Cloning, Expression, and Functional Characterization of the β Regulatory Subunit of Human Methionine Adenosyltransferase (MAT II)*

MAT II, the extrahepatic form of methionine adenosyltransferase (MAT), consists of catalytic α2/α2′ subunits and a noncatalytic β subunit, believed to have a regulatory function. The full-length cDNA that encodes the β subunit of human MAT II was cloned and found to encode for a 334-amino acid protein with a calculated molecular weight of 37,552. Analysis of sequence homology showed similarity with bacterial enzymes that catalyze the reduction of TDP-linked sugars. The β subunit cDNA was cloned into the pQE-30 expression vector, and the recombinant His tagged protein, which was expressed in Escherichia coli, was recognized by antibodies to the human MAT II, to synthetic peptides copying the sequence of native β subunit protein, and to the rβ protein. There is no cross-reactivity between the MAT II α2 or β subunits. None of the anti-β subunit antibodies reacted with protein extracts of E. coli host cells, suggesting that these bacteria have no β subunit protein. Interestingly, the rβ subunit associated withE. coli as well as human MAT α subunits. This association changed the kinetic properties of both enzymes and lowered theK m of MAT for l-methionine. Together, the data show that we have cloned and expressed the human MAT II β subunit and confirmed its long suspected regulatory function. This knowledge affords a molecular means by which MAT activity and consequently the levels of AdoMet may be modulated in mammalian cells.

source of the propylamine moiety used in polyamine biosynthesis, and it serves as co-factor for other key enzymes in the one-carbon metabolism pathway (3)(4)(5). MAT is present in all living species, including thermophilic archaebacteria, plants, yeast, and mammals (reviewed in Refs. 4 and 6 -8). Interestingly, most species have more than one MAT isozyme (6).
In mammals, it is now established that there are at least two MAT isozymes (9 -12). MAT I/III is expressed only in liver and has a catalytic subunit designated ␣ 1 that is encoded by the MAT1A gene (8,9,(13)(14)(15)(16). MAT I and MAT III represent different oligomeric forms of the ␣ 1 subunit -MAT III is a dimer, and MAT I is a tetramer of the ␣ 1 subunit (9,(17)(18)(19). MAT I and MAT III differ considerably in their physical, kinetic, and regulatory properties (8,9,20). The MAT II isozyme is expressed in all tissues, including the liver, and has been studied in many tissues including erythrocytes, lymphocytes, brain, kidney, testis, and fetal liver (11, 20 -27).
We have been characterizing the human MAT II from human lymphocytes (22, 28 -31) 2 and were able to show that the form present in activated lymphocytes consists of distinct subunits (22,29). The catalytic MAT II ␣ 2 subunit, which is encoded by the MAT2A gene, was cloned and characterized and found to be homologous but different from the catalytic ␣ 1 subunit of the liver MAT I/III isozyme (13-15, 26, 30). The MAT II ␣ 2 subunit, which has a calculated molecular weight of 43,600, migrates on SDS-PAGE gels as a 53-kDa protein and is posttranslationally modified to generate MAT II ␣ 2Ј subunit (22). The catalytic ␣ 2 /␣ 2Ј subunits are found in native MAT II associated with a catalytically inactive subunit designated MAT II ␤, which migrates on SDS-PAGE as a 38-kDa protein (22,29,31). Inasmuch as the MAT II ␤ subunit had no homology to the ␣ 2 subunit (22,33), it was postulated that it is encoded by a different gene, which was putatively designated MAT2B (12). In this study, we report the complete sequence of cDNA encoding the entire MAT II ␤ subunit, and we show that the protein expressed in Escherichia coli associates with the E. coli as well as the human catalytic ␣ subunits of MAT. The association of ␤ and ␣ subunits changes the kinetic properties of MAT, thereby providing direct evidence for the regulatory role of the MAT II ␤ subunit.

MATERIALS AND METHODS
Cloning of the Human MAT II ␤ Subunit cDNA-Degenerate primers were designed based on partial amino acid sequence of two tryptic peptides (generated by partial digestion of pure MAT II ␤ protein as described previously (29 -31). The forward 5Ј DEG primer (5Ј-GTNG-GNMGNGARAARGARYTNWSNATHCAYTTYGTNCC) was based on the sequence of the N-terminal peptide (VGREKELSIHFVPGSNELV), and the reverse 3Ј DEG primer (5Ј-GTYTGYTCRTTNCCNSWCCART-* This work was supported by National Institutes of Health Grant GM-54892-09 and by Merit Review Award Funds from Veterans Affairs. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF182814.
