Messenger RNA Sequence Complexity and Homology 1n Developmental Stages of Drosophila

Total polysomal RNA and polyadenylated mRNA from third instar larvae, pupae, and adults of D. melanogaster were hybridized in vast excess to labeled single-copy DNA in order to measure the sequence complexity of each RNA popula­ tion. Then, to measure the sequence homology between the populations, each was hybridized to DNA enriched for messen­ ger coding sequences in third instar larvae and to DNA depleted of these sequences. Our results show that a similar num­ ber of genes, approximately 16,000, is expressed in larvae, pupae, and adults, and that only one-third of these is expressed as polyadenylated mRNA. Further, the composition of both polyadenylated and nonpolyadenylated mRNA classes is shown to change very little between these three stages of development. Finally, the head of adult Drosophila is shown to contain 11,000 RNA species, approximately 70% of the number contained in the entire adult.


INTRODUCTION
The marked morphological and physiological changes which occur during animal development are considered to be accompanied by or to result from changes in gene expression. Indeed, different sets of genes are expressed during development and in different adult tissues of the sea urchin (Galauet al., 1974;. In Drosophila, Levy and McCarthy (1975) used cDNA hybridization kinetics to demonstrate a high degree of sequence homology among the polyadenylated RN As most abundant in lar vae, adults, and Schneider's cells. However, a proportion of the polyadenylated RNA sequences from third instar larvae were shown to be poorly represented or absent in the RN A populations from adults and Schneider's cells. These data are consistent with some measure of differen tial gene expression, especially among rare-class RN As, but the technique employed does not allow quantitation of such changes. More recently, Izquierdo and Bishop (1979) employed the same techniques to demonstrate > 90% sequence homology between polyadenylated RN A molecules in the cytoplasm of L3 cultured cells and those in larvae, pupae, and adults. Eighty-five to one hundred percent sequence homology was observed when the poly adenylated RN A population from embryo was compared to that from L3 cells and third larval instar. The mea surements of sequence homology among polyadenylated RN A populations reported in both of these studies repre sent minimum estimates, their accuracy limited by the technique employed. Specifically, it was not possible in either study to determine the number of different RN A species represented in a population of molecules shown 141 by the assay to be nonhomologous. Further, both studies investigated the sequence homology only between popu lations of polyadenylated RNA molecules. Recently, Zimmerman et al. (1980) have demonstrated the pres ence of a very complex population of RN A molecules as sociated with polysomes which lacks 3'-polyadenylation. Approximately two-thirds of the single-copy genes ex pressed in third instar larvae of D. melanogaster were shown to be expressed as nonadenylated RN A mole cules, none of which were included in the measurements of sequence homology described above.
In the study reported here, we have compared the set of single-copy sequence transcripts, adenylated and non adenylated, present on polysomes of third instar larvae of D. melanogaster with those present on polysomes of pupae and adults. We have approached this problem first by measuring the number of diverse sequences present in each RNA population using RNA-excess saturation hybridization (Galau et al., 1974). Similar experiments were then carried out using a population of DNA mole cules greatly enriched for sequences expressed on poly somes of third instar larvae to determine what propor tion of this DNA is expressed on polysomes of pupae and adults. DNA greatly depleted of sequences expressed on polysomes of third instar larvae was also obtained to de termine what proportion of these sequences is expressed in the later stages. In addition, we examined the RNA sequence diversity on polysomes in the adult head, an en richment for neural tissue.
We demonstrate here that a similar number of genes is expressed on polysomes of third instar larvae, pupae, and adults, and that approximately one-third of these is expressed as polyadenylated RNA. Furthermore, we observed a high degree of sequence homology between polyadenylated mRNAs of larvae, pupae, and adults, and between the nonadenylated RNAs of these stages. A small qualitative difference was detected between the polyadenylated mRNA populations, representing approximately 10% of the total number of genes expressed in pupae and adults.

Rearing of Organisms and Collection of Timed Stages
Drosophila melanogaster, strain Oregon R, was maintained in population cages at room temperature, 60% humidity. Eggs were collected for 2 hr on a cornmeal agar food source spread with yeast paste and were incubated at 25°C. Late third instar larvae were collected on Day 5 (120-125 hr). Pupae were collected synchronously 5-7 hr after puparium formation by flotation (Mitchell and Lipps,19'78). Adults were collected within the first day after eclosion, and adult heads were prepared as described (Schmidt-Nielsen et al., 1977).

Preparation of RNA and DNA
The experimental methods used for the preparation of larval polysomal RNA, poly(A+)RNA, sizing of RNA by electrophoresis in agarose gels in the presence of methyl mercury hydroxide and isolation and labeling of single-copy DNA are as previously described (Zimmerman et al., 1980). For preparation of RNA from isolated polyribosomes, a standard SDS-phenol/chloroform procedure was employed (Galau et al., 1976).
Preparation of mDNA, Null mDNA, (A+)mDNA, Null (A +)mDNA 3H single-copy DNA was fractionated into subsets as illustrated in Fig. 1. As shown, labeled single-copy DNA (-3 pg of which -10% encodes mRNA) was hybridized to termination (Rot 30,000) with a 50-fold sequence excess of total polysomal RNA (7 mg) from third instar larvae. RNA sequence excess calculations assume that messenger RNA comprises 2% of the mass of total polysomal RNA and that complex mRNA occupies only 10% of the total mRNA mass (Zimmerman et al., 1980). Thus, even rare mRNA species are in 50-fold excess over the coding DNA. The RNA-DNA mixture was then divided into two fractions. Ninety percent of the hybridization mixture was digested with nuclease S, (1.5 x lo5 units in 5 ml for 30 min at 37°C) and undigested material collected by Sephadex G-100 chromatography. RNA was then removed by hydrolysis with 0.1 N NaOH at 60°C for 1 hr. The remaining single-stranded DNA was further enriched for coding sequences by a second round of hybridization. The DNA prepared in this manner represents mDNA, a population greatly enriched for messenger RNA coding sequences in third instar larvae. The remaining 10% of the original hybridization mixture was centrifuged twice to equilibrium in a neutral CsCl buoyant density gradient, and after each centrifugation the band of unhybridized DNA was removed. The band was comprised of single-stranded noncoding DNA (94.5%) and DNA-DNA duplex (4.5%). After the second centrifugation the banded DNA was treated with 0.1 N NaOH at 60°C for 1 hr and was neutralized with 2 M Hepes, pH 4. The DNA prepared in this manner represents null mDNA, a population greatly depleted of messenger RNA coding sequences in third instar larvae. (A+)mDNA and null(A+)mDNA were prepared analogously by hybridizing labeled single-copy DNA (--0.8 pg) with a 40-fold sequence excess of polyadenylated mRNA (-10 pg) to termination (Rot 1300). Three rounds of hybridization were required to provide an A+mDNA population sufficiently enriched for coding sequences (see Results).
Hybridization of RNA with [3H]DNA and Assay of Hybridization 3H-Labeled single-copy DNA, mDNA, null mDNA, (A+)mDNA and null(A+)mDNA were incubated with a 2 loo-fold sequence excess of RNA in 0.75 M NaCl, 0.003 M Pipes, pH 6.8, at 65°C in sealed capillary pipets. Yeast tRNA (10 pg) was added to samples containing poly(A+) polysomal RNA. Hybridization samples were incubated from 1 hr to 4 days, and the R$ value achieved was calculated based on the concentration of D. melanogaster RNA in each sample. The contribution of DNA-DNA self-reassociation to the total amount of hybridization was determined by conducting parallel reactions in which an identical amount of yeast RNA was substituted for Drosophila total polysomal or poly(A+)RNA.
Samples were analyzed for hybrid content as described by Zimmerman et al. (1980) using TCA precipitation after S, nuclease treatment. The percentage DNA-RNA hybridization was calculated by subtracting the percentage S, nuclease resistance of the samples containing heterologous RNA (i.e., resistance due to DNA-DNA reassociation and to the inherent resistance of the nonhybridized tracer) from the percent S, nuclease resistance of samples containing D. melanogaster RNA. The amount of S, nuclease-resistant material in control samples containing heterologous RNA was consistently 2-3% of the input DNA.

RESULTS
Current estimates of the number of diverse mRNA species present in larvae, pupae, and adults of D.  nogaster are based largely upon analysis of cDNA hybridization kinetics (Levy and McCarthy, 1975;Izquierdo and Bishop, 1979). The results of these studies indicate that in these developmental stages there are between 4900 and 7000 different mRNA molecules. These estimates are derived solely from measurements of poly(A+)RNA, since poly(A-)RNA is refractory to analysis by the cDNA techniques employed. In view of the recent observation (Zimmerman et al., 1980) that much of the complexity of total polysomal RNA in third instar larvae is not adenylated, it was desirable to reexamine the sequence complexity of total polysomal mRNA in both pupae and adults. In these studies, the RNA of interest was hybridized in sequence excess to 3H singlecopy DNA (see Materials and Methods), and the amount of hybridization observed at saturation was used as a measure of the sequence complexity of the RNA.
