The Plant Cell, Vol. 5, 311-319, March 1993 O 1993 American Society of Plant Physiologists Transposon-Mediated Mutations in the Untranslated Leader of Maize Adhl that lncrease and Decrease Pollen-Specific Gene Expression R. Kelly Dawe,' A. Rosanna Lachmansingh, and Michael Freeling Department of Plant Biology, University of California, Berkeley, California 94720 The unstable mutation Adhl-Fm335 contains a Dissocialion (Dsl) transposable element at position +53 in the untrans- lated leader of the maize Alcohol dehydrogenase-1 (Adhl) gene. Excision of Dsl is known to generate new alleles with small additions and rearrangements of Adhl DNA. We characterized 16 revertant alleles with respect to ADHl activity levels in scutellum (nutritive tissue of the seed), anaerobic root, and pollen. Whereas gene expression was not different from the wild type in the sporophytic tissues of the scutellum and anaerobic root, there were strong allelic differences in pollen. One allele underexpressed pollen ADH1 at 48% of the wild-type level, and another overexpressed pollen ADHl at 163% of the wild-type level. Quantitative RNase protection assays demonstrated that the mutant phenotypes reflected changes in the levels of steady state mRNA in pollen. These data provide a definitive demonstration of an overexpression mutant in plants and further show that marked increases in mRNA levels can follow minor alterations in central untrans- lated leader sequences. The nucleotide sequence of 12 new revertant alleles and the molecular mechanisms responsible for pollen-specific gene expression are discussed. INTRODUCTION Little is known of how gene expression changes during the alternating diploid and haploid generations of vascular plants. Studies indicate that most plant genes are expressed in both the diploid stage, or sporophyte, and the haploid stage, or gametophyte (Ottaviano and Mulcahy, 1989). In maize, ~ 72% of all isozymes (Gorla et al., 1986) and 85% of all mRNAs (Mascarenhas et al., 1984) are.expressed in both the sporo- phyte and male gametophyte. One way to study the genetic basis of gametogenesis isto analyze genes that are expressed specifically in pollen. This method has been used to identify 5'sequences required for pollen expression in avariety of spe- cies (van Tunen et al., 1990; Albani et al., 1991; Twell et al., 1991; Hamilton et al., 1992). Transient expression assays have been used to identify minimal pollen-specific promoters, en- hancer sequences, and potential negative regulatory elements (Twell et al., 1991; Hamilton et al., 1992). Sequence compari- sons (Albani et al., 1991) and the fact that a pollen-specific promoter region from maize faithfully directs pollen expres- sion in tobacco indicate that the cis-acting sequences involved in pollen expression are functionally conserved among diver- gent species (Guerrero et al., 1990). Another strategy for analyzing gametophytic gene expres- sion involves mutational analysis of genes that are expressed 1 To whom correspondenceshould be addressed at the Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720. throughout the life cycle. The maize Alcoholdehydrogenase-7 (Adhl) gene is particularly well suited to genetic studies in both sporophyte and gametophyte (see Freeling and Bennett, 1985; Bailey-Serreset al., 1988; Williams et al., 1991). Although most mutations reduceAdh7 expression to the same extent in both the sporophyte and gametophyte (Freeling and Bennett, 1985), mutations can also affect either one independently. This is il- lustrated by the mutant allele Adhl-3F7724, which carries a transposon in the promoter region that is flanked on both sides by a TATA box. Enzyme activity and Adhl-3F7724 transcript lev- els are reduced in root and scutellum but not in pollen (Chen et al., 1987). Excision of the transposon and both TATA boxes leads to restored root expression but significantly reduced pol- len expression (Kloeckener-Gruissem and Freeling, 1987; Kloeckener-Gruissemet al., 1992). Thus, pollen expression is unaffected by transposon insertion at the TATA box but dra- matically affected by the removal of the consensus promoter region. In a similar study of the bronze4 gene, transposon- mediated lesions at position -63 in the promoter reduced en- dosperm and husk expression but left pollen expression unaffected (Sullivan et al., 1989). These studies suggest that sporophyte-and gametophyte-specifying elements coexist in plant promoters, similar to other control elements (Benfey and Chua, 1990). The unstable mutation Adhl-Fm335 provides an opportu- nity to further analyze the sporophyticlgametophytic regulation of maize Adhl. Adhl-Fm335 expresses a low level of ADHl
Transcript
The Plant Cell, Vol. 5, 311-319, March 1993 O 1993 American Society of Plant Physiologists
Transposon-Mediated Mutations in the Untranslated Leader of Maize Adhl that lncrease and Decrease Pollen-Specific Gene Expression
R. Kelly Dawe,' A. Rosanna Lachmansingh, and Michael Freeling Department of Plant Biology, University of California, Berkeley, California 94720
The unstable mutation Adhl-Fm335 contains a Dissocialion (Dsl) transposable element at position +53 in the untrans- lated leader of the maize Alcohol dehydrogenase-1 (Adhl) gene. Excision of Dsl is known to generate new alleles with small additions and rearrangements of Adhl DNA. We characterized 16 revertant alleles with respect to ADHl activity levels in scutellum (nutritive tissue of the seed), anaerobic root, and pollen. Whereas gene expression was not different from the wild type in the sporophytic tissues of the scutellum and anaerobic root, there were strong allelic differences in pollen. One allele underexpressed pollen ADH1 at 48% of the wild-type level, and another overexpressed pollen ADHl at 163% of the wild-type level. Quantitative RNase protection assays demonstrated that the mutant phenotypes reflected changes in the levels of steady state mRNA in pollen. These data provide a definitive demonstration of an overexpression mutant in plants and further show that marked increases in mRNA levels can follow minor alterations in central untrans- lated leader sequences. The nucleotide sequence of 12 new revertant alleles and the molecular mechanisms responsible for pollen-specific gene expression are discussed.
