Alternative 3’ Splice Acceptor Sites Modulate Enzymic Activity in Derivative … · The Plant Cell, Vol. 4, 1453-1462, November 1992 O 1992 American Society of Plant Physiologists

The Plant Cell, Vol. 4, 1453-1462, November 1992 O 1992 American Society of Plant Physiologists

Alternative 3’ Splice Acceptor Sites Modulate Enzymic Activity in Derivative Alleles of the Maize bronzel-mutable 13 Allele

Ron J. Okagaki,’ Thomas D. Sullivan, John W. Schiefelbein,2 and Oliver E. Nelson, Jr. Laboratory of Genetics, University of Wisconsin, Madison, Wisconsin 53706

The defective Suppressor-mutator (dSpm)-induced allele bronzel-mutable 73 (bzl-ml3) and many of its derivative alleles are leaky mutants with measurable levels of flavonol O3-gIucosyltransferase activity. This activity results from splicing at acceptor site-1, one of two cryptic 3’splice sites within the dSpm insertion in bzl-ml3- In this study, splicing in bz7-11173 change-in-state (CS) alleles CS-3 and CS-64 was shown to be altered from bz7-m73; previous work found altered splicing in CS-9. CS-64 is a null allele and lacks the acceptor site-1-spliced transcript because this site is deleted. CS-3 and CS-9 had increased levels of the acceptor site-1 transcript relative to bz7-ml3 and increased enzymic activities. A deletion in CS-9 altered splicing by eliminating acceptor site-2. Both acceptor sites were intact in CS-3, but a deletion removed most of a 275-bp GC-rich sequence in dSpm. This suggests that GC-rich sequences affect splicing and is consistent with models postulating a role for AU content in the splicing of plant introns. Splicing does not necessarily occur, however, at the junction of AU-rich intron sequences and GC-rich exon sequences.

INTRODUCTION

lnsertion mutants in maize often exhibit partia1 enzymic activity when transposable element sequences are spliced out of pre-mRNAs (reviewed in Weil and Wessler, 1990). One such mutant is bronzel-murable 73 (bzl-ml3). This allele contains a 2241-bp defective Suppressor-muraror (dSpm) element inserted 38 bp downstream of the sole intron in Bz7 (Schiefelbein et al., 1988a). One unspliced and two spliced transcripts have been detected in bzl-ml3. Spliced transcripts utilize the normal 5’splice donor site and one of two cryptic 3’acceptor sites, AS-1 and AS-2, within the dSpm insertion. Figure 1A depicts the structure of a Bzl allele along with the dSpm insertion in bzl-m73.

Change-in-state (CS) alleles of bzl-m13 are derivatives in which the dSpm insertion is altered; many contain deletions within the dSpm insertion (Schiefelbein et al., 1985a). These alleles condition increased or decreased enzymic activities relative to bzl-m13 (Raboy et al., 1989). Previous work determined that bzl-m13 had 13% of nonmutant enzymic activity in husk tissue (Kim et al., 1987; Raboy et al., 1989), and that splicing at AS-1, but not AS-2, produces an in-frame transcript that may encode the functional enzyme, flavonol 03-glucosyltransferase (UFGT EC 2.4.1.91). This suggests that these alleles possess splicing mutations in bzl-m13 acting to increase or

To whom correspondence should be addressed at the Department of Horticultural Sciences, University of Florida, Gainesville, FL 32611. * Current address: Department of Biology, University of Michigan, Ann Arbor, MI 48109.

decrease the amount of AS-1-spliced in-frame transcript and thereby the functional enzyme. Here, we present an analysis of severa1 bzl-m13 CS alleles and the effect of the mutations on splicing and enzymic activity.

RESULTS

Structure of the Alleles

DNA gel blot analysis identified a deletion of approximately 400 nucleotides in CS-3 (Schiefelbein et al., 1985a). This in- formation was used to design a polymerase chain reaction (PCR) cloning strategy. Figure 1A shows the position of primers, ‘!A“ and “D,” used to amplify CS-3. DNA sequence analysis of amplified products showed that the 5‘ break point of the CS-3 deletion is at nucleotide 2465 and the 3‘ break point falls between 2904 and 2907. Between the break points are 21 bp of filler sequence. The last 19 bp of filler sequence plus the next 4 bp of dSpm sequence, CCAT, are identical to an upstream dSpm sequence located at 2266 to 2288 (Figure 1).

Differences between CS-64 and bzl-m13 were not detected by DNA gel blot analysis (J. W. Schiefelbein and T. D. Sullivan, unpublished data). However, the null phenotype of CS-64 in the absence of Spm, coupled with the late excision events detected when Spm is present, suggested a small lesion in the downstream terminal inverted repeat (TIR). A 2-bp deletion

Dow

nloaded from https://academ

ic.oup.com/plcell/article/4/11/1453/5984423 by guest on 08 August 2021

1454 The Plant Cell

A

bzlml3 dS

V 0.25 D

B 748

I B Z 1 ffCCCACCCGGGCCPCGCCAG

As-l (29871 n bzl -m13 CrCCCACmcactacaaga . (dspm) - tcttgtagtgSfSGZCffCGCCAG

CS - 64

As-? (3003)

tcttg----gSCGGGCCTCGCCAG n

1515 I cs-9 aggatatcaa cttaa cgtactgatc upstream ..................

aggatatcaa ctcaa atcaagagtg CS-9

ccccgacact cttaa attaagagtg downstream ................

I 1 8 8 s

2465 I cs-3 gctgg cagggcgaggccgcccgcccg acgc cgccc upstream . . . . . . . . . . . . .

gctgg gcgtgagaaaccctgtactaa ccat taaga cs-3

gtgca gcaaaacccacactttcacct ccat taaga domscream ..................

