Purification and characterization of mouse hypoxanthine-guanine phosphoribosyltransferase.

Hypoxanthine-guanine phosphoribosyltransferase (HGPR transferase) (EC 2.4.2.8) has been purified approximately 4500-fold to apparent homogeneity from mouse liver. The procedure involves the use of affinity chromatography and was designed to be readily adaptable to small scale isolations. The enzyme appears to be composed of 3 subunits of identical molecular weight (27,000 per subunit). The subunit molecular weight has also been determined by the analysis of radioactively labeled HGPR transferase immunoprecipitated from wild type and mutant (HGPR transferase) mouse tissue culture cell lines.

nH 7.4. the samnle was filtered through the DEAE-disc, and the disc was washed four times with 5-ml al&ots of 10 mM T&-Cl, pH 7.4. The discs were then placed in scintillation vials. the 13HlIMP eluted with 0.5 ml of 3% NaCl and counted in 5 ml of'Patt&son Greene (5) scintillation fluyd. A unit of enzyme is defined as that amount required to convert 1 nmol of hypoxanthine to IMP per hour under standard assay conditions. Bugera-The buffer solutions used in these experiments are listed in Table I. For convenience, they will be referred to by letter.
Dithiothreitol was added just before the buffer was used. Afinity Chromatography Column-The 3,3'-iminobispropylamine-GMP agarose affinity chromatography column used for the purification of mouse liver HGPR transferase was prepared as follows.
3,3'-Iminobispropylamine agarose was prepared from Sepharose 4B using the CNBr technique of Cuatrecasas (7). Sepharose 4B, washed with 10 volumes of deionized water, was activated with CNBr at a ratio of 300 mg of CNBr per ml of packed beads.
The activation was terminated by washing with 20 volumes of ice-cold deionized water.
The beads were washed briefly with 2 M 3,3'-iminobispropylamine adjusted to pH 10, placed in 1 volume of this solution, and allowed to react overnight at 4O. The trinitrobenzene sulfonic acid color test (7) indicated that a high degree of coupling had been achieved.
A lOO-ml packed volume of 3,3'-iminobispropylamine agarose was washed with 5 liters of HSO, 100 ml of Buffer I, and 100 ml of Buffer J. The washed beads were suspended to a total volume of 130 ml in Buffer J. Other amine agaroses were prepared identically.
Oxidized GMP was coupled to 3,3'-iminobispropylamine agarose using modifications of the methods developed by Gilham for cou- and allowed to react in the dark at 0" for 30 min. 3,3'-Iminobispropylamine agarose (100 ml) in Buffer J was added to the oxidized GMP solution and stirred gently for 1 hour. Two sequential additions of NaBHa were then made: the first was 0.83 g dissolved in 75 ml of Buffer J, and the second 1.2 g of solid NaBHa. Each was allowed to react for 1 hour.
The reaction was stopped by filtration and washing with 6 liters of HZO. An appropriate aliquot was poured into a column and washed with > 10 volumes of Buffer G, then >2 volumes of Buffer G containing 1 mg per ml of bovine serum albumin, > 10 volumes of Buffer H Font&&g 1.5 M KC1 instead of 1.2 M KCl. and finallv >lO volumes of Buffer G. The 3,3'-aminobispropylamine-GMPaagarose prepared in this way was tested on a small scale to ensure that it would retain greater than lo4 units of HGPR transferase per ml of packed column volume.
Each purification required a freshly prepared affinity column.
Purzficatzon of Mouse Laver HGPR Transferase-One hundred Swiss white mice, 6-to 7.weeks-old, were killed by decapitation and their livers rapidly removed.
The gall bladders were excised and the livers were washed twice in ice-cold Buffer A. Unless otherwise noted, all subsequent steps were done at O-4". Two and one-half volumes of Buffer A were added and the liver cells were broken in a Dounce homogenizer by six strokes with a loose and six strokes with a tight pestle. The lysate was spun at 10,000 X g for 30 min in a Sorvall GSA rotor and the supernatant removed with as little of the overlying lipid as possible.
This supernatant (S-10) was centrifuged at 165,000 X u for 4 hours in an International A-170 and the supernatant was removed from under the overlying lipid. This material (S-165) could be stored at -20" with little loss of activity for several weeks.
The S-165 was diluted at 0" with 2 volumes of ice-cold HzO. The pH was rapidly lowered to 5.0 by dropwise addition of 1 N acetic acid.
The resulting precipitate was removed by centrifugation at 10,000 X g for 15 min in a Sorvall GSA rotor.
The pH of the supernatant was readjusted to 7.6 with 1 N KOH. This step does not give a large purification, but it is known (9) that a number of guanine binding proteins involved in protein synthesis are removed.
The pH 5 supernatant was then heat treated. Aliquots (150 ml) of DH 5 sunernatant at 2530" were diluted with 1 volume of Buffer B preheated to 85". This mixture was placed in an 85" bath and vigorously agitated. After approximately 2.5 min, the temperature reached 70" and was maintained at 70" by withdrawal and immersion for approximately 2.5 min. The total time was carefully monitored to be 5 min. The samples were rapidly chilled in an ice salt bath at -15 to -17".
