Evidence for heterogeneity of ribosomes within the HeLa cell.

The proteins of several subsets of ribosomal subunits isolated from HeLa cells were separated by high resolution electrophoresis in polyacrylamide slab gels containing sodium dodecyl sulfate. Comparisons were made among native subunits, and subunits derived from single ribosomes, free polysomes, and membrane-bound polysomes. At least 10 heterogeneously distributed proteins were identified among the 40 S subunit proteins, compared wtih 17 to 20 proteins that occurred in all 40 S subunit classes. All 60 S subunits shared 34 separable proteins. One protein was missing only from membrane-bound 60 S subunits and another protein was present only in membrane-bound 60 S subunits. It is suggested that these two proteins regulate the binding of 60 S subunits to the endoplasmic reticulum.


SUMMARY
The proteins of several subsets of ribosomal subunits isolated from HeLa cells were separated by high resolution electrophoresis in polyacrylamide slab gels containing sodium dodecyl sulfate.
Comparisons were made among native subunits, and subunits derived from single ribosomes, free polysomes, and membrane-bound polysomes. At least 10 heterogeneously distributed proteins were identified among the 40 S subunit proteins, compared with 17 to 20 proteins that occurred in all 40 S subunit classes.
All 60 S subunits shared 34 separable proteins.
One protein was missing only from membrane-bound 60 S subunits and another protein was present only in membrane-bound 60 S subunits. It is suggested that these two proteins regulate the binding of 60 S subunits to the endoplasmic reticulum.
One of the major unresolved problems about the nature of ribosomes is the question of heterogeneity versus homogeneity.
In bacteria, the existence of "fractional" ribosomal proteins (i.e. those not present on every ribosome) has been interpreted as evidence of dynamic heterogeneity, which may reflect distinct steps in the protein synthesis process (1, 2). Other types of evidence suggest bacterial ribosomed may be functionally differcntiated, intracellularly (3,4) or in different growth states (40). In cukaryotes, a wide variety of observations have pointed to the same possibilities (S-14), but there is no direct rigorous proof of heterogenous ribosomes in a single eukaryotic organism as yet, although the existence of fractional proteins has been suggested (15,16).
Our interest in the biosynthesis of mammalian ribosomes has involved us in a long-term study of ribosomal proteins in IIeLa cells. Although the cell populations are homogeneous, several subsets of intracellular ribosomes are well known. These are native subunits, single ribosomes, free polysomes, and bound (membrane-associated) polysomes. We thought it advisable to In the transition area of 30,000 to 40,000 the curve is somewhat uncertain, but reproducibility of results has been very good.

RESULTS
Purification of Subunits-HeLa cells growing in fresh medium have very few single ribosomes (Fig. 1A).
Even so, a second low salt sucrose gradient was usually necessary to separate native subunits completely from each other and from single ribosomes. The final stage, sucrose gradients containing 0.5 M KCl, produces very homogeneous peaks (Fig. 1B)

