Characterization of globin messenger ribonucleic acids in membrane polysomes of mouse reticulocytes.

Between 20 and 30% of the polysomes in mouse reticulocytes are associated with the cell membrane fraction. These polysomes are not liberated by washing with 0.5 m KCl and are therefore thought to be attached to membranes. They have the same percentage of polyadenylic acid-containing RNA, as determined by oligo(dT)-cellulose affinity chromatography, as cytoplasmic polysomes. Analysis of the RNA by polyacrylamide gel electrophoresis in aqueous solutions shows that at least 95% of the RNA migrates identically with the cytoplasmic polysomal globin mRNAs. Electrophoresis in 99% formamide resolves the membrane mRNAs into two bands which migrate identically with the cytoplasmic alpha- and beta-globin mRNAs. The molar ratio of the alpha- and beta-globin mRNAs, as determined by quantitating the bands on the formamide gels, is similar to that of the cytoplasmic mRNAs. There is also no difference in biological activity between the two messenger preparations. The polyadenylic acid in the membrane mRNAs, isolated after labeling mice in vivo for 20 hours with (32P)orthophosphate, migrates with the three broad size classes previously shown to be present in the cytoplasmic globin mRNAs. The 32P specific activity of the membrane mRNAs and ribosomal RNA fractions after different times of labeling with 32Pi are similar to those from cytoplasmic polysomes. These observations show that the reticulocyte membranes contain approximately 20% of the cellular globin nRNAs, and that these mRNAs are similar to those from cytoplasmic polysomes.


SUMMARY
Between 20 and 30% of the polysomes in mouse reticulocytes are associated with the cell membrane fraction.
These polysomes are not liberated by washing with 0.5 M KC1 and are therefore thought to be attached to membranes. They have the same percentage of polyadenylic acid-containing RNA, as determined by oligo(dT)-cellulose affinity chromatography, as cytoplasmic polysomes. Analysis of the RNA by polyacrylamide gel electrophoresis in aqueous solutions shows that at least 95 % of the RNA migrates identically with the cytoplasmic polysomal globin mRNAs. Electrophoresis in 99% formamide resolves the membrane mRNAs into two bands which migrate identically with the cytoplasmic (Y-and P-globin mRNAs.
The molar ratio of the oy-and @-globin mRNAs, as determined by quantitating the bands on the formamide gels, is similar to that of the cytoplasmic mRNAs. There is also no difference in biological activity between the two messenger preparations. The polyadenylic acid in the membrane mRNAs, isolated after labeling mice in uiuo for 20 hours with (32P]orthophosphate, migrates with the three broad size classes previously shown to be present in the cytoplasmic globin mRNAs.
The 32P specific activity of the membrane mRNAs and ribosomal RNA fractions after different times of labeling with 3*Pi are similar to those from cytoplasmic polysomes. These observations show that the reticulocyte membranes contain approximately 20% of the cellular globin mRNAs, and that these mRNAs are similar to those from cytoplasmic polysomes.
In many animal cells, polysomes exist either free in the cytoplasm or bound to the endoplasmic reticulum (l-3).
Although there is evidence that free and membrane-bound polysomes synthesize different classes of protein (4-6), it is not clear whether this differential activity is a result of mRNA distribution per se or is due to other factors (7-9).
Attempts to find differences between free and membrane-bound mRNAs have been unsuc-* This work was supported by United States Public Health Grant GM-10999 from National Institute of General Medical Sciences, National Science Foundation Grant GB-41204, and American Cancer Society Grant NP-59. cessful. One possibility, that of very different sizes of poly(A) regions, has been eliminated (10, 11).
It has been reported that 20% of the polysomes in reticulocytes are membrane-bound (12-15). These polysomes may be bound to the cell membrane rather than the endoplasmic reticulum as the latter structures are riot readily detectable in reticulocytes. Analysis of the membrane-bound RNA showed that it contains a 9 S RNA fraction which co-migrates with the cytoplasmic 9 S globin mRNA on 3y0 polyacrylamide gels (16). This RNA represents about 1% of the membrane ribosomal RNA and is not derived from the pelleted white cells (16). These preliminary studies therefore suggested that the membrane-bound polysomes contain globin mRNA-like RNA. However, it is not known what percentage of the mRNA in these membrane-bound polysomes is globin mRNA or if the globin mRNA present is identical with that found in the free cytoplasmic polysomes.