It took approximately 80 s to reach 37" and 3.25 min to reach 15". The resulting precipitate was removed by centrifugation at 10,000 X g for 15 min in aSorval1 GSA rotor.
The supernatant was diluted with 1 volume of Buffer C and the nH adjusted to 5.8 with 1 N HCl.
The conductivity of the diluted supernatant was always checked and shown to be well below that of Buffer D. The diluted suaernatant was then loaded onto a 500.ml CM-50 Sephadex column equilibrated in Buffer D. The column was loaded at its maximal flow rate. After the sample had been loaded, the column was washed with 2 volumes of Buffer D and 2 volumes of Buffer E. The HGPR transferase activity was eluted by 2 volumes of Buffer F, assayed, and pooled. If the goal is to obtain an enzyme of moderate purity ("60%) but of higher yield, the following Amicon ultrafiltration and gel filtration steps may be omitted. If a highly purified enzyme is desired, they should be included.
The enzymatic activity was pooled and concentrated to 14.8 ml using an Amicon PM-30 ultrafiltration membrane. This material was loaded on a Sephadex G-150 column (2.5 X 100 cm) equilibrated with Buffer G. The column was eluted in an ascending direction with Buffer G. Five-milliliter samples were collected. The HGPR transferase was located by enzymatic assay and pooled.
The pooled material was loaded slowly onto a 30-ml3,3'iminobispropylamine-GMP agarose column equilibrated in Buffer G. After  The cells were incubated with the radioactive amino acids for 2.5 hours at 37" followed by two 5-ml washes with Buffer K.
The cells were then extracted with 250~1 of Buffer N as previously described (5)) followed by centrifugation at 1000 X g for 20 min to sediment the cell debris and nuclei.
Approximately 3.5 mg of a HGPR transferase-CRM-cell extract was added to compete with radioactively labeled non-HGPR transferase protein.
Forty-two micrograms of an&HGPR transferase serum were added followed by incubation at 4" for 12 hours. Goat antirabbit r-globulin (1.5 mg) was then added and incubation continued for an additional 4 hours at 0". The resulting immunoprecipitates were sedimented by centrifugation at 1000 X g for 20 min. The supernatants were removed and the precipitates were washed twice with 400 ~1 of Buffer L, with centrifugation at 1000 X g for 20 min after each washing.
The precipitates were washed for a final time with 400 ~1 of Buffer M and the pellets were sedimented as before.
The pellets were then processed and subjected to electrophoresis as described above. Nondenaturing Gels-The methods and buffers of Hedrick and Smith (3)

RESULTS
PuriJication-The purification of HGPR transferase from mouse liver is summarized in Table II. The method includes a series of batch steps, Sephadex gel filtration, and affinity chromatography.
The over-ail yield of 17% results in a homogeneous enzyme as judged by electrophoresis on sodium dodecyl sulfate-urea polyacrylamide gels (see Fig. 2).   Table III shows that the ability of the column to bmd HGPR transferase 1s very dependent on the nature of the arm linking the oxidized GMP to the agarose.
The ethylenediamine-GMP-agarose column has a much lower capacity for binding HGPR transferase than does the 3,3'-iminobispropylamine-GMP agarose column. 3,3'-Iminobispropylamine agarose does not bind HGPR transferase unless it is reacted with oxidized GMP.
The choice of an appropriate arm, however, is dependent not only on the length, but also on the chemical composition. For example, l,Sdiaminooctane agarose was found to have the unfortunate property of binding many proteins before or after coupling with oxidized GMP. Because this effect was not observed with ethylenediamine or 3,3'-iminobispropylamine, we have attributed it to the increased hydrophobicity of the long hydrocarbon chain.
The binding specificity of 3,3'-iminobispropylamine-GMP agarose is demonstrated by its ability to retain HGPR transferase from mouse, rabbit, goat, and pig, but not the related enzyme adenine phosphoribosyltransferase.
The yield of HGPR transferase from the affinity column was dependent on the purity of the ldaded sample. Affinity chromatography of crude supernatant fractions results in a lOO-fold purification of HGPR transferase with quantitative recovery of activity.
On the other hand, the passage of highly purified samples such as the G-150 eluant through the column affords lower yields (see Table II). As observed in Fig. 3, HGPR transferase from both mouse liver and L cells sediments slightly faster than horse alcohol dehydrogenase (mol wt = 83,000, (13)) with an ~~0,~ of 5 S. The position of the enzymes was determined by activity assays. Mouse liver HGPR transferase was run on a Sephadex G-200 column with proteins of known Stokes radii (see Fig. 4). The enzymatic activity eluted just ahead of bovine serum albumin at a position indicating a Stokes radius of 36 A, precisely the same value reported for human HGPR transferase (15). Care must be taken in such experiments because at high dilution and low ionic strength the enzyme will interact with Sephadex and be retarded from this position.