** To whom correspondence should be addressed: University of GRAANGTNCCYTT) was based on the sequence of an internal peptide (LDPSIKGTFHWSGNEQT). Total RNA was extracted from the human T cell leukemia lines, Jurkat (ATCC-8163), or MOLT4 (ATCC-1582) using RNazol B (Tel-Test, Inc.). The RNA (2 g) was reverse transcribed into cDNA using 25 units of avian myeloblastoma virus reverse transcriptase (Promega), oligo(dT) 15 (Promega), and 1 mM dNTPs (Amersham Pharmacia Biotech). The cDNA was amplified using Taq DNA polymerase and the 5Ј DEG and 3Ј DEG primers. After 30 cycles of denaturation at 95°C for 30 s, annealing at 48°C for 1 min, and extension at 72°C for 1 min, a 765-bp PCR product was generated. The amplified PCR product was purified from agarose gel using QIAquick gel extraction spin columns (Qiagen), ligated into the pGEM-T easy vector (Promega), and used to transform E. coli strain JM109 (Promega). Positive clones were selected on LB agar (Difco) containing 100 g/ml ampicillin, 40 g/ml X-gal (5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside), and 100 M isopropyl-␤-D-thiogalactoside (all chemicals were from Sigma). Plasmid DNA was purified from positive clones and checked by restriction digestion with EcoRI (Promega) to verify the presence of the cloned PCR product. The cloned DNA was sequenced using the fmol cycle sequencing system (Promega). Clone 1, which contained sequences matching both degenerate primers, as well as sequences that encode the remaining amino acid sequence in the Nterminal and internal ␤ subunit peptides, was selected for further sequencing (see Fig. 1). Based on the sequence of Clone 1, nondegenerate primers were synthesized.
3Ј and 5Ј RACE-Because Clone 1 was a partial sequence, the RACE methodology was employed to obtain overlapping 5Ј and 3Ј clones from human Jurkat and MOLT4 cell cDNAs. The missing 5Ј sequence of the ␤ subunit cDNA was obtained by 5Ј RACE technique. Marathon Ready MOLT4 cDNA (CLONTECH) was amplified using Marathon Ready primer AP1 and a 3Ј MAT II ␤ subunit-specific (3Ј␤) primer (5Ј-CAT-TCTCTTCTCTGCTAGCTGCC). The 776-bp PCR product was purified using QIAquick gel purification spin columns (Qiagen) and subjected to nested PCR using the AP1 primer and another ␤ subunit-specific (3Ј␤ 2 ) primer (5Ј-CTCTCTGTAAGGTGGATTTGT). The 508-bp nested PCR product was purified and cloned, and 10 clones were sequenced as described above. The sequence provided the putative ATG initiation signal and verified the previous sequence obtained with the degenerate primers. Clone 2, which represents the consensus sequence, was selected for further studies.
To obtain the DNA sequence that encodes the C-terminal end of the ␤-subunit, we utilized 3Ј RACE technique. RNA from Jurkat cells was reverse transcribed into cDNA as described above, except that a Ttailed primer (5Ј-GGCCACGCGTCGACTAGTACT 17 ) was used in place of oligo(dT) 15 . The cDNA was amplified using the 5Ј MAT II ␤ subunitspecific forward (5Ј␤) primer (5Ј-CTGTCGGCTGGTGGAGGAGGAA), and a 3Ј-primer identical to the primer used to make the cDNA, but lacking the T tail. The resulting 981-bp PCR product was subjected to a nested PCR reaction using a 5Ј␤ 2 forward primer complementary to the internal region of the cloned DNA (5Ј-GCAACAAGTCAGCAAA-CATGG). A 433-bp PCR product was generated, ligated into the pGEM T-easy vector, and used to transform E. coli strain JM109. Positive clones were selected as described above, and the cloned DNA was sequenced using Thermo-Sequence (Amersham Pharmacia Biotech) and the Visible Genetics cycle sequencing system utilizing fluorescent Cy5 labeled M13-forward and M13-reverse primers. A total of 10 clones were sequenced to verify the in frame sequence from base 765 to base 1005, the last base in the stop codon (TAG). Clone 3 was selected for further studies (Fig. 1).
