Functional Mapping of Destabilizing Elements in the Protein-coding Region of the Drosophila fushi tarazumRNA*

The instability of the fushi tarazu(ftz) mRNA is essential for the proper development of the Drosophila embryo. Previously, we identified a 201-nucleotide instability element (FIE3) in the 3′ untranslated region (UTR) of the ftz mRNA. Here we report on the identification of two additional elements in the protein-coding region of the message: the 63-nucleotide-long FIE5-1 and the 69-nucleotide-long FIE5-2. The function of both elements was position-dependent; the same elements destabilized RNAs when present within the coding region but did not when embedded in the 3′ UTR of the hybrid mRNAs. We conclude that ftzmRNA has three redundant instability elements, two in the protein-coding region and one in the 3′ UTR. Although each instability element is sufficient to destabilize a heterologous mRNA, the destabilizing activity of the two 5′-elements depended on their position within the message.

Drosophila embryonic development depends on the precise temporal and spatial expression of maternal and zygotic pattern-forming genes (1). Maternal pattern-forming genes are transcribed during oogenesis, and their mRNA abundance decreases rapidly in the early embryo. Moreover, many mRNAs encoded by zygotic pattern-forming genes undergo dramatic changes in abundance and spatial distribution during early embryogenesis. To achieve these rapid changes, especially for rapid down-regulation, transcriptional control alone is insufficient, and regulation at the level of mRNA stability is essential. For instance, the maternal bicoid mRNA is completely stable during the first 2 h of embryogenesis but is rapidly destabilized at cellularization of the blastoderm (2). As discussed in the following text, the zygotic fushi tarazu (ftz) mRNA is one of the most unstable eukaryotic mRNAs known. Given that most mRNAs in the Drosophila embryo are constitutively stable (30), the question arises how selected mRNAs in the same embryo cytoplasm are targeted for degradation. Recognition of the targeted RNAs by the RNA degrading machinery must involve cis-acting sequences. These sequences are the focus of the present study.
ftz is a member of the pair-rule class of segmentation genes and one of the best characterized early zygotic genes. In early embryos ftz mRNA is detected only from about 1.5 to 4.5 h after fertilization. When first expressed, ftz mRNA is uniformly distributed through the embryo (4). As development progresses, its distribution first becomes restricted to a region comprising from 15 to 65% of egg length, then to four broad bands, and finally to seven narrow stripes that encircle the embryo (4 -6). The seven stripes are short-lived, and no ftz mRNA is detected by 5 h after fertilization. This rapid change of expression pattern and formation of stripes in a short time span can be attributed to the termination of transcription in interstripe regions coupled with rapid mRNA turnover. The need for rapid mRNA turnover is emphasized by the fact that the FTZ protein activates its own transcription in a positive feedback loop (7). Thus, it is important during the evolution of the spatial pattern of ftz expression that both mRNA and protein be rapidly cleared from the interstripe regions. Edgar et al. (8) measured ftz mRNA stability in embryos and found that its half-life changes from 14 min at 2.5 h to 6 min at 4 h after fertilization. This makes ftz mRNA one of the shortest-lived mRNAs among higher eukaryotes (3). Stabilization of ftz mRNA and FTZ protein results in developmental delay and defects, suggesting that ftz mRNA and FTZ protein instability are crucial for normal development (9 -11).
Earlier studies from this laboratory using hybrid genes that fuse ftz sequences to the stable ribosomal protein A1 (rpA1) mRNA provided evidence for at least two destabilizing elements in ftz mRNA (12). One consisted of a 201-nucleotide element, including an essential 68-nucleotide sequence, located in the 3Ј UTR 1 and termed FIE3 (ftz instability element 3Ј). The other element(s) were assigned to the 5Ј-one-third of the ftz mRNA but remained otherwise uncharacterized. Here we report on the identification of two separate 5Ј-instability elements within the first 600 nucleotides of the ftz mRNA. Both 5Ј-elements are located within the protein-coding region of the message. Each 5Ј-instability element is sufficient to destabilize the stable rpA1 mRNA, but the destabilizing activity is dependent on their position within the mRNA.

Construction of Transgenes
The first letter of the name of each construct designates the promoter that drives it. Thus, F ϭ ftz, r ϭ rpA1, and S ϭ sgs-3 promoter attached to the hsp26 nurse cell enhancer.

