Efficient production and isolation of recombinant amino-terminal half-molecule of human serum transferrin from baby hamster kidney cells

PROTEIN

EXPRESSION

AND PURIFICATION

2, 214-220

(19%)

Efficient Production and Isolation of Recombinant Amino-Terminal Half-Molecule of Human Serum Transferrin from Baby Hamster Kidney Cells Anne

B. Mason,*

Walter

D. Funk,t,l

Ross T. A. MacGillivray,?

and Robert

*Department Biochemistry,

of Biochemistry University of Vermont College of Medicine, Burlington, Vermont University of British Columbia, Vancouver, British Columbia, Canada V6T 1 W5

Received

12, 1991,

June

and in revised

form

July

The transferrins are a group of glycosylated metalbinding proteins that function in the transport of iron to cells and as bacteriostatic agents in a variety of biological fluids (l-6). Examination of the primary structures of various transferrins indicates that a gene duplication event occurred in evolution resulting in the present 80kDa, two-sited protein that binds two metal ions in homologous binding domains. X-ray crystallographic studies (7-10) reveal the presence of two globular lobes each made up of two domains which define a deep cleft conaddress: of Texas

Department of Cell at Dallas, Southwestern

05405; and TDepartment

of

12, 1991

Expression of the amino-terminal lobe of human serum transferrin secreted into the culture medium by transformed baby hamster kidney (BBR) cells has been increased from the levels reported originally of lo-15 Fg/ml to 55-120 fig/ml. Use of the serum substitute, Ultraser G, has facilitated isolation of the recombinant protein, resulting in -80% recovery of expressed hTF/ 2N from the culture medium. In the three experiments described, 300-760 mg of recombinant protein was collected over a period of 25 days from five expanded surface roller bottles each containing 200 ml of medium (seven to nine collections). The use of alginate beads to encapsulate the transformed BIIK cells provided no advantage over normal culturing over 25 days. A lag in production resulting in 30% less recombinant protein over this time period was observed. The production and isolation procedures described are easily handled by one person. The system is amenable to incorporation of isotopically substituted amino acids useful in NMR studies. 0 1991 Academic Press, Inc.

’ Present University TX 75235.

C. Woodworth*

Biology and Neuroscience, Medical Center, Dallas,

taining the binding site for each metal ion and a synergistic anion. Segregation of the two 40-kDa, iron-binding lobes of ovotransferrin has proved useful for studies by high-resolution NMR spectroscopy (llJ2). Although both the amino- and carboxyl-terminal lobes of human serum transferrin have been prepared proteolytically (13-16), the yields are low and the preparations are not always homogeneous. In contrast to ovotransferrin in which trypsin cleavage of the diferric protein yields both halfmolecules (17,18), no proteolytic scheme has been found which renders both lobes of human transferrin simultaneously. The presence of extra disulfide bonds that restrict the hinge region of human transferrin to seven amino acid residues probably accounts for this finding (19,20). Additional rationale for pursuing a molecular biological approach for preparation of the two ironbinding domains of human transferrin has been given in detail (21). This approach allows preparation of homogenous, well-defined fragments into which isotope-substituted amino acids useful in NMR studies can be incorporated easily and economically. An added advantage is the possibility of pursuing site-directed mutagenesis experiments aimed at further elucidation of the role of certain amino acids which lie in or near the iron-binding site. Clearly the signals for synthesis, processing, and secretion of proteins reside in the synthetic machinery of mammalian cells. Thus, although bacterial systems have been widely used as expression systems for recombinant human proteins, they are not always successful. After efforts to express the human amino-terminal half-molecule lobe in Escherichia coli failed, successful expression of functional hTF/2N was achieved using baby hamster kidney (BHK) cells stably transfected with the pNUT-hTF/2N vector (21). Expression of hTFI2N is driven by a mouse metallothionein 1 pro-

214 Copyright 0 1991 All rights of reproduction

1046-5928/91$3.00 by Academic Press, Inc. in any form reserved.

RECOMBINANT

TRANSFERRIN

moter, while a cDNA encoding a mutant form of dihydrofolate reductase is driven by the SV40 promoter, allowing selection of transfected cells in the presence of 500 pM methotrexate. The slower growth and lower total mass of cells that grow in an anchorage-dependent manner make it more difficult to achieve high levels of expression using mammalian cells. In addition, the requirement for animal serum in the culture medium can complicate the subsequent isolation of the recombinant protein. The present study sought to optimize conditions under which the BHK cells were cultured both to maximize expression levels and to make the subsequent isolation of hTFI2N faster, more efficient, and more reproducible. MATERIALS

AND

METHODS

Materials Dulbecco’s modified Eagle’s medium-Ham’s F-12 nutrient mixture (DMEM-F-12) was obtained with and without phenol red from Sigma. Defined and supplemented newborn calf serum (Cat. A-2151-D) was obtained from Hyclone and pretested to assure adequate growth of BHK cells. The serum replacement Ultraser G came from either GIBCO or Serva. A penicillin/streptomycin sulfate solution was from GIBCO. Low-viscosity sodium alginate was purchased as a sterile solution from Bellco Glass. Cellift antifoam reagent was from Ventrex Lab. Corning expanded surface roller bottles and Dynatech Removawells were obtained from a local distributer. All of the chromatographic resins, DEAESephacel, Sephacryl S-lOOHR, and Polyanion SI, were from Pharmacia. Methotrexate from Cetus was purchased at a local hospital pharmacy. Immersible-CX Ultrