Expression of the Silent Hemoglobin Gene in Sheep STUDIES Op‘ THE MESSENGER

Abstract We have analyzed the messenger ribonucleic acids from reticulocytes of adult sheep carrying different globin genes and synthesizing correspondingly different hemoglobins. Hemoglobin A (α2β2a) is synthesized by A/A homozygotes in the early stages of anemia, whereas these same animals switch over to hemoglobin C (α2β2c) synthesis when they are severely anemic. On the other hand, only hemoglobin B (α2β2b) is synthesized by normal or by anemic B/B homozygotes. All four types of sheep globin messenger ribonucleic acids (i.e. for α, βa, βb, or βc chains) are successfully translated in a protein-synthesizing lysate from rabbit reticulocytes. The newly synthesized radioactive sheep globin chains are separated from the rabbit globin chains by chromatography on a carboxymethylcellulose column. The results suggest that a switch from βa to βc messenger ribonucleic acid accompanies the development of anemia in A/A sheep. However, the rabbit reticulocyte lysate system synthesizes sheep globin chains in different proportions than they are synthesized intracellularly; in the rabbit lysate, α and βa chains constitute a lower percentage and βc chains a larger percentage of the synthesis. The method of Lodish (Lodish, H. F. (1971) J. Biol. Chem. 246, 7131) was used to quantitate the relative levels of α, βa, and βc messenger ribonucleic acids in reticulocytes from an A/A homozygote. The proportion of βc messenger is approximately twice as high as the proportion of βc synthesis in sheep reticulocytes. On the contrary, the proportion of α messenger is lower than the proportion of α chain synthesis in sheep reticulocytes.

is synthesized by A/A homozygotes in the early stages of anemia, whereas these same animals switch over to hemoglobin C (o(~&') synthesis when they are severely anemic.
On the other hand, only hemoglobin B ((y2PZB) is synthesized by normal or by anemic B/B homozygotes.
All four types of sheep globin messenger ribonucleic acids (i.e. for o(, PA, /3", or PC chains) are successfully translated in a protein-synthesizing lysate from rabbit reticulocytes. The newly synthesized radioactive sheep globin chains are separated from the rabbit globin chains by chromatography on a carboxymethylcellulose column.
The results suggest that a switch from /3" to PC messenger ribonucleic acid accompanies the development of anemia in A/A sheep. However, the rabbit reticulocyte lysate system synthesizes sheep globin chains in different proportions than they are synthesized intracellularly; in the rabbit lysate, o( and PA chains constitute a lower percentage and PC chains a larger percentage of the synthesis.
The proportion of /3' messenger is approximately twice as high as the proportion of PC synthesis in sheep reticulocytes.
On the contrary, the proportion of 01 messenger is lower than the proportion of o( chain synthesis in sheep reticulocytes. Domestic sheep and r&ted species have an unusual polymorphism at the g , r11e loc*us which specifies the P-chain of hemo-* This work was supported 1)~. United States Yllblic Realth Service Grants CA-11347 alltl JIL-CA-13060-04. Marian Baldy was supported by a postdoctoral !'elloMship from the United States Pltblic HexIt h Service. A preliminary report of this work was presented at the 25th annllal meeting of the Society of (ieneral Physiologists (srlhsession on Hemoglobin Synthesis), Woo&s TTole, l\Iass., September, 1971. 1 To whom al1 corl,espolltlerlce should be addressed.
Alt,hough shrclp have 011ly olw kllowvll cy chain gene, there arc two common alleles: ,4 inld I{, of the /3 chail, locus. The A and Ii alleles are quite dissimilar since the fi.1 and 0" polypcptides differ by seven amino acid rc~placemcnts (6). Furthermore, when adult sheep with the .\ all& arc n~atlc anemic, they cease producaing p,' chains and they switch 011 the synthesis of a different type of chain, the /3" chain. 011 t1w contrary, expression of the 1% allele is unaffcctcd by allenlix. Moreover, the PC chain differs markedly from p,' and fl" chains both in its amino acid sequence and in its length (6). We will refer to this switch from fi" to /3" synthesis in anemic sheep as "silent gent" expression.
The mechanism responsible f'o~, this switch in /3 chain synthesis has not been clucidatcd.
This paper describes some of our studic>s of tllc, control of sheep hemoglobin synthesis.
Wc have tried to Icarn whethc~r the fl^ ---f PC synthesis swit('h in anemic shec1) is :Lc*c9rnpaliicd by a corresponding switch in types of mRN.\ prcscxllt in reticulocytes, or whether the mRS.\ remains constalrt during thr synthesis switch.
Our data suggest that :I switch from p" to @( mRNA does accompany anc>mia in A/-\ homozygotrs. Ilow ever, shcrp hemoglobin synth&s is also c*ontrollctl at the translational level, since the different globin mRN':\s arc translated intracellularly with very different efhcicricics. Furthcrmorc~, silent gene expression is a~c*ompanied by c*hangc~ in t,hc prc,portions of scvcral tRS;\s (7). Preparation and Labeling OJ Sheep ZZeticulcc!/les-~~(lult \Yillamettc sheep w+hing 60 to T2 kg wcrc matl~~ atlemic and were bled according to the I'ollowing schedule. ISlood was always collected into heparin and was kept chilkd at 0" until used. 011 1)ay 0, 300 ml of blood were collcctcd alid the allimals were injected subcutaneously with 10 mg per kg of ph~ylhydrazine.
One milliliter of the packed cells n-as routin+ rc>suspendcd at 1 x log cells per ml in nutrient mc,dium as describ(,tl by I-Iori and Rabinovitz (8). 'I'hr, ccl1 suspension was swirled gently ilr :I water bath at 37" for 10 min bc9'ore th(t addition of 2 PCi per ml of :I uiist,urc of uniformly lY~-labr~l~d L amino acids (New Engluirtl Diuclcar Corp., SEC-445).
Ii1 sonic csl)eriments, various coiicrlitrntioirs of cycloheximidr (Calbiochem) were added to aliquots of the cell suspension to inhibit protein synthesis. A2ftcr 60 min of incorporation, the cells wcrc chilled by dilution with 4 volumes of cold physiological salt solution. The cells \vrre lyscd with 1 volume of water and 0.4 vrolume of toluene. After ccntrifugation at 10,000 X g for 30 min, the hcmolgsate was dialyzed thoroughly with 0.01 M r\'aCl.
The gels were stained with boirzidine (IO). Electrophoresis of the hemoglobins in polyacryhrmide gels was by the method of Smith and Evatt (II). 'I'hc unstaim~d lrolgacrylamide gels ww scanned for absorbance at 540 nm OII a gel scanner in a Gilford model 222 spectro-l~hotomttcr.
This method resulted iu a good separation of hc~moglobins -1, R, and C and allowcd an accurate (&5'i;) and ral)id quaiititntivc analysis of their lnoportions. ('hromatographic separation of shc~cp globin chains on a ca~~l~os~methylccllulose column was carried out at room temperaturn, by a modification of the method of .2dams et al. (12). Globin (75 mg) was prepared from labeled sheep hemoglobins with cold acid acetone according to the procedure of Anson :IIK~ Mirskg (13). It was redissolved in Ruffer A (8 M urea, 0.05 M P-~nercaptocthanol, 0.00125 M NasHPOd, adjusted to pH 7.0 with II31'04) :urd was dialyzed overnight at 4" against the satnc buffer. L1 stock solution of 8 M urea was deionized just More use by passage through a miscd bed ion exchange column (Rio-Rnd, XG 501-X81), 20 to 50 mc~sh); the coilductivity of the olutrtl urea solution was always less thaii IO-" ohm-'.
When the globin chains were labeled with both 3H a~rd 14C, counting channels were c,nilrloyed in which the $H window contained 14 'i;, r*C comit:: and the W window contained essentially no 31-I counts.

Preparation oj Sheep Globin Jlessenger
Ribonucleic Acids-Slrcrp rct,iculocytes which had bcc11 washed three times with cold physiological salt solution were lysed by the addition of an c~qual volume of lysis buffer (2 IIIM Mg('ly, 1 111~ dithiothreitol, a~rtl 0.1 nihr T<I)T*1, 1rH 7.5). The lysate was centrifuged for 10 miir at 30,000 x g. The sediment was estracted with a volume of lysis buffer equal to that of the packed cells originally used, and tlie estracst was again ccntrifugcd for 10 min at 30,000 X g. The two 30,000 x g supernat'ants were combined and were ccntrifugcd at 105,000 X g for 3 hours in the Spinco 30 rotor. 'l'hc pellet, which contained, ribosomcs and some more slowly sedimcnting materials, was used for the preparation of mes-sc~~grr RNA.