1313
of the 40 S subunits by co-sedimenting forms of the 60 S particle, and vice versa, has been found to be undetectable by CsCl gradient analysis (not shown) and by the absence of characteristic proteins on the electropherograms (see below). Proteins Common to All 40 S subunits-All subunits were subjected to dodecyl sulfate gel electrophoresis as described under "Experimental Procedures," with the use of 14%, IS%, IS%, and 20% acrylamide in the gels. This permitted us to detect several components that separate poorly or not at all in gels of one concentration, but are clearly resolved in gels of a different concentration.
We find 17 electrophoretically identifiable components in all 40 S subunits, regardless of the cell fraction from which the subunits were obtained.
An example of a complete gel pattern of 40 S subunit protein is given in Fig. 3. We shall refer to these 17 components as "primary ribosomal proteins" because of their ubiquity.
Molecular weights of the 40 S subunit proteins are summarized in Table I. The range of molecular weights represented by the 17 primary proteins is 17,000 to 33,700, which agrees well with the measurements of Bickle and Traut (23) on mouse cell ribosomal proteins.
Two other features of the data in Table I are worth noting.
First, the molecular weight estimates were remarkably reproducible, as indicated by the narrow range of values found for all components.
Even in the range of 30,000 to 40,000, where the slope of the calibration curve changes radically (Fig. 2), reproducibility was excellent. Absolute errors of measurement appear to be more likely in this molecular weight range than relative errors. Secondly, the resolving power of this gel electrophoresis technique is very high. Components 1 and 2, for example, differ by at most five amino acids, but they are almost completely separated.
Components 13 and 14 or components 15 and 16 differ by only three to four amino acids out of a total of more than 250 amino acids, but they are distinguishable. We conclude that very few proteins can have escaped detection in this survey, although the number is probably not zero (see "Discussion").
Proteins That Have Heterogeneous 40 S Subunit Distributions-The simplest gel pattern is shown in Fig. 3. It is characteristic of 40 S subunits from single ribosomes.
The two small bumps just above the 40,000 mark occur frequently, but are extremely variable.
Otherwise, there are no protein components that distinguish this class of 40 S subunits.
Instead, the single ribosome 40 S subunits are characterized by the simplicity of the gel pattern in the high molecular weight range; the densitometer tracing is literally at the baseline level almost everywhere above 33,000.
All four types of 40 S subunits give exactly the same gel pattern for components 1 to 11. However, the gel pattern of the native 40 S subunit proteins contrasts strikingly ( Fig. 4A) in Subunit Classes single ribosome pattern in the high molecular weight region. r\'umerous proteins larger than component 17 are present. These vary from one preparation to other, and therefore have not been assigned numbers.
Sane is quantitatively significant indivitlually, but the total constitute a substantial fraction of the protein in the sample. JTc suspect, although vx cannot prove, that this heterogcncous group of proteins consists mostly of co-scdimentirig nonribosomal materials. Native subunits arc collected from a low salt sucrose gradicnt, concentrated by centrifugatiori, and sediment& into another sucrose gradient containing 0.5 M KU.
An:-aggregates or naturally occurring groups of proteins that wcrc not disrupted by the 0.5 RI KC1 would accompany the 40 S subunits.
Some possibilities are particles containing elongation factors (24) or aminoacyl-tRSA synthetases (25,26). III addition, thcrc may be remnants of initiation factors, most of which are removed by the KC1 (27,28). Our technique is sensitive enough to detect proteins that may be present on fewer than one subunit in 20. iYo unique, major protein component is found in native subunits.
The situation is quite different for native subunits that have been exposed only to low salt buffer (29).
Components 16a to l&l occur only in native 40 8 subunit preparations.
They are very minor components, in a quantitative sense, but their occurrence is reproducible.
The esistencc of a minor subset of native 40 S subunits, the newly synthesized subunits, which contain more protein than the major variety of 40 S subunits, has been known for a long time (30) have shown that the estrw protein of newly synthesized 40 S subunits is not removed by 0.5 M KCl. Some or all of components liia to 1Gd might be that extra protein.
The fret polysomc JO S profiles (Fig. 4Bj also sho~v a large number of high molecular neight cornpone~~ts, none of which can definitely be identified with the cornponcnts of the 40 8 native subunit profile.
Four small peaks, &Agnatcd 17:~ to Ed, occur regularly in the free polgsome 40 S profiles; all higher molecular weight peaks occur erratically.
The origin 01 the variable high molecular ncight proteins cannot easily bc ascribed to co-scdirnciitilig contaminants in this case, which would have to combine the unuhual propertics of scdimenting \vith polybomes through a low-salt sucrose gradient and thtn changing to a 40 S sedimentation rate upon exposure to 0.5 AI KCI.
It is also unlikely that they rcprc,sent contaminants adsorbed during homogenization, bccaust they do not occur with single ribosome 40 S part&x.
ITc suggest that small amounts of elongation factors, anlinoacyl-tI<NA~ synthetases, and other proteiris associated with active ribosomes arc attached to the free polysome 40 S in such a way that they resist our usual purification proccduro.
This group of high molecular ITveight contaminants may bc partly responsible for the xpparcnt diffcrcncc in protein content between euknryotic and proknryotic small ribosomal subunits (31).
Two additional component s sliolild lx: notccl in tlic free polysome 40 S subunit profile; these arc tlesignated lla and 12a. They do not occur in the 40 S native subunit or single ribosome 40 S subunit patterns, but they are present in subunits from bound polysomes (Fig. 4C). Thus, they may be proteins found only on ribosomes that were active in protein synthesis at the time the cells were harvested.
It should be noted that components I la and 12a cannot be dismissed on the grounds of quantitative insignificance, although they appear as minor shoulders on peaks 12 and 13, respectively.
When proteins overlap like this, it is impossible to estimate how much of a compound peak is due to any individual component.
A different separation technique would be required to settle this question.  Fig. 4C). All three panels show densitometer tracings from the same 16% polyacrylamide gel slab. Only the higher molecular weight proteins are shown; no differences in components 1 to 11 were observed.

