Age-dependent Regulation of the Polymorphic Forms of a2u-Globulin*

Hepatic synthesis of a2,-globulin in the male rat shows a gradual decline and ultimate loss during aging and senescence. Northern blot analysis with a cloned cDNA probe showed that the decrease in az,-globulin synthesis during aging is associated with a corresponding decrease in the concentration of its hepatic mRNA. By means of two-dimensional gel electrophoresis, aZU-globulin is resolved into a family of proteins. A mono- clonal mouse antibody can identify at least five major isoelectric variants of aZu-globulin within the total pro- tein synthesized by rat hepatocytes. These isoelectric variants are also present in the in vitro translation products of hepatic mRNA. An examination of the hepatic synthesis of the isoelectric variants of az.-glob-ulin during aging showed a differential regulation of the variant forms of this protein. Variant 2 (PI 6.1, the most prominent form) is the first to appear at puberty (40 days). The weakest member of the five major iso- electric forms (variant 4, PI 4.1) is the last to disappear at senescence. Although the overall decline in a2,-glob- ulin synthesis during aging seems to be due to an age-dependent decrease in the androgen responsiveness of hepatocytes, it is postulated that the differential regulation of the isoelectric variants may represent changes at the level of the genes coding for this protein. The mammalian liver sexual More 80%

Hepatic synthesis of a2,-globulin in the male rat shows a gradual decline and ultimate loss during aging and senescence. Northern blot analysis with a cloned cDNA probe showed that the decrease in az,-globulin synthesis during aging is associated with a corresponding decrease in the concentration of its hepatic mRNA. By means of two-dimensional gel electrophoresis, aZUglobulin is resolved into a family of proteins. A monoclonal mouse antibody can identify at least five major isoelectric variants of aZu-globulin within the total protein synthesized by rat hepatocytes. These isoelectric variants are also present in the in vitro translation products of hepatic mRNA. An examination of the hepatic synthesis of the isoelectric variants of az.-globulin during aging showed a differential regulation of the variant forms of this protein. Variant 2 (PI 6.1, the most prominent form) is the first to appear at puberty (40 days). The weakest member of the five major isoelectric forms (variant 4, PI 4.1) is the last to disappear at senescence. Although the overall decline in a2,-globulin synthesis during aging seems to be due to an agedependent decrease in the androgen responsiveness of hepatocytes, it is postulated that the differential regulation of the isoelectric variants may represent changes at the level of the genes coding for this protein.
The mammalian liver shows a considerable degree of sexual dimorphism and is recognized increasingly as an important target organ for sex hormones (1). More than 80% of the liver tissue is composed of one cell type, i.e. hepatocyte, which is a reverting postmitotic cell that normally survives as long as the animal itself. Thus, hepatocytes can serve as a convenient cellular model for the study of changes in endocrine responsiveness during aging (2). It is also generally believed that the age-dependent decrement of cellular function involves an interaction between the intrinsic genetic program and various extrinsic cellular and extracellular regulatory factors among which hormones play a predominant role (3,4). The molecular basis of these functional changes is not clearly understood and many theories on aging have been proposed ( 5 ) .
In the differentiated postmitotic cell, only a certain part of the genome is normally utilized for transcription and expression of some of these genes shows considerable changes during aging (6). Investigation of the age-dependent changes in the synthesis of these "senescence marker proteins" is expected to provide important information concerning the molecular basis of aging. In our laboratory, we have been interested in studying the effect of aging on the androgen-dependent syn-* This investigation was supported by National Institutes of Health Grant AM-14744. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. thesis of a male urinary protein called a2,,-globulin (7). In this article, we describe multiple isoelectric forms of a,,-globulin and the effect of aging on the regulation of the various polymorphic forms of this protein.

MATERIALS AND METHODS
Animals and Hormone Treatment-Albino rats of the Fisher strain (F-344) were obtained from the colony of the National Institute on Aging, maintained at Charles River Breeding Laboratories (Wilmington, MA). These animals generally attain puberty around 40 days of age and are considered at a state of senescence beyond 650 days. The animals were housed in an air-conditioned animal room with 12 h of light and 12 h of darkness and were kept on observation for 10 days before any experimentation.
After they were killed, the internal organs were carefully examined and experiments were terminated if any sign of tumor or other abnormal changes were detected.
Primary Culture of Rat Hepatocytes-Liver cells were cultured in monolayer by the modified collagenase perfusion method as described before (8). After dispersion, cells were cultured for 6 h in Eagle's minimum essential medium, subsequently placed in a methioninefree minimum essential medium, and labeled for 2 h with ["SI methionine (1200 Ci/mmol; New England Nuclear, Boston, MA; 100 pCi/flask). The labeled cells were harvested, washed, and immediately lysed in a buffer containing 9.2 M urea for analysis by two-dimensional gel electrophoresis (9).
Two-dimensional Gel Electrophoresis of Hepatic Proteins-Twodimensional gel electrophoretic separation of hepatic proteins was performed following the protocol of O'Farrell (9). The hepatocytes were lysed with the sample buffer and labeled proteins (lo6 cpm of radioactivity) were separated in the first dimension by isoelectric focusing on a cylindrical gel. Separation in the second dimension was performed on a 12% polyacrylamide slab gel containing 0.1% sodium dodecyl sulfate. After electrophoresis, slab gels were treated with ENHANCE (New England Nuclear), dried, and fluorographed for 7 days a t -70 "C. "C-labeled molecular weight markers (Bethesda Research Laboratories, Gaithersburg, MD) were included during electrophoresis in the second dimension. The isoelectric points of various hepatic proteins were determined in reference to a mixture of proteins with known PI (United States Biochemical Corp., Cleveland, OH) which was labeled with Iz5I. The '2sII-labeled pl markers were added to hepatic proteins ("S-labeled) and were analyzed by two-dimensional gel electrophoresis followed by autoradiography. In order to avoid variations that could result from differences in the uptake of labeled methionine by different batches of hepatocytes, for the purpose of comparative analysis between two experimental groups, the same amount of radioactive proteins was used for gel electrophoresis.
Western Blotting of Proteins on Nitrocellulose Papers-Total protein from the cultured liver cells of adult male rats, separated by twodimensional gel electrophoresis, was transferred from the polyacrylamide slab gel into a nitrocellulose filter by the procedure of Burnette (10). Immunoreaction was also carried out according to the procedure described by the same author. The gist of the procedure is as follows.
The nitrocellulose filter was first incubated a t 37 "C for 30 min with a buffer containing bovine serum albumin (10 mM Tris-HC1, pH 7.4, 0.15 M NaC1, 50 mg/ml of bovine serum albumin, and 0.2% v/v Nonidet P-40). The paper was then incubated a t room temperature for 120 min with a mouse monoclonal antibody for cus,-globulin in the presence of the albumin-containing buffer (characteristics of this monoclonal antibody will be published elsewhere). Subsequently, the paper was extensively washed a t room temperature. The washed paper was incubated for 30 min with 1251-labeled anti-mouse goat IgG (10' cpm). (The anti-mouse IgG was purchased from the Cappel Labora-Regulation of the Polymorphic Forms of cu2,-Globulin during Aging tory, Cochranville, PA.) The paper was then washed, air-dried, covered with a plastic film (Saran Wrap), and autoradiographed at "io "C.

Extraction o f Hepatic m R N A and Northern
Analysis of crz,,-Globulin mRNA-Total hepatic RNA was extracted with phenol-sodium dodecyl sulfate according to the procedure of Rosenfeld et al. (11). Poly(A)-containing hepatic mRNA (20 p g ) , isolated by affinity chromatography on oligo(dT)-cellulose, was electrophoresed on a 1.5% agarose slab gel containing 10 mM methylmercury hydroxide. Electrophoret ically separated RNA was transferred to diazohenzyloxymethylcellulose paper (12). The paper was subsequently hybridized to a '"P-laheled cr2,,-glohulin cDNA prohe according to Wahl et al. (13) and autoradiographed on Kodak x-ray film.
('ell-frcc Translation of Rat Liver Messenger R N A and Immunochemical Identification of tr?,-Glohulin-In vitro translation of messenger RNA was performed in a rabbit reticulocyte lysate in the presence of dog pancreas microsomal membrane with [:"S]methionine as the laheled amino acid. The dog pancreas membrane was prepared according to Blohel and Dohherstein (14). The translation products (10" cpm of protein radioactivity) were incubated for 15 min at room temperature with 5 pI of the mouse ascites fluid containing monoclonal antihodv for tr?,-glohulin. The antigen-antibody complex was ahsorhed on IgCsorh (The Enzyme Center, Boston, MA) and was pelleted hy centrifugation a t 10,000 X g for 5 min. The pellet was washed several times with a solution containing 10 mM sodium phosphate, pH 7.6, 0.14 M NaCI, 1% (v/v) Triton X-100, and 10 mM unlaheled methionine. The pellet was then dissolved in the sample huffer of O'Farrell (9) containing 9.2 M urea and analyzed by twodimensional gel electrophoresis and fluorography.

Sucrose Iknsity Gradient Analvsis of the Cytosol Androgen Binding I'rotein o f Male Hat
Liwr-After cervical dislocation of the animal, the liver was immediately flushed via the portal vein with 50 ml of 0.15 M NaCl at 4 "C. The perfused liver was cut into small pieces and homogenized in a solution containing 10 mM Tris-HCI, pH 7.5, 15 mM disodium EDTA, 6 mM dithiothreitol (1 ml of huffer/g of tissue) (1.5). Aliquots (100 p l ) of the cytosol were incubated at 0 "C for 60 min in glass tubes containing tritium-labeled 5cu-dihydrotestosterone (Amersham/Searle). The labeled cytosol was fractionated through a .5-20rE linear sucrose density gradient (centrifuged for 22 h at 47,000 rpm and 4 "C, using a Beckman SW 50.1 rotor).
Radioimmunoassay o f cr~.-Globulin-Radioimmunoassay of cy?,,globulin was performed according to the procedure described earlier (16). Total protein in the liver cytosol was determined according to Lowry e t a/. (17) and the results were expressed as nanograms of trsuglobulin per mg of hepatic protein.

Changes in the Hepatic Levels of a,,,-Globulin and Its mRNA
during Aging-In the normal male rat, as,,-globulin begins to appear at puberty, reaching a peak level at about 100 days of age, and its hepatic synthesis ceases at senescence. Fig. 1 shows changes in the hepatic content of n,,,-globulin during aging of the male rat. Approximately a 50% reduction in the hepatic level of tu-,,-globulin is observed a t 600 days of age as compared to 100-day-old rats and synthesis of this protein almost ceases at late senescence ( X 0 0 days).
Northern blot of the electrophoretically separated total hepatic mRNA from 75-, 750-, and 900-day-old males, hybridized with a cloned cDNA probe for a,,,-globulin, shows that the decline in the hepatic synthesis of as,,-globulin during aging is associated with a corresponding decrease in the hepatic content of its mRNA (Fig. 2). Although the hepatic tissue of the 750-day-old male rat contains some a,,,-globulin mRNA, the liver of the 900-day-old male rat is almost devoid of this mRNA. Results obtained through this semiquantitative but highly sensitive technique are in agreement with the conclusion that a decrease in the hepatic synthesis of a?,,globulin during aging is a general reflection of the corresponding decrease in the hepatic concentration of n,,-globulin mRNA.
Electrophoretic Separation and Immunochemical Identification of the Polymorphic Forms of n2,-Globulin- Fig. 3 shows the electrophoretic distribution of labeled hepatic proteins synthesized by hepatocytes derived from a 70-day-old male rat (young adult). Comparison of this protein pattern with a similar autoradiogram of labeled proteins synthesized by the hepatocytes obtained from a 70-day-old female rat (data not shown) shows several male-specific proteins, which are circled in white and pointed with arrows. Results that follow demonstrate that the cluster of male-specific proteins migrating between the 18-and 21-kilodalton range (white circles mithin the boxed area in Fig. 3) shows immunochemical reactivity with antibody against tu,,-globulin. Five major isoelectric variants of tu,,-globulin are numbered according to their decreasing isoelectric points (PI 7.8, 6.1, 4.9, 4.1, and 3.7). The different molecular weight variants of a,,-globulin have been described in an earlier publication (18).
