequence similarity to those of Ac-like TEs (Kunze, 1996) and has an 11-bp TSD. Batuta, on the other hand, appears to be a solo long terminal repeat of a retroelement (LTR). Solo LTRs are presumably generated by homologous recombination of the two LTRs of an ancestral retroelement (Mager and Goodchild, 1989). Like retroelement LTRs (Bingham and Zachar, 1989), Batuta has a 5-bp TSD and contains potential promoter (e.g. multiple TATA-, CAAT-, and GC-like sequences), polyadenylation, and enhancer sequences. The existence of LTR fragments within or near genes has been documented in several eukaryotes (White et al., 1994; Noma et al., 1997; Goodchild and Poulter, 1998; Liao et al., 1998).

Abp1 Transcriptional Initiation Is Conserved among Maize and the Teosintes

The Abp1 core promoter was previously reported to be contributed by sequences present in Tourist-Zm11. This is difficult to reconcile with our results for two reasons. First, some teosinte Abp1 5'-flanking types (Fig. 2, type B) lack Tourist-Zm11. In this case, the core promoter would have to be provided by other sequences. Second, the Abp1 5'-flanking regions of some teosintes contain not only Tourist-Zm11, but also Batuta (Fig. 2, type C) or Batuta and Pilgrim (Fig. 2, type D) further downstream. If the core promoter is provided by Tourist-Zm11, the resulting transcripts will contain unusually long 5'-UTRs.

Although the maize Abp1 +1 was previously mapped to 320 nt upstream of the ATG, we provide strong evidence that the Abp1 +1 for both maize and the teosintes is in fact located 110 nt upstream of the ATG. First, we were unable to amplify, by RT-PCR, sequences upstream of the ATG (the 5'-UTR) using primers positioned more than 110 nt upstream of the ATG. However, amplification products were obtained using primers located within 110 nt upstream of the ATG. Second, we investigated the possibility that Abp1 may be differentially transcribed from two promoters; one that generates the 1-kb transcript that starts 110 nt upstream of the ATG, whereas the other generates a transcript that starts 320 nt upstream of the ATG. Northern-blot analysis using a full-length Abp1 cDNA probe and total RNA extracted from 18 different tissues and developmental stages did not reveal the presence of a transcript originating 320 nt upstream of the ATG (data not shown). Likewise, RT-PCR products from seven of the RNA samples only revealed the transcript predicted to originate 110 nt upstream of the ATG. Third, primer-directed RNaseH digestion of Abp1 mRNA followed by northern hybridization indicated that the 5' end of the transcript is located approximately 110 nt upstream of the ATG. Fourth, using a probe upstream of 110 nt from the ATG (5NT), we were unable to detect any hybridization signal on northern blots. However, a band of the expected size corresponding to a transcript initiating 110 nt upstream of the ATG was detected using the EX1 probe. Fifth, using two methods, 5' RACE and primer extension, to determine the precise position of the Abp1 transcript 5' end in maize and the teosintes, we found that the Abp1 transcript starts at two positions located at 109 and 110 nt upstream of the ATG. When a primer (ZMAUX969A) located 28 nt downstream from what was previously reported to be the Abp1 +1 (i.e. 320 nt upstream of the ATG; Schwob et al., 1993), no extension products were obtained. Our mapping of +1 to 110 bp upstream of the ATG fits well with previous (Inohara et al., 1989) and our determination of transcript size.

We also show that the Abp1 +1 is conserved between maize and the teosintes despite the structural diversity (conferred mostly by different TE insertions) of their 5'-flanking regions. The sequence upstream of the ATG up to the position of the first insertion is highly conserved between maize and the teosintes, including the region immediately upstream of +1 (approximately 170 bp; >95% identity). In addition, although Tourist-Zm11 and Batuta both contain putative TATA box motifs, these motifs are located much further upstream (> 180) than the canonical position of approximately 25 to 40 for a typical TATA box. Therefore, we propose that the sequences immediately upstream of +1 provide the Abp1 core promoter. This proposal has two implications. First, the TE insertions in the Abp1 upstream region do not contribute to the Abp1 core promoter sequence. Second, since the region from +1 up to the first TE insertion lacks any putative TATA box motifs, this suggests that Abp1 transcription may be derived by a TATA-less promoter.