I 2925

Figure 1. Alleles Characterized in This Study

(A) Shown on the top line is the Bz7 allele. Open boxes represent exons, whereas lines denote the intron and flanking sequences. lmmediately below, a dotted line marks the bz7-R deletion. A Smal site in Bz7 is disrupted by the bzl-ml3 insertion. The dSpm insert in bzl-ml3 is represented by a triangle. Open vertical arrows mark the positions of the cryptic 3‘ splice acceptor sites AS-I and AS-2, and the filled vertical arrow identifies the upstream Pstl site depicted in Figure 2. Horizontal arrows show oligonucleotide primers used for PCR amplification, namely A, B, C, and D. Sequences deleted in CS-3 and CS-9 are denoted by breaks in lines. CS-64 contains a 4-bp deletion that disrupts AS-I as shown by the plus (+) symbol. Restriction enzymes are B, BstYI; N, Ncol; P, Pstl; S, Sall; Sm, Smal; X, Xbal. Numbering is based on bzl-ml3, where position 1 is at the transcription start site and 2989 is the last nucleotide in the dSpm insertion. (6) The exon 2 sequence of 6z7 flanking the insertion site is shown. Underlined is the host sequence duplicated by the dSpm insertion; note that the insertion disrupts a Smal site (CCCGGG). lmmediately below is shown the sequence from bzl-ml3; capital letters denote the Bz7 sequence and lowercase letters denote the dSpm sequence. A vertical arrow marks the 3’ splice site AS-I. The dSpm sequence is presented 5’ to 3’ as oriented in bzl-ml3, which is opposite to the

within the upstream TIR of bZl-ml3 produced an allele with raie excisions, CS-6 (Schiefelbein et al., 1988a). A small deletion in the downstream TIR could similarly reduce transposition frequency, and in addition a deletion here may disrupt the AS-1 splice site producing the null phenotype. PCR was used to amplify and clone this region of CS-64 using primers “B’hand “ C (Figure 1A). Based on the sequence of two clones, we found a 4-bp deletion removing bases 2985 to 2988 including the AG dinucleotide in AS-1 (Figure 1B).

Splicing in bzl-mlt and CS Alleles

Splice sites in bzl-ml3, CS-3, and CS-9 were determined by RNA gel blots, analysis of cDNA clones, and S1 nuclease map- ping (Kim et al., 1987; Raboy et al., 1989). These results were used to design nuclease protection experiments to quantify relative amounts of transcripts in these alleles. CS-64 was studied in a separate set of experiments. Figure 26 diagrams the probes that were used.

Nuclease protection experiments with bzl-ml3, CS-3, and CS-9 used RNA isolated from two or more plants. Most RNA samples were assayed in independent experiments, and relative proportions of transcripts between experiments remained within a few percent, as shown in Table 1. Detected radioac- tivity was approximately linear when RNA input was increased from 10 to 40 pg (data not shown). Routinely 10 to 20 pg of RNA was used.

Pattern of Splicing in bzl-ml3

bzl-m73 RNA produced three labeled fragments using probe bzl3P-Nco (data not shown). The AS-1-spliced transcript protects a 595-nucleotide fragment, a 1721-nucleotide fragment is protected by the AS-2-spliced transcript, and the unspliced transcript protects a 2640-nucleotide fragment. Average values are 81% k 3%, 14% f 2%, and 4% f 1% for the AS-1- spliced, AS-2-spliced, and unspliced transcripts, respectively (Table 1). Averages here and elsewhere were rounded off and may not add up to 100%.

direction of transcription in Spm. The 4-bp deletion in CS-64 is shown by dashes. A potential3‘splice site that may be used in CS-64 is marked with an arrow. CS-9 and CS-3 have interna1 deletions of the dSpm sequence, and sequences flanking the deletion break points are shown. Asterisks mark bases that are in common with bzl-ml3 sequences at the upstream and downstream break points; upstream sequences are presented above and downstream sequences are below the sequence of the alleles. Deletion break points in CS-9 occur within a direct repeat sequence, cttaa. lnserted between CS-3 break points are 21 bases of filler sequence. Nineteen of the inserted bases are identical with a nearby sequence located from positions 2266 to 2285 followed by four additional bases overlapping the 3‘ break point.

Dow



Splicing in bronzel-mutable 73 Derivatives 1455

Probe cs3XB-Bst was used with one bzl-m13 sample giv- ing an estimate of 63% AS-1 transcript (Table 1); the difference between this measurement and the previous one is discussed later. Lane 3 of Figure 3 shows the two fragments from cs3XB- Bst protected by bzl-m13 RNA. The AS-1 transcript protects the 188-nucleotide fragment, while AS-2 and unspliced transcripts protect a 251-nucleotide fragment. A larger, approximately 375- nucleotide band is thought to be an artifact occurring when two FINA molecules hybridize to the probe.

CS-64: Splicing in the Absence of AS-1

The recently isolated CS-64 allele has not been incorporated into the 8 PI genetic background necessary for strong expression of the Bzl allele in husk tissue; therefore, poly(A)+ RNA was isolated from seedling stem tissue of CS-64 and bzl-m13 plantlets (Sullivan et al., 1989). RNA gel blot analysis of this RNA detected an unspliced transcript and an AS-2-spliced transcript in CS-64 (data not shown). These transcripts comigrated with their counterparts from bzl-m73. As expected, the AS-1-spliced transcript is absent in CS-64, although during long exposures there is evidence for a transcript of

P X AS-2 S AS-1 B N P Sm A

C rbzl3P

rD3S1 0.5 kb

Figure 2. Probes and in Vitro-Transcribed RNAs Used in This Study.