1315
The pattern for 40 S subunits from bound polysomes (Fig.  4C) differs conspicuously from the free polysome 40 S pattern in the absence of nearly all the proteins of molecular weight greater than 35,000. This is actually a consequence of technique.
If free polysomes are treated with deoxycholate in the same manner as the bound polysomes, the protein patterns subsequently obtained from the 40 S subunits are essentially the same as those for the bound polysome 40 S particles (not illustrated).
Apparently the detergent treatment makes the purification procedure more rigorous.
It is worth noting, however, that the occurrence and proportions of the primary 17 proteins, plus lla and 12a, are indistinguishable in the bound and free polysome 40 S subunits, whether the latter are treated with detergent or not (not shown).
Thus this group of proteins is resistant to deoxycholate, which strengthens the evidence that they are genuine ribosomal proteins.
The occurrence of component 17~ in the 40 S pattern from bound polysomes is reproducible, but its significance is unknown. 60 S Particle-All four classes of 60 S particles gave identical protein patterns on dodecyl sulfate gels, with one exception described below. Therefore, we have illustrated only one complete pattern, that of free polysome 60 S particles (Fig. 5). We find 35 regularly occurring components, in good agreement with the results obtained by two-dimensional electrophoresis of rat liver 60 S ribosomal proteins by Sherton and Wool (32) and others (8,9,33,34). The uniformity of this pattern implies that none of the numbered proteins is a co-sedimenting contaminant, but it does not rule out the possibility that some of these proteins are ubiquitous, adsorbed contaminants that are resistant to 0.5 M KCl. If this occurs, it cannot involve many proteins.
For the present, we shall consider all 35 components as "primary" proteins of the large subunit.
The sole evidence of heterogeneity among 60 S subunits that we have found is illustrated in Fig. 6. The particles from bound polysomes lack component 35, and contain a component, 31a, not found elsewhere.
The illustration shows tracings obtained from 14% acrylamide gel patterns, which present the difference most clearly.
Components 32 and 34 are not distinguished on such gels; we are not certain whether they occur in bound polysome 60 S subunits.
We have demonstrated, however, that treatment of free polysome 60 S subunits with deoxycholate does not alter the gel pattern of the proteins in any detectable way (not illustrated).
In particular, components 32 to 35 remain in exactly the proportions illustrated in Fig. 5, and component 31a does not appear.
Molecular weights of the 60 S subunit proteins are given in  require no special comment.
Components 31a to 35 behaved in an unusual manner; their apparent molecular weights increased as the concentration of acrylamide in the separation gel increased. Component 31a, for example, migrated in 147, gels as though its molecular weight were 36,400, in contrast to the estimate of 43,500 obtained from 20% gels, given in Table II. The comparable numbers for component 35 were 41,000 in 1470 gels and 58,800 in 20% gels. This type of electrophoretic behavior suggests some unusual feature of the composition of these proteins (22), and this possibility is considered under "Discussion." DISCUSSION We have compared the proteins found in four major intracellular subsets of ribosomes: native subunits, single ribosomes, free polysomes, and membrane-associated ('.bound") polysomes. Seventeen electrophoretically separable components are found in all four classes of 40 S subunits, and these are designated "primary" ribosomal proteins.
It is probable that several ot the electrophoretic components include more than one protein (e.g. 5, 7, and 15). Thus, there are approximately 20 primary proteins in the KeLa cell small subunit, essentially the same as the number of Escherichia coli 30 S subunit proteins (1, 35).
All classes of 40 S subunits are not identical, however. Co- of proteins from 60 S subunits The numbers given for average molecular weights and range of molecular weights refer to all classes of 60 S subunits and to data obtained from 14$&, 16%, IS%, and 207, sodium dodecyl sulfatepolyacrylamide gels as described under "Experimental Procedures." The exception is that only data from 20y0 gels are given for the proteins marked with an asterisk (see text).

Protein
Average Range -scdimcnting or loosely bound 'Lcontaminants" complicate the patterns for the native and free polysomc subunits.
These may be, in part, remnants of salt-sensitive populations of proteins functionally associated with ribosomes, such as initiation and elongation factors. A more distinctive type of heterogeneity is represented by a variety of minor components, such as 16a to 16d in native 40 S subunits, 17a to 17d in free polysome 40 S particles, and lla and 12a in both free polysome and bound polysome 40 S particles.
The functional significance of these cornponents cannot be assessed at present, but operationally they show that the four classes of ribosomal subunits studied here are distinguishable in terms of their minor proteins. It should also be noted that the total number of 40 S subunit proteins revealed by this analysis is approximately 30 (I7 to 20 primary proteins and 10 heterogeneously distributed proteins), which agrees well with two-dimensional electrophoretic studies (e.g. Ref. 32).