Immunological characterization of rabbit hemoglobin α and β chain synthesizing polysomes. Estimation of relative numbers of active α and β messenger ribonucleic acid

Abstract Anti-hemoglobin chain antibodies, purified by immunoadsorption to Sepharose-linked antigen and thereby made both RNase-free and immunologically specific for either β or α chain, are readily bound by nascent hemoglobin chains attached to rabbit reticulocyte polysomes. Antibodies to other antigens, e.g. anti-immunoglobulin G, are not bound. Use of chain-specific antibodies enabled us to determine quantitatively the mean number of ribosomes associated with rabbit hemoglobin β- and α-messenger RNA. Such mean sizes of β- and α-polysomes could be calculated, for example, from density gradient profiles of polysomes previously coated with either 125I-labeled anti-β chain or 125I-labeled anti-α chain. Complementary estimates were obtained, pre- and postimmunoprecipitation, from (a) the A260 density gradient profiles of polysomes and (b) the profiles of β and α polysomes differentially labeled with [3H]isoleucine and [14C]valine. All estimates were similar; viz. the β:α ratio for mean number of ribosomes per mRNA ranged from 1.5 to 1.86. Among unaltered polysomes from the same rabbits, the β:α ratio for total relative numbers of synthetically active β and α ribosomes was calculated from y intercepts of Dintzis-Naughton plots to be ∼1.04 to 1.17. With these two kinds of β:α ratios as a basis, the relative numbers of active α-mRNA and β-mRNA were found to be somewhat greater than previously described by Lodish and Jacobsen ((1972) J. Biol. Chem. 247, 3622). For individual animals, the α-mRNA:β-mRNA ratio ranged from 1.28 to 1.78. The addition of the further 25% of α-mRNA which is unattached to polysomes (Jacobs-Lorena, M., and Baglioni, C. (1972) Proc. Nat. Acad. Sci. U. S. A. 69, 1425) suggests that total α-mRNA is 2-fold more abundant than β-mRNA.

Immunological methods have recently been applied by Schimke and his associates (l-4) to the characterization and * This work was supported in part by United States Public Health Service Grant 9 It01 HL 15026-07. purification of hen oviduct polysomes uhich synthesize ovalbumin.
The approach is straightforward; viz. antibodies developed against a completed polypeptide can react with nascent chains, i.e. species of polypeptide still attached to polysomes. The antibodies thus bound may be isolated by either immunoprecipitation or immunoadsorption and thereby messenger RNA for a particular polypeptide separated from other forms of mRNA.
Although the same notion has occurred to many in the past (see Ref. 4 for brief review), Schimke and his colleagues were the first to bring analytical rigor to the procedure and its products.
In this report we apply immunological methods to the characterization of rabbit hemoglobin a! and 0 chain-synthesizing polysomes and describe, first, the preparation and properties of the requisite immunological reagents; second, the binding between antibodies and reticulocyte polysomes, including, as a result, direct evidence that a! polysomes (as previously supposed (5-7)) are considerably smaller than /3 polysomes; and, third, some derivative calculations indicating that the numbers of active cr-mR1VA exceed /3-mRNA by a factor of at least 1.28 to 1.78. Elsewhere (S), we use the latter result to create an enlarged vien of ways in which biosynthesis of hemoglobin a and /3 chains and their evolutionary changes are interrelated.

RESULTS
Purification and Immunological Assay of Antibodies-RNase activity and attendant polysome breakdown were detectable whenever -20 ~1 or more of serum were mixed with 1 ml of polysomes containing 10 A260 units. Removal of RNase from antisera was accomplished via immunoadsorption.
The extent of purification of desorbed antibodies RBS assessed through comparison of quantitites of protein desorbed from antigell-bearing Sepharose versus quantities desorbed from antigen-free Sepharose (Table 1). Purification R-as usually several hundred-fold, and it is not surprising that all antibodies isolated by immunoadsorption were free of detectable RNase. 111 Table I, recoveries of purified antibodies against hemoglobin LY chains are approximately comparable to recoveries of various purified anti-IgG.