(A) Diagram of the 3' portion of bzl-ml3. Restriction enzyme sites marked are B, BstYI; N, Ncol; P, Pstl; S, Sall; Sm, Smal; X, Xbal. Ar- rows mark splice sites. The exon sequence is represented by an open box, dSpm by a filled box, and lines represent flanking untranscribed sequences. Deleted sequences in CS-3 are indicated by the dotted line. (B) Three probes used in nuclease protection experiments. bzl3P-Nco and cs3XB-Bst are end-labeled probes, whereas cs64PB-Sal is a uniformly labeled riboprobe. The gap in cs3XB-Bst denotes sequences present in bz7-ml3 but absent in this CS-3-derived probe. Angled lines represent the vector sequence included in the probes. (C) In vitro-transcribed RNAs used to standardize nuclease protection experiments. rbzl3P represents unspliced transcripts and rD3S1 represents AS-1-spliced transcripts.

'

Table 1. Results from Nuclease Protection Assays Measuring Relative Levels of Transcripts

Percent Transcript Spliced At Allelel Plant No. AS-1 AS-2 Unspliced Probe

bzl-m13 15 78

80 18 83

84 39 84

79 63

1 86 20 95

97 21 99

97 96

31 92 96 92

CS-3

cs-9 32 95

92 94

38 93 97 97

16 6 16 4 13 4 12 4 12 4 17 4 - 37a -

4 10 1 4 1 2 O 1 - 3a -

4a - 2 6 1 3 2 6

-

~~~~~~

-, Proportion of transcript spliced at the indicated splice acceptor site was not measured. For bzlm73, plant number 39, and CS-3, plant number 21, the probe used measured the sum of unspliced and AS-2 spliced transcripts. In CS-9, AS-2 is not present. a Total of AS-2 and unspliced transcripts.

approximately the same size as the AS-1-spliced transcript in CS-64 (data not shown).

Nuclease protection experiments characterized these transcripts using a 499-nucleotide, uniformly labeled riboprobe, cs64PB-Sal (Figure 28). bzl-m13 RNA spliced at AS4 protected the 188-nucleotide fragment seen in Figure 4, lane 3. No RNA corresponding to splicing at AS-1 was detected in CS-64 RNA (Figure 4, lane 2). On longer exposure, a faint signal was detected; this may correspond to an RNA spliced at an AG dinucleotide 16 nucleotides downstream of AS-1 (data not shown) or could be a gel artifact.

CS-9: Splicing in the Absence of AS-2

AS-2 was deleted in CS-9, and only two transcripts were expected. With the cs3XB-Bst probe, protected fragments of 188 and 299 nucleotides corresponding to the AS-1 and unspliced transcripts were detected as shown in lane 8 of Figure 3. Probes were protected up to the CS-9 deletion break point. Both RNA gel blot experiments and nuclease protection experiments

Dow



1456 The Plant Cell

1 2 3 4 5 6 7 8

deletion<f== breakpoint

AS-1

Figure 3. Nuclease Protection Experiment with cs3XB-Bst.

In lane 1, Bz1 RNA protects a fragment up to the dSpm insertion twonucleotides downstream from AS-1. bzR RNA does not protect any end-labeled fragments (lane 2). Two fragments are protected by the threetranscripts from bz1-m13 (lane 3). Both the unspliced transcript andthe AS-2 transcript were cleaved at the deletion break point in theCS-3-derived probe. CS-3 (lanes 4 to 7) and CS-9 (lane 8) contain ahigher proportion of RNA spliced at AS-1 than bz1-m13. Total RNA(20 ng) from husk was used in this experiment.

failed to show additional transcripts in this allele. In this re-spect, CS-9 differs from the loss of AS-1 in CS-64. With probebz13P-Nco, 95% ± 4% of CS-9 transcripts were spliced atAS-1 (Table 1).

CS-3: Deletion of a GC-Rich Sequence betweenAS-1 and AS-2

In contrast with CS-64 and CS-9, both cryptic splice acceptorsites are intact in CS-3. The AS-1-spliced, AS-2-spliced, andunspliced transcripts produced three protected fragments withprobe cs3XB-Bst that were of 188, 375, and 446 nucleotides,respectively. A 646-nucleotide fragment was produced whenthe probe hybridized to its complementary strand. Figure 3shows results from an RNase protection experiment with RNAfrom four different CS-3 plants and one bz1-m13 plant. Visualinspection of Figure 3 indicates that a greater fraction of AS-1transcripts is found in CS-3 than in bz1-m13 (lane 3 versus lanes4 through 7). Quantification of signals produced with cs3XB-Bst

verified this increase, and in CS-3 the average estimatesare 94% ± 4%, 2% ± 1%, and 4% ± 3% for AS-1-spliced,AS-2-spliced, and unspliced transcripts, respectively.

Standardization of Results

One anomaly was found. The proportion of AS-1 transcript inbzl-m13 was 81% ± 3% using bz13P-Nco and 63% withcs3XB-Bst. The accuracy of quantification by nuclease protec-tion assays was assessed using two in vitro-transcribed RNAsdepicted in Figure 2C. RNA rbz13P represents an unsplicedtranscript, and rD3S1 represents the AS-1-spliced transcript.These RNAs were mixed in molar ratios of 3:2, 3:10, and3:20 (rD3S1 to rbz13P) and used in S1 nuclease protection

(CS_-S4splice site)

probeAS-2 andunspliced

C&J34deletion,"AS-1"

Figure 4. RNase Protection Experiment of CS-64.

Under conditions used, nucleases cut inefficiently at the four-nucleotidegap between bz1-m13 RNA and the CS-64-derived cs64PB-Sal probe.This permitted detection of unspliced and AS-2-spliced transcripts frombz1-m13. Lane 1 contains the tRNA control. AS-2 and unspliced tran-scripts were detected in seedling stem RNA from CS-64 (lane 2) andbz1-m13 RNA from seedling stems (lane 3) or husk (lane 4), but theAS-1-spliced transcript was not detected with CS-64. Lane 5 contain-ing BzT-3 husk RNA shows the expected band. Two micrograms ofpoly(A)+ seedling stem RNA and 0.2 ng of total husk RNA were usedin the experiments.

Dow



Splicing in bronzel-mutable 13 Derivatives 1457

Table 2. Bias in the Detection of Transcripts by NucleaseProtection AssaysInput RatiorD3S1:rbz13P3:23:103:20