Transcription from TATA-less promoters depends on the presence of an initiator element (Inr). The Inr is a strong core promoter element that is functionally analogous to the TATA box (Smale, 1997). The Inr extends from nucleotide position 6 to +11 with the consensus Py Py A N T/A Py Py (underlined residue = +1). Although the Py residues are not needed in all four positions, they must be present in at least a few positions. The most critical residues are central CA, A at +1, T or A at +3, and Py at 1 (Smale, 1997). As seen in Figure 7, the sequence flanking the maize and teosinte Abp1 +1 (110 nt from the ATG) is GGA CA GTC, which conforms to the Inr consensus. Sequence conservation at this region between maize and the teosintes provides additional support that it functions as an Inr (Fig. 7). Many TATA-less promoters are also characterized by a GC-rich 5'-UTR and 5'-flanking region that contains binding sites for transcription factors such as Sp1 and AP (Dufau et al., 1995; Liu and Hla, 1997; Smale, 1997). The Abp1 5'-UTR and 5'-flanking region (1 to 100) are similarly GC-rich (GC content: 60% and 66%, respectively). In summary, the Abp1 core promoter is conserved between maize and the teosintes, is TATA-less, and is located downstream of the TE insertions.

View larger version (22K): [in this window] [in a new window]   Figure 7.   Alignment of the consensus Inr and Abp1 sequences flanking +1. The critical Inr residues are enclosed in the box and the four Abp1 transcriptional start sites are in bold.

Potential Significance of the Abp1 5'-Flanking TE Insertions in Abp1 Gene Expression

The finding that the Abp1 core promoter sequence is not provided by the TE insertions in the Abp1 5'-flanking regions raises questions about whether they play any role in Abp1 expression. Schwob et al. (1993) investigated the ability of different Abp1 promoter deletions to promote transient chloramphenicol acetyl transferase (CAT) reporter gene expression in maize leaf protoplasts. They describe a region from 684 to 449 bp upstream of the ATG that contains a negative regulatory element. When the sequence from 449 to 256 bp upstream of the ATG was deleted, expression declined almost 4-fold. It is interesting that this region contains the entire Tourist-Zm11 element plus an additional 66 bp. In addition, the sequence from 256 to 17 bp upstream of the ATG (which we suggest to contribute the core promoter) promotes expression levels higher than those obtained from full-length Abp4 and Abp5 promoters (other members of the ABP gene family; Lazarus et al., 1991; Schwob et al., 1993).

In light of our mapping of the Abp1 transcript 5' end and the presence of a TATA-less promoter, we suggest that Tourist-Zm11 may act as a positive regulatory element and that the sequence 684 to 449 bp upstream of the ATG may contain a negative regulatory element. Abp1 minimal expression would be conferred by the core promoter, and different expression levels and possibly inducible, tissue-specific, and/or temporal expression would be the net action of both the positive and negative regulatory elements. Furthermore, eukaryotic promoter database (Cavin Périer et al., 1999) searches using the Abp1 5'-flanking regions as queries revealed that Tourist-Zm11, Batuta, and Pilgrim all contain several putative cis-acting regulatory sequences (data not shown). This opens the possibility that differential expression may be determined by cis-acting regulatory elements contributed by these TEs. In view of the fact that Abp1 is differentially expressed in coleoptiles and female reproductive organs (Hesse et al., 1993; N. Elrouby and T.E. Bureau, unpublished data), the presence of a TATA-less core promoter may further indicate a possible function of the upstream cis-acting regulatory sequences. Many TATA-less promoters have been shown to confer a greater potential for differential expression (e.g. Dufau et al., 1995; Liu and Hla, 1997) than the tightly controlled TATA-containing promoters.