Cloning and Expression of MAT II ␤ Subunit-The cDNA prepared from resting peripheral blood mononuclear cells (PBMC) total RNA was amplified by PCR using Pfu DNA polymerase (Promega) and the MAT II ␤ 5Ј subunit-specific primer, 5Ј ␤r (5Ј-CGAGCTCATGGTGGGGAGG-GAGAAAGAACTGT) and 3Ј subunit-specific primer, 3Ј ␤r (5Ј-CCCAAGCTTAACCCAACACAAATAAACTAATGA). For cloning purposes, the 5Ј primer contained a SacI restriction site, and the 3Ј primer had a HindIII site. The PCR product was purified using QIAquick gel extraction (Qiagen) as described above. Using 200 nM dATP and 2.5 units of Taq DNA polymerase (Promega), A-overhangs were added to the purified PCR product in a reaction carried out at 72°C for 2 h. Following cleanup of the modified PCR product, the product was ligated into the pGEM-T-easy vector (Promega) and subsequently used to transform E. coli strain JM109. Ten positive clones were fully sequenced using the fmol cycle sequencing system (Promega) to rule out any possible base substitutions, and one representative clone, Clone 4, which represented the full-length ␤ subunit cDNA, was used for further studies. The sequence of the ORF was identical for cDNA representing RNA from Jurkat cells, MOLT4 cells, and normal human PBMC. The MAT II ␤ cDNA was excised from the pGEM T-easy vector using SacI and HindIII and purified. Expression vector pQE-30 (Qiagen), which is designed to express proteins containing a His 6 tag at the N-terminal was used. The vector was linearized by digestion with the restriction enzymes SacI and HindIII for 1 h at 37°C, then purified from an agarose gel, and incubated at 37°C with 5 units of alkaline phosphatase (Promega) to enhance the efficiency of ligation. SacI-and HindIII-digested Clone 4 was directionally cloned into unique SacI and HindIII sites in the multiple cloning site of the pQE-30 vector. Following ligation of the MAT II ␤ cDNA into the prepared pQE-30 vector, the ligated vector was used to transform E. coli expression strain M15 (Qiagen). Cursory sequencing of 20 positive clones was carried out using the fmol cycle sequencing system (Promega), and the clones were verified for proper frame and orientation of the MAT II ␤ cDNA. A representative clone, Clone 5, was inoculated into LB broth containing 25 g/ml kanamycin and 100 g/ml ampicillin. The culture was incubated, with shaking, at 37°C until an A 600 of 0.5-0.7 was reached, and then isopropyl-␤-D-thiogalactoside (Sigma) was added to a final concentration of 1 mM, and the culture was allowed to incubate for an additional 4 h under the same conditions. Bacterial pellets were disrupted using sonication and then analyzed via Western blot and MAT II ␤-specific polyclonal antibodies that were generated to the two MAT II ␤ peptides (29 -31).
Purification of Recombinant MAT II ␤-The MAT II ␤ expression clone, Clone 5, was inoculated into 1 liter of LB broth containing 25 g/ml kanamycin and 100 g/ml ampicillin and then incubated with shaking at 37°C until an A 600 of 0.5-0.7 was reached. isopropyl-␤-Dthiogalactoside (Sigma) was added to a final concentration of 1 mM, and the culture was allowed to incubate for an additional 4 h under the same conditions. Initial attempts to purify His-tagged MAT II ␤ subunit under native conditions via sonication and nickel-nitrilotriacetic acidagarose resulted in a complex of the E. coli MAT ␣ subunit and the recombinant human ␤ subunit. Therefore, subsequent purification of MAT II ␤ away from the endogenous E. coli MAT ␣ was performed by two methods. The first involved purification of the His-tagged protein under denaturing conditions of 8 M urea, and the second involved separation of SDS-PAGE and elution of the ␤ subunit protein from the gel.
To purify recombinant MAT II ␤ protein under denaturing conditions, the cell pellet was lysed at room temperature by stirring the pellet in a buffered solution containing 8 M urea, pH 8.0 (denaturing conditions). Once the solution became translucent, the cellular debris The first amplification was conducted using generate primers (5Ј DEG and 3Ј DEG), which were designed based on the sequence of the two tryptic peptides of native human MAT II ␤ subunit protein. Subsequent sequencing was performed using the sequence specific primers indicated.
were removed by centrifugation. The clarified supernatant was loaded onto a nickel-nitrilotriacetic acid-agarose column (Qiagen) to capture the His-tagged protein. The column was washed with several volumes of buffered 8 M urea, pH 6.3, until a A 280 of 0.001-0.005 is reached. Elution of the recombinant MAT II ␤ (rMAT II ␤) was carried out using buffered 8 M urea, pH 5.9 and 4.5. The eluted protein was extensively dialyzed against 50 mM Tris-HCl, pH 7.5, and then concentrated and analyzed by Western blots and silver-stained SDS-PAGE for determination of size and purity. The pure recombinant human MAT II ␤ subunit protein was used to immunize rabbits and to generate polyclonal antibodies as detailed below.