The measurements of sequence complexity of total polysomal RNAs are shown in Fig. 2 and are summarized in Table 1. At saturation (R,,t > 25,000), 10.6% of the single-copy DNA hybridized to polysomal RNA of larvae, a value in excellent agreement with that previously reported by Zimmerman et al. (1980). Interestingly, the saturation values for polysomal RNA from pupae and adult (10.3 and ll.O%, respectively) were approximately those observed for larval polysomal RNA. These percentages correspond to RNA sequence complexities of 1.9 x lo7 to 2.0 x 10' nucleotides and represent between 15,000 and 16,000 diverse RNA molecules of average size 1250 nucleotides. A comparison of these measurements with the sequence complexity of poly(A+)mRNA from both pupae and adult (see Table 1; Fig. 3) (Levy and McCarthy, 1975;Izquierdo and Bishop, 1979)  Trace quantities of 3H single-copy DNA (specific activity -6 x 10s cpm/pg) were hybridized with a 2 RIO-fold sequence excess of total polysomal RNA from third instar larvae (A), 5 to 7 hr pupae (B) or adults (C). Hybridization was monitored by resistance to S, nuclease digestion. Data points have been corrected for the contribution of tracer self-reassociation and for tracer reactability (-90% when reassociated with an excess of total D. melanogaster nuclear DNA). nonadenylated RNA to the total complexity of polysomal RNA in both pupae and adult is similar to that previously observed for larvae, and is approximately 60%.

Sequence Overlap of Total Polysomal RNAs from Different Stages of Development
The measurements of sequence complexity described above indicated that a similar proportion of single copy DNA, 10.3-ll.O%, is represented in the polysomal RNA of larvae, pupae, and adults of D. melanogaster. We next wished to determine what proportion of these RNA sequences is shared among the stages, and what proportion is unique to a particular stage. To evaluate this, a population of DNA enriched for single-copy sequences ex-pressed on polysomes of third instar larvae was prepared (mDNA), as was a population of DNA depleted of these sequences (null mDNA). r3H]mDNA hybridized with polysomal RNA from larvae to a maximum saturation value of 60%, representing a sixfold enrichment for mRNA coding sequences (Fig. 4A). In order to determine the amount of overlap between the mRNA sequences present in larvae and those found on polysomes in pupae and adult, polysomal RNA from pupae and adult was hybridized in sequence excess (-150 pg) to trace quantities (-5 x 10m4 pg) of the [3H]mDNA. The amount of [3H]mDNA which hybridized with the RNAs at saturation may be directly compared with the 60% value observed for the self-reaction (i.e., larval polyso-ma1 RNA x [3H]mDNA) to provide a measurement of the amount of sequence overlap. The proportion of the mDNA, 40%, which does not hybridize to larval polyso-ma1 RNA is assumed to represent random contamination with the noncoding portion of single-copy DNA. Since at most 11% of this random contamination could represent coding sequences for pupal or adult polysomal RNA, a maximum value of 4.4% (0.11 x 0.4 x 100%) of the hybridization observed between the mDNA preparation and polysomal RNA from pupae or adults could be ascribed to this DNA population. The results of this experiment are presented in Fig. 4 and summarized in Table 2.
At saturation, 63 and 59% of the [3H]mDNA hybridized with polysomal RNA from pupae and adults, respectively (Figs. 4B and C). Therefore, within the limitations of the measurements, it is clear that these RNA populations contain most if not all of the diverse RNA sequences present on polysomes of third instar larvae. Likewise, since the sequence complexity of the pupae and adult polysomal RNAs are approximately equivalent (10.3 + 0.8 and 11.0 + 0.8%, respectively), the vast majority of diverse mRNA sequences present in pupae and adults must also be held in common between these two developmental stages.