In general, TEs may influence the evolution of wild-type genes and gene expression in a variety of mechanisms (for review, see McDonald, 1995; Britten, 1996, 1997; Kidwell and Lisch, 1997). TE insertions may induce changes in chromatin structure that may, for example, insulate a gene's promoter from enhancer sequences located distal to the site of insertion (Gerasimova et al., 1995). TEs inserted in the 5'-flanking regions of genes may alternatively influence gene expression by altering transcriptional initiation in a temporal or tissue-specific manner. This change may result from the generation of a novel transcript initiated from the TE promoter. For example, a retrotransposon insertion in the upstream region of the maize 19-kD zein genes results in the generation of an additional zein transcript initiated from the TE promoter (Kriz et al., 1987; White et al., 1994). TEs may also influence normal gene expression by contributing positive or negative regulatory sequences (McDonald, 1995). For instance, an enhancer sequence contained within a cryptic retrotransposon located in the 5'-flanking region of the murine slp gene confers androgen-specific regulation (Stavenhagen and Robins, 1988). Likewise, a retroelement insertion in the 5'-flanking region of the human Amy1 gene confers salivary gland-specific expression (Meisler and Ting, 1993). In addition, an endogenous retroviral LTR inserted in intron 19 of the human leptin receptor gene OBR induces steroid-mediated alternative splicing of its mRNA (Kapitonov and Jurka, 1999). Selection on short sequences within TEs associated with the regulatory regions of normal genes may provide a mechanism for the evolution of eukaryotic enhancers (McDonald et al., 1997). Whether TEs, namely Tourist, Batuta, and Pilgrim, are involved in the expression or regulation of the maize and/or teosinte Abp1 genes will require fine-scale promoter characterization. This type of study will ultimately help determine if TEs are for the most part "junk DNA" or, alternatively, are fundamental generators of phenotypic variation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material and Nucleic Acids Extraction

Maize (Zea mays subsp. mays cv W22 and cv Seneca Horizon) germplasms were obtained from Susan Wessler (University of Georgia, Athens) and Stokes Seeds (Ontario, Canada), respectively. The teosinte germplasm was obtained from John Doebley (University of Minnesota, St. Paul). For genomic DNA isolation, seedlings were grown under greenhouse conditions for 3 weeks and genomic DNA was extracted as described (Dellaporta et al., 1985). Total RNA was isolated from 7- to 10-d-etiolated seedlings following a LiCl precipitation method (Ausubel et al., 1996). Poly(A⁺) RNA was purified using the PolyAtract mRNA Isolation System III kit (Promega, Madison, WI).

PCR Amplification, RT-PCR, and 5' RACE

PCR amplifications were performed in 10 mM Tris [tris(hydroxymethyl)aminomethane]-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 0.001% (w/v) gelatin, dNTPs (200 µM each), 35 ng of each primer, 75 ng of genomic DNA, and 1 unit of AmpliTaq Gold DNA polymerase (Perkin Elmer, Foster City, CA). PCR conditions were as described (Bureau and Wessler, 1994b) except that the annealing step was performed for 2 min at 65°C. First-strand cDNA synthesis for RT-PCR and 5' RACE was performed using AMV RT (Pharmacia Biotech, Piscataway, NJ) and 20 to 30 ng of poly(A⁺) RNA. We used the Abp1 genomic sequence (Schwob et al., 1993, accession no. L08425) to design the primers used in this study. The primer name consists of two parts; ZMAUX for Z. mays ABP 1, and a number that corresponds to the position of the first nucleotide of the primer sequence in the L08425 sequence followed by the suffix "S" for the sense sequence or "A" for antisense sequence. The primers used to amplify the maize and teosinte 5'-flanking region were ZMAUX574S (5'-AGTAAGACCGCCCCAACAGG-3') or ZMAUX563S (5'-CATATTTTCTTAGTAAGACCGCCCCAACAGGTTACCC-3') and ZMAUX1316A (5'-GAAGGAGGCAGCGAGGAGGAC-3'). The primers used for RT-PCR and 5' RACE were ZMAUX920S (5'-ACTGGAACCGAACAAGCCC-3'), ZMAUX949S (5'-GCCGGTTGGATTCTT-3'), ZMAUX1040S (5'-CTGGCAAGGCGAGGCTTGTCCTGCGTG-3'), ZMAUX1078A (5'-GCCCAGAGACAAGCCGCACC-3'), ZMAUX1197S (5'-CTCATTCCACTCCGACATTC-3'), ZMAUX1242A (5'-ATTGCCTGCTCCCGACAGCTGC-3'), ZMAUX1316A, ZMAUX2993A (5'-CTCTCACCAATGAGTTATCTCGCAC-3'), ZMAUX3435A (5'-GCCACACTTCCACCTCCTTCATC-3'), ZMAUX3442S (5'-CAATAAGTCCAGGTCAAAGGACGCCAATC-3'), and ZMAUX5490A (5'-GGAAACACTTGTGACCTAGAG-3').