'1 he exceptionally low yields associated with purified goat anti-rabbit hemoglobin @ (Table 1) Deere earlier forecast by two findings: namely, (a) rabbit /3 chain provoked precipitating antibodies in only two of 12 immunized goats, in contrast, for example, to five responders of five goats immunized with rabbit cr; and (b), \lhhen detectable in u hole sera, precipitating antibodies against rabbit 0 M-ere evanescent during the course of serial bleedings.
Following immunoadsorptive purification, anti-rabbit hemoglobin /3 preparations became totally unreactive on Cuchterlony plates, even uhen assayed at antibody concentrations of 1 mg per ml and, in t.his respect, were quite unlike anti-rabbit (Y. Purified anti-rabbit fl preparations from six of 12 goats immunized did, however, contain nonprecipitating antibodies. These were demonstrated by L251-labeled anti-P binding ( Fig.  1) to Sepharose-linked @ under conditions in which nonimmunological binding of trace quantities of protein was precluded by prior saturation of Sepharose with normal goat serum.

Specijkity of Antibodies against Hemoglobin
Chains--Although the precipitating reactions found for anti-rabbit hemoglobin a! chain on Ouchterlony plates were immunologically specific (i.e. anti-a did not react with fl chain), the possible existence of cross-reactive nonprecipitating antibodies was uot excluded.
Exclusion was essential because our later studies depended OII complete immunological specificity.
Hpecificitics of both anti-a! and anti-p were therefore further tested by binding iodinated antibodies to excess antigen coupled to Sepharose. Testing was performed, once again, under circumstances in which nonimmunological binding was prevented.
We were surprised to find that binding patterns of purified iodinated anti-rabbit a! and anti-rabbit p were marred by considerable immunological nonspecificity.
For example, when assayed in the presence of a 100 M excess chain antigen, 18 to 23% of l*SI-labeled anti-a, despite extensive prior exposure to Sepharose-/3 chain during purification, is bound by /3 chain-bearing Sepharose (cf. 27y0 to a! chain-bearing Sepharose), even when the /3 chains come from a succession of chromatographic fractions increasingly remote from QI. Such immunological nonspecificity could be explained either by residual cross-reacting antibody in purified preparations or by cross-contamination of a and /3 antigens with one another. The latter possibility was tested by repeated isolation and re-chromatography on CM-cellulose of individual hemoglobin chains. The section of chain peak re-chromatographed in each cycle was varied in an attempt to remove trace cross-contamination of (Y and B chains with one another. Following five such cycles, the now highly purified rabbit hemoglobin (Y and fl chains were coupled to Sepharose and tested in antigen excess for their ability  Anti-g (591D). 0 28 to bind each of several lots of iodinated ant,i-rabbit a! and anti-& The original lots of iodinated antibodies, which had been immunologically nonspecific when assayed against hemoglobin chains isolated once, were entirely specific when tested (Table  II) against Sepharose linked to repeatedly chromatographed chains. Clearly, cross-contamination of antigens rather than residual unwanted antibodies was the source of prior nonspecific- and, separately, '2'1~labeled anti-rabbit IgG (9.4 pmoles of active antibody protein, 15 nCi per pmole) from Goat 596 (O---0) to rabbit reticulocyte polysomes (~100 pmoles of ribosomes as polysomes in each instance).
The point is an important one because the use of small quantities of highly purified individual chains as final test antigens enables us to show not only that each of the antibodies later used for reactions with reticulocyte polysomes is entirely specific for a particular hemoglobin chain but also that hemoglobin chains chromatographed once, even though contaminated \\ith -1 y. of unwanted chain,2 are satisfactory for both immunization and immunoadsorption.
The example is not unique; a similar situation occurred with immunoadsorptively purified antibodies against mouse cy and /3 chains as uell as against human a, /3, and y chains, when in each case, immunological nonspecificity uas seen with hemoglobin chains chromatographed once but vanished when highly purified, re-chromatographed chains were used as antigens.* Binding of Iodinated Antibodies to Polysomes-The differential binding of purified lalI-labeled anti-rabbit p and 1251-labeled antirabbit IgG to rabbit reticulocyte polysomes sedimented in individual sucrose density gradients is shonn in Fig. 2. In this instance, 62% of the immunologically active anti-p applied was bound and resulted in a recognizable polysome profile, whereas only 2.6% of applied anti-IgG appeared in the polysome region. From this and related analyses, including many sedimentation assays of polysomes allowed to react with iodinated anti-rabbit a! (e.g. in Fig. 3), it is evident that antibodies developed against completed hemoglobin chains can react with nascent hemoglobin chains. In contrast, nonimmunological associations between reticulocyte polysomes and purified antibodies to antigens presumably absent, typified by 1251-labeled anti-lgG binding in Fig. 2, are negligible and usually involve <l% of applied antibody.