INTRODUCTION
Little is known of how gene expression changes during the alternating diploid and haploid generations of vascular plants. Studies indicate that most plant genes are expressed in both the diploid stage, or sporophyte, and the haploid stage, or gametophyte (Ottaviano and Mulcahy, 1989). In maize, ~ 7 2 % of all isozymes (Gorla et al., 1986) and 85% of all mRNAs (Mascarenhas et al., 1984) are.expressed in both the sporo- phyte and male gametophyte. One way to study the genetic basis of gametogenesis isto analyze genes that are expressed specifically in pollen. This method has been used to identify 5'sequences required for pollen expression in avariety of spe- cies (van Tunen et al., 1990; Albani et al., 1991; Twell et al., 1991; Hamilton et al., 1992). Transient expression assays have been used to identify minimal pollen-specific promoters, en- hancer sequences, and potential negative regulatory elements (Twell et al., 1991; Hamilton et al., 1992). Sequence compari- sons (Albani et al., 1991) and the fact that a pollen-specific promoter region from maize faithfully directs pollen expres- sion in tobacco indicate that the cis-acting sequences involved in pollen expression are functionally conserved among diver- gent species (Guerrero et al., 1990).
Another strategy for analyzing gametophytic gene expres- sion involves mutational analysis of genes that are expressed
1 To whom correspondence should be addressed at the Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720.
throughout the life cycle. The maize Alcoholdehydrogenase-7 (Adhl) gene is particularly well suited to genetic studies in both sporophyte and gametophyte (see Freeling and Bennett, 1985; Bailey-Serres et al., 1988; Williams et al., 1991). Although most mutations reduceAdh7 expression to the same extent in both the sporophyte and gametophyte (Freeling and Bennett, 1985), mutations can also affect either one independently. This is il- lustrated by the mutant allele Adhl-3F7724, which carries a transposon in the promoter region that is flanked on both sides by a TATA box. Enzyme activity and Adhl-3F7724 transcript lev- els are reduced in root and scutellum but not in pollen (Chen et al., 1987). Excision of the transposon and both TATA boxes leads to restored root expression but significantly reduced pol- len expression (Kloeckener-Gruissem and Freeling, 1987; Kloeckener-Gruissem et al., 1992). Thus, pollen expression is unaffected by transposon insertion at the TATA box but dra- matically affected by the removal of the consensus promoter region. In a similar study of the bronze4 gene, transposon- mediated lesions at position -63 in the promoter reduced en- dosperm and husk expression but left pollen expression unaffected (Sullivan et al., 1989). These studies suggest that sporophyte- and gametophyte-specifying elements coexist in plant promoters, similar to other control elements (Benfey and Chua, 1990).
The unstable mutation Adhl-Fm335 provides an opportu- nity to further analyze the sporophyticlgametophytic regulation of maize Adhl. Adhl-Fm335 expresses a low level of ADHl
312 The Plant Cell
(20% of wild type) and in the presence of the regulatory ele- ment Acfivator (Ac), reverts to wild type in -5% of progeny (Osterman and Schwartz, 1981). Molecular analysis proved that Adhl-Fm335 contains a small Dissociafion (Ds7) element at position +53 in the untranslated leader of Adhl (Sutton et al., 1984). The sequence data from eight revertants of Adhf- Fm335 (Dennis et al., 1986) was used in the formulation of molecular models for transposition in plants (Peacock et al., 1984; Saedler and Nevers, 1985). It was found that an 8-bp region of host DNA was duplicated upon the insertion of an Ac or Ds element and that parts of the duplication were often left behind when the transposon was excised. We demonstrate that such small insertions in the untranslated leader of theAdh7 gene affect mRNA accumulation specifically in pollen and that mutation in this region can result in a significant increase in gene expression over wild-type levels.
RESULTS
Identification of the Progenitor Allele
To provide an accurate description of Adhl-Fm335 and its de- rivatives, it was necessary to make comparisons to the progenitor allele from which the mutant was derived. Unfortu- nately, the progenitor allele for Adhl-Fm335 is not known and the original strains are no longer available. Because the 5‘se- quences flanking the maize Adhf gene are exceptionally polymorphic (Johns et al., 1983), it was possible to identify the progenitor allele by comparing the 5’ restriction map of Adhl- Fm335 to other strains of similar descent (see Methods). Using this approach, we identified the progenitor of Adh7-Fm335 as the same progenitor that produced another unstable mutation called Adh7-2F77 (which has a Ds2 element in the fourth exon; Chen et al., 1986). Both unstable mutations were derived from the unusual Adh7-2Fallele (Doring et al., 1984). Knowing this; we were able to utilize a full revertant of Adh7-2F77, Adh7-2F77- r57, as our wild-type strain (this revertant was isolated by B. DeFrancisci, B. Kloeckener-Gruissem, and M. Freeling; others were described by Chen et al., 1986).
ldentification and Phenotypic Characterization of Revertants
The Adhl-Fm335 mutant and its revertants produce an elec- trophoretically fast-migrating (F) enzyme, whereas Adhl-S produces an electrophoretically slow-migrating (S) enzyme. Ac- tiveADH1 is a molecular dimer in which the F and S monomers randomly associate; thus, extracts from heterozygous diploid tissues produce three electrophoretically distinguishable bands representing the F.F, F5 , and S S ADHl dimers. During mei- osis, the two alleles segregate away from each other such that only the F.F and S.S homodimers are present in pollen extracts.
In either diploid or haploid tissue, the distinct Adhf-S counter allele provides a sensitive interna1 control, serving as a refer- ente for studying quantitative changes in Adhl-F expression (for review, see Freeling and Bennett, 1985).
Ten plants carrying Adhl-Fm335 and Ac were crossed as males to plants carrying Adhl-S or AdhW3034 (another slow- migrating allele). From these crosses, a total of 1286 kernels were screened for ADH1-F activity in scutellum (nutritive tis- sue that surrounds the embryo) that deviated from the 20% of wild-type activity that is typical of Adhl-Fm335. There were 43 revertants with approximately wild-type expression and no partia1 revertants or null mutants. Based on preliminary ob- servations, 12 revertants were subsequently chosen for detailed study. Each of the revertants was confirmed as a derivative of the Adh7-2F allele by the presence of two diagnostic 5’ re- striction sites (data not shown). In addition, we included severa1 previously described revertants of Adhl-Fm335 in our analy- sis (with the prefix RV). These were RWO, RV26, RV37, and RV46, all of which are known to be unique at the nucleotide sequence level (Dennis et al., 1986; seeds were generously supplied by D. Schwartz, Indiana University, Bloomington).