In some experiments, the nickel-agarose purified proteins were separated on SDS-PAGE and visualized by impregnation in cold 300 mM KCl, and the band corresponding to the 38 -39-kDa protein was excised, and the protein was electroeluted from the gel into Tris-glycine buffer, pH 8, dialyzed against 10 mM ammonium bicarbonate and then lyophilized. The lyophilized protein was reconstituted in 50 mM Tris-HCl, pH 7.5, and analyzed by Western blotting and silver staining for size and purity. The purified ␤ subunit protein was also analyzed for functional activity in MAT assays containing E. coli MAT ␣ or recombinant human MAT II ␣ protein. The nickel-agarose and SDS-PAGE purified ␤ subunit protein retained functional activity.
Production of Polyclonal Antibodies to MAT II ␤ Subunit-Polyclonal antibodies were generated to the ␤ subunit synthetic peptides (VGREKELSIHFVPGSNELV and LDPSIKGTFHWSGNEQT; Refs. 29 -31), as well as to the purified rMAT II ␤ protein. Both antibodies were generated in male New Zealand White rabbits (Myrtle's Rabbitry), weighing 5-6 lb. Prior to initial immunizations, a preimmune blood was drawn, and sera were used as a negative control in subsequent enzymelinked immunosorbent assay titer determinations and in Western blots. The two synthetic ␤ subunit peptides were combined, emulsified in an equal volume of Freund's Adjuvant Complete (Sigma), and injected into rabbits. The purified rMAT II ␤ subunit was similarly treated. Initial injections were made in three sites: each hindquarter and the back of the neck. Boosters were made every 2 weeks following the removal of 10 cc of blood from the ear vein for the determination of antibody titers by enzyme-linked immunosorbent assay. Freund's incomplete adjuvant (Sigma) was used in all booster injections.
Antibody titers were determined in an enzyme-linked immunosorbent assay using the rMAT II ␤ protein diluted in a 50 mM NaHCO 3 solution, pH 9.6, and incubated overnight at 37°C in enzyme-linked immunosorbent assay titer plates. Following a wash with 1ϫ phosphate-buffered saline (PBS) containing 0.05% Tween (PBS-T), the wells were blocked for 30 min at 37°C with a solution containing 1ϫ PBS and 1% bovine serum albumin (1% BSA-PBS). After washing, serial dilutions of rabbit antiserum in 1% BSA-PBS were added to wells containing the bound rMAT II ␤ protein and incubated at 37°C for 1.5 h. Following a wash with 1ϫ PBS, secondary antibody (goat anti-rabbit/ horseradish peroxidase diluted 1:1000 in 1% BSA-PBS) was added to each well and then incubated for 45 min at 37°C. The final wash was made with 1ϫ PBS-T, followed by a wash with 1ϫ PBS. Color reagent (5 mM 5-aminosalicylic acid, pH 6.0, and 10 l/ml 0.5% H 2 O 2 ) was added to each well and then incubated at room temperature for 30 min. Titers were determined by measuring absorbance at 478 nm with a cut-off of 0.1 over control wells containing preimmune serum. After the titers had reached a plateau (12-14 weeks), the final rabbit serum was collected.
SDS-PAGE and Western Blotting-E. coli cell extracts containing recombinant MAT II ␤ subunit or pure recombinant MAT II ␤ subunit protein were diluted in loading buffer (60 mM Tris-HCl, pH 6.8, 2% SDS, 5% 2-mercaptoethanol, and 5% glycerol), heated in a boiling water bath for 4 min, and analyzed by SDS-PAGE (10% total acrylamide, 2.7% bisacrylamide) as previously (29). After electroblotting the proteins onto the nitrocellulose for 1 h at 25-30 V/cm, the blots were blocked overnight with 6% nonfat dry milk in TBS (50 mM Tris, pH 7.5, and 150 mM NaCl). Following the removal of the blocking solution, the blot was washed in TBS and incubated with primary polyclonal anti-holoenzyme antisera (33) or polyclonal antisera generated either against pure human recombinant MAT II ␣2 or ␤ subunits or against the two synthetic peptides copying sequences of the ␤ subunits. The blots were developed with secondary anti-rabbit antibodies conjugated to horseradish peroxidase and the luminol-chemiluminescence reagents (ECL; Amersham Pharmacia Biotech) as described previously (29). The processed blots were exposed to Kodak X-Omat film, and the autoradiograms were analyzed. For some experiments, the autoradiograms were scanned using a Howtek Scanmaster-3 scanner (Protein Data Base, Inc., Huntington Station, NY), and the intensity of the desired band was integrated and expressed in arbitrary units.