The results of the above experiments suggest that among the three developmental stages studied, only a small number of mRNA sequences may be unique to any one of the individual stages. To determine more accurately what proportion of the mRNA sequences present in pupae and adult are not present in larvae, trace quantities of 3H-null mDNA were hybridized with a sequence excess of polysomal RNA from either larvae, pupae, or adult. The third instar larval polysomal RNA reacted with the null mDNA preparation to a saturation value of only 2.0% as compared to a saturation value of 10.6% for the reaction of this RNA with the starting DNA population. Polysomal RNA preparations from pupae and adults drove 3.0 and 3.2%, respectively, of null mDNA into hybrid form, demonstrating the presence of a small .6 x lo6 6.6 x 10-S 6.9 x 1O-2 9.6 x 1O-2 6,100 a Terminal hybridization values described by a least-squares computer solution of the data shown in Figs. 2, 3, and 6A. * Complexity = saturation value X 2 (assuming assymmetric transcription) X (9.1 X lo'), where 9.1 x 10' nucleotide pairs is the complexity of single-copy DNA from D. melanoguster (Manning et al., 1975).
c Pseudo-first-order rate constant predicted from an RNA population of known complexity. The predicted rate is calculated from the relationship K 5374 x 200 em = RNA complexity x where 5374 is the complexity of 0X174 RNA and 200 M-l set-1 is the pseudo-first-order rate constant for an RNA excess hybridization between 0X174 RNA and 300 nucleotide driver RF DNA (Galauet al., 1977). L is the mass average length of the Drosophila driver RNA (1250 nucleotides). It should be noted that the correction for length of Drosophila driver RNA is based on results derived from studies on DNA-DNA reassociation. Similar studies on RNA-DNA reassociation have not been reported.
d Fraction of the RNA mass which is driving the reaction as calculated from the ratio of Kobs to K,,, (Galau et al., 1974;Hough et al., 1975). e Data for the RNA sequence complexity of polyadenylated mRNA from larvae are taken from Zimmerman et al. (1980). number of gene sequences (-1600) expressed in these stages which is not expressed in third instar larvae.

Sequence Overlap of Pol y(A +)mRNA from D$ferent Developmental Stages
The data above indicate that most if not all of the DNA sequences expressed on polysomes of larvae are also expressed on polysomes of pupae and adults, but that some additional sequences are expressed in the later stages.
They do not, however, demonstrate whether the sequences present as polyadenylated RNA on larval polysomes are also present as polyadenylated sequences on the polysomes from pupae and adults. To evaluate this, a DNA population greatly enriched for single-copy sequences expressed as poly(A+)RNA on larval polysomes was prepared [(A+)mDNA], as was a population of DNA depleted of these sequences [null(A+)mDNA]. The logic for utilizing the (A+ )mDNA and null(A+ )mDNA for determining the amount of overlap in the (A+)mRNAs FIG. 3. Saturation hybridization of 3H single-copy DNA to D. melanoguster polyadenylated messenger RNA. Trace quantities of 3H single-copy DNA (specific activity -6 x 106 cpm/pg) were hybridized with a 2 lOO-fold sequence excess of polyadenylated messenger RNA from 5 to 7-hr pupae (A) or adults (B). Hybridization was monitored by resistance to S, nuclease digestion. Data points have been corrected for the contribution of tracer self-reassociation and for tracer reactability (-90% when reassociated with an excess of total D. melanogaster nuclear DNA). The results of the experiments presented above clearly indicate that a similar set of diverse sequences exists on polysomes in three morphologically and physiologically distinct stages of development of Drosophila. The question of whether the total mRNA sequence complexity of the whole organism represents the sum of the different mRNA sequence complexities of various organs and tissue types, or whether all organs and tissues contain mRNA with a sequence complexity equal to that of total is not addressed by these studies. Some insight into this question can, however, be obtained from the studies of Galau et al. (1976). Here, it has been clearly demonstrated that several adult tissues in the sea urchin share a substantial subset of their mRNA population. However, it is apparent that individual tissues of the adult exhibit a population of mRNA unique to that tissue. In view of these results, we chose to measure and subsequently contrast the RNA sequence complexity of an adult structure significantly enriched for a specific tissue type to that observed for the total organism.
We chose to investigate the sequence complexity of the mRNA population present in the head of adult Drosophila for the following reasons. Although the head from the three stages as well as the number of stage specific (A+)mRNAs is identical to that described above for total polysomal RNA. 3H-(A+)mDNA hybridized with poly(A+)mRNA from larvae to a saturation value of 38%, representing approximately a lO-fold enrichment of coding sequences over the 3.3% reported by Zimmerman et al. (1980) (Fig. 5A; Table 2). At saturation, 38 and 40% of (A+)mDNA hybridized with poly(A+)mRNA from pupae and adults, respectively (Figs. 5B, C). This result shows that most if not all of the polyadenylated RNA sequences present on polysomes of third instar larvae are also present as polyadenylated RNA on the polysomes of pupae and adults. 3H-Null(A+)mDNA hybridized with poly(A+) mRNA from third instar larvae to saturation value of 1.6%. Poly(A+)mRNA from pupae and adults drove 3.3% of 3H-null(A+)mDNA into hybrid form at saturation, indicoding sequences in third instar larvae. c 3H-(A+)mDNA represents a radioactively labeled population of single-copy DNA greatly enriched for polyadenylated messenger .._. .