Discrepancies in the ftz Nucleotide Sequence
Numbering in the following text uses ϩ1 as the position of transcription initiation, which corresponds to position 901 in the sequence deposited in GenBank TM (accession numbers X00854 and K01951). We found that ftz genomic clone A439 (13) had a nine-nucleotide deletion in the protein-coding region (from 260 to 268) when compared with the sequence deposited in the data base. To resolve this discrepancy, the ftz 5Ј-region was amplified by PCR using DNA from yellow-white flies, the recipient strain for P-element transformation. Sequencing revealed that the same nine nucleotides were missing. Moreover, with the exception of the nine-base deletion and four one-base polymorphisms (C *This work was supported by Grant IBN-9630369 from the National Science Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
instead of T at 126 and 222, C instead of A at 516, and G instead of C at 551), the sequence of the entire 5Ј-one-third of the ftz mRNA (from 1 to 636) matched the sequence deposited in GenBank TM .

FftzϩaldB and FftzϩaldB-FIE3 (see Fig. 1B)
The 0.7-kb fragment (from 856 to 1557, where ϩ1 is the transcription initiation site) of rat aldolase B (aldB) cDNA (15) was synthesized by PCR with primers ald1 and ald2 (Table I). This PCR fragment contained the 3Ј-one-third of the protein-coding region and entire 3Ј UTR of the aldB cDNA, including the polyadenylation signal (16), provided by the primer ald2. The PCR fragment was cloned into the pGEM-T Easy vector (Promega), digested with EcoRI, and inserted into the EcoRI site of the 3Ј UTR of f3. In turn, this f3ϩaldB fragment was fused to Ff5 to produce FftzϩaldB. A f3-FIE3 fragment that lacks the 201-base pair FIE3 sequence was obtained as previously described (12). FftzϩaldB-FIE3 was constructed as described for FftzϩaldB, except that f3-FIE3 was used instead of f3.

Deletion Constructs
Rr Constructs (see Fig. 1C)-The PCR products synthesized from Ff5r3 with primers 1 and 2, 1 and 4, and 3 and 2 (Table I) were digested with NgoMI and inserted into the NgoMI site in the 3Ј UTR of the rpA1 gene in plasmid pD5. They were named Rr-abc, Rr-ab, and Rr-bc, respectively.

Synchronized Embryo Collections
Embryos were collected on fresh yeast plates at 25°C for 1 h and left at 25°C to age for different lengths of time. Embryos were collected, washed with embryo washing buffer (0.5% Triton X-100, 300 mM NaCl, 10 mM Tris, pH 7.5), frozen in liquid nitrogen, and kept at Ϫ80°C. Embryo developmental synchrony was checked by staging an aliquot of each collection at 2-3 h of development. Collections containing more than 10% older embryos (retained) or unfertilized eggs were discarded. RNA Extraction and Northern Blot Hybridization-Total RNA was isolated from frozen embryos by homogenization in TRI REAGENT (Molecular Research Center, Inc.) following the manufacturer's recommendations. Northern blot analysis was performed as previously described (22). Briefly, 10 g of total RNA was fractionated by electrophoresis in 1-2% agarose gels containing 18% formaldehyde, and RNA was transferred to a Hybond-Nϩ membrane (Amersham Pharmacia Biotech) (23). The RNAs were fixed to the membrane by UV crosslinking and hybridized with ␣-32 P-labeled probes synthesized by random primer labeling. For hybridization with a different probe, the membrane was boiled 5 min to remove the first probe, following the manufacturer's recommendations. The plasmid pD5, containing the rpA1 gene, Ff5, and f3 were used to prepare probes for Northern blot hybridization. Estimates of the half-life of mRNAs were obtained from the quantification of the radioactive signal with the Molecular Imager system (Bio-Rad). For each transgenic strain, Northern blot hybridization was repeated at least three times with RNA from separate embryo collections.