cDNA and Predicted Protein Sequence of the Human MAT II
␤ Subunit-The cloning strategy was based on the design of degenerate primers representing partial amino acid sequence of an N-terminal (19-mer) and an internal (17-mer) peptides of the trypsin digested ␤ subunit protein, purified from human lymphocytes (29,30). Complementary DNA prepared from Jurkat cells mRNA was amplified with the degenerate primers, 5Ј DEG and 3Ј DEG, that would encode partial sequences of the two ␤ subunit peptides. The cDNA from positive clones were isolated at the preparative level and entirely sequenced in both directions. The cDNA clones that contained sequences corresponding to both peptides were selected for further sequencing, FIG. 2. Nucleotide sequences of cloned cDNA for the ␤ subunit of human lymphocyte MAT II. The sequence was determined from 10 clones that were amplified using sequence specific primers. The deduced amino acid sequence is shown for the ORF. Nucleotides are numbered beginning at the first position of the ORF. Amino acid positions are shown in parentheses. Shaded amino acid residues correspond to tryptic peptide sequences that were chemically determined from purified ␤ subunit. The arrows indicate the position and direction of primers used to elucidate the sequence of ␤ subunit cDNA.
and Clone 1 (765 base pairs), which represented the consensus sequence, was fully characterized for the design of nondegenerate primers. Clone 1, which represents a partial sequence of the ␤ subunit cDNA, would be expected to encode for 255 amino acids. However, because the ␤ subunit migrates on SDS-PAGE as a 38-kDa protein, we estimated that at least 60 -80 amino acids or 180 -240 base pairs were still unaccounted for. Accordingly, sequence-specific primers based on the sequence of Clone 1 were synthesized and used in 3Ј and 5Ј RACE to obtain the complete ORF of the ␤ subunit cDNA (Fig. 2).
A minimum of 10 positive clones from 3Ј or 5Ј RACE were sequenced in both directions, and Clone 2 represented the consensus of the 5Ј RACE clones, whereas Clone 3 represented that of the 3Ј RACE clones. The sequence of Clone 2, which was generated by 5Ј RACE, showed that Clone 1 was only missing the ATG start codon at its 5Ј end. The authenticity of this ATG codon as the initiator codon was confirmed in several ways. First, this ATG is in frame with sequences that encode for amino acids found in the N-terminal ␤ subunit peptide (Fig. 2). Second, the sequence of bases flanking this ATG codon are in accordance with sequences identified by Kozak (34) as consensus sequences preceding or following an authentic initiator codon. For example, the sequence (Ϫ3)GnnATGG(ϩ4) showed the second highest incidence of functional initiator codons (130 of 699 tested) in vertebrate mRNAs analyzed (34). In addition, the presence of a C at position Ϫ1 and the presence of a G at positions Ϫ3 and Ϫ6 provide further support that the ATG in the 5Ј-GACGGCGGGCATGG sequence of Clone 2 is the initiator codon for the ␤ subunit cDNA. Clone 2 also provided 60 base pairs sequence of the 5Ј-untranslated sequence of the cDNA.
Clone 3, which was generated by 3Ј RACE and represented the consensus sequence of 10 sequenced clones, provided the remaining 3Ј sequence of the ␤ subunit cDNA. This clone provided an additional 288 base pairs, which included the remaining 3Ј 237 base pairs of the ORF, followed by the termination codon, TGA, at position 1003 (Fig. 2). Clone 3 also provided and an additional 48 base pairs of 3Ј-untranslated sequence.
Based on the sequence of Clones 1, 2, and 3, specific primers were synthesized to amplify the full-length cDNA of the ␤ subunit 1005 base pairs, including the initiator and terminator codons using normal human PBMC RNA as template. Ten clones containing the full-length cDNA of the ␤ subunit were sequenced in both directions, and the consensus sequence, represented by Clone 4, is shown in Fig. 2.
No sequence differences were noted between the sequence of Clone 4 and cDNA prepared from normal human PBMC of several individuals, or cDNA from Jurkat or MOLT-4 cells. Further analysis of the 3Ј-untranslated region of the ␤ subunit cDNA provided 1807 base pairs of untranslated sequence, which included the polyadenylation signal sequence AATAAA (35,36) at position 1766, beginning 24 bp upstream of the poly(A) region (Fig. 2).
Sequence Comparisons-The cloned full-length cDNA contained standard 5Ј-and 3Ј-flanking regions and an open reading frame of 1002 base pairs. The cDNA encodes for 334 amino acids with a calculated molecular weight of 37,551.81 and has a pI of 6.90. Thus, unlike the MAT II ␣2 and ␣ 2Ј subunits, which migrate at higher than expected molecular weight on SDS-PAGE (22,29,30), the MAT II ␤ subunit migrates at its expected molecular mass of 38 kDa.