Observed RatiorD3S1:rbz13P3:1.53:6.53:10

Input/Observed0.750.650.50

experiments. These experiments measured ratios of 3:1.5,3:6.5,and 3:10, as shown in Table 2, with cs3XB-Bst. An autoradio-gram from this experiment is shown in Figure 5. Nucleaseprotection experiments apparently underestimated actualamounts of unspliced transcript. The deviation ranged from0.75 to 0.50 in control experiments with an average value of0.629. Results with probe cs3XB-Bst, therefore, were adjustedby 0.629 as described in Methods. Adjusted results arepresented in Table 3.

The small amount of protected transcript and backgroundnoise prevented accurate quantification of results with probebz13P-Nco. Visual inspection of autoradiograms suggests thatoverestimation of AS-1 transcript is greater than with cs3XB-Bst (Figure 5; compare lanes 3 and 7). This possibly accountsfor different results using bz13P-Nco and cs3XB-Bst withbz1-m13 RNA.

DISCUSSION

Structure of Derivative Alleles

Deletion break points in derivative alleles resemble break pointsfound in spontaneous mutations. Filler DMA related to nearbysequences is inserted between deletion break points in sev-eral maize deletions (Wessler et al., 1990). This occurs in CS-3.Short, direct repeats at deletion break points are anothercommon structure (Albertini et al., 1982). Deletion break pointsin CS-9 occur within the direct repeat CTTAA (Figure 1B).Direct repeats are also found at the deletion break pointsfrom the CS-72 derivative of bz-m13 (G. Bunkers, S. G. Pickett,and V. Raboy, unpublished data), derivative dSpm-7995 ofa-m2 (Masson et al., 1987), and the dSpm in bz1-m13 itself(Schiefelbein et al., 1988a). Resemblance of these deletionbreak points with break points in spontaneous deletions sug-gests that double strand breaks within dSpm are initiated bytransposase and enlarged by exonuclease activity. Repair

bz13P-Nco probe cs3XB-Bst probe

unspllcedand rbz13P ,

1 2 3 4

I I

5 6 7 8

Enzymic AssaysAS-1 andrD3S1

3= unspJIced

<^= AS-2

UFGT enzymic activity was determined for BzT-3, bz1-m13,CS-3, and CS-9. Tissues assayed were the same as used forRNA studies. Plants containing the bz1-m13 allele were het-erozygous with bz1-R allele, whereas other alleles were inthe homozygous condition. Measured enzymic activities forbz1-m13, therefore, were doubled to make results comparable(Dooner and Nelson, 1977a). Activity was not assayed for CS-64.CS-64 is phenotypically null, and previous work with null alleleshas shown that these alleles have undetectable (less than 1%of nonmutant) activity (Dooner and Nelson, 1977b; Schiefelbeinet al., 1988b; Raboy et al., 1989).

The nonmutant allele BzT-3 had 436 units per mg protein,and bz1-m13 has 23% of nonmutant activity (101 units per mgprotein). CS-3 and CS-9 are intermediate with approximately50% of nonmutant activity (216 and 227 units per mg pro-tein, respectively). Analysis of variance indicated statisticallysignificant differences between alleles, and simultaneousconfidence intervals for differences between alleles werecalculated. Enzymic activity of Bz7'-3 was significantly differ-ent at the 95% confidence level from CS-3, CS-9, and bz1-m13;bz1-m13 and CS-3 also differed at the 95% confidence level.The bz1-m13 and CS-9 alleles differed at the 90% confidencelevel. These results were similar to those previously obtainedby Raboy et al. (1989), who reported 13% of nonmutant activ-ity for bz1-m13, 39% for CS-3, and 49% for CS-9.

bzR band

AS-1 and'rD3S1

Figure 5. Standardizing Nuclease Protection Results Usingin Vitro-Transcribed RNAs.Lanes 1 and 5 contain 15 ̂ g of total RNA from bz1-m13 and CS-3, respec-tively. The remaining lanes contain 2 ng of in vitro-transcribed RNA;molar ratios of rD3S1 to rbz13P are 3:2 (lanes 2 and 6), 3:10 (lanes3 and 7), and 3:20 (lanes 4 and 8). For bz13P-Nco, arrows mark bandsproduced by bzR RNA, AS-2-spliced RNA, and AS-1-spliced RNA inlane 1; for lanes 2 to 4 bands were produced by hybridization withRNAs rbz13P and rD3S1. Arrows for cs3XB-Bst mark bands producedby the AS-1-spliced, AS-2-spliced, and unspliced transcripts in lane5, and bands from rbz13P and rD3S1 in lanes 6 to 8. This autoradio-gram was underexposed to highlight differences in lanes 6 through8. For this reason, the AS-2-spliced and unspliced transcripts expectedwith CS-3 are not visible. Difficulties with gel conditions and backgroundprevented quantification of results with bz13P-Nco. However, detec-tion of rbz13P here is reduced compared with cs3XB-Bst.

Dow



1458 The Plant Cell

Table 3. Relationship between the Fraction of Transcripts Spliced at AS-1 and Enzymic Activity

Allele AS-1 (YO) (Ulmg Proteiny

BZ 1‘-3 - b 436 bz1 -m 13 51 101 cs-3 91 216 cs-9 93 227

Spliced at UFGT

a Flavonol 03-glucosyltransferase activity is in units per mg protein. Splicing occurs at the normal 3’ splice site in the nonmutant allele,

not at AS-I.

would preferentially occur at short, direct repeats (Roth and Wilson, 1986) or be mediated by nearby sequences producing filler sequence.

AS-1 Splicing 1s Responsible for Enzymic Activity

CS-64 provided direct evidence that the AS-1 transcript encodes a functional enzyme. Earlier work with bzl-ml3 and its derivatives suggested that AS-1 splicing produced functional enzyme (Kim et al., 1987). This was based on the observation that splicing at AS-1 maintains the open reading frame. Thirty- eight nucleotides in exon 2 are deleted, two nucleotides of dSpm sequence are inserted, and three nucleotides from the host sequence duplication are gained for a net loss of 33 nucleotides. CS-64 has lost the AS-1 transcript and results in the null phenotype. This strongly suggests that the AS-1 transcript encodes a functional enzyme.