does not represent an anatomical structure containing an individual organ or tissue type, it does represent a significant enrichment for neural tissue. This is evident from the fact that, although the head comprises only 10% of the body mass, it contains -50% of the total neural tissue of the adult (Demerec, 1950). In keeping with this anatomical distribution, 56% of the acetylcholine receptor sites in adult flies have been found to be present in head tissue (Schmidt-Nielsen et al., 1977). Furthermore, certain tissue types and organs are clearly absent in the head, such as the reproductive organs and the vast majority of the digestive and excretory systems. Thus, the head represents a structure which is devoid of many tissue types while containing a significant enrichment for neural tissue. The sequence complexity of head polysomal RNA was measured by hybridization of an excess of polysomal IO FIG. 5. Saturation hybridization of aH-(A+)mDNA to D. melanogaster polyadenylated messenger RNA. Trace quantities of 3H-(A+)mDNA were hybridized with a 2 lOO-fold sequence excess of polyaenylated messenger RNA from third instar larvae (A), 5-to 7-hr pupae (B), or adults (C). Hybridization was monitored by resistance to S, nuclease digestion.
EQUIVALENT Rot Ix~O-~I FIG. 6. Saturation hybridization of rH]DNA to D. melanogaster polysomal RNA from adult heads. Trace quantities of 3H single-copy DNA (A) or [3H]mDNA (B) were hybridized with a 2 loo-fold sequence excess of total polysomal RNA from adult heads. Hybridization was monitored by resistance to S, nuclease digestion. Data points in (A) have been corrected for the contribution of tracer self-reassociation and for tracer reactability (-90% when reassociated with an excess of total D. melanogaster nuclear DNA).
RNA to 3H single-copy DNA (Fig. 6A). At saturation, 7.7% of the single-copy DNA hybridized to the polysomal RNA, representing an RNA sequence complexity of 1.4 x 10' nucleotides, or about 11,000 diverse RNA species. This represents -69% (11,000/16,000) of the RNA sequence complexity observed for total adult polysomal RNA. Hybridization of [3H]mDNA with excess head polysomal RNA (Fig. 6B) showed that at saturation only 46% of the mDNA hybridized with the RNA from adult heads. These two observations confirm that the particular set of tissues found in the head contain only a subset of the RNA sequence complexity of the entire adult animal and agree in establishing that approximately 70-75% of the RNA diversity of the entire adult is expressed in the head.

DISCUSSION
In this report, measurements of sequence complexity using RNA excess saturation hybridization demonstrate that 15,000-16,000 diverse RNA sequences of average size are present on the polyribosomes of third instar larvae, 5-to 7-hr pupae and adults of D. melanogaster, a number consistent with that previously measured for third instar larvae (Zimmerman et al., 1980). Interest-ingly, a similar number of genes is expressed in a large lack of it, is a nonrandom, constant characteristic of a variety of eukaryotic organisms whose genomic sequence particular RNA species expressed on polyribosomes durcomplexities range in size from 0.3 to 20 times that of ing these stages of development in Drosophila. Drosophila (Galau et al., 1976;Hastie and Bishop, 1976; Hybridizations using null mDNA indicate that a small Van Ness and Hahn, 1980;Lanar, Levy, andMannnumber of new sequences, -2000, are expressed on ing, 1981).
polysomes of pupae and adults which were not expressed As reported previously for larvae (Zimmerman et al., in larvae. Furthermore, these new sequences appear in 1980), the poly(A+)mRNA of pupae and adults consti-the polyadenylated class as evidenced by an increased tutes an unusually small proportion of the mass of polyso-hybridization of larval null(A+)mDNA with ma1 RNA, 0.2-0.5%, and contributes only approxi-poly(A+)mRNA from pupae and adults. We report that mately one-third of its sequence diversity. Specifically, the number of genes expressed in poly(A+)mRNA of our measurements show that pupal poly(A+)mRNA conpupae and adults is 6500 and 6100, respectively, an intains approximately 6500 different sequences, while crease of 700-1100 over the value for larvae reported by 15,000 diverse sequences are present in total polysomal Zimmerman et al. (1980) from experiments performed RNA. Similarly, adult poly(A+)mRNA contains 6100 exactly as those reported here. Kinetic measurements different sequences, with 16,000 present in total polyso-which employed [3H]cDNA to poly(A+) cytoplasmic ma1 RNA. Therefore, in larvae, pupae, and adults, non-RNA have reported the number of genes expressed in adenylated RNA molecules associated with polysomal larvae to be 24900 (Izquierdo and Bishop, 1979) and RNA constitute approximately 60% of the RNA se-6900 (Levy and McCarthy, 1975) and the number exquence diversity.