The capacity of reticulocyte polysomes to bind rabbit antihemoglobin p chain, as seen in Fig. 4, can be compared with the binding capacity of /3 chain-bearing Sepharose (seen in Fig. 1 3. Comparative binding profiles of lZ61-labeled anti-rabbit hemoglobin p (0---0) (12 pmoles of active antibody protein, 60 nCi per pmole from Goat 591) (upper panel) and l*'I-labeled anti-rabbit hemoglobin a (0---0) (16 pmoles of active antibody protein, 33 nCi per pmole) from Goat 398 (lower panel) to rabbit reticulocyte polysomes (117 pmoles of ribosomes as polysomes from Rabbit 90 in each case). Conditions differed from those described for Fig. 2 in that anti-8 was from a different iodination lot, ribosomes came from a different rabbit, and gradients (fractionated here by upward displacement) were monitored continuously for Az~,, before radioactive assay. Counts are omitted from fractions from the top of gradient (lefl). Percentages of labeled antibody bound to polysomes (counts to the right of Fraction 15) represent 28% of applied active anti-fi protein and 18% of applied active anti-a.
The As5,, plot (O-O), corrected for rising base-line during monitoring and asynchrony with fractions, was virtually identical in a third gradient (not shown) containing untreated 3H-labeled polysomes. Numbers above absorbance points here and in Figs. 5 and 6 refer to expected (22) number of ribosomes per mRNA in each polysome size class. The exact location of each size class depended on coincident examination (not shown) of sharp peaks associated with [aH]lysine-labeled polysomes.
The principal Az~,, peak (Fraction 10) was positively identified as 80 S, i.e. containing 1 whole ribosome, rather than the 60 S fragment free of nascent chain; identification depended on demonstration of ~2: 1 A260 ratios between 28 S and 18 S rRNA in derivatives (25) of the 80 S peak. hemoglobin 0 chain to reticulocyte polysomes. 0 and 0. 14 pmoles of ribosomes as polysomes were used; 0, 41 pmoles; n , %? pmoles. All polysomes came from the same batch used for analvsis denicted in Fig. 3. Each point was obtained after sedimentation of an antibodi-polysome mixture through 4.4 ml containing discontinuous concentrations of sucrose. TWO different lots of '2SI-labeled anti-@ were used. Data shown as 0 and 0 were obtained with the *25I-labeled anti-p lot (45y0 maximal immunological activity after iodination) used in Fig. 1 for measurement of antibodv bindine to 13 linked to Seaharose. This lot I ,> was used here in varying amounts at constant specific activity (24 nCi per pmole). A different lot of labeled antibody (34% maximal immunological activity after iodination) was applied at constant specific activity (30 nCi per pmole) for 0 point assays, whereas for n point assays it was diluted with unlabeled Goat 591 anti-p from the same batch of purified antibody (-90% maximal immunological activity before iodination). n points are not connected because the estimate of 90% maximal immunological activity in unlabeled anti-0 (on which point locations depend) is only approximate (q.v. legend to Fig. 1). The percentage of active antibody bound/applied (right ordinate) is shown only for 0 and 0; a similar plot can be constructed for n and 0.
for the same lot of iodinated antibody.
For any particular load, the moles of anti-/3 bound per mole of polysomal ribosomes represent slightly more than one-half of the quantity bound per mole of p chain linked to Sepharose. This result is unespected, for it suggests that nearly all of the polysomes engaged in /3 synthesis (approximately one-half of all polysomes; see below) bear nascent chains which can bind antibody raised against completed /3 chains. Alternatively, it is possible that a few nearly completed /3 chains can each bind 2 or more antibody molecules.
A more likely explanation is that the binding capacity of /3 chain-bearing Sepharose has been somewhat underestimated, because not all of 0 measured as a protein remains active as an antigen.