Detection of Transcripts in RNA Protection Assays

Nuclease protection assays identified an increase in the proportion of AS-1-spliced transcript from CS-3 and CS-9 relative to bZl-ml3. 60th probes used gave this result, although estimates of the actual increase differed with the probe used. Notably, the proportion of AS-1 transcript in bzl-ml3 was 81% 2 3% when measured with probe bzl3P-Nco versus 63% with probe cs3XB-Bst. Synthetic RNAs corresponding to the AS-1 transcript and the unspliced transcript, and control experiments using these RNAs confirmed that these RNAs were not detected with equal efficiency. There are severa1 possible causes for this result. Differences in the length of hybridizing regions between probes and transcripts may affect renaturation kinetics (Chamberlin et al., 1978), or differences in sequence composition of RNA-DNA duplexes could lead to differences in hybridization.

We attempted to correct for this discrepancy by standardizing results using in vitro-transcribed RNAs. Lacking a synthetic RNA corresponding to the AS-2 transcript, the correction is probably incomplete. Despite this, the approximate twofold increase of UFW enzymic activity in CS-3 and CS-9 over bzl-m13

was approximately matched by increases in the estimated fraction of AS-1-spliced transcript from 51% in bzl-m13 to 91% and 93% in CS-3 and CS-9, respectively.

Splicing in bzlml3 and CS Alleles

These derivative alleles of bZl-ml3 effectively represent a col- lection of splicing mutants identifying sequences important for proper splicing of pre-mRNAs. Splicing in bZl-ml3 occurs at two cryptic 3’ splice sites as determined by RNA gel blot analysis (Raboy et al., 1989). Nuclease protection experiments examining the 3’half of the dSpm insertion confirmed this. This analysis does not eliminate the possibility of transcripts spliced beyond the ends of the probes, but RNA gel blot experiments of Raboy et al. (1989) suggest that such transcripts either do not exist or are in low abundance. Deleting AS-1 (CS-64) or AS-2 (CS-9) eliminates one of the transcripts as expected, and in the case of CS-64 may activate a cryptic splice acceptor site.

Splicing is altered in CS-3, although splice sites are not disrupted. Deletion break points are 105 and 63 nucleotides away from AS-2 and AS-1, respectively. lncreased stability of the AS-1 transcript in CS-3 versus bzl-m13 cannot explain increase of enzymic activity and the AS-1 transcript, because the sequence of the AS-1 transcript is identical in bzl-m13 and CS-3. De- creased stability of the AS-2 and/or the unspliced transcript in CS-3 could explain a proportional increase in the AS-1 transcript. This would not account for increased enzymic activity unless there were an increase in transcription sufficient to double enzymic activity. Such a coincidence seems implausible, and the recent derivation of CS-3 from bZl-ml3 coupled with the inability of DNA gel blot analysis to detect changes other than the described deletion makes this possibility unlikely (Schiefelbein et al., 1985a). The simplest interpretation of results is that deleting a sequence between AS-2 and AS-1 increases the use of the AS-1 splice site.

In CS-3, the loss of sequence brings AS-1 and AS-2 closer together, thus either the spatial proximity of AS-1 and AS-2 or the missing sequences themselves determine where splicing occurs. Currently, we have no direct evidence to eliminate either possibility. This question has, however, been studied elsewhere. Reed and Maniatis (1986) measured the use of two identical 3’splice sites separated by either exon or nonexon sequences. They found that inserting varying lengths of nonexon sequence between splice sites did not alter the choice of splice sites. The downstream site was used. In contrast, when the exon sequence normally found next to the splice site was present, splicing occurred at the upstream splice site, and they con- cluded that exon sequences influence splice site selection.

AU Content and Splicing

Wiebauer et al. (1988) observed that exon sequences flanking splice junctions have lower AU content than adjacent intron sequences. For dicots, exon sequences have an A + T content

Dow



Splicing in bronzel-mufable 13 Derivatives 1459

of 55.Oo/o, whereas adjacent intron sequences have an A + T content of 74.3%. The percentages are 42.7 and 58.7, respectively, for monocots. This transition is not pronounced in mammals or Saccharomyces cerevisiae. However, abrupt tran- sitions in AU content were subsequently noticed in a ciliate, Tetrahymena thermophila; a slime mold, Dictyostelium dis- coideum; a nematode, Caenorhabditis elegans; an insect, Drosophila; and the fungi Neurospora crassa and Schizosac- charomyces pombe (Csank et al., 1990).

Further evidence for the importance of AU content in splicing in plants carne from experiments using artificial introns. These experiments demonstrated that splicing efficiency drops when GC-rich sequences are added to introns (Goodall and Filipowicz, 1989). Splicing appears then to remove AU-rich sequences at loose consensus sequences near junctions of AU-rich and GC-rich sequences.

Application of this model predicts that AS-2 is favored in bzl-ml3, whereas AS-1 is favored in CS-3. The 5' 821 splice site, together with AS-2, defines atypical AU-rich intron. Figure 6A shows a 275-nucleotide stretch that has a G + C content of 83% and is located between AS-2 and AS-i in bzl-m13; therefore, the intron defined by the 5' BZI splice site and AS-1 produces an intron containing GC-rich sequences in bzl-ml3. This GC-rich sequence may inhibit splicing at AS-1.