In previous studies of Adhl-Fm335 revertants, emphasis was placed on the DNA alterations caused by Dsl excision and not on the effects that excision may have had on Adhl expres- sion (Dennis et al., 1986). Hence, neither the gametophytic expression of revertant alleles nor a rigorous description of their sporophytic expression is available. Using starch gels such as those shown in Figure 1, we analyzed the progenitor allele and each of 16 revertant alleles in scutellum, anaerobic root, and pollen. Statistical analysis of the quantitative data, shown in Table 1, indicated that there was no difference in scutellum or anaerobic root expression; however, pollen gene expression varied widely among alleles. Three categories of revertants were established: (1) a single low allele with reduced pollen expression (RV26); (2) 13 revertants with pollen expres- sion levels similar to the AdhF2F progenitor (RV70, RV37, RV46, d778, d795, d805, d807, d808, d870, d827, d825, d827; and d836); and (3) two revertants with increased pollen expres- sion (d807 and d876). The category designations are illustrated in Figure 2. Relative to wild-type pollen expression (Adh7-2F), the alleles ranged in pollen expression from a low of 48% (RV26) to a high of 163% (d876).
The pollen-specific phenotypes described here could con- ceivably be the result of mutations that alter the viability of pollen. For instance, if the RV26 allele (or a closely linked mu- tation) selectively reduced the viability of Adhl-F pollen, then the amount of ADH1-F enzyme in pollen might be reduced. Likewise, d807 may have provided for an increased pollen via- bility. This was determined by crossing RV26IAdhl-S and d8071 AdhlS pollen to females carrying a third allele. In such crosses, half the progeny should carry ADH1-F and half should carry ADH1-S. Testcross pollinations of this type produced ADH1-S and ADH1-F in a 1:l ratio, showing that theAdh7-Fm335 rever- tants affect the level of ADHl product without affecting pollen viability (for RV26lAdhl-S, progeny were 87(S):107(F); for d8071 Adhl-S, 85(S):65(F); x2 = 2.1 and 2.7, respectively).
Pollen Gene Expression 313
F.FF.Ss.s
Adh2
F.FF.SS-S
SCUTELLUM ANAEROBIC ROOT POLLEN
Figure 1. Starch gels of Adh1-2F, Ffl/26, and d816 in Scutellum, An-aerobic Root, and Pollen.
Data shown are from AdhT-S heterozygotes. In diploid tissues of scutel-lum and anaerobic root, three MM bands are present representingthe two homodimer bands (S-S and F-F) and the heterodimer band(F-S). The isozyme ADH2 is also strongly induced in anaerobic root,so that two additional bands representing the ADH1-S ADH2 andADH1-F-ADH2 heterodimers are visible (for simplicity these are col-lectively labeled "ADH2"). In pollen, because AdM is not expresseduntil the products of meiosis have separated, the ADH1 (S-F) hetero-dimer is absent (Adh2 is not expressed in pollen).
Sequence Analysis
The progenitor allele and the 12 revertant alleles were clonedand the DMA sequence determined around the site of Ds7 in-sertion; the progenitor sequence was derived from Adh1-2F11.
In addition, the RV26 allele was cloned and the published se-quence was verified. As shown in Table 2, all of the revertantsdiffered from the progenitor allele by minor alterations of the8-bp host duplication sequence. The alterations are typical ofthe excision events generated by other plant transposons: partsof the host DNA sequence were deleted, and base pair trans-versions occurred around the original position of the Ds1element (Peacock et al., 1984; Saedler and Nevers, 1985). DNAgel blot analysis further suggested that there were no largescale changes within 10 kb upstream of the Adh1 start site (tothe Hindlll site; data not shown). The sequence data revealedthree unique alleles that have not been previously described.The novel alleles included two category 2 alleles, d795 andd807, and the category 3 alleles, d801 and d816, which areidentical in sequence (Table 2). The remaining alleles repre-sent new isolates of previously described revertants (Denniset al., 1986). The d836 allele is identical to RV31 and RV43,d808 is identical to RV46, and six new alleles (o778, d805, d810,d821, 0-825, and d827) are identical to RV1 and RV56.
ADH1 expression levels were not significantly differentamong alleles with the same sequence. Individual compari-sons among the six alleles identical to o778 produced nosignificant differences at the 5% confidence level. Addition-ally, there were no significant differences between two otherpairs of identical alleles, d801 versus d816 and RV46 versusd83& The strong correlation between nucleotide sequence andADH1 expression levels confirms the overall accuracy of using
Table 1. Scutellum, Root, and Pollen Expression of Adh1-FM335, Adh1-2F, and Various Adh1-Fm335 Revertants8
a Calculated from Adh1-FIAdh1-S heterozygotes and expressed as a percent of ADH1-S enzyme activity.b Data from seeds and plants that lacked Ac.0 Gel band too faint to be quantified.o Alleles identified by Dennis et al. (1986).• Derived from the same male parent.' Derived from the same male parent.
314 The Plant Cell
. , 3 200
190 -
9 110 w 100 R P J A 80
70 060 ' 50 se 40
30 20 10
n
L'
7
Adhl-Fm335 derivatives Figure 2. Pollen Expression of Adhl-2F and Its Derivatives.
Data on pollen expression from Table 1 are plotted as a percent of the ADH1-2F (wild-type) level. Standard deviations are shown as ver- tical lines at the top of each bar, and three categories of alleles (1, 2, and 3) are denoted.
allozyme ratios to measure quantitative gene expression (Freeling and Bennett, 1985).
Analysis of mRNA
In pollen, RV26 expresses ADHl at 48% of the wild-type level and d807 expresses ADHl at 148% of the wild-type level. To address the question of whether the mutations affect the level of steady state mRNA or the efficiency of mRNA translation, mRNA levels were analyzed in mutant pollen using RNase pro- tection assays. Plants either homozygous for RV26, d778, and d807 or heterozygous for d778lRV26 and d778ld801 were used to prepare anaerobic root and tassel spikelet poly(A)+ mRNA. Spikelets, which include pollen, anther walls, and leaflike struc- tures that surround the anthers, were used as a source of pollen mRNA because they produce more abundant and generally better qualityAdh7 mRNA than from mature pollen itself. Adhl mRNA is not found in leaf tissue, and previous studies indi- cate that, if present, Adh7 mRNA derived from anther walls is minimal (Kloeckener-Gruissem et al., 1992).