Rationale-
The stability of transgenic mRNAs in developing embryos was measured by a strategy previously developed for this purpose (12). The critical feature is the use of promoters that are strongly expressed during oogenesis but silent in early embryogenesis. The two promoters used in this study are rpA1 (14) and hsp26/sgs3 (24,25). The latter promoter contains the nurse cell-specific enhancer from the hsp26 gene linked to the basal sgs3 promoter. Another key feature is that the ftz mRNA is stable in ovaries and is destabilized at fertilization. 2 As a consequence, transcripts containing ftz sequences accumulate to high abundance during oogenesis and start decaying when the egg is fertilized. 2 Because the rpA1 and hsp26/sgs3 promot-  ers are silent in early embryos, decay of transgenic mRNA abundance (as measured by Northern blot analysis) serves as a direct measure of mRNA stability. This strategy avoids the use of drugs that may cause artifacts (27). The rpA1 mRNA is stable in both ovaries and embryos and served as an internal loading control.
The ftz mRNA Contains Multiple RNA-destabilizing Sequences-In an earlier study (12), the stability of the 5Ј-onethird (here called f5; nucleotides 1-636) and the 3Ј-two-thirds (here called f3) of the ftz mRNA was investigated separately. One of these fragments (f5) was fused to the 3Ј-two-thirds (r3) and the other (f3) to the 5Ј-one-third (r5) of the stable rpA1 mRNA to yield the Ff5r3 and Rr5f3 hybrid mRNAs, respectively (Fig. 1A). Both hybrid transcripts were unstable ( Fig.  2A). Rr5f3 is transcribed maternally, and its mRNA decayed rapidly after fertilization. Ff5r3 is transcribed only transiently from the ftz promoter during early embryogenesis, and its rapid decay after 4 h of development (compare 3-4 h and 4 -5 h in Fig. 2A) indicates that this mRNA is highly unstable. We concluded that in addition to the previously characterized 201nucleotide FIE3 in the 3Ј UTR, the 5Ј-one-third of the ftz mRNA also contains destabilizing sequences.
The Fftzϩald-FIE3 construct provided further evidence for the presence of RNA-destabilizing sequences in f5. This construct consisted of a deletion of FIE3 from the intact ftz mRNA (Fig. 1B). This construct also contained a rat aldolase B mRNA (aldB) tag, to allow the transgenic mRNA to be distinguished on Northern blots from the endogenous ftz transcript. To verify that insertion of the aldolase tag does not affect ftz mRNA stability, a control construct (Fftzϩald; Fig. 1B) containing the intact ftz mRNA tagged with the aldB fragment was analyzed in parallel. As shown in Fig. 2B, both constructs were unstable in Drosophila embryos, indicating that ftz mRNA contains instability elements other than FIE3. The destabilizing element of f3 is FIE3 (12). The identification and characterization of the f5-destabilizing sequences (termed FIE5) was the object of this study.
The Destabilizing Activity of FIE5 Is Position-dependent-Previously, we have inserted FIE3 into the 3Ј UTR of the stable rpA1 gene to demonstrate that this element is sufficient for mRNA destabilization (12). We created similar constructs to investigate whether FIE5 is also sufficient for RNA destabilization. The entire f5 sequence (1-636), the 5Ј-two-thirds (1-423), and the 3Ј-two-thirds (211-636) were inserted into the 3Ј UTR of rpA1 to yield constructs Rr-abc, Rr-ab, and Rr-bc, respectively (Fig. 1C). Transcription in all constructs was driven by the rpA1 promoter. Surprisingly, the resulting three transcripts were stable in transgenic embryos (Fig. 3). These results suggest that FIE5 destabilizing activity is position-dependent and that unlike FIE3, FIE5 is not functional when located in the 3Ј UTR.
Initial Mapping of the 5Ј-Destabilizing Sequences-Next, we initiated a deletion analysis by placing the f5 fragment at its original position 5Ј of the rpA1 sequences. Construct Ss-abc contained the whole f5 sequence (from 1 to 636), Ss-ab contained the 5Ј-two-thirds (from 1 to 423), and Ss-bc contained the 3Ј-two-thirds (from 217 to 636) (Fig. 1D). All transcripts decayed rapidly during early embryogenesis (Fig. 4). These results suggest that FIE5 destabilizing sequences are likely to be present in the Ss-ab/Ss-bc overlap region. Alternatively, more than one destabilizing sequence may occur within f5. The results also indicate that the maternal hsp26/sgs3 promoter can be used for measurement of mRNA stability despite the presence of 41 additional sgs3-encoded nucleotides at the 5Јend of all transcripts driven by this promoter. Separate experiments showed that these 41 nucleotides contain no destabiliz-ing sequences (cf. constructs Ss-a, Ss-b2, Ss-c3, and Ss-c5).