Because the elution position of a protein from Sephadex is very sensitive to shape as well as molecular weight (16) an equivalent unhydrated sphere of the same surface area (see Fig. 5). Both methods of calculation gave similar results showing that native HGPR transferase is slightly larger than bovine serum albumin.
This result is in good agreement with the size estimate made on Sephadex G-200.
It has been pointed out that Sephadex gel filtration is much more sensit,ive to the shape of the molecule than is electrophoresis on nondenaturing polyacrylamide gels (16). Consequently, the similar behavior of HGPR transferase on Sephadex and polyacrylamide gels indicates that the enzyme is not highly asymmetrical.
Because the molecule appears to be reasonably spherical, the elution data from gel filtration (Fig. 4) and the retardation data from polyacrylamide gels (Fig. 5)  on sucrose gradients is that of a molecule slightly larger than 80,000. We feel that these data indicate that HGPR transferase is not highly asymmetrical, is of greater than normal density, and has a molecular weight of 80,000 f 4,000.
Subunit Structure of HGPR Transjerase-The subunit structure of purified mouse liver HGPR transferase was investigated by sodium dodecyl sulfate and sodium dodecyl sulfate-urea polyacrylamide gel electrophoresis. As observed in Fig. 2, the purified protein migrated as a single band on such gels. Comparison of the electrophoretic mobility of this band with proteins of known subunit molecular weight used as internal standards indicates that it has a molecular weight of 27,000 f 1,000 (see Fig. 6).
The higher background seen in this experiment relative to the ones depicted in Fig. 7 resulted from the increased amount of anti-HGPR transferase serum used.
for microscale enzyme isolations. Characterization of the physical properties of the wild type enzyme was also required. The key to the isolation procedure is the affinity column. It can be used to purify HGPR transferase approximately 100. fold from crude supernatants.
The "arm" linking the G&II' to the agarose was carefully chosen in order to obtain a column which bound HGPR transferase strongly and selectively. Both length and chemical composition of the arm were found to be important.
The methodology used to build the 3,3'-iminobispropylamine-GMP agarose column should be applicable to the construction of any nucleotide affinity column. The procedure is rapid and does not demand extensive organic synthesis.
Our studies of the physical characteristics of undenatured HGPR transferase have led us to t'he conclusion that it has a molecular weight of 80,000 f 4,000, is not highly asymmetrical, and is probably of greater than average density. The shape of the enzyme was deduced from the observation that HGPR transferase behaves similarly (in terms of size) on both Sephadex G-200 gel filtration and nondenaturing polyacrylamide gel electrophoresis.
If HGPR transferase were highly asymmetrical, these two techniques would not be in agreement due to the greater sensitivity of gel filtration to molecular shape (16). On both Sephadex G-200 and polyacrylamide gels, HGPR transferase behaves as a molecule smaller than 80,000.
In contrast to these results, the enzyme sediments slightly faster than a molecule of 80,000 on calibrated sucrose velocity gradients. These data suggest t,hat, t,he molecule is of greater than average density.
A combined biochemical and genetic analysis has been used to determine the subunit structure of HGPR transferase. Electrophoresis of the purified mouse liver enzyme on calibrated sodium dodecyl sulfate and sodium dodecyl sulfate-urea gels resulted in a single band migrating at a position corresponding to 27,000 =t 1,000 daltons.
Genetic evidence indicates that this band is the HGPR transferase subunit. Extracts of wild type L cells (HGPR transferase+) and two types of mutant L cells (HGPR  transferase-CRP\I-,  HGPR  transferase-CRM+), labeled with radioactive amino acids, were immunoprecipitated with highly specific HGPR transferase antiserum and the immunoprecipitates were analyzed on calibrated sodium dodecyl sulfate-urea gels. The HGPR transferase+ and HGPR transferase-CRM+ precipitat'es contained radioactive material which migrated with a sharp peak at 27,000, but no such material was derived from HGPR transferase-CRM-precipitates. Furthermore, this peak of radioactivity was reduced if purified mouse liver HGPR transferase was added prior to immunoprecipitation. From these results, we are confident that mouse HGPR transferase (liver and L cell) is composed of subunits with a molecular weight of 27,000.
Our results do not agree with the data reported for human IIGI'R transfcrase (15). We have pointed out that human and mouse HGl'R transferase behave identically on Sephadex G-200. 1 he disagreement arises over the subunit molecular weight of 34,000 reported by Arnold and Kelley (15 The data are most consistent with the interpretation that the enzyme is a trimer.
We cannot, however, completely rule out the possibility that the native state of the molecule is a higher multimer which has dissociated on sucrose gradients, polyacrylamide gel electrophoresis, and Sephadex gel filtration, Human and mouse HGPR transferase presumably have the same subunit composit'ion due to their identical Stokes radii and subunit molecular weights.