A uniformly labeled RNA probe complementary to Adh7 mRNA was prepared from a plasmid carrying the d778 clone. When a d778 probe is hybridized with mRNA from a homozy- gous d778 plant and digested with RNase A, a large portion of the probe homologous to theAdh7 leader is protected from
digestion. As shown in Figure 38, the predicted size of the protected fragment is 139 nucleotides. In contrast, when the radioactive d778 probe is annealed to the heterologous RV26 and 19801 mRNAs, RNase-sensitive loops are formed in the nonhomologous regions (Figure 3B). The predicted sizes of the fragments following RNase digestion of the heterologous hybrid molecules are 51 and 84 nucleotides for RV26, and 49 and 86 nucleotides for d807 (Figure 38). As shown in Figure 3A, the bands generated empirically by RNase protection are within 2 or 3 nucleotides of the predicted sizes. Such devia- tions from the predicted lengths are probably caused by the variable melting of double-stranded molecules at ends and within loop regions and the fact that sizes were estimated from DNA standards, which differ slightly in molecular weight from the RNA molecules under study. On the whole, the sizes of the protected fragments from all three alleles, d778, RV26, and d807, are consistent with the previously suggested site of tran- scription initiation in pollen (Kloeckener-Gruissem et al., 1992).
Unexpected high molecular weight bands were also ob- served following RNase protection of spikelet mRNAs (Figure 3A). For instance, an additional 145-nucleotide band was ob- served in spikelet mRNA derived from d807 plants (Figure 3A, lanes 8 and 9). This 145-nucleotide band could be interpreted as representing a second class of d807 mRNA, perhaps with a nove1 initiation site upstream. However, the 145-nucleotide band from either homozygous or heterozygous plants was not cleaved into two smaller sized fragments (Figure 3A, lanes 8 and 9), suggesting that it was not derived from a legitimate d807 transcript. To reproduce the 145-nucleotide band under different conditions, RNase protection of the d807 mRNA was performed using an homologous d807 probe, and only a sin- gle band with the predicted size of 135 nucleotides was produced (data not shown). These data suggest that the un- expected large bands observed in Figure 3A are primarily the result of experimental artifacts and are not factors in the tissue-
Table 2. Sequence of the Dsl-lnduced Host Duplication Region from Eiqht Unique Adhl-Fm335 Revertant Alleles
Allele Sequencea
Adhl-2F GGTGAGGGACTGA--------GGGTC Fm335b GGTGAGGGACTGAGGGACTGAGGGTC d778 and 7 othersC GGTGAGGGACTGTCGGACTGAGGGTC RVlO GGTGAGGGACTGTCCGACTGAGGGTC d795 GGTGAGGGACTGCCGGACTGAGGGTC d807 GGTGAGGGACTC CGGACTGAGGGTC d808, RV31, and RV43 GGTGAGGGACTG GGACTGAGGGTC d836 and RV46 GGTGAGGGACT GGACTGAGGGTC R V26 GGTGAGGGACTGT CACTGAGGGTC d80l and d876 GGTGAGGGAC GGACTGAGGGTC
a Bases given in italic represent base pair transversions from the host duplication sequence. b The underlined region was duplicated upon insertion of Dsl, which lies between the central A and G bases (Dsl is not shown). C ldentical to RVl, RV56, d805, d870, d825, d827, and d836.
Pollen Gene Expression 315
DNA1 2 3 4 5 6 7 8 9 standards
145139
8684
specific expression of the mutant phenotypes. The observa-tion that unexpected bands were found in spikelet mRNA, butnot root mRNA, may reflect the fact that significantly more spike-let mRNA was used in the hybridization reactions (seeMethods).
Quantitative differences in steady state mRNA levels be-tween the RV26 and d801 alleles were analyzed by comparingplants heterozygous for d778/RV26 and d778/d801. Extractsfrom these plants produced three protected fragments (Fig-ure 3A, lanes 5 to 8), making it possible to use the d778 mRNAas an internal standard for the quantification of RV26 and d801Adh1 mRNAs. In root, the RV26-\o-d778 and d801-\o-d778mRNA ratios were 1.04 ± 0.1 and 1.17 ± 0.04, respectively,whereas in pollen, the ftV26-to-o778 and d807-to-d778 mRNAratios were 0.96 ± 0.11 and 2.81 ± 0.08 (mean ± SD), respec-tively. Thus, in both organs, the levels of MM mRNA closelyparallel the level of the gene product. In root, where RV26 andd801 express the same level of ADH1 (Table 1), the mRNA lev-els are nearly identical; whereas in pollen, where the levelsof ADH1 in RV26 and d801 plants differed by a factor of 3.1(Table 1), there was a 2.9-fold difference in the relative quanti-ties of spikelet mRNA.
DISCUSSION
-51-49
B139nt
d778 (+44nt)-/-GCCACTGTCGGACTGA-^-(+79nt)
Figure 3. RNase Protection of RV26 and d801 mRNA in AnaerobicRoots and Spikelets.(A) Autoradiograph of one RNase protection experiment. A DNA se-quencing reaction was used for approximate length standards, whichare given at right in nucleotides. A uniformly labeled antisense RNAprobe from the cloned d778 allele (lane 1, at a 1/500 dilution) was an-nealed to anaerobic root and spikelet poly(A)+ mRNA purified fromplants of five different gentoypes (shown above each lane). When theo778 probe was hybridized with root mRNA from a homozygous o778plant, a single band was produced with the expected length of 139nucleotides (lane 2). However, when the probe was hybridized to theheterologous RV26 and dflOJ mRNAs, two bands of the (roughly)predicted lengths of 51 and 84 nucleotides (RV26) and 49 and 86 nucleo-tides (d801) were protected. In heterozygous plants that contain boththe homologous d778 mRNA and the heterologous RV26 or d801mRNA, RNase protection produced all three bands (lanes 5 to 8). Otherhigher molecular weight bands are explained in the text.