Deletion Analysis Identifies Two Separate Destabilizing Elements in the 5Ј-Coding Region-A second generation of constructs placed each third of f5 next to 5Ј-rpA1 sequences (Ss-a, Ss-b, and Ss-c; Fig. 1D). As shown in Fig. 5, s-a mRNA was stable, whereas s-b and s-c mRNAs decayed rapidly during early embryogenesis. These results suggest that f5 has at least FIG. 1. Summary of constructs used to identify ftz mRNA instability elements. The first letter in the construct name identifies the promoter: R for rpA1, F for ftz, and S for hsp26/sgs3 (cf. "Experimental Procedures"). mRNA stability in transgenic embryos is indicated to the right. A, stability of reciprocal hybrid mRNAs containing ftz and rpA1 sequences. Rr5f3 has the 5Ј-one-third of the rpA1 gene, including the promoter region, and the 3Ј-two-thirds of the ftz gene, including the 3Ј-flanking region. Ff5r3 has the 5Ј-one-third of the ftz gene, including the promoter region, and the 3Ј-two-thirds of the rpA1 gene. B, deletion of the ftz 3Ј-instability element (FIE3). Both constructs are driven by the ftz promoter and have a rat aldolase B (aldB) tag to allow the transcripts to be distinguished from the endogenous ftz mRNA. Note that the second construct lacks FIE3. C, overlapping deletions of the 5Ј-one-third of the ftz mRNA inserted into the rpA1 3Ј UTR. The first construct, with an FIE3 insert, summarizes the results of a previous study (12). The next three constructs contain the first 636 nucleotides of the ftz mRNA (or deletions thereof) inserted into the same position of the rpA1 3Ј UTR. D, overlapping deletions of the 5Ј-one-third of the ftz mRNA, driven by the hsp26/sgs3 promoter. The promoter contains the nurse cell-specific hsp26 enhancer upstream of the sgs3 basic promoter. Each transcript contains 41 nucleotides of sgs3 followed by 5Ј-ftz sequences fused in frame to 3Ј-rpA1 mRNA sequences. For constructs lacking the ftz translation initiation site, a six-base pair sequence containing an ATG was added to confer similar translational efficiency (cf. "Experimental Procedures"). Horizontal lines, untranslated regions; V, intron; open boxes, ftz protein-coding regions; closed boxes, rpA1 protein-coding regions; stippled boxes, aldB protein-coding region; boxes with vertical lines, rpA1 promoter region; boxes with partial black shading, ftz promoter region; boxes with diagonal lines, sgs3 promoter region. Promoter regions are not drawn to scale. two destabilizing elements, one (FIE5-1) within fragment b (from 217 to 423) and the other (FIE5-2) within fragment c (from 406 to 636). Each element is sufficient for destabilization of an otherwise stable mRNA, and both elements are located in the protein-coding region of ftz mRNA.
Further Mapping of FIE5-1-Additional deletion constructs (Ss-b1 to Ss-b3; Fig. 1D) were generated to further map FIE5-1. Initial analysis of embryos carrying two overlapping constructs, Ss-b1 and Ss-b2, indicated that the s-b1 mRNA was unstable whereas s-b2 mRNA was stable (Fig. 6). These results suggested that FIE5-1 is located in a region of Ss-b1 (250 -300) that does not overlap with the two stable sequences, SS-a and SS-b2 (Fig. 1D). This assumption was confirmed with a third construct, Ss-b3 (235-306), which encodes an unstable mRNA (Fig. 6). Thus, FIE5-1 is located within a 63-nucleotide sequence of the ftz protein-coding region (the stated length takes into account that our construct and yellow-white flies have nine fewer nucleotides than the sequence deposited in the data base; cf. "Experimental Procedures").