Since the portion uhich is antigenically inactive cannot be directly assessed, the actual fraction of /? nascent chains which can bind anti-/3 is uncertain.
Elots like Fig. 4 nonetheless retain their principal value, which is to serve as a gauge for estimating optimal quantities of antibody later required for immunoprecipitation of polysomes.
Reactions between anti-chains and polysomes are also useful in assessing an aspect of immunological specificity which has until now been ignored, viz. the possibility that purified antibody preparations might react with nonhemoglobin antigens. In general, the hemoglobin chains used for both immunization and 7215 later immunoadsorption were chromatographed only once. Trace quantities of nonhemoglobin antigens were undoubtedly present.
Antibodies produced against these might be recovered during adsorption-desorption from Sepharose bearing contaminated antigen.
That such unwanted antibodies are absent from purified preparations can be shown by comparisons of the binding of antibody to Sepharose-linked antigen and the binding of the same antibody to polysomes.
For example, under conditions in which molar ratios of antigen to antibody are > 100 : 1, 34y0 of the radioactivity in iodinated anti-0 preparation 591B (Table  II) was bound to purified /3 chain linked to Sepharose and 34% to polysomes.
Similar equivalence was seen, at different percentages, with other iodinated anti-@ preparations.
We interpret these findings as inconsistent with an appreciable contamination of anti-/3 with antibodies to nonhemoglobin antigens. Our interpretation rests on the obvious differences between the kinds and concentrations of nonhemoglobin proteins present in /3 chain preparations isolated from the blood of nonanemic rabbits and those present in reticulocyte polysomes.
Because of such differences in antigenic profiles, any antibodies developed against nonhemoglobin antigens should bind more or less well to polysomes than they do to isolated hemoglobin p chain. Yet, even when anti-/3 preparations are tested against gross excesses of antigen, so that non-0 antigens might have every chance for expression, there is no difference between the quantity of antibody bound to polysomes and that bound to chain. Accordingly, these anti-/3 preparations are free of unwanted antibody. Similar analytical precision was not possible for lz51-labled anti-a reactions in which polysome-bound anti-a is incompletely resolved from the tailing of unbound anti-a! (q.v. below) during density gradient sedimentation.
Nonetheless, under conditions of gross antigen excess, the percentage of a given lot of iodinated anti-a bound to Gepharose-cr and the percentage bound to nascent a chains is approximately comparable. This, once again, suggests that antibodies to nohemoglobin proteins are absent. Contrasting Sizes of (Y and fl Chain-synthesizing Polysomes-Differences in rabbit (Animal 90) reticulocyte polysome profiles formed by binding 1251-labeled anti-a and 1251-labeled anti-0 are evident in Fig. 3. lodinated anti-a, on the average, binds to considerably smaller polysomes than does anti-o.
In contrast, the AzcO patterns are identical for each analysis and are indistinguishable from that given by antibody-free polysomes; degradation by RNase is thus excluded as an explanation for the differences between the two iodinated antibody-labeled polysome profiles.
Although precise estimates of the mean sizes of antia-labeled polysomes and anti-@-labeled polysomes are obtunded by the tailing of unbound 1*51-labeled antibodies into the region of polysomes containing 1 and 2 ribosomes per mRNA, approximations are still possible.
13~ inspection, the mean size of polysomes (Fig. 3) labeled by 1251-anti-@ is somewhat greater than 6 ribosomes per mRNA and, in any case, slightly larger than the mean of untreated polysomes (~5 ribosomes per mRNA) and substantially larger than the mean size of polysomes labeled with lz51-anti-a (~4 ribosomes per mRNA).
In Fig. 5, a similar @:a, ratio for mean polysome sizes was adduced in a second rabbit (Rabbit 89) from (a) the (Y polysome Atso profile (mean size, ~4 ribosomes per mRNA) remaining after immunoprecipitation of the bulk of /3 polysomes by SUCcessive use of goat anti-rabbit /I and rabbit anti-goat IgG, and (b) the /3-polysome A260 profile (mean size, -6 ribosomes per mRNA) left after immunoprecipitation of most a! polysomes through seriatim use of goat anti-rabbit (IL and rabbit anti-goat In all three tubes, the i.8-ml supernatant loaded atop the 28-m] gradient (q.0. Fig. 3  reaction and 57% of total Atso present in anti-a reaction were precipitated.