The answers to these questions may be helpful in understanding the significance of the membrane-bound polysomes found in reticulocytes. In this paper we show that 20% of total mRNA in mouse reticulocytes is located on reticulocyte membrane polysomes, and that this mRNA is identical with the free cytoplasmic polysomal mRNA in size and biological activity.
The poly(A) distribution in the two mRNAs after labeling mice in tivo with [321']orthophosphate for 20 hours was also similar.

Isolation of Free and Membrane-bound
Polysomes-Reticulocytes were collected from mice made anemic with six daily injections of phenylhydrazine hydrochloride (17,18). The cells were washed by the method of Lingrel et al. (17,18) and lysed by the addition of 4 volumes of 10 mM KCI, 1.5 mM MgCl,, and 10 mM Tris-HCl (pH 7.6). The membranes were removed by centrifugation at 15,000 X g for 15 min (19,20). Free polysomes were isolated from the supernatant fraction by a further centrifugation for 3 hours at 150,000 X g. Membrane-bound polysomes were isolated by the method of Bulova and Burka (15). The membranes were washed four times with 4 volumes of lysis buffer and pelleted white cells were discarded.
Polysomes were dissociated from the membranes by treatment with 0.2'% sodium deoxycholate (15). After centrifuging down the membranes at 15,000 X g for 15 min, the polysomes were isolated by centrifugation at 150,090 X g for 3 hours.
Isolation of 9 S RNA-RNA was isolated from free and membrane-bound polysomes by extraction with phenol-chloroformisoamyl alcohol (20-22). The polysomes were taken up in 1% sodium dodecyl sulfate, 100 mM NaCl, 1 mM EDTA, 100 mM Tris (pH 9.0) (approximately 100 Azso units per ml), and incubated for 5 min at room temperature to dissociate the proteins from the RNA. After shaking at room temperature with an equal volume of phenol-chloroform-isoamyl alcohol, the phases were separated by centrifugation at 15,000 X g for 15 min at room temperature. The globins were extracted with acid-acetone apd the products were separated on carboxymethylcellulose (18).

Percentage of RNA in Membrane Polysome Fractions-When membrane-bound
and free polysomes were prepared from several batches of reticulocytes, the percentage of RNA in the membrane fraction was between 15 and 237, of that in the free polysomes ( Table 1). The recovery of poly(A)-containing mRNA from both fractions varied between 0.9% and 1.2% of the non-polp(A)containing RNAs (Table 1). The membrane-bound polysomes therefore contain between 15 and 23% of the total cellular poly(A)-containing mRNA population. As treatment with high salt has been shown to preferentially release those polysomes which are only loosely bound to the endoplasmic reticulum in other cell types (26,27), the recovery of reticulocyte membrane-bound RNA was compared with or without washing in 0.5 M KCI. Only a negligible loss of RNA occurred from the membranes which were washed in high salt (Table I, Experiment  4). Size Characterization of Membrane-bound mRNA--The RNAs isolated from membrane-bound polysomes migrated similarly to the globin mRNAs from cytoplasmic polysomes when they were subjected to electrophoresis on sodium dodecyl sulfate 3% polyacrylamide gels (Fig. 1, a and b). Samples of both RNAs subjected to electrophoresis on the same gel showed no greater spread in band width than did the separate samples (Fig. lc). Scanning the unstained gels at 260 nm showed that greater than 95% of the membrane mRNA migrated in the 9 S RNA region.
Although electrophoresis in aqueous gels showed that there was negligible contamination of the membrane mRNAs by RNAs with sizes other than 9 S, the presence of mRNAs corresponding exactly in size to both a-and /3-globin mRNAs could not be determined.
When the membrane mRNAs were subjected to electrophoresis on 7.5% polyacrylamide gels in 99% formamide, conditions in which the cytoplasmic mRNAs are resolved into their (Y-and /Lglobin mRNA components (20), the membrane mRNAs were also resolved into two bands (Fig. 2b). The migration of these bands relative to the 4 S and 5 S RNAs was the same as for the cytoplasmic mRNAs (Fig. 2a). Electrophoresis of both cytoplasmic and membrane mRNAs on the same gel (Fig. 2~) showed no broadening of the bands. By both these criteria, the membrane mRNAs migrate identically with the cytoplasmic LY-and /3-globin mRNAs. Fig. 3 shows a stained gel of the membrane mRNAs after electrophoresis on 10% polyacrylamide gels in the presence of formamide.