Further Mapping of FIE5-2-Additional deletion constructs (Ss-c1 to Ss-c5; Fig. 1D) were also generated to further map FIE5-2. As shown in Fig. 7, the s-c1 and s-c2 mRNAs were both unstable in early embryos, indicating that an instability element is located in the region of overlap or that multiple insta-bility elements are located in ftz fragment c. To clarify these issues, three more constructs (Ss-c3 to Ss-c5; Fig. 1D) were analyzed. Of the three constructs, only Ss-c4 (from 520 to 588) encoded an unstable mRNA (Fig. 7), indicating that FIE5-2 FIG. 2. Evidence for destabilizing elements in the 5-region of the ftz mRNA. Each construct was transformed into the Drosophila germ line. Total RNAs from synchronized embryos were analyzed by Northern blot hybridization with ftz and rpA1 probes. The structure of each construct is illustrated next to its name. Embryo age ranges (in h) are indicated below the autoradiograms. The rpA1 mRNA served as a loading control. A, hybrid ftz/rpA1 mRNAs. The r5f3 construct is driven by the rpA1 promoter, and the f5r3 construct is driven by the ftz promoter. mRNA sizes are as follows: ftz, 1.8 kb; r5f3, 1.4 kb; f5r3, 0.6 kb; rpA1, 0.6 kb. B, deletion of the FIE3 sequences from the ftz mRNA. A 700-nucleotide rat aldolase B fragment was inserted in the ftz 3Ј UTR to distinguish the transgenic mRNA from the endogenous ftz mRNA. The second construct differs from the first by the absence of the 201nucleotide FIE3 element. Both constructs are driven by the ftz promoter. mRNA sizes: ftzϩaldB, 2.5 kb; and ftzϩaldB-FIE3, 2.3 kb.

FIG. 3. First generation deletion constructs for the identification of ftz 5-instability elements (FIE5s).
The entire 5Ј-one-third of the ftz mRNA (r-abc, 636 nucleotides) or deletions thereof (r-ab and r-bc) were inserted into the 3Ј UTR of the rpA1 mRNA. All constructs are driven by the rpA1 promoter. mRNA sizes are as follows: r-abc, 1.2 kb; r-ab, 1.0 kb; r-bc, 1.0 kb; and rpA1, 0.6 kb. Additional information can be found in the legend to Fig. 2.   FIG. 4. Second generation deletion constructs for the identification of FIE5s. s-abc, s-ab, and s-bc mRNAs are similar to those described in the legend to Fig. 3 except that 5Ј-ftz sequences were fused in frame with 3Ј-rpA1 sequences. A translation initiation site was added at the beginning of the translated sequence of the s-bc mRNA. All constructs are driven by the hsp26/sgs3 nurse cell-specific promoter. mRNA sizes are as follows: s-abc, 1.1 kb; s-ab, 0.9 kb; s-bc, 0.9 kb; s-abcϩG, 1.1 kb; and rpA1, 0.6 kb. Additional information can be found in the legend to Fig. 2. must reside within this 69-nucleotide region. In some experiments, the s-c3 mRNA was observed to decline slightly from 0 -1 to 1-2 h but remained stable during the remaining time points (data not shown).

DISCUSSION
In an earlier study (12), we developed a new method for the in vivo analysis of mRNA stability in early Drosophila embryos that does not require the use of drugs or any experimental interference. This method led to the identification of a mRNA-destabilizing element (FIE3) in the ftz 3Ј UTR. In this work the method was used to identify two additional instability elements, both in the 5Ј-protein-coding region. Hence, ftz mRNA has three redundant destabilizing elements, each of which is sufficient to promote mRNA degradation in early embryos. Although the significance of the occurrence of three redundant elements in the same message can only be speculated on, redundancy may be tied to the fact that ftz mRNA instability is crucial for normal embryonic development (see the Introduction).