In CS-3, both AS-2 and AS1 define AU-rich introns. All except 40 nucleotides of the GC-rich sequence has been deleted. Under the suggested model, this would permit more splicing at AS-1 in CS-3. This can also be interpreted as the presence of a downstream GC-rich sequence makes AS-2 an efficient splice site.

It is surprising that AS-2 is used at all, because AS-2 is em- bedded in AU-rich sequence. The transition to GC-rich sequence begins approximately 50 nucleotides downstream of AS-2, and there are severa1 sequences resembling the 3' splice consensus sequences in the vicinity. One such sequence occurs at the junction between AU-rich and GC-rich sequences (Figure 6B). This sequence (GTGAG-G) matches four of six nucleotides of the consensus (TTCAG-G) proposed by Goodall and Filipowicz (1991). AS-2 (CACAG-G) also matches four nucleotides in the consensus sequence.

One exception where splicing occurs without a transition between GC-rich and AU-rich sequences is GC-rich maize introns (Goodall and Filipowicz, 1991); this does not appear to explain the use of AS-2. Experimental evidence argues that near perfect matches with splice consensus sequences are required for splicing GC-rich introns, and AS-2 is a poor match to the consensus sequence (Figure 6B). Neither does the pres- ente nor absence of potential branch sequences provide a satisfactory explanation for the choice of 3' splice sites. Se- quences resembling the mammalian branch consensus sequence are found upstream of actual and potential splice sites in bzl-ml3, and in addition plant and mammalian introns lacking normal branch sequences are spliced (Padgett et al., 1985; Goodall and Filipowicz, 1989). It is possible that transi- tions from AU-rich to GC-rich sequences specify a locality where splicing will occuc This appears to occur in mammals,

A

K GC

80

70

60

50

40

30

20

10

íf As-'

B

Proposed 3'-Consensus Splice Sequences

Hanley and Schuler 1988

TITIT mTT TGCBG CCGAC ACGNN N a1 ternate

Goodall and Filipowicz 1991

T T C X G AU-rich introns TGCX G GC-rich introns

Sequences of 3'-Splice Sites

B z l TTCCA TCGTT CGCX C

AS-1 GACGT TITCT TGTX T

AS-2 TGCAT TTCAT CACAG G

AS- ' C'ITAC GCTAA G T G Z G

Figure 6. GC Content of dSpm lnsertion in bz-m73 and 3'Splice Sites.

(A) The GC content of dSpm and adjacent 827 sequences is graphed in 50-nucleotide blocks. lndicated at the top are the positions of AS-1, AS-2, CS-3 deletion, and BZ7 sequences. Below, an arrow marks a potential 3'splice site, AS-', that is not used; this site is at the begin- ning of the GC-rich region in dSpm. (E) The 3' splice consensus sequences proposed by Hanley and Schuler (1988) and Goodall and Filipowicz (1991) are presented along with 3' splice sequences of the 8Z1 intron, AS-I, AS-2, and the un- used splice site. For the Hanley and Schuler consensus sequence, the favored nucleotide is shown abwe, with the most common alter- nate given below.

Dow



1460 The Plant Cell

although the distinction there between intron versus exon sequences is not based on AU or GC content (Nelson and Green, 1988). Therefore, whereas current models of plant intron splicing explain decreased utilization of AS-2 in CS-3 versus bzl-ml3, the use of AS-2 remains difficult to explain.

METHODS

Plant Materials and Nucleic Acid lsolation

All Bronze7 (BZ7) alleles used in this study, except for the reference allele, bzl-R, were derived from the bronzel-mutable 73 (bzl-ml3) allele. The bz7-m13 allele contains an insertion of a 2241-bp defective Suppmssor-mutator (dSpm) element inserted into exon 2 of the BZ7 locus (Schiefelbein et al., 19%). 8zP3 is a revertant allele from bZl-ml3 with the nonmutant Bzl sequence restored (Schiefelbein et al., 1985b). lsolation of CS-3 (change-in-state) and CS-9 is described by Schiefelbein et al. (1985a); CS-64 was subsequently isolated in the same manner. Sequences from CS-9 (Kim et al., 1987) and bzl-R (Ralston et al., 1988) have been reported. Alleles used for flavonol 03-glucosyltransferase (UFGT) assays were in a genetic background containing B and PI to boost expression of BZ7 in husk tissue.

DNA was isolated from maize seedlings or mature leaves using a modification of the technique of Shure et al. (1983). Total ANA from husk tissue and poly(A)+ RNA from seedling stems were isolated by the techniques of Sullivan et al. (1989).

UFGT Assays

Husk tissue 15 days after pollination was collected and immediately frozen in liquid nitrogen. Enzyme assays were done as described by Raboy et al. (1989). Activity from the null allelebzl-R was determined, and residual UFGT activity representing a minor non-Bzl-encoded activity was subtracted from total UFGT activity in the other alleles. Each husk sample was assayed once, except for sample 39, where results from two assays were averaged.

Cloning and Sequencing of CS-3 and CS-64

Genomic sequences from CS-3 and CS-64 were amplified using the polymerase chain reaction (PCR). Products were extracted with phenol- chloroform and then chloroform before ethanol precipitation. Sequences were cloned into plasmid vectors using interna1 restriction enzyme sites in the amplified DNA. Conditions used for PCR were taken from Kim and Smithies (1988) and Sullivan et al. (1989), and oligonucleotide primers were synthesized by the University of Wisconsin Biotechnol- ogy Center (Madison); approximate positionsof theprimersareshown in Figure 1.