IgG. The resulting /3:a ratio of 6:4 for mean polysome sizes is essentially the same as the prior ratio realized from coating of polysomes with 12ZI-labeled antibodies. 111 itself, such concordance is indirect proof of the capacity of immunological methods to separate chain-specific polysomes.
Finally, in Fig. 6 the same sort of fl:cu ratio for mean polysome sizes is obtained with material from a third rabbit (Animal 109) by yet another method.
Here the analysis of polysomes remaiIIing soluble after immunoprecipitation is combined with identification of (Y and 0 chains by differential labeling of nascent peptides nith [3H]isoleucine and [Wlvaline.
1 he method of differential labeling, developed by Hunt et al. (5), depends on (a) the occurrence and known position (11) of 3 isoleucine and 11 valine residues in completed rabbit a! chains, (b) 1 isoleucine and 18 valines in fl chains, (c) measurement of 3H:14C ratios in completed (Y and /3 chains, and (d) algebraic calculation for each fraction of the relative contributions of nascent (11 and @ chains to 'H radioactivity.
The result is that polysomes bearing Q! chains are marked by considerably higher *H:% ratios than are /? chain-bearing polysomes. These ratios, given the variables and manipulations just described, can be converted to Wlabeled a: andp polysome profiles. Although Hunt et al.'s method (5) is subject to errors and uncertainties in each variable (e.g. allelic variations between individuals in isoleucine (26) and valine (26,27) residues present at particular positions) and furthermore is potentially affected by trace contributions from isoleucine-rich nonhemoglobin nascent chains, it is nonetheless grossly valid, as evident from comparisons of Fig. 6 mith Figs. 3 and 5. Among untreated polysomes in the top panel of Fig. 6, the derivative (Y polysome mean size is -3.5 ribosomes per mRNA, whereas B mean size is -6.5 ribosomes per mRNA. Following removal of the majority of a: polysomes through use of anti-a and antiantibody (center panel, Fig. 6), /? mean polysome size by both AzGO and 3H radioactivity remains -6.5 ribosomes per mRNA. In the bottom panel of Fig. 6, the mean size of a polysomes remaining after immunological removal of p polysomes is -4 ribosomes per mRNA, whether judged by At60 or 3H radioactivity. Thus in the control (top) panel of Fig. 6, /3:cu mean polysome size ratio is ~6.5:3.5 = 1.86, whereas after immunoprecipitation (center and bottom panels) it is ~6.5:4 = 1.62, as measured by two different criteria (A160 and radioactivity).
Relative Numbers of cx and p Chain-synthesizing Ribosomes-The foregoing estimates of mean (Y and /3 polysome size are later used, along with estimates of the relative numbers of ribosomes, to calculate the numbers of active (Y-and &mRNA.
Relative numbers of ti and p chain-synthesizing ribosomes were deduced from Dintzis-Naughton plots (12, 13) of [aH]lysine-labeled nascent peptides (Fig. 7). In each section of Fig. 7, y axis intercepts, produced by regression slopes of relative specific activity along or&ales versus chain sequence position (12, 13) of peptide lysine residues along abscissae, represent the hypothetical sum of peptide lysine residues and thus the sum of nascent chains, i.e. ribosome-attached chains. Since each synthetically active ribosome is associated with only one nascent chain, the y intercept of each a and /3 regression line is the sum of ribosomes actively involved in the synthesis of that chain. The ratio of y intercepts (where z = 1) is a measure (5) of the relative numbers of ribosomes engaged in (Y and p chain synthesis.
However, before calculating these numbers, y intercept ratios (which depend on 3H contributions from mascent peptides and 14C contributions from uniformly labeled hemoglobin of the same rabbit) must be corrected for any (Y versus 0 imbalance in the radioactivity of the batch of uniformly labeled [WIlysine chains used as an index of recovery. Such calculations and corrections can be illustrated by reference to the Dintzis-Naughton plots shown in Fig. 7. The p:cr ratio of y intercepts for Rabbit 90 (upper panel) is 1.015. This ratio, hhen corrected for a p:a ratio of 1.153 in the uniformly labeled 14C chains used, gives a fl :a ratio of 1.17 for the relative numbers of synthetically active ribosomes.