The half-life of a hybrid mRNA containing FIE3 was previously estimated to be about 50 min (12). In this study we estimated the half-life of the FIE5-1-containing s-b mRNA and the FIE5-2-containing s-c mRNA to be about 51 and 65 min, respectively. 3 Thus, each element appears to have similar "strength." Each element can act independently, and we found no evidence that the destabilizing activity of these elements is additive or synergistic. The estimated half-lives cited above are significantly longer than the 14 -6-min half-life reported previously for the endogenous ftz mRNA (8). One reason for this difference may be that our measurements started early during embryonic development (from fertilization to 4 h), whereas endogenous ftz transcription occurs between 1.5 and ϳ4.5 h. The gradual decrease of ftz mRNA half-life from 14 to 6 min between 2.5 and 4 h of development (8) suggests that degrading activity increases as embryonic development progresses. Thus, the degrading activity may be low at the very beginning of embryogenesis. Another conceivable reason for the difference in estimated half-lives might be the cytoplasmic localization of the mRNAs (28,29). Endogenous ftz mRNA is located in the 3 J. Ito and M. Jacobs-Lorena, unpublished observations. FIG. 5. Third generation deletion mutants for the identification of FIE5s. The s-a, s-b, and s-c constructs contain the 5Ј-one-third, the middle-one-third, and the 3Ј-one-third of f5, respectively. A translation initiation site was added to all constructs. All constructs are driven by the hsp26/sgs3 nurse cell-specific promoter. mRNA sizes are as follows: s-a, 0.7 kb; s-b, 0.7 kb; s-c, 0.7 kb; rpA1, 0.6 kb. Additional information can be found in the legend to Fig. 2. FIG. 6. Fourth generation deletion mutants for the mapping of FIE5-1. A translation initiation site was added to all constructs. All constructs are driven by the hsp26/sgs3 nurse cell-specific promoter. mRNA sizes are as follows: s-b1, 0.6 kb; s-b2, 0.6 kb; s-b3, 0.5 kb; rpA1, 0.6 kb. Additional information can be found in the legend to Fig. 2.   FIG. 7. Fourth generation deletion mutants for the mapping of FIE5-2. A translation initiation site was added to all constructs. All constructs are driven by the hsp26/sgs3 nurse cell-specific promoter. mRNA sizes are as follows: s-c1, 0.6 kb; s-c2, 0.6 kb; s-c3, 0.6 kb; s-c4, 0.5 kb; s-c5, 0.5 kb; rpA1, 0.6 kb. Additional information can be found in the legend to Fig. 2. apical cytoplasm, whereas the distribution of the hybrid mRNAs in the transgenic embryos is unknown (30). Apical localization requires the last 53 nucleotides of the ftz 3Ј UTR (31,32), 43 nucleotides of which overlap with FIE3. Deletion of FIE3 (which comprises most of the apical localization sequence) in ftzϩaldB-FIE3 did not seem to substantially affect stability when compared with the FIE3-containing ftzϩaldB mRNA, suggesting that mRNA stability and localization are independent of each other.
The r-abc and s-abc mRNAs both contain the identical 5Ј-ftz sequence; yet their stability in the early embryo differs dramatically. The main difference between the two mRNAs is the position of the ftz sequences within the mRNA: in the 3Ј UTR for r-abc and in the original 5Ј-position for s-abc. Therefore, the structure and sequence of the RNA elements are not sufficient for destabilizing activity, and position within the mRNA is crucial. Note that when FIE3 was inserted at the same position in the rpA1 3Ј UTR as were the FIE5s in r-abc, FIE3 had full destabilizing activity. Thus, FIE3 is active, and FIE5s are inactive when inserted at the identical position of the rpA1 mRNA. These results suggest that FIE5s and FIE3 destabilize mRNAs by different mechanisms. This position dependence of the FIE5 elements suggests that translation is required for degradation to occur. Precedents exist for a connection between mRNA stability and translation (30,(33)(34)(35)(36). The suggested dependence of FIE5 activity on mRNA translation is consistent with the results of Edgar et al. (8), who reported that general inhibition of embryonic protein synthesis by cycloheximide injection stabilizes the ftz mRNA. However, these results do not rule out the alternative possibility that cycloheximide prevents the synthesis of an unstable protein required for mRNA degradation.
The three cis-acting ftz instability elements are likely to act by providing a binding site for a factor or a protein complex that mediates mRNA degradation. A sequence comparison among the three elements and a search for similarity with sequences deposited in data bases did not yield any significant homologies. Binding sites could be recognized as secondary structures rather than primary sequences (37). However, no stable secondary structure that has more than a four-base straight stem or a common secondary structure among the three elements was predicted when a computer program of Zuker (26,38) was used to fold these sequences. Moreover, site-directed mutagenesis of certain nucleotides within FIE3 did not alter the destabilizing activity of this element. 2 Thus, it is presently unclear how these cis-acting elements are recognized in the embryo. Recently, a protein that binds to the ftz apical localization element was identified (32). The identification of proteins that recognize the three instability elements characterized in this and in previous work will bring significant insights to the questions of sequence recognition and mechanism of RNA degradation.