The molar ratio of fl-mRNA to a-mRNA in the membrane RNA preparations was determined by scanning the stained 7.5% and 10% formamide gels at 260 nm and quantitating the area under each peak (20  2. The 7.5% polyacrylamide gel electrophoresis of membrane and cytoplasmic mRNAs in the presence of formamide. The mRNAs were isolated as outlined under "Methods." Gels were electrophoresed for 30 min at 1 ma per gel, then for 3 hours at 4 ma per gel. To each gel, along with the mRNA samples, 15 Mg of 4 S and 5 S RNA markers were applied. a, 6 rg of mRNAs from free cytoplasmic polysomes; b, 6 pg of mRNAs from membrane polysomes; c, 3pg of membrane mRNAs + 3 pg of cytoplasmic mRNAs. The gel was electrophoresed for 30 min at 1 ma per gel, then for 3 hours at 4 ma per gel. To the gel were applied 6 pg of mRNAs. 1.2; this ratio is identical with that previously found for the messengers in free cytoplasmic polysomes (20).
Biological A&&--Although the size estimates show that the membrane mRNAs are very similar in size to those in the cytoplasm, they provide no information on the biological activity of the membrane mRNAs.
In order to determine whether the membrane mRNAs were as active in globin synthesis as the cytoplasmic polysomal mRNAs, both were assayed for cr-and fi-globin chain synthesis in duck reticulocyte lysates.
The biological activity of the membrane mRNAs was similar to that of cytoplasmic mRNAs prepared from polysomes isolated in the presence of 0.2% sodium deoxycholate (Table II).
The mRNAs prepared from membranes treated with high salt showed a comparable globin synthetic activity to those prepared in the usual manner.
In all cases, the mRNAs isolated from deoxycholate-treated polysomes showed approximately one-half of the globin synthetic capacity of cytoplasmic polysomal mRNAs (Table II).
We have previously shown that the cytoplasmic mRNAs with different size classes of poly(A) can be fractionated on Millipore filters (22). The mRNAs containing the larger size classes, 55 to 65 and 75 to 120 nucleotides in length, are retained on the filters while the mRNAs containing poly(A)s 35 to 45 nucleotides in length do not bind to the filters. These experiments proved that the different size classes of poly(A) were found on different messenger populations and were present in both LY-and @mRNAs (22). A sample of membrane-bound mR.NA which was similarly fractionated on Xlillipore filters showed the same poly(A) distribution between the Millipore-bound and unbound mRNAs as that shown for the cytoplasmic globin mRNAs (Fig. 4b). The Millipore-bound and unbound mRNAs also have similar /3-mRNA to cr-mRNA ratios as determined by electrophoresis on polyacrylamide gels in the presence of formamide (results not shown).
The distribution of poly(A) lengths between the two messengers is therefore similar for the cytoplasmic and membrane globin mRNAs.
The comparable /3 to ac-globin synthesis in both membrane and cytoplasmic fractions shows that there is no preferential accumulation of /3-or a-globin mRNA activity in the membrane-bound polysomes.
Specijic Activity of RNAs after Diflerent Labeling Times-Although the chemical and biological characterizations of the mRNAs from free and membrane-bound polysomes show no difference between them, it is still possible that the membranebound RNAs are synthesized at different times than those in the cytoplasm.
The specific activities of RNAs isolated from animals which received [3*P]orthophosphate at various times prior to collection of reticulocytes were therefore examined.

Poly(A) Content of mRNAs Labeled with [W]Orthophosphate
When an anemic animal is injected with [3zP]orthophosphate, 20 Hours before Collection of Reticulocytes-In order to compare the "Pi is incorporated into RNAs synthesized by all the prethe poly(A) content of free and membrane-bound mRNAs, mice cursor cells. Since the precursor cells only appear in the circulawere injected with [""PI orthophosphate 20 hours before collecting tion when they have matured into reticulocytes, the labeled RNA the reticulocytes, and mRNAs were prepared as under "Meth-present in circulating reticulocytes originates only from those The mRNAs were isolated as under "Methods." Part of the cytoplasmic polysomal mRNA was isolated from polysomes pelleted in the presence of 0.2% sodium deoxycholate.