The pellets were dissolved in a buffer containing El>TA, and the ribonucleoprotein parti& that coiltains 10 S nicssenger RIY.1 was separatrd from ribosomal subunits on a su(~rose density gradient (14). The Rx.1 was then precipitated by adjusting the appropriate fractions (srdirnrliting betvveen 10 and 30 S) to 0.15 hr NaCl and to 0.5r; sodium dodecyl sulfate, folloncd by the addition of 3 volumes of ethanol.
-1fter 2 hours at -2O", the precipitate was collectrd by centrifugation; it was dissolved in buffer containing sodium dodecyl sulfate and was extracted with pherlol-chlorofol,nl, as dcscribcd by Penman (15). lifter precipitation with ethanol, the RD;:L was washed at least three times with ethauol at -20" to remove detergent, was dissolved in water at a concentration of 1 mg per ml, and was stored frozen at -70".
This method gave biologically active messengers for all of the sheep globin chains (see below). Although less thorough than the methods which have been used to Innify rabbit and human globin mRS.1 (16-a]), this method has the advantage that it does not require assuming that all of the sheep globin mRNAs have precisely the same sedimentation cocfhcient, an assumption which would have been unsuitable for this project.
The present method was considered adequate because impure mRNA preparations have becu successfully assayed in the rabbit reticulocyte lysatc (22) and because further purification would have resulted in a lowering of giclds.
Cell-free Protein Synthesis in a Lysate from Rabbit Reticulocyfes--'l'hc methods used for preparing the rabbit reticulocyte lysate and for the incorporation of radioactive amino acids into hemoglobin arc described elsewhere (16, 23). We employed 0.4 ml of the lysate per ml of the cell-free iiicorl)oration mixture. Incorporation of L- [4,lcucine (40 PCi per ml, 690 PCi per PM) was done in the absence of added messenger RSA, whereas a l)arallel incorporation of uniformly labrltd Q4C]leucine (5 &'i l)er ml, 259 PCi per PM) was done in the prcscnce of 25 pg of ~rIcsscrrger RXA. Following iricoi~l~oration for 90 min at 29", the two incorlroration tubes were mixed and 60 mg of nonradioactive carrier sheep hemoglobin were added as described for each experiment.
The lrreparation of globin, the separation of globin c~hains on a cal,bos?-nlrt,h?cellulosc column, and the methods for 31I am1 1% radioact,ivity measurement arc described above.
Iii slieel), the rcsultirig severe aucmia is accompanied by a complctc replac~cmcnt of Hb ' A (a&') by IIb C (c&c). A t _ 8' 1 :I y pic~i runnplc of this replacement in an .1/X homozggote is shown in the starch gel electrophoresis analysis in Fig. 1. The hcmglobins ar(' stained with bcnzidine. We have obscrvcd very similar kinetics for the switch from p* to 0" chains in four A/.\ aud iu one :1/l< shc~~p (e.g. see Referemrs 5 a1~1 i).
Following a lag of 6 to 7 days, Hb C appears in the blood.
Hb 9 disappears very ral)idly after this time because it is no longer synthesized aird because the phenylhydrazinc: is causing ver>-rapid destruction of circulating erythrocytrs. illustrates that Hb C stains more weakly with benzidine than the same amount of IIb A.
The normal sheep, bled before onset, of anemia, had hemntocrits of 50%, contained few reticulocytes (less l(r,), and their blood was inactive in incorporation of radioactive amino acids into hemoglobin.
However, by Day 7 of anemia, the hematocrits had fallen to 25 to 30% and the blood contained approximately 15yG reticulocytes.
On Day 21, hematocrits were between 13 and 20% and the blood contained approximately 90(,> reticulocytes. Translation of Sheep Globin dlessenger RNA in a Rabbit Reticulocyte Lysate-We have assayed the sheep reticulocyte mRNA in a protein synthesizing lysate from rabbit reticulocytes (see "Materials and Methods"). One aliquot of the rabbit reticulocyte lysate system was incubated with [3H]leucine, and a second aliquot was incubated with both [*4C]leucine and with sheep reticulocyte mRNiZ.
The two aliquots were then mixed with each other and with nonradioactive carrier sheep hemoglobins.
The resulting mixture of materials was then converted to globin and was analyzed by chromatography on a carboxymethylcellulose column.
A typical analysis using mRNA from a B/B homozygote is shown in Fig. 2. The optical density tracing at 280 nm shows the positions of elution of the unlabeled sheep LY and ,@ chains which were added as carrier.
The rabbit (Y and p chains cochromatograph as a single peak which elutes slightly ahead of the sheep a chains. Generally, there is no clear separation between the rabbit globin chains and the sheep a: chains. The aH and 14C radioactivity profiles are also plotted. Clearly, there is a peak containing 14C label which co-chromatographs with the carrier sheep PB chains. On the contrary, 3H label is relatively absent from this region of the eluate. This 14C peak was absent when sheep mRNA was not added to the rabbit lysate system. We conclude that pB mRNA from the B/B sheep was successfully translated in the rabbit reticulocyte lysate. As expected, the lysate system also incorporated a relatively large amount of both t3H]leucine and [14C]leucine into the runoff proteins which elute rapidly from the column and into the peak which contains the rabbit a and /3 chains. Fig. 3 shows the results obtained when the messenger RNA was from a severely anemic A/A sheep. The reticulocytes used as the mRNA source were synthesizing only 0" and Q chains, but were not producing any p* chains. The optical density tracing shows the positions of elution of the carrier @A, PC, and 01 chains. Clearly, there is a peak containing 14C label which co-chromatographs with the carrier PC chains, but there is no such peak which co-chromatographs with the carrier p* chains. Furthermore, the material which co-chromatographs with the carrier PC chains is relatively deficient in aH, suggesting that its synthesis was caused by the added sheep messenger RNA.
We conclude that the PC mRNA from the severely anemic A/A sheep was successfully translated in the rabbit reticulocyte lysate. Fig. 4 shows the results obtained when the sheep messenger RNA was from an A/A homozygote at an earlier stage of anemia; the blood was taken on Day 7 of anemia.
As will be shown below, the reticulocytes which served as the mRNA source were synthesizing ,P chains as well as PC and cx chains. As can be seen from the optical density tracing, the separation of globin chains was especially good on this column and a resolution of the sheep cy chains from the peak of rabbit o( plus /3 chains is clearly apparent.
There In the control cells (Fig. 6a), the relative incorporation ratio of flr:fl.&:cr is 1.0:0.32:1.27, ~l~~rcas in Fig. 6b  is the same as in Fig. 6, except that t,he sheep trwtl been anemic for 21 days. The reticulocytes were labeled in the absence (a) or presence (h) of 0.5 pa per ml of cycloheximide. 0 -0 0.1~. 280 nrn; A---A, I'C. nrR?S.\ than fir mRNh, although the 0-l mRNA was translated rcllatively efficiently in the control ~11s. Furthermore, the ~11s contained considerably less cx rnKX.1 than /3 mRNA. A\ similar experiment, using cells from a later stage of anemia is sl~own in Fig. 7. Th r relative Icvcl of incorporation into OL vhwitrs is markedly reduced in tllr I)resenc'e of cyclohesinlide. III the control, the incorporation ratio is /3':0( equal to 1.0:0.68, whereas in 0.5 pg per ml of cyclohcsimidc the ratio is 1 .O :0.51, rcsprctivcly. 111 agreement with Fig. 6, these data suggest that sheep reticulocytes c*ontain relativcxly less mRNA for a! chains tjlim for /3 chains, an d that tlrc 01 rrrRSA is more efficiently translated.
In all experiments, the types of globin chains made in the cell-free system were the same as were synthesized int~racellularly in tile sheep reticulocgtes used as the rnRKh sour('e. Thus, the type of p chains programed by mRX.2 from X/A sheep changed during the course of anemia. hIessengcr RNA taken early itI auc,mia programed the synthesis of both fl* and 8' chains (Fig. 4), whereas that taken later in 6669 anemia programed the synthesis of only /S chains (Fig. 3). This sugges;ts that a switch in fi chaiii mRSA accompanies silent gene expression in sheep. An alternative line of data which supports this conclusion is discussed below.
However, some problems with this interpretation of the cellfree synthesis data must be considered.
Conceivably, nonutilized mRNA might occur in the sheep cells in an "inactive" form. If such mRKC'h was labile to our extraction procedure or if it was bound to cell membranes or mitochondria, it would have escaped our detection.
Unfortunatrly, we have not succeeded in extracting active sheep globin mRN;\ directly from intact cells. The subcellular fraction we have utilized for mRNA extraction contained all of the cellular ribosomes and approximately 407; of the globin mRK\;h-protein complex which sediments in the "informosome" region at about 20 S (30-32) (see "Materials and Methods").