pressed in pupae and adults to be 6900 and 24900, re-The heads of adult flies were prepared to represent an spectively (Izquierdo and Bishop, 1979). In addition, the enrichment for neural tissue (Demerec, 1950). Polysomal fact that we observe little if any increase in sequence di-RNA from heads has a complexity which corresponds to versity of total polysomal RNA from the later stages is 11,000 different RNA sequences. The sequence diversity evidence that the number of new sequences detected by represents a subset, -7O%, of that observed in the en-our overlap studies is quite small. tire adult. This result suggests that different adult tis-Thus, very little qualitative change occurs in the adesues may contain somewhat different subsets of se-nylated and nonadenylated classes of polysome assoquences on polysomes, all of which total the 16,000 ciated RNAs during Drosophila development. Such exsequences observed on polysomes of the entire adult. In-tensive RNA sequence homology has been previously deed, qualitative differences in the RNA sequences on reported to exist between tissues which differ dramatipolysomes from several adult tissues of the sea urchin cally in morphology and physiology. For example, culhave been well characterized (Galau et al., 1976). The tures of undifferentiated chick myoblasts appear to share distribution of the polysomal sequence complexity from the same set of approximately 17,000 different messenadult heads among adenylated and nonadenylated classes ger RNA sequences with differentiated myofibrils in was not measured because of scarcity of material. vitro (Paterson and Bishop, 1977). Two highly differen-Thus far, the results indicate that a similar number of tiated chicken tissues, liver and oviduct, have been diverse sequences exists on polyribosomes in three shown by hybridization kinetics to contain 12,000-15,000 stages of development of Drosophila, and suggest that diverse mRNA species, of which at least 85% are held in there may be little qualitative difference between the common (Axe1 et al., 1976). In addition, extensive homo-RNA populations in these organisms. Our sequence logy between the mRNAs of normal and chemically or overlap studies using mDNA demonstrate that, indeed, virally transformed cells has been reported, although virtually all of the DNA coding sequences expressed on transformation results in a large number of phenotypic polysomes in larvae are also expressed on polysomes in changes (Getz et al., 1977;Rolton et al., 1977;Williams et pupae and adults. Within the limits of detection of our al., 1977). technique, therefore, coding sequences represented on In contrast to the studies discussed thus far, a variety polysomes during larval development continue to be of sea urchin embryo and adult tissues share only a small present throughout the life of the animal. It is possible proportion of their total mRNA sequence diversity. For that during development, such sequences change their example, of the 14,000 mRNA species found in gastrula, distribution among the adenylated and nonadenylated 11,000 are also expressed in pluteus, and only 1000-1500 classes of RNA; however, our overlap studies using in three adult tissues (Galau et al., 1976). Thus, it is clear (A+)mDNA show this not to be the case. In fact, all of that qualitatively distinct sets of single-copy sequence the polyadenylated sequences on polysomes of larvae transcripts appear on polyribosomes during sea urchin also appear as adenylated molecules on polysomes of development. pupae and adults. Therefore, polyadenylation, or the Although large qualitative changes in mRNA popula- 1976; Paterson and Bishop, 1977;Wilkes et al., 1979). Our study does not address this question, although recent reports have demonstrated that changes in the relative abundance of individual polyadenylated RNAs accompany development in D. melanogaster (Biessman, 1981) and in Xenolyus Lewis (Dworkin and Dawid, 1986). Such quantitative changes may influence the phenotype of the developing system more profoundly than the presence on polyribosomes of an apparently constant sequence diversity.
In summary, our studies confirm and extend previous reports that the majority of polyadenylated mRNA sequence diversity in Drosophila is retained throughout development (Levy and McCarthy, 1975;Izquierdo and Bishop, 1979;Arthur et al., 1979). Further, we report that the nonadenylated sequences present on polyribosomes represent a discrete class of molecules present continually during three stages of development in Drosophila. Thus, it is clear from our studies and others that development or differentiation is not necessarily accompanied by sweeping qualitative changes in gene expression.