5' RACE was performed as described (Troutt et al., 1992). For this purpose, first-strand cDNA was ligated to an anchor oligonucleotide (5'-TTTAGTGAGGGTTAATAAGCGGCCGCGTCGTGACTGGGAGCGC-3') using T4 RNA ligase (Epicentre Technologies, Madison, WI), and then amplified using a gene-specific primer (ZMAUX3435A or ZMAUX2993A) and an oligonucleotide complementary to the anchor oligonucleotide (5'-GCGGCCGCTTATTAACCCTCACTAAA-3').

Cloning, Sequencing, and Sequence Analysis

All PCR fragments were cloned into pCR2.1 as a part of the Original or Topo TA cloning kit (Invitrogen, Carlsbad, CA). Clones were given names that describe the species or subspecies and a number that corresponds to the clone number (Table I). Sequencing was performed following the dideoxy termination method using the SequiTherm Excel II DNA sequencing kit (Epicentre Technologies). Database searches were performed using the algorithm BLAST (version 2.0; National Center for Biotechnology Information, Bethesda, MD). Sequence analysis was performed using the Gap and PileUp programs as part of the UWGCG program suite (version 9.0, University of Wisconsin Genetics Computer Group, Madison).

Probes

Abp1 exon 1 (EX1) and 3PE cDNAs were isolated by RT-PCR using the primers ZMAUX1197S and ZMAUX2993A, and ZMAUX3442S and ZMAUX5490A, respectively (Fig. 1), and cloned into pCR2.1 as described above. To generate the DNA probes used for northern hybridizations, the EX1 and 3PE cDNAs were released from pCR2.1 by EcoRI digestion. The 5NT probe, which covers genomic sequences from 110 to 676 bp upstream of the ATG (Fig. 1), was released from ZmW22 by EcoRI-SalI digestion. After gel purification, all DNA fragments were radiolabeled by random priming using the Oligolabelling kit (Pharmacia Biotech).

Northern Analysis

Total RNA(10-12 µg) from each species or subspecies was separated on 1.5% (w/v) formaldehyde agarose gels, transferred to nylon membranes (GeneScreen Plus, NEN Life Science Products, Boston), and hybridized to the EX1, 3PE, or 5NT probes in a solution of 0.25 M Na₂HPO4 (pH 7.4), 1 mM EDTA, and 7% (w/v) SDS at 60°C overnight. Membranes were then washed in 0.1× SSC, 1% (w/v) SDS for 30 min at 60°C and subjected to autoradiography. To generate the run-off transcripts used as controls in Figure 5, pCR2.1 clones containing ZmW22 (in vitro transcript no. 1) or a full-length Abp1 cDNA (in vitro transcript no. 2) were linearized with BamHI and used in transcription reactions following the RiboMax RNA production system (Promega).

Primer-Directed RNaseH Digestion of Abp1 mRNA

Site-specific enzymatic cleavage of RNA was as previously described (Donis-Keller, 1979). For this purpose, 12 µg of total RNA from each species or subspecies was incubated with 50 ng of the antisense primer ZMAUX3435A at 90°C for 5 min, chilled on ice for 5 min, and then allowed to anneal for 15 min at 42°C in 50 mM Tris-HCl pH 8, 10 mM MgCl₂, and 100 mM NaCl. Five units of RNaseH (Epicentre Technologies) was added and the reaction was incubated at 37°C for 30 min. Nucleic acids were ethanol-precipitated, separated on a 1.5% (w/v) formaldehyde gel, transferred onto nylon membranes, hybridized to the EX1 probe, and subjected to autoradiography. The same blot was then hybridized to the 3PE probe and re-exposed.

Primer Extension

Primer extension was performed as described (Ausubel et al., 1996). -³²P end-labeled primers (ZMAUX1242A, ZMAUX1078A, or ZMAUX969A [5'-GGCAACAAGAATCCAACCGGC-3']) were mixed individually with 20 µg of total RNA and extended using AMV RT (Pharmacia Biotech). The extension products were separated on 8% (w/v) polyacrylamide/urea gels alongside with ³⁵S-labeled sequencing reactions primed with the same primer used for the extension reaction.