In the lower panel of Fig. 7, the p:a ratio of y intercepts for Rabbit 109 is 0.87, which, when corrected for a /3:cu ratio of 1.19 in 14Clabeled chains, yields a /?:a ratio of 1.04 for synthetically active ribosomes.
In deriving the preceding estimates of relative numbers of cr and p chain-synthesizing ribosomes, we have assumed that the completed 0 chains, shown as stippled areas in Fig. 7, are covalently bonded to ribosomes.
'I he existence of relatively constant proportions (-10%) of completed fl chains and an attendant lack of significant amounts of completed (Y chains are universal, as inferred from the location of 5 intercepts in every one of eight Dintzis-Naughton plots. 'Ihese include two plots obtained after preparative sucrose density gradient partition of previously pelleted and rehydrated l-olysomes. Although the I f!C FIG. 6. Sucrose density gradient in an SW 27 rotor of rabbit reticulocyte polysomes from Rabbit 109 labeled with [*H]isoleu-7217 latter results effectively preclude contamination of nascent peptides with completed chains from hemoglobin tetramers (where both (Y and fl chains are found to be equally well labeled), we cannot at this time exclude the possibility that ribosome-bound, completed p chains have been secondarily attached.
If completed 0 chains are secondarily attached, it follows that they are probably attached to ribosomes which are also synthesizing a nascent chain and must therefore be excluded from calculations of synthetically active ribosomes based on y intercepts.
In later calculation of relative numbers of synthetically active a-and &mRNA (Table III), we shall use both kinds of estimates of @:a ribosome ratio, i.e. with and wit.hout the contribution of completed p chains.

Mean Sizes of /5' Chain-synthesizing
Polysomes versus Q! Chainsynthesizing Polysomes-Each of the approaches used for comparing the mean number of ribosomes attached to &mRNA, i.e. mean /3 polysome size, with mean (Y polysome size is subject to one or another kind of bias. In general, such biases are in the direction of producing underestimates in the /3:a! ratio for mean sizes. For example, /3:cr estimates of >6:4 (Fig. 3), which depend on sedimentation of polysomes previously coated with iodinated antibodies to hemoglobin chains, are prejudiced by differences in the target polysomes.
Because (Y polysomes are on the average smaller than /3 polysomes, it is expected that the proportion of (Y polysomes in the single ribosome per mRNA category will be greater than the proportion of /3 polysomes in this class. Consequently, the proportion of single-ribosome polysomes which bear a nascent chain too short to react with antibodies developed against whole chains will be relatively greater among (Y polysomes than among /3 polysomes.
To a lesser extent, the same will be true for polysomes bearing 2 or 3 ribosomes per mRNA.
1 he resulting skewed distribution of unlabeled ribosomes leads to overestimation of the mean size of QI polysomes and corresponding underestimation of the /3:a! ratio for mean sizes of polysomes.
The ratio p:a = 6:4 between mean numbers of ribosomes per mRhA, adduced by sucrose gradient sedimentation (Fig. 5) of what remains after polysome precipitation using anti-a (i.e. 0 polysomes) and uhat remains after precipitation using anti-8 (i.e. cr polysomes), is also an underestimate.
In this case, residual (Y polysomes are undoubtedly left in the anti-a! postprecipitation supernatant, whereas residual /3 polysomes are left in the anti-/3 supernatant.
1 he sources of potential underestimation of the p :CY ratio for mean polysome sizes, intrinsic to the methods used in Figs. 3 and 5, do not influence estimates (Fig. 6)  '1 his may help to explain why the p:a! ratios for mean sizes found in Fig. 6 are the largest which we observe, i.e. m-6.5:3.5 for untreated polysomes in the top panel of Fig. 6 and ~6.5:4 for composite analysis of postprecipitation remainders in the center and bottom panels. Such isoleucine-valine-dependent estimates, nonetheless, have their own kind of flaw. For example, the calculated partition of (Y and /3 polysomes in each fraction of Fig.  6 depends on prior measurement of aH:14C ratios in uniformly labeled (IL and /3 chains obtained from whole hemoglobin molecules (5). These latter measurements Eere invariably marred (in each of seven rabbits) by an aberrant cx chain which, during CM-cellulose chromatography at pH 2.7, lies under the Feak and descending limb of the /3 chain. Although this abberant (Y chain represents only -1 To of all cy chain globin, it is nonetheless richer in isoleucine residues and poorer in valine residues than /3 chain per se and thus leads to a spurious elevation of the (aH]isoleucine to ['*C]valine ratio under the nominal P-peak. The effect of this alteration on the calculated partition of counts between a and fl polysomes has undoubtedly contributed to uplifting the P-plateau underlying the otherwise isolated (Y polysome peak in the lower panel of Fig. 6. While such effects are probably too small to have much influence on the positions of (Y and /3 polysome peaks as such, they nevertheless can influence the shape of the polysome profiles.