Our data suggest that no mR?;h for 0" chains was l)resent ill this subcellular fraction isolated from severely anemic A1/X sheep (Fig. 3). Nonetheless, it is obvious that this type of objection can never be fully excluded.
In any case, we believe that our mRNA purification method makes fewer rest,rictive assumptions about the globin rnRnTA size and sedimentation properties than others (16-21) which have been previously used.
-Uthough the rabbit reticulocyte lysate can synthesize all types of sheep globins, the relative synthesis levels differ from those made intracellularly.
Several possible explanations could account, for this result.
(a) The various mRK-\s were isolated in different yields.
(b) The mRNAs were isolated in equal yields but were translated with different efficiencies in the rabbit lysate.
(c) The mRNXs Tvere translated with different efficiencies in the intact sheep reticulocytes.
Although we cannot exclude any of these possibilities, t'he dat'a discussed below support the conclusion that, the third erplanntion is one contributing factor.

Evidence
for DiJ'erential Translaiion oj Sheep Dlobin mRNA in Sheep Reticulocytes-Olle possible explanation of silent gene cxpression is that the regulntiou occurs solely at the translational level, i.e. that both PA and /3" mRNX are present in roughly equal amounts in normal or in anemic animals and that a selective translational process is responsible for thr PJ' + PC synthesis s\vitcBh. According to this csplanat~ion, fit\ rnRNA would be translated relatively efficiently in normal animals and rclativcly inrfficiently in heavily anemic animals; the l)rol)ortion of synthesis of either chain would be prol)ortional to the relative efficiency of translation of its mRNA%.
Our data with cycloheximidc (Figs. 6 and 7) argue strongly against this interpretation.
For example, /3' mRKA is translated only about one-half as cfficirntly as p-4 mKNA, even when the /3":@" synthesis ratio is 1.0 :0.32 (Fig. 6). :\ccording to the translational regulation hypothesis, the 6" mRSX should have been more efficiently translated in these cells. In addition, these data suggest that the /3* and /3(' mRNhs occur intracellularly in grossly different amounts which change during the course of silent gene expression.
We have 1~0s used this method to analyze the development of anemia in two different A/A sheep, and the results imply strongly that the proportion of PC mRNh increases during anemia.
In this respect, the conclusion from these studies is the same as that arrived at frorn the cellfree protein synthesis data.
However, these data do suggest that the different sheep globin 1nRiT~1 are translated intraccllularly with very different efficiencies.
,Uthough present in relatively low levels, the cx mRXh is relatively efficiently trallslated. When protein by guest on March 24, 2020 http://www.jbc.org/ synthesis is inhibited with cyclohesimide, there OCCUE a reduction of the o( :/3 irrrorporation ratio (Figs. 6 and 7). As can bc seen from Figs. 5 to 7, we have observed that the level of incorporation into o( chains is frequently much less than that into fl chairrs, even in the absence of cycloheximide. This disparity is heightened by cycloheximide. One possible esplanation is that, prot,ein synthesis in our conditions of cell incubation may be already suboptimal, even in the absence of added inhibitors of protein synthesis.
The (Y:P synthesis ratio is likely to bc sensitively dependent upon the conditions of the cells and upon the incubation media. Clearly, the i?z viz10 synthesis ratio may be quite different from the ratio obt'ained by measuring incorporation of radioactive amino acids by cell BORDI,E~, J., & N.\T;GHTON, ill. in reticulocytrs. This implies that the regulation occurx at suspensions.
the level of trarrscription of the genes to form mRXA, at the level of processing the rnRXA prior to its utilization, or at the preceding qualification, our data suggest that silent gene es-On the other hand, the synthesis of sheep globins is also regulated at the translational level.
pression is accompanied by a switch in types of mR1;-\ present The different mRNXs are translated intracellularly with different efficiencies. For example, the (Y mRNA is present in low amounts but is relatively efficiently translated. .% similar conclusion was made by Lodish with rabbit reticulocytes (24), except in that cast the /I mRNA was more efficiently translated.
We have also slIowrI that the relative proportions of different tRN-!.s change during silent gene expression (7). These studies are consistent with the view that the relative levels of synthesis of different proteins may be influenced at several different levels of regulatiorr.
In the case of silent gene expression our data suggest that the prirnary regulation occurs at the level of production, processing, or degradation of mRN;Z and that the translational effects occur secondarily to modulat'e the final result.