Amplification of CS-3 sequence used primers D (5‘-CGGTGTCGG- CGTCGTCCTCG-33 and A (5‘-GGCGGCATGCCGAACTGAGT-33. The temperature cycle began with 5 min at 94oC. then 30 cycles of 94°C for 30 sec and 72OC for 5 min, followed by 10 min at 72%. Amplified sequences were digested with Xbal and Mbol and ligated to thevec- tor pUC119 to give clone pcs3XB. Double-stranded DNA from three clones and single-stranded DNA from one clonewere sequenced using Sequenase II (United States Biochemical Corp.).

The CS-64 sequence was amplified using primers B (5’-TGGGCA- GGATCTCCGCGAGC-3’) and C (5’-GAGACGGEGCGGTCGACTG-3’). Approximately 200 ng of genomic DNA from a plant heterozygous CS-64/bzl-R was amplified by PCR: 94OC for 2 min, 40 cycles of 92% for 30 sec and 72OC for 5 min, and a final 15 min incubation at 72%. Amplification products were digested with Sal1 and BstYI, and subcloned into pBluescriptll KS+ (Stratagene) to yield pcs64SR Single- stranded DNA was prepared from two clones and sequenced.

RNA Blot Analysls

Two micrograms of poly(A)+ RNA from seedling stems or 0.2 pg of total RNA from husk tissue was denatured, fractionated on formaldehyde gels, blotted, and probed as described by Sullivan et al. (1989). Mo- lecular weights were estimated using commercially obtained RNA markers (Bethesda Research Laboratories).

Nuclease Protection Assays

Procedures for S1 nuclease protection experiments and 5’end labeling were described earlier (Sullivan et al., 1989). Briefly, RNA was hybridized with denatured probe in 80% formamide overnight. Prod- ucts were digested with S1 nuclease (Pharmacia, Piscataway, NJ) at 37% for 30 min; under these conditions cutting at single base differences and short gaps was limited. After nuclease treatment samples were fractionated on a denaturing gel. Gels were then dried and ra- dioactive signals directly measured using a blot analyzer (Betascope 603; Betagen, Waltham, MA).

Probe bzl3P-Nco was prepared from plasmid pbzml3P (Figure 28). The plasmid was digested with restriction enzyme Ncol prior to 5’end labeling using T4 polynucleotide kinase and Y.~~P-ATP (3000 Cilmmol); the labeled fragment was digested with Pvull, and a 1.7-kb fragment was purified using a denaturing gel. To prepare cs3XB-Bst (Figure 2B), plasmid pcs3XB was cut at Pvull sites in pUC119, and a OBkb fragment was isolated from an agarose gel. This fragment was digested with BstYl and 5‘ end labeled; a 0.8-kb fragment was isolated from a denaturing gel. Specific activity of probes was approximately 106 cpm/pmol.

The RNase protection procedure used for CS-64 was based on a modification of Zinn et al. (1983). RNA-RNA hybrids were treated with 2 pg of RNase T1 (Sigma) and 20 pg of nuclease P1 (Calbiochem, San Diego, CA) in 320 pL of digestion buffer at 30% for 60 min and further treated with 50 pg of prdeinase K (Boehringer Mannheim) and SDS toO.6% for 15 min at 37%. Samples were processed for denaturing acrylamide gel electrophoresis as was done for nuclease protection experiments.

A uniformty labeled RNA probe was used for analysis of CS-64. This probe wasmstructed by inserting a Sal1 fragment from pbzml3P into the Sal1 site of lhe CS-64 plasmid. This added upstream dSpm sequences to pcs64SB creating the plasmid pCS64PB; this plasmid was digested with Sal[ and RNA was transcribed with T7 RNA polymerase. Transcriptions in the presence of 50 pCi of U-~~P-CTP (3000 Cilmmol) were done according to the supplier’s directions (Promega, Madison, WI). The DNA template was removed using RNase-free DNase (Promega), and labeled RNA was precipitated and gel purified. Hybridizations of the riboprobe with RNA were done as described for SI nuclease protection experiments; specific activity of the probe was approximately 3.5 x 107 cpm/pmol.

Dow



Splicing in bronze7-mutable 73 Derivatives 1461

Standardization of Nuclease Protection Experiments

In vitro-transcribed RNAs were,used for standards to assess hybridization of probe to transcripts. RNA rbzl3P was prepared by subcloning the Pstl fragment from pbzl3P into pGEM-3Z (Promega). To produce rD3S1, a Smal fragment from the cloned 627'4 allele was subcloned into pGEM-3Z. The synthetic RNA rbzl3P corresponded to the unspliced transcript, and rD3S1 corresponded to the 3' side of an AS-I-spliced transcript (Figure 2C). Plasmids were linearized at the Hindlll site, and T7 polymerase was used to transcribe RNA using a Promega kit according to the supplier's directions. DNA was removed from RNA by two rounds of LiCl precipitation. These two RNAs were used in SI nuclease protection experiments. Each assay contained 2 ng of in vitro-transcribed RNA with 15 Bg of tRNA.

A correction factor for differences in detecting transcripts was em- pirically determined. The arithmetic mean of underrepresentation in signal from rbzl3P was calculated, 0.629, and the signal for the AS-1 transcript was multiplied by 0.629 to correct for detection bias prior to calculating relative transcript levels.

ACKNOWLEDGMENTS

We thank Glen A. Galau, L. Curtis Hannah, and Don McCarty for helpful discussions concerning the research and writing of this manuscript. Help with field workwas provided by Russell Huseth and Wayne Smith, and technical assistance by Shawn Rogers. This research was sup- ported by the College of Agriculture and Life Sciences, University of Wisconsin-Madison, and Grant No. DMB-8811036 from the National Science Foundation. This is journal paper No. 3233 of the Laboratory of Genetics, University of Wisconsin-Madison.