This distortion will be especially manifest in areas u here polysomes for the two kinds of chain overlap, viz. in the 3 to 5 ribosomes per mRNA region. l3y contrast, distortion should be negligible in those regions which are largely occupied by /3 polysomes, i.e. the >6 ribosomcs per mRNA region.
The resultant skewing of /3 polysome profiles to the lighter side may lead to mean sizes of p polysomes slightly smaller than actually present and, to this degree, result in a modest underestimation of the difference between the average sizes of /3 and (Y polysomes. Despite the uncertainties which beset each method (Figs. 3, 5, and 6) for estimating mean polysome size, the resulting /3:a ratios for mean sizes are remarkably similar (-6:4 to -6.5: 3.5, i.e. 1.5 to 1.86) and, even without the real possibility of ratio underestimation, somewhat greater than the previously reported (6, 7) ratios of 1.3 to 1.4. In any case, we corroborate the mean size differences between (Y and @ polysomes which heretofore had been known (5-7) only by the indirect and partially flawed method of differential labeling of isoleucine and valine residues. 1 he immunological methods exemplified in Figs. 3 and 5 should be especially useful for studies in those species in which isoleucine residues are absent.
Relative Numbers of Synthetically Active QI-and fl-mRNA---1 he two kinds of p:cu ratios, i.e. ratios for (a) relative numbers of ribosomes engaged in synthesis of each chain (Fig. 7) and (b) mean numbers of ribosomes attached to each type of mRKA (Figs. 3,5,and 6), are used in Table 111 to calculate the relative numbers of (Y-and P-mRNA active in each of several reticulocyte preparations. In Table Ill relative numbers of (Y-and &mRKA are, for example, estimated for Rabbit 90 by substituting the relative numbers of OL Versus p ribosomes (1.17 uersus 1, upper panel in Fig. 7) and mean polysome sizes (4 ribosomes per a-mRNA and 6 ribosomes per fi-mRNA, Fig. 3 viz. ($9)/(1.17/6) = 1.28. In making this estimate we suppose that completed 0 chains (stippled areas in Fig. 7) are covalentl> bonded to polysomes rather than, as discussed earlier, secondarily bound.
If completed /3 chains are assumed to be secondarily attached, the number of ribosomes active in /3 synthesis is reduced from 1.17 to 1.08, and the (Y :p ratio for calculated numbers of active mRNA is increased to 1.39. Similar calculations are also shown (Table III) for Rabbit 109, for which the a:fi ratio for calculated numbers of mRNA ranges, depending upon assumptions, between 1.56 and 2.04. Since the relative excess of a-mRNA will increase according to the degree to which the difference between CY and /!? polysome mean size has been underestimated, we believe that the actual excess is somewhat greater than that given in Table III and larger than the 1.3-to 1.4-fold excess indirectly adduced by Lodish (6) and Lodish and Jacobsen (7).
Estimates in Table III  Whether there is an equivalent excess of cY-mRKA among inactive templates is uncertain.
The only straw in the wind is the finding of translatable rabbit cr-mRNA in supernatants remaining after reticulocytic ribosomes have been pelleted (28) ; little translatable /3-mRNA is evident in this fraction.
Such nonpolysomal cr-mRPu A represents -25% of all detectable (Y chain template (28). This contribution, when added to the roughly 1.5-fold excess of (Y-mRNA found for a composite of values in Table III, suggests that the actual proportion of a-mRNA, in all its forms, may be approximately twice that of &mRNA. We have used this presumptive 2-fold CY excess as part of a hypothesis (8) for the ways in which differential restriction to initiation of LY-and &mRNA