Analysis of the DNA and the protein sequences on data bases from GenBank TM and the Biologiocal Information Resource site using, among others, the Advanced Blast and SwissPort data bases, revealed up to 28% homology between the MAT II ␤ subunit and a family of bacterial enzymes that catalyze the reduction of TDP-linked sugars such as dTDP-4-dehydrorhamnose reductase, several nucleoside-diphosphate-sugar epimerases (37), and other proteins involved in the synthesis of polysaccharides (38). The alignment with two representative proteins, dTDP-6-deoxy-L-mannose-dehydrogenase and UDPglucose 4-epimerase, is shown in Fig. 3. Analysis for DNA sequence homology revealed identity between several segments of the MAT II ␤ cDNA and several human, mouse, and rat DNA sequences, which were part of contigs that were sequenced for the human genome project or cDNA libraries from specific tissues or tumors. None of these sequences contained the entire MAT II ␤ ORF, and no function was ascribed to any of the genes, but the majority of these sequences were designated as being similar to bacterial dTDP-4-dehydrorhamnose reductase genes.
Expression of the Recombinant ␤ Subunit Protein in E. coli-Clone 4, which represented the full-length cDNA encoding the complete ORF (1002 base pairs) for the ␤ subunit, was directionally cloned into the pQE-30 expression vector, which was designed to express the ␤ subunit protein with a poly-His tag at the N-terminal end of the molecule. Clone 5, which represented 20 sequenced clones, provided expression of the ␤ subunit protein in E. coli. Antibodies to synthetic peptides copying the sequence of the N-terminal and internal ␤ subunit protein peptides were found to recognize the recombinant 38-kDa protein in transfected E. coli extracts (Fig. 4). In concordance with our previous studies (33), there was no immunological crossreactivity between the MAT II ␣ and ␤ subunits. Furthermore, no reactivity was detected between the anti-␤ peptides and protein extracts of E. coli host cells that were either untransfected or that were transfected with the same vector and expressing the human recombinant MAT II a 2 subunit cDNA (Fig. 4B) (30).
When the His 6 -tagged recombinant MAT II ␤ subunit was purified on nickel-nitrilotriacetic acid-agarose column and eluted with 300 mM imidazole, analysis of the dialyzed protein on silver-stained SDS-PAGE showed that several protein bands co-purified with the recombinant protein (Fig. 4C, lane  1). Further analysis by Western blot, which was probed with  3). B is a Western blot of the same crude material analyzed in A. The blot was probed with antibodies to the ␤ subunit protein. C is a silver-stained SDS-PAGE of His-tagged recombinant ␤ subunit protein purified from E. coli cells that were transfected with the pQE-30 vector harboring the MAT II ␤ subunit cDNA, using Ni-agarose column purification (lane 1) and His-tagged recombinant ␤ subunit protein purified on nickel-agarose column then further purified and eluted from a preparative SDS-PAGE (lane 2) as detailed under "Materials and Methods" and "Results." D is a Western blot of the same material analyzed in C and probed with antibodies to both the MAT ␣ 2 and ␤ subunits. C, lane 1 is the His-tagged recombinant ␤ subunit protein eluted from the nickel-agarose column, and lane 2 is the eluted material subjected to further purification and electroelution from a preparative SDS-PAGE. The antibodies to the MAT II ␣ 2 subunit recognized the cross-reactive E. coli MAT ␣ subunit, which co-purified with the recombinant His-tagged ␤ subunit (lane 1), but there was no contaminating ␣ subunit in the SDS-PAGE eluted material (lane 2). E and F represent the same blot shown in D, stripped and reprobed with antibodies to the MAT II ␣ 2 subunit or with antibodies to the MAT II ␤ subunit, respectively. antibodies to either the ␣ or the ␤ subunits or both, showed that the endogenous E. coli MAT ␣ protein was co-purifying with the recombinant human MAT II ␤ subunit protein (Fig. 4, D and E,  lanes 1). To purify the recombinant MAT II ␤ subunit protein away from the endogenous E. coli MAT ␣ protein, the nickelagarose purified proteins were separated on SDS-PAGE, the protein bands were visualized by impregnation in cold 300 mM KCl, and the band corresponding to the 38 -39-kDa protein was excised. The protein was electroeluted from the gel slice into Tris-glycine buffer, pH 8, dialyzed, lyophilized, reconstituted in 50 mM Tris-Cl buffer, pH 7.5, and reanalyzed. Analysis of this purified protein on silver-stained SDS-PAGE showed the migration of a single 38 -39-kDa protein (Fig. 4, C, lane 2), and the Western blot revealed that only the recombinant human MAT II ␤ subunit protein was present in this fraction (Fig. 4, D and  F, lanes 2). As will be discussed below, the gel-purified ␤ subunit protein retained functional activity.
The identity of the immunoreactive band with the MAT II ␤ subunit protein was verified by the fact it was recognized by antibodies generated to either the ␤ subunit partial peptides as well as by antibodies to the whole recombinant human MAT II ␤ subunit protein (data not shown). Again, neither antibody to the ␤ subunit showed any reactivity with any E. coli protein (Fig. 4).
Functional Analysis of the Recombinant Human MAT II ␤ Subunit Protein-The recombinant human MAT II ␤ subunit protein was found to have no MAT catalytic activity; however, it modulated the kinetic properties of the MAT II ␣ 2 catalytic subunit. The nickel-agarose and SDS-PAGE-purified recombinant MAT II ␤ subunit protein associated spontaneously with E. coli MAT ␣ subunit as well as with the recombinant human MAT II ␣ 2 subunit. The effect of the MAT II ␤ subunit protein on the kinetic activity of the E. coli MAT or the recombinant human MAT II ␣2 was analyzed. In the absence of the MAT II ␤ subunit, both the E. coli MAT and the recombinant human MAT II ␣ 2 exhibited normal Michaelis and Menten kinetics with the apparent presence of a single catalytic form (Figs. 5 and 6).
The K m for L-Met of the E. coli MAT ␣ subunit was 80 -90 M at high L-Met concentrations (Fig. 5A) and 65-80 at low L-Met concentrations (Fig. 5B). In the presence of the MAT II ␤ subunit, however, two kinetic forms appeared one with a K m for L-Met of 80 -90 M (Fig. 5C) and another with a K m of 30 -38 M (Fig. 5D). Similarly, the K m for L-Met of the human MAT II ␣ 2 catalytic subunit alone was 60 -100 M (Fig. 6), and in the presence of the r␤ subunit two kinetic forms were evident with K m values of 76 M (Fig. 6C) and 22 M (Fig. 6D). Together, these observations confirm previous suggestions (31) that one of the functions of the ␤ subunit is to lower the K m of MAT for L-Met.

DISCUSSION
The existence of multiple isozymes of MAT in mammalian tissues is well established (reviewed in Refs. 4 and 6). Whereas, the liver-specific enzyme appears to be a homodimer or tetramer of a single ␣ 1 subunit, the extrahepatic MAT II enzyme, which is expressed in all tissues, appears to consist of nonidentical subunits, ␣ 2 and ␤. In 1985, Kotb and Kredich (22) reported that native MAT II from human leukemic cells has a molecular weight of 185,000 and consists of two related subunits ␣ 2 and ␣ 2Ј , which migrated on SDS-PAGE as 53-and 51-kDa proteins and had an identical V8-protease peptide banding pattern. By contrast, the ␤ subunit migrated on SDS-PAGE as a 38-kDa protein and had a peptide banding pattern that was quite distinct from that of the two ␣ 2 subunits, suggesting that ␤ is distinct from the ␣ 2 subunit. Antibodies to the pure MAT II enzyme from human leukemic cells, whose purity was verified by analytical ultracentrifugational analysis, recognized both ␣ 2 and ␤ subunits in human lymphocytes and showed cross-reactivity with E. coli and yeast MAT ␣ subunits, which is not surprising given the now known high degree of primary sequence homology of MAT catalytic subunits in the many species analyzed; however, there was no indication that the anti-human MAT II antibodies were recognizing a protein similar to the ␤ subunit in either E. coli or yeast cell extracts (33). Moreover, polyclonal antibodies to E. coli or yeast MAT reacted with the human MAT I, II, and III ␣ subunits but failed to recognize the MAT II ␤ subunit protein. Together, these findings strongly suggested that the ␤ subunit of MAT II is distinct from the MAT ␣ 2 /␣ 2Ј subunits, and we proposed that it is probably encoded by a distinct gene, which we putatively designated MAT2B (12).
Several complementary forms of evidence indicate that the cDNA we have characterized from Jurkat T-cells, MOLT-4 cells, and normal peripheral blood mononuclear cells is that of the human MAT II ␤ subunit. First, the deduced amino acid sequence contains sequences that are identical to those of the two tryptic human lymphocyte ␤ subunit peptide sequences (29). Second, expression of this cDNA in E. coli gave a protein band that migrated at the expected size of the authentic ␤ subunit in SDS-PAGE (Fig. 5) and reacted with antiserum to the human lymphocyte MAT II, to the tryptic peptides, and to the recombinant MAT II ␤ subunit protein. More important, the fact that the cloned protein lowered the K m (L-Met) of the recombinant human MAT II ␣ 2 subunit from 80 to 34 M confirms the previously described regulatory function of the MAT II ␤ subunit (29,31).
Detailed kinetic analysis of MAT II from human leukemic cells expressing both the ␣ and ␤ subunits of the enzyme revealed strong regulation of enzyme activity by its products, P i , PPi, and AdoMet (22,28). Specifically, the activity of the enzyme was inhibited by almost 2-fold in the presence of 25 M AdoMet and by 3-fold in the presence of 25 M AdoMet and 30 M PPi. In 1995, De La Rosa et al. (30) reported that the ␣ 2 subunit was responsible for the catalytic activity of MAT, whereas the ␤ subunit lacked activity. However, the studies revealed that recombinant MAT ␣ 2 expressed in E. coli had a K m for L-Met of 60 -80 M compared with a K m of 4 M in leukemic cells or 15-20 M in resting lymphocytes. These observations suggested that MAT II ␤ subunit may have regulatory properties. This notion was further emphasized by the studies of LeGros et al. (31), who in 1997 demonstrated that physiological stimulation of human lymphocytes with superantigen resulted in down-regulation of ␤ subunit expression. The disappearance of the ␤ subunit was accompanied by a change in MAT kinetic properties with the MAT II ␣ 2 /␣ 2Ј enzyme form exhibiting a 3-fold higher K m for L-Met and, more importantly, showing resistance to feedback product inhibition by AdoMet and resulting in a 5-6-fold increase in intracellular AdoMet levels (31). These studies led us to propose that the ␤ subunit regulates MAT II activity by lowering the K m for L-Met and rendering the enzyme more resistant to feedback inhibition by its products (31). The data presented here, as well as those presented in Halim et al., 2 provide direct evidence that these hypotheses are correct and confirm the regulatory role of the ␤ subunit of MAT II.
The sequence of the ␤ subunit cDNA was identical for Jurkat T cells, MOLT-4 cells, and PBMC from different individuals. The ORF for the MAT II ␤ subunit, which begins with the Met-Val sequence, encodes for a 334-amino acid protein and in concordance with the calculated molecular weight, and the recombinant protein migrates on SDS-PAGE as 38-kDa protein. Thus, the MAT II ␤ subunit, unlike the ␣ 2 subunit, migrates at its expected molecular size. The MAT II holoenzyme is known to be very hydrophobic inasmuch as it binds very strongly to phenyl-Sepharose columns and can only be eluted with 40% Me 2 SO (22). The Kyte-Doolittle hydrophobicity plots of human MAT II ␤ subunit protein (data not shown) revealed two prominent hydrophobic segments, whereas the ␣ 2 subunit has three minor and one major hydrophobic segment. Further studies are required to determine whether the oligomerization of the MAT II ␣2 and ␤ subunits is responsible for the strong hydrophobic property of the enzyme.
The presence of the ␤ subunit protein was previously detected in human erythrocytes, lymphocytes, bovine brain, Ehrlich's ascites tumor, and calf thymus (22,25). Inasmuch as antibodies to the ␤ subunit did not react with any E. coli protein, it appears that these bacteria lack this subunit. Similarly, we were unable to detect the ␤ subunit in yeast cell extracts (data not shown). Although the presence of the ␤ subunit in other species needs to be further investigated, it is possible that this protein is only present in mammalian tissue. In fact, we have evidence that the ␤ subunit is expressed in mouse lymphoid tissues, and the presence of DNA homology between segments of the MAT II ␤ cDNA and sequences found in randomly sequenced DNA contigs of cDNA libraries from mouse and rat tumors confirm the presence of this protein in other mammals. However, future studies will determine the species and tissue distribution of this protein.
The homology of the MAT II ␤ subunit with bacterial enzymes that catalyze the reduction of TDP-linked sugars (38) is remarkable and may provide clues as to the mechanism by which the ␤ subunit alters the kinetics of the ␣ subunit or may reveal additional functions for this protein. Despite the stretches of identity and strong sequence similarity between these proteins, the anti-MAT II ␤ antibodies does not react with either E. coli or yeast extracts. Further, there is no evidence that other proteins present in E. coli extract interact with the MAT ␣ subunit or with His-tagged recombinant human MAT II ␣ 2 subunit expressed in E. coli.
In summary, we have cloned and expressed the ␤ subunit of human MAT II and confirmed its regulatory function. The mode of interaction between the MAT II ␣2 and ␤ subunits remains to be determined; however, our recent studies suggest that the two subunits spontaneously associate. 2 The association between the human MAT II ␤ subunit and the E. coli MAT ␣ subunit is quite interesting because it may provide important clues as to the structural requirements for the association between the ␣ and ␤ subunits of MAT. The association between the MAT II subunits has a strong effect on the MAT kinetic properties, and we believe that this may be an important means by which the enzyme activity is regulated in vivo and a way to alter AdoMet levels to meet cellular requirements and regulate cell function.