The presence of various isoelectric forms of n,,,-globulin within the in uitro translation products of the hepatic mRNA was examined by immunoprecipitation of labeled tu2,-glohulin followed by two-dimensional gel electrophoresis. The result presented in Fig. 4A shows that isoelectric variants 2, 3, 4, and 5 can be clearly seen in the autoradiogram obtained after electrophoretic separation of the immunoprecipitated translation product of the hepatic mRNA. Variant 1, however, is not very distinct in this autoradiogram. A comparative analysis of the variant forms of a,,,-globulin that can be transferred to a nitrocellulose filter after two-dimensional gel electrophoresis of total hepatic protein is shown in Fig. 4R. For this  FIG. 3 . Two-dimensional  Upper frame ( A ) shows autoradiogram of the immunoreactive tr2,-glohulin derived from the laheled in uitro translation products of hepatic mRNA ohtained from young adult rats. Immunochemical isolation of ct2,,-gIohulin was carried out with a monoclonal mouse antibody. The positions of different isoelectric variants of tr2,-glohulin are marked with numbered arrows. The l o u w frame ( R ) shows an autoradiogram of the immunoreactive tu2,,-glohulin within the total liver cell protein transferred to nitrocellulose paper from the slab gel by electrohlotting. The nitrocellulose paper was first reacted with a monoclonal mouse antibody against cuy,,-glohulin and subsequently with laheled polyclonal goat antihody (IgG fraction) against mouse IgG. In frame R, the most acidic isoelectric variant (5) is not very distinct and a minor hasic variant (marked with an upward arrow) is visihle. experiment, the lysate of male rat hepatocytes was first separated by two-dimensional gel electrophoresis and the proteins from the gel were electroblotted onto a nitrocellulose filter. The filter was then immunoreacted with a monoclonal antibody made against cu,,-globulin and the antigen-antibody complex was subsequently detected with "'Ianti-mouse yglobulin. This procedure can clearly identify variant forms 1, 2 , 3 , and 4 within the liver cell lysate obtained from male rats (young adult). In addition, this "immunoblotting" procedure 9 . 7 8.3 6.4 4.9 4 3 P I also identified a minor basic (PI 9.2) isoelectric variant (marked by an upward arrow in Figs. 3 and 4B). However, the variant 5 is rather indistinct in the autoradiogram of the "immunoblot" of the total hepatic protein. A lack of efficient transfer of this acidic form (PI 3.7) from the sodium dodecyl sulfate-slab gel onto the nitrocellulose filter may account for this anomaly. The immunoblotting data also show that the monoclonal antibody reacts with the minor glycosylated forms (18) of the isoelectric variants of a,,,-globulin. Differential Regulation of the Variant Forms of a,,-Globulin during Aging-Changes in the hepatic synthesis of the different polymorphic forms of a,,,-globulin during maturation and aging were examined by two-dimensional gel electrophoresis of the total hepatocytic protein after pulse-labeling with ['''SI methionine. Fig. 5 shows segments of the autoradiograms of the two-dimensional slab gel that contains the n2,-globulin cluster. Results show that the gradual loss of a,,-globulin synthesis in the aging male rat is not associated with an equivalent decline in the synthesis of all of the various isoelectric forms of tu,,,-globulin. The age-dependent decline in the hepatic synthesis of a,,,-globulin seems to be associated with a differential regulation of the isoelectric variants of this protein. Changes in the synthesis of the various forms of cu2,,globulin in the male rat during aging are summarized in Fig.  5. Among different isoelectric variants of a,,,-globulin, the most prominent one with PI 6.1 (isoelectric variant 2) is the first to appear at puberty (40 days old). During the transition from young adulthood to middle age (70 days to 450 days), the variant 1 (PI 7.8) is first to disappear. In early senescence (600 days), two variants (i. e. 2 and 4) remain. The last one to disappear is the isoelectric variant 4. It is of interest that variant 4 is the least prominent form in the young adult.