The mRNAs were assayed for activity in a duck reticulocyte lysate and the amounts of (Y-and p-globin synthesis were determined as under "Methods." To each cell RNAs dere digested with RNase T1 and RNase A (see "Methods").
The poly(A)-containing fragments were isolated by oligodeoxythymidylate cellulose affinity chromatography, desalted, and lyophilized prior to electrophdresis (see "Methods").
Reticulocvte 4 S and 5 S RNAs were used as standards.
cells which mature in the time interval between labeling and collection. The labeled RNA entering the circulation shortly after [321']orthophosphate injection represents newly synthesized RiYA while that entering at later times represents older RNA (22).
The specific activities of the rRNAs from both free cytoplasmic and membrane-bound polysomcs arc very similar in the 6-to 26-hour time interval studied (Fig. 5). The specific activity of the membrane poly(A)-containing RNAs also followed that of the cytoplasmic poly(A)-containing RNAs, although some differences were noted at the 5-hour and 26.hour time points.

I)ISCUSSION
The membrane-bound polysomes contain approximately 20% of the total cellular poly(A)-containing mRNA.
It is likely that these polysomes arc truly membrane-bound as they are not dissociated by washing with high salt. This procedure has been shown to dissociate ribosomes which are loosely bound while leaving tightly bound ribosomes still attached to the membranes (25, 26). O---O, membrane 9 S RNA; l ---0, cytoplasmic 9 S RNA; n---A, membrane 18 S and 28 S RNAs; ApA, cytoplasmic 18 S and 28 S RNAs, Over 95% of the membrane-bound mRn'A migrates in the 9 S RNA region on sodium dodecyl sulfate 3% polyacrylamide gels. This 9 S RiXA is identical in size with the 01. and P-globin mRNAs isolated from cytoplasmic polysomcs as determined by elcctrophoresis on 7.5% polyacrylamide gels in the presence of formamide. Electrophoresis of the membrane mRiYAs and marker ribosomal RNA species on 3.67, polyacrylamide gels in the presence of formamide also showed that the molecular weights of the (Y-and fl-globin mRNAs from membranes were identical with those from the cytoplasmic polysomes.l There is therefore no difference in the lengths of the noncoding regions between the cytoplasmic and membrane LY-and P-globin mRNAs (20).
The biological activity of the membrane mHXAs is also similar to that of the fret polysomal mRNAs.
This result agrees with the finding that the amount and nature of the nonglobin proteins synthesized by free cytoplasmic polysomes in reticulocytes are similar to those synthesized by the membrane polysomcs (28,29). There is, therefore, no evidence for a preferential localization of nonglobin mRiYAs on the membrane-bound polysomes such as proposed by Rulova and Rurka (15), although our studies do not rule out the synthesis of a small amount of nonheme protein.
We have previously shown that the the molar ratio of p-and a-mRNA in the postmembrane supernatant fraction of rcticulocytes is approximately 1 (20). The globin n&NA localized in the membrane polysomcs also has a ~:CI ratio close to unity. Thus, there is no preferential localization of tither messenger in the membrane fraction.
The molar amounts of /3-and Lu-mKn'As in the total cellular mRNA population is therefore close to 1.
There has been some disagreement about the synthetic products of membranebound polysomes. In rabbits, for example, the proteins synthesized by these polysomes do not comigrate with globin chains (15,29).
However, the tryptic peptide analysis showed that greater than 957, of the protein synthesized by the mcmbranc-bound polysomcs was globin (29). The anomalous migration of the rabbit globins may be an artifact introduced in the isolation of globins synthesized by membrane-bound polysomes in cell-free systems as the globins synthesized by isolated mouse membrane globin mRKAs in a reticulocyte lysate migrate identically with mouse CC and P-globin chains.
Although our results show that approximately 200/ of the cellular globin mRNAs are present in membrane-bound ribosomes, this does not imply that all the membrane-bound mRNAs are active in globin synthesis.
Although Woodward et al. (29) found that bound ribosomes contained about 12 y0 as much radioactivity as did free ribosomes after cells were incubated with [3H]tyrosine for 10 min, we find that approximately 20% of the total globin mRNAs are located on the membranes.
Thus, it is possible that, although the isolated membrane-bound globin mRNAs are as active in cell-free systems as the cytoplasmic messengers, their activity may be modified in z&o. Such a modification of activity has been shown for ferritin mRNAs localized on membrane-bound polysomes and albumin mRKAs localized on free cytoplasmic polysomes (9).