Received August 3, 1992; accepted September 18, 1992.

REFERENCES

Albertini, A.M., Hofer, M., Calos, M.P., and Miller, J.H. (1982). On the formation of spontaneous deletions: The importance of short sequence homologies in the generation of large deletions. Cell29,

Chamberlin, M.E., Galau, G.A., Britten, R.J., and Davidson, E.H. (1978). Studies on nucleic acid reassociation kinetics. V. Effects of disparity in tracer and driver fragment lengths. Nucl. Acids Res. 5,

Csank, C., Taylor, F.M., and Martindale, D.W. (1990). Nuclear pre- mRNA introns: Analysis and comparison of intron sequences from Tetrahymena thermophila and other eukaryotes. Nucl. Acids Res.

Dooner, H.K., and Nelson, O.E. (1977a). Genetic control of UDPglu- cose:flavonol3-Qglucosyltransferase in the endosperm of maize. Biochem. Genet. 15, 509-519.

Dooner, H.K., and Nelson, O.E., Jr. (1977b). Controlling element- induced alterations in UDPg1ucose:flavonoid glucosyltransferase, the enzyme specified by the bronze locus in maize. Proc. Natl. Acad. Sci. USA 74, 5623-5627.

319-328.

2073-2094.

18, 5133-5141.

Goodall, G.J., and Filipowicz, W. (1989). The AU-rich sequences present in the introns of plant nuclear pre-mRNAs are required for splicing. Cell 58, 473-483.

Goodall, G.J., and Filipowicz, W. (1991). Different effects of intron nucleotide composition and secondary structure on pre-mRNA splicing in monocot and dicot plants. EM60 J. 10, 2635-2644.

Hanley, B.A., and Schuler, M.A. (1988). Plant introns sequences: Evi- dencefor distinct groupsof introns. Nucl. Acids Res. 16,7159-7176.

Kim, H.-S., and Smithies, O. (1988). Recombinant fragment assay for gene targeting based on the polymerase chain reaction. Nucl. Acids Res. 16, 8887-8903.

Kim, H.Y., Schiefelbein, J.W., Raboy, V., Furtek, D.B., and Nelson, O.E., Jr. (1987). RNA splicing permits expression of a maize gene with a defective Suppressor-mutator transposable element insertion in an exon. Proc. Natl. Acad. Sci. USA 84, 5863-5867.

Masson, P., Surosky, R., Kingsbury, J.A., and Fedoroff, N.V. (1987). Genetic and molecular analysis of the Spm-dependent a-m2 alleles of the maize a locus. Genetics 117, 117-137.

Nelson, K.K., and Green, M.R. (1988). Splice site selection and ribo- nucleoprotein complex assembly during in vitro pre-mRNA splicing. Genes Dev. 2, 319-329.

Padgett, R.A., Konanka, M.M., Aebi, M., Hornig, H., Weissmann, C., and Sharp, P.A. (1985). Nonconsensus branch-site sequences in the in vitro splicing of transcripts of mutant rabbit P-globin genes. Proc. Natl. Acad. Sci. USA 82, 8349-8353.

Raboy, V., Kim, H.Y., Schiefelbein, J.W., and Nelson, O.E., Jr. (1989). Deletions in a dSpm insert in a maize Bronze-7 allele alter RNA processing and gene expression. Genetics 122, 695-703.

Ralston, E.J., English, J.J., and Dooner, H.K. (1988). Sequence of three bronze alleles of maize and correlation with the genetic fine structure. Genetics 119, 185-197.

Reed, R., and Maniatis, T. (1986). A role for exon sequences and splice- site proximity in splice-site selection. Cell 46, 681-690.

Roth, D.B., and Wilson, J.H. (1986). Nonhomologous recombination in mammalian cells: Role forshort sequence homologies in the join- ing reaction. MOI. Cell. Biol. 6, 4295-4304.

Schiefelbein, J.W., Raboy, V., Fedoroff, N.V., and Nelson, O.E., Jr. (1985a). Deletions within a defective Suppressor-mutator element in maize affect the frequency and developmental timing of its excision from the bronze Iocus. Proc. Natl. Acad. Sci. USA 82,4783-47W.

Schiefelbein, J.W., Furtek, D.B., Raboy, V., Banks, J.A., Fedoroff, N.V., and Nelson, O.E. (1985b). Exploiting transposable elements to study the expression of a maize gene. In Plant Genetics, M. Freeling, ed (New York: Alan R. Liss, Inc.), pp. 455-459.

Schiefelbein, J.W., Raboy, V., Kim, H.-Y., and Nelson, O.E. (1988a). Molecular characterization of suppressor-mutator (Spm)-induced mutations at the bronze-7 locus in maize: The b Z 4 " alleles. In Proc. lnternational Symposium on Plant Transposable Elements, O. Nelson, ed (New York: Plenum Press), pp. 261-278.

Schiefelbein, J.W., Jr., Furtek, D.B., Dooner, H.K., and Nelson, O.E., Jr. (1988b). Two mutations in a maize bronze-7 allele caused by transposable elements of the Ac-Ds family alter the quantity and quality of the gene product. Genetics 120, 767-777.

Shure, M., Wessler, S., and Fedoroff, N. (1983). Molecular identifi- cation and isolation of the Waxy locus in maize. Cell 35, 225-233.

Sullivan, T.D., Schiefelbein, J.W., Jr., and Nelson, O.E., Jr. (1989). Tissue-specific effects of maize Bronze gene promoter mutations induced by Ds7 insertion and excision. Dev. Genet. 10, 412-424.

Dow



1462 The Plant Cell

Weil, C.F., and Wessler, S.R. (1990). The effects of plant transposable element insartion on transcription initiation and RNA processing. Annu. Rev. Plant Physiol. Plant MOI. Biol. 41, 527-552.

Wessler, S., Tarpley, A., Purugganan, M., Spell, M., and Okagaki, R. (1990). Filler DNA is associated with spontaneous deletions in maize. Proc. Natl. Acad. Sci. USA 87, 673143735.

Wiebauer, K., Herrero, J.J., and Filipowicz, W. (1988). Nuclear pre- mRNA processing in plants: Distinct modes of 3'-splice site selection in plants and animals. MOI. Cell. Biol. 8, 2042-2051.

Zinn, K., DiMaio, D., and Maniatis, T. (1983). ldentification of two distinct regulatory regions adjacent to the human e-interferon gene. Cell 34, 865-879.

Dow



Documents

Alternative 3’ Splice Acceptor Sites Modulate Enzymic Activity in Derivative … · The Plant Cell, Vol. 4, 1453-1462, November 1992 O 1992 American Society of Plant Physiologists