Correlation between cu,,-Globulin Synthesis and Hepatic Level of Androgen Binding Protein during Aging-Androgendependent synthesis of np,-globulin in rat liver has been found to be correlated with the presence of a cytoplasmic androgen binding protein (1,6). The possible correlation between the gradual loss of cup,,-globulin synthesis in aging male rats and the cytoplasmic androgen binding activity was examined by sucrose density gradient analysis of the hepatic cytosol prelabeled with ["H]5a-dihydrotestosterone (Fig. 6). The results of androgen binding activity show a strong correlation with the age-dependent decline of n2,,-globulin synthesis in these animals.

DISCUSSION
A gradual decline in the hepatic synthesis of cu2,,-globulin during aging is associated with a corresponding decrease in the hepatic content of cu2,,-globulin mRNA. By means of twodimensional polyacrylamide gel electrophoresis, cu2,,-globulin has been resolved into a family of proteins. Members of the n,,,-globulin protein family vary both in molecular weights and in isoelectric points. Earlier studies in our laboratory have shown that primary translation products of tu2,,-globulin mRNA can be resolved into two molecular weight forms ( M , = 18,800 and 18,100). Furthermore, approximately 3% of these molecular weight variants undergo glycosylation, producing two additional forms with M , = 21,200 and 20,600 (18). Because of the technical constraints of sample loading, the degree of resolution in the second dimension of the twodimensional gel electrophoresis is not sufficient for a clear separation of all of the molecular weight variants of n2,,globulin, especially the two minor glycosylated forms, which appear indistinguishable. On the other hand, incomplete resolution of the two major nonglycosylated molecular weight forms can be seen. Identification of different isoelectric variants of cup,,-globulin, not only within the liver cell lysate but also in the in vitro translation products of hepatic mRNA, indicates that these isoelectric variants are coded by different species of cu2,,-globulin mRNA. Genetic analysis of the electrophoretic variants of the analogous mouse urinary protein, MUP, has indicated that the polymorphic forms of this protein are also coded by different genes (19). In this regard, it is of interest that both n2,,-globulin and MUP are coded by a gene family consisting of 16-20 genes (20-22). The present results show five major and possibly two minor isoelectric variants of cup,,-globulin. If each of these isoelectric variants also contains an additional molecular weight variant, as re-ported earlier (18), they can account for all of the 16 putative genes for a2,-globulin.
Among the five major isoelectric variants, variant 2 is the most predominant one and is the first to appear during maturation. However, variant 4, which is the weakest of the five major isoelectric variants, is the last to disappear. The multigene family of a2,-globulin has been localized on a single chromosome (chromosome 5) and may represent a distinct cluster of genes (21). Because of this possible clustering of the a,,-globulin gene family, the differential regulation of the polymorphic forms of a2,-globulin during aging is of considerable interest. Recent studies on the DNase I sensitivity of the cluster of ovalbumin-related genes (X, Y, and OV) have shown that the "ovalbumin domain" containing all three of these genes undergoes a coordinate conformational transition associated with transcriptional activation and yet the rates of transcription of these genes are markedly different from each other (23). It is possible that a similar situation exists in the case of the a2,-globulin gene family and each member of the aZu gene cluster is preceded by promoter sequences and other regulatory elements of different efficiency. Age-dependent changes in the promoter efficiency of the individual members of the cu2,,-globulin gene family may provide a possible molecular basis for the differential expression of the polymorphic forms of a2,-globulin during aging. Such an effect may be due to a direct modification of the gene structure possibly by the addition of hydroxyl free radical to the 5-6 double bond of thymine (24) or other base modification, e.g. changes in the methylation of certain strategically located cytosine residues (25). In addition to this type of direct effect on gene structure that may result in the differential regulation of gene expression, the principal cause for the overall decline in a2,-globulin seems to be the changing androgen responsiveness due to an age-dependent decline in the hepatic hormone receptivity. However, the likely mechanism for such a decline in receptor activity can also be due to a decreased expression of the gene coding for the receptor protein.