By studying revertants of the DsWnduced mutation Adh1-Fm335, we have directly demonstrated a naturally occurringoverexpression mutant in maize and have shown that over-expression is caused by increased steady state levels of mRNA(Figures 1 and 3A). To our knowledge, no similar c/s-actingoverexpression mutation has been identified in plants. Thenearest examples are in Drosophila, where transposon inser-tion has been shown to cause mutations with increased levelsof steady state mRNA. In the Hairy-wing mutations, transpo-sons within a structural gene cause an increased level oftruncated transcripts (Campuzano et al., 1986). In three high-G6PD activity mutants (Ito et al., 1989) and the Om(1D) muta-tions (Tanda and Corces, 1991), overexpression of the geneproduct is caused by the activity of a nearby transposon. Al-though ourAdhl mutations were generated by a transposon,they differ from the Drosophila mutants by the fact that a trans-poson is no longer present at the locus. Thus, in addition tothe previously described transposon-associated phenomena,
(B) Line drawing illustrating the regions of nonhomology between o778and the two alleles RV26 and d801. The alleles are identical exceptfor the region affected by Ds1 excision. When the d778 probe is hy-bridized to d778 mRNA, a single band of 139 nucleotides is expected.Because the d778 allele is longer than either of the alleles RV26 andd801, nonhomologous loops formed in the probe strand of the hybridsare digested by RNase A, yielding two unequal sized fragments (theexpected sizes in nucleotides (nt) are indicated for each allele).
316 The Plant Cell
which include genetic instability, the generation of stable null mutants, the alteration of enzymatic properties, and genetic suppression (reviewed by Wessler, 1988; Gierl et al., 1989), we describe a clear case where the excision of a transposon has the positive effect of creating a new allele with a level of expression that is higher than that of the wild-type progenitor.
Our revertants identify a new region of maize Adhl between positions +45 and +52 that regulates transcript accumula- tion in pollen. Mutations at this site can both increase and decrease pollen-specific expression (Figure 2, categories 1 and 3). Pollen-specific information in the Adhl gene has pre- viously been localized to the TATA box at position -31 using the unstable mutant Adhl-3Fll24 (Chen et al., 1987; Kloeckener-Gruissem et al., 1992). While the TATA box has a known function in promoting transcription, the untranslated leader has not been shown to regulate organ-specific expres- sion at a quantitative level. The 8-bp sequence GGGACTGA was duplicated by the insertion of Dsl into Adhl-Fm335 (Ta- ble l). The last seven bases of this region, GGACTGA, are conserved among four different Adhl alleles for which se- quence is available (Sutton et al., 1984; Chen, 1986; the alleles are Adhl-PF, Adhl-S, Adhl-F, and Adhl-3F). Further, as shown in Figure 4, this 7-bp motif is found twice in the untranslated leader of Adhl, once at position +19 and once at +46.
At least four lines of evidence suggest a specific role for the duplicated GGACTGA motif in promoting pollen-specific ADHl expression. (1) RV26, the only revertant with reduced pollen expression (Figure 2) is also the only revertant that lacks a complete GGACTGA motif at position +46 (Table 2). (2) A gene encoding a closely related maize ADH isozyme, Adh2, lacks any sequence homology to GGACTGA in the untranslated leader (Dennis et al., 1985) and is not expressed in pollen (Schwartz, 1971). (3) The maize Waxy locus contains a single CGACTGAsequence beginning at position +21 in the untrans- lated leader (Klosgen et al., 1986), which is very similar in position and content to the first repeat of the Adhl motif, ex- cept for a C residue substituted for a G residue in the first base pair. Waxy is the only other well-studied maize gene that is
+1 +19 ACAGGCTCATCTCGGTTTGGACTGATTGGTTTCGTAA
CTGGTGAG-GGGTCTCGGAGTGGATCGATTT +108
GGGATTCTGTTCGAAGATTTGCGGAGGGGGGCAATG
+4 6
Figure 4. The Untranslated Leader Sequence of AdhlPF:
The transcription start site (+1) for Adhl was determined by Howard et al. (1987) and Kloeckener-Gruissem et al. (1992). The sequence is redrawn from Sutton et al. (1984) to highlight the duplicated GGACTGA motif, which occurs once at position +I9 and again at position +46. In Adhl-Fm335, the insertion of Dsl produced a duplication of the se- quence GGGACTGA at position +45, placing the Dsl element at position +53. The ATG initiation codon occurs at position +108.
expressed in both the sporophyte and male gametophyte. (4) There is near match to the GGACTGA motif in the pollen- specific clone BP79, which was derived from oilseed rape. Po- sition +25 in the untranslated leader of BP79 contains the sequence GGTCTGA, which differs from the Adhl sequence by a T residue substituted for an A residue at the third base pair (Albani et al., 1991).
Among the known pollen-specific genes, BP79 from oilseed rape is the only other gene demonstrated to have the same developmental expression profile as maize Adhl. Both Adhl and BP 79 are expressed throughout pollen development, be- ginning at a stage immediately following the completion of meiosis (Stinson and Mascarenhas, 1985; Albani et al., 1991). This pattern differs from most pollen-specific genes, which lack the GGACTGA motif and are not expressed until the first micro- spore mitosis or later (van Tunen et al., 1989; Twell et al., 1991; Baltzet al., 1992; Carpenter et al., 1992; Hamilton et al., 1992). Thus, it may be that the GGACTGA motif is involved in the early expression pattern typical of Adhl and BP79. We also note that the GGACTGA sequence shows similarity to a motif demon- strated to enhance gene expression in pollen, the 56/59 box, which has a critical sequence of GTGA (Twell et al., 1991). Whether the GGACTGA sequence is an indispensable feature of pollen-specific expression or acts by enhancing the expres- sion provided by another region(s) remains to be seen. The availability of a simple and accurate transient assay system (Hamilton et al., 1992) should make it possible to address this question.
We provide the sequence of 12 newly identified revertants of Adhl-Fm335 (Table 2). Each can be fully explained by previ- ously proposed models for the mechanism of transposition in plants (Peacock et al., 1984, Saedler and Neves, 1985). The most unique feature of this large group of alleles is the fact that one of the alleles was recovered eight times (those identi- cal to d778). These alleles were derived from different male parents, ruling out the possibility that they were clonal deriva- tives of the same excision event. It appears that although the Ds element generates important genetic diversity, the range of variation may be limited by recurring classes of excision products.
At the level of ADHl activity, we characterized 13 alleles with wild-type expression levels (Figure 2, category 2), one allele with reduced pollen expression (category l), and two alleles with increased pollen expression (category 3). The revertant phenotypes were entirely pollen specific: all of the 16 rever- tants analyzed had wild-type expression levels in the sporophytic tissues of the scutellum and anaerobic root. DNA sequence analysis and RNase protection experiments were used to understand the molecular mechanism(s) underlying these pollen-specific phenotypes. Based on RNase protection data, we argue that the mRNAfrom three alleles (d778, RV26, and d807) is initiated at the same position as wild-type mRNA. Furthermore, because the steady state levels of mRNA paral- lel the changes in ADHl enzyme activity for each allele (Figure 3A), we suggest that the translatability of the mRNAs is not
Pollen Gene Expression 317
altered. The overall length of the untranslated leader appears to have no effect on gene expression, because the progenitor allele is shorter than both the underexpressing and overex- pressing alleles (Table 2, RV26 and d876). Finally, we know that the DNA sequences involved are small in scale; the d836 allele differs from the d876 allele by only one thymidine nucleo- tide, yet they differ in pollen ADHl levels by a factor of 1.8.
It is possible that changes in the central 5'untranslated leader of Adh7 have the effect of altering mRNA stability. The stabil- ity of Adh7 mRNA has been discussed in a previous report on Adhl-Fm335 (Dennis et al., 1988). In roots, the Os7 element is spliced from the mRNA, leaving 14 nucleotides of Ds7 se- quence and the 8-bp host duplication sequence. This 22-bp insertion contains some of the same sequence found in the revertants. Whereas the Adhl-Fm335 mutation reduced steady state mRNA to 1% of wild-type levels, the rate of transcription from the gene remained unaltered (Dennis et al., 1988). It is possible that the stability of the mRNA is reduced due to the 22-bp insertion, but this effect is probably a result of improper processing of the Ds"intron"(as noted by the authors). Indeed, in most eukaryotic genes the sequences that regulate mRNA decay are found in 3' untranslated regions (see Brawerman, 1987; Gruissem et al., 1988).
Alternatively, the small insertions conditioned by Os exci- sion might affect transcription. Although the Adhl GGACTGA motif between +46 and +52 is downstream of the region nor- mally associated with promoting transcription, promoter elements have been shown to occur in intragenic sequences (Hutmarket al., 1986; Nakatani et al., 1990). The best charac- terized example is an element between +10 and +50 of the human glial fibrillary acidic protein (gfa) gene that interacts with the TATA binding factor TFllD (Nakatani et al., 1990). In both Arabidopsis and maize, there are two genes that encode TFllD (Gasch et al., 1990; J. M. Vogel and M. Freeling, manu- script in preparation), raising the possibility that specific expression patterns may involve different TATA binding fac- tors, perhaps with different recognition sequences. High resolution run-on transcription studies or biochemical analy- ses of TFllDlpromoter interactions will help to clarify how the revertants of Adhl-Fm335 affect mRNA accumulation.
METHODS
Selectlon of Revertants
Seeds homozygous for Adhl-fm335 and heterozygous for Activator (Ac) were provided by John Osterman, University of Nebraska at Lin- coln. Ten plants of this genotype were crossed as a male to plants either homozygous for Adhl-S or Adhl-S3034 (for description of Adhl alleles, se8 Freeling and Bennett, 1985). Scutellar slices were pre- pared and Adhl allozymes separated by starch gel electrophoresis (Freeling and Schwartz, 1973). Seeds showing wild-type Adhl-F lev- els were grown and either self-crossed or crossed by a strain homozygous for Adhl-S.
ADHl Activity Measurements
All assays of revertant expression were carried out on Adhl-fm335- (derivathe)/Adhl-S heterozygotes. Extracts from mature (aerobic) scutel- lum and anaerobically induced root (submerged in water for 24 hr) were prepared as described by Woodman and Freeling (1981). P o l h was collected and stored dry at -8OOC (ADH1 activity lasts for at least 1 year under these conditions). Approximately 10 mg of pollen was ground in 20 mg alumina (Sigma) and 100 pL of extraction buffer (50 mM Tris-HCI, 5% [vhr] 2-mercaptoethanol, 15% [vhr] glycerol, and 0.01% [whr] bromophenol blue) using plastic mortars designed to fit into microcentrifuge tubes (Kontes Pellet Pestles; Fisher Scientific). The extracts were incubated on ice for 15 min, and centrifuged at 14,000 rpm for 15 min. lhe Adhl allozymes were separated by either 5% stack- ing and 8% separating native polyacrylamide gel electrophoresis (Sachs et al., 1980) or by starch gel electrophoresis as described above.
ADHl expression levels were quantified using an integrating densi- tometer (lransidyne General 2970; see Woodman and Freeling, 1981). l h e percent contribution by either allele was determined by summing the integrated density of the appropriate homodimer band (FF or SS) and one-half the density of the hetercdimer band (FS). Because ADHl is not expressed until the products of meiosis have separated (Stinson and Mascarenhas, 1985), Adhl-S and Adhl-F are expressed indepen- dently in pollen; thus, in heterozygous stocks, it is possible (and rnost accurate) to calculate ADH1-F activity as a percentage of ADH1-S. Statistical comparisons (Student's t tests) were made computationally using the Statworks program (Cricket Software, Inc., Philadelphia, PA).
DNA Methodology
DNA was prepared by the method of Murray and lhompson (1980). Unless otherwise noted, we followed standard procedures for DNA manipulation (Sambrook et al., 1989). Restriction mapping was car- ried out by DNA gel blot analysis using a 750-bp radioactive probe homologous to the 5' untranscribed region of Adhl (BamHI to Banll fragment). Five enzymes (Hindlll, Sacl, Kpnl, Bg111, and Bcll) were used to identify Adhl-2F as the progenitor of Adhl-Fm335.
The revertants were cloned by amplifying the untranslated leader of the respective alleles using a polymerase chain reaction (PCR) kit (Perkin Elmer Cetus) and 21 nucleotide primem (primem OAand UG06; see Kloeckener-Gruissem et al., 1992) in a DNA thermal cycler (Perkin Elmer Cetus). The DNA was denatured for 1 min at 93OC, annealed for 2 min at 55OC. and extended for 3 min at 72%, 25 times each. For d778, d795, d807, and Adhl-2F77, the DNA was digested with Haelll and Bcll. The resulting 218-bp fragment was ligated into pBluescript KS- (Stratagene), digested with EcoRV and BamHI, and transformed into fscherichia coli DH5a (Bethesda Research Laboratories). PCR products from the remaining revertants (RV26, d805.4807, d808, d870, d876, d827, d825, d827, and d836) were cloned directly using the TA cloning kit from lnvitrogen (San Diego, CA). The colonies obtained from transformations were screened by colony hybridization with a probe specific to the amplified region.
For each allele, plasmids from two independent colonies (only one for Adhl-2Fll) were purified for sequencing. The double-stranded DNA of each of 27 plasmids was sequenced in one direction using Se- quenase V.2 (U. S. Biochemicals) and a primer homologous to the T7 promoter. Only the sequence duplicated by the Dsl element and 20 bp on each side were recorded and analyzed. l h e sequences did not vary between duplicate cloned alleles except for two single nucleo- tide differences, which were assumed to be PCR errors (in one d795
318 The Plant Cell
and one RV26 clone). Otherwise, except for the regions affected by Ds excision, all revertants were identical to the Adhl-2Fll sequence and to the relevant sequence of Adhl-Fm335 (Sutton et al., 1984).
RNase Protection Assays
PoIy(A)+ mRNA was prepared from'anaerobic roots (submerged in wa- ter for 24 hr) and mature spikelets as described by Kloeckener-Gruissem et al. (1992). RNase protection was essentially as described in Kloeckener-Gruissem et al. (1992) with the following modifications. The DNA templates for probe synthesis were the cloned d778 and d807 alleles described above. The plasmids were linearized with Hindlll and a 270-nucleotide radioactive probe prepared with T7 RNA polymer- ase (Promega). The probes were annealed to ~0.01 pg of the root mRNAs and 1.0 pg of the spikelet mRNAs (Adhl mRNA is highly abun- dant in anaerobic roots) and digested with 80 NglmL RNase A (Sigma) for 60 min at 21OC. RNase was removed by a 40-min digestion with protease K (0.25 pg/mL) at 21OC. DNA sequencing reactions were used as approximate size markers. Quantitation (by densitometry of two ex- periments) was carried out using a video camera and National lnstitutes of Health software in the Center for Plant Developmental Biology at University of California, Berkeley. Only the 49-, 51-, 84-,8&, and 139- nucleotide bands shown in Figures 3A and 38 were used in quantitation.
ACKNOWLEDGMENTS
We thank Marguerite C. Brickman for help with data analysis. We also thank Barbara Kloeckener-Gruissem and Julie M. Vogel for helpful discussions and for critically reading the manuscript. This work was supported by a U.S. Department of Agriculture National Needs Fel- lowship to R.K.D. and a Department of Energy grant to M.F.
Received October 29, 1992; accepted January 5, 1993.
REFERENCES
Albani, D., Altosaar, I., Arnison, P.G., and Fabijanski, S.F. (1991). A gene showing sequence similarity to pectin esterase is specifi- cally expressed in developing pollen of Brassica napus. Sequence in its S'flanking region is consewed in other pollen-specific pmmoters. Plant Moi. Biol. 16, 501-513.
Bailey-Serres, J., Kloeckener-Gruissem, B., and Freeling, M. (1988). Genetic and'molecular approaches to the study of the anaerobic response and tissue specific gene expression in maize. Plant Cell Environ. 11, 351-357.
Baltz, R., Domon, C., Plllay, D.T.N., and Stelnmetz, A. (1992). Char- acterization of a pollen-specific cDNA from sunflower encoding a
,zinc finger protein. Plant J. 2, 713-724. Benfey, P.N., and Chua, NrH. (1990). The cauliflower mosaic virus
35s promoter: Combinatorial regulation of transcription in plants. Science 250, 959-966.
Brawerman, G. (1987). Determinants of messenger RNA stability. Cell 48, 5-6.
Campuzano, S., Balcells, L., Vlllares, R., Carramolino, L., Garcia- Alonso L., and Modolell, J. (1986). Excess function Hajfy-wing mu- tations caused by gypsy and copia insertions within structural genes of the achaete-scufe locus of Dmsophila. Cell 44, 303-312.
Carpenter, J.L., Ploense, S.E., Snustad, D.P., and Sllflow, C.D. (1992). Preferential expression of an a-tubulin gene of Arabidopsis in pol- len. Plant Cell 4, 557-571.
Chen, C.-H. (1986). Mutational analysis of maize Adhl by the use of transposable elements. Ph.D. Thesis (Berkeley: University of California).
Chen, C.-H., Freeling, M., and Merckelbach, A. (1986). Enzymatic and morphological consequences of Ds excisions from Adhl. Maydica
Chen, C.-H., Oishi, K.K., Kloeckener-Gruissem, B., and Freeling, M. (1987). Organ-specific expression of maizeAdh7 is altered after a MU transposon insertion. Genetics 116, 469-477.
Dennis, E.S., Sachs, M.M., Gerlach, W.L., Flnnegan, E.J., and Peacock, W.J. (1985). Molecular analysis of the alcohol dehydmg- enase 2 (Adh2) gene of maize. Nucl. Acids Res. 13, 727-743.
Dennis, E.S., Gerlach, W.L., and Peacock, W.J. (1986). Excision of the Ds controlling element from the Adhl gene in maize. Maydica
Dennis, E.S., Sachs, M.M., Gerlach, W.L., Beach, L.,and Peacock, W.J. (1988). The Dsl transposable element acts as an intron in the mutant allele Adhl-Fm335 and is spliced from the message. Nucl. Acids Res. 16, 3815-3828.
D6ring, H.-P., Freeling, M., Hake, S., Johns, M.A., Kunze, R., Merckelbach, A., Salamlni, F., and Starllnger, P. (1984). ADs-mu- tation of the Adhl gene in maize. MOI. Gen. Genet. 193, 199-204.
Freeling, M., and Schwartz, D. (1973). Genetic relationships betweeri the multiple alcohol dehydrogenases of maize. Biochem. Genet. 8,
Freeling, M., and Bennett, D.C. (1985). Maize Adhl. Annu. Rev. Ge- net. 19, 297-323.
Gasch, A., Hoffmann, A., Horlkoshi, M., Roeder, R.G., and Chua, N.-H. (1990). Arabidopsis thaliana contains two genes for TFIID. Na- ture 346, 390-394.
Gierl, A., Saedler, H., and Peterson, P.A. (1989). Maize transposa- ble elements. Annu. Rev. Genet. 23, 71-85.
Gorla, M.S., Frova, C., Blnelll, G., and Ottaviano, E. (1986). The extent of gametophytic-sporophytic gene expression in maize. Theor. Appl. Genet. 72, 42-47.
Gruissem, W., Barkan, A., Deng, X., and Stern, D. (1988). Transcrip- tional and posttranscriptional control of plastid mRNA levels in higher plants. Trends Genet. 4, 258-263.
Guerrero, F.D., Crossland, L., Smutzer, G.S., Hamllton, D.A., and Mascarenhas, J.P. (1990). Promoter sequences from a maize pollen- specific gene direct tissue-specific transcription in tobacco. MOI. Gen. Genet. 224, 161-168.
Hamilton, D.A., Roy, M., Rueda, J., Sindhu, R.K., Sanford, J., and Mascarenhas, J.P. (1992). Dissection of a pollen-specific promoter from maize by transient transformation assays. Plant MOI. Biol. 18,
Howard, E.A., Walker, J.C., Dennls, E.S., and Peacock, W.J., (1987). Regulated expression of an elcohol dehydmgenase 7 chimeric gene introduced into maize protoplasts. Planta 170, 535-540.
Hutmark, D., Klemenz, R., and Gehring, W.J. (1986). Translational and transcriptional control elements in the untranslated leader of the heat-shock gene hsp22. Cell 44, 429-438.
31, 93-108.
31, 47-57.
27-36.
211-218.
Pollen Gene Expression 319
Ito, H., Yoshida, K., and Horl, S.H. (1989). Positive regulation of the Drosophile melanogaster G6PD gene by an insertion sequence. Bio- chem. Genet. 27, 379-391.
Johns, M.A., Strommer, J.N., and Freeling, M. (1983). Exception- ally high levels of restriction site polymorphism in DNA near the maize Adhl gene. Genetics 105, 733-743.
Kloeckener-Gruissem, B., and Freellng, M. (1987). Relationships be- tween anaerobic inducibility and tissue-specific expression for maize anaerobic genes. In Plant Molecular Biology, D. von Wettstein and N.-H. Chua, eds (New York: Plenum Press), pp. 293-303.
Kloeckener-Grulssem, B., Vogel, M., and Freellng, M. (1992). The TATA box promoter region of maize Adhl affects its organ-specific expression. EMBO J. 11, 157-166.
Klosgen, R.B., Gierl, A., Schwarz-Sommer, i!., and Saedler, H. (1986). Molecular analysis of the waxy locus of Zea mays. MOI. Gen. Genet. 103, 237-244.
Mascarenhas, N.T., Bashe, D., Elsenberg, A., Wllling, R.P., Xiao, C.M., and Mascarenhas, J.P. (1984). Messenger RNA in corn pol- len and prdein synthesisduring germination and pollen tube growth. Theor. Appl. Genet. 68, 323-326.
Murray, M.G., and Thompson, W.F. (1980). Rapid isolation of high molecular weight plant DNA. Nucl. Acids Res. 8, 4321-4362.
Nakatanl, Y., Horikoshl, M., Brenner, M., Yamamoto, T., Besnard, F., Roeder, R.G., and Freese, E. (1990). A downstream initiation element required for efficient TATA box binding and in vitro function of TFIID. Nature 348, 86-88.
Osterman, J.C., and Schwartz, D. (1981). Analysis of a controlling- element mutation at the Adhl locus in maize. Genetics 99,267-273.
Ottaviano, E., and Mulcahy, D.L. (1989). Genetics of angiosperm pol- len. Adv. Genet. 26, 1-64.
Peacock, W.J., Dennis, E.S., Gerlach, W.L., Sachs, M.M., and Schwartz, D. (1984). lnsertion and excision of Ds controlling ele- ments in maize. Cold Spring Harb. Symp. Quant. Biol. 49,347-354.
Sachs, M.M., Freeling, M., and Okimoto, R. (1980). The anaerobic proteins of maize. Cell 20, 761-767.
Saedler, H., and Nevem, P. (1985). Transposition in plants: Amolecu- lar model. EMBO J. 4, 585-590.
. .
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Clon- ing: A Laboratory Manual. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press).
Schwartz, D. (197l). Genetic control of alcohol dehydrcgenase-A com- petition model for gene action. Genetics 67, 411-425.
Stinson, J., and Mascarenhas, J.P. (1985). Onset of alcohol de- hydrogenase synthesis during microsporogenesis in maize. Plant Physiol. 77, 222-224.
Sullivan, T.D., Schiefelbein, J.W., and Nelson, O.E., Jr. (1989).Tissue- specific effects of maize Bronze gene promoter mutations induced by Dsl insertion and excision. Dev. Genet. 10, 412-424.
Sutton, W.D., Gerlach, W.L., Schwartz, D., and Peacock, W.J. (1984). Molecular analysis of Ds controlling element mutations at the Adhl locus of maize. Science 223, 1265-1268.
Tanda, S., and Corces, V.G. (1991). Retrotransposon-induced overex- pression of a homeobox gene causes defects in eye morphogenesis in Drosophila. EMBO J. 10, 407-417.
Twell, D., Yamaguchi, J., Wing, R.A., Ushlba, J., and McCormick, S. (1991). Promoter analysis of genes that are coordinately expressed during pollen development reveals pollen-specific enhancer se- quences and shared regulatory elements. Genes Dev. 5,496-507.
van Tunen, A.J., Hartman, S.A., Mur, L.A., and MOI, J.N.M. (1989). Regulation of chalcone flavanone isomerase (CHI) gene expression in Petunia hybrida: The use of alternative promoters in corolla, an- thers and pollen. Plant MOI. Biol. 12, 539-551.
van Tunen, A.J., Mur, L.A., Bmuns, G.S., Rienstra, J.-D., Koes, R.E., and MOI, J.N.M. (1990). Pollen- and anther-specific chi promoters from petunia: Tandem promoter regulation of the chiA gene. Plant Cell 2, 393-401.
Wessler, S.R. (1988). Phenotypic diversity mediated by the maize trans- posable elements Ac and Spm. Science 242, 399-405.
Williams, C.E., Kloeckener-Gruissem, B., and Freeling, M. (1991). Naturally occurring variants of maize Adhl differ in organ-specific expression both in quantity and developmental timing. Maydica 36,
Woodman, J.C., and Freeling, M . (1981). ldentification of a genetic element that controls the organ specific expression Of Adhl in maize. Genetics 98, 357-378.
115-128.
DOI 10.1105/tpc.5.3.311 1993;5;311-319Plant Cell
R K Dawe, A R Lachmansingh and M Freelingdecrease pollen-specific gene expression.
Transposon-mediated mutations in the untranslated leader of maize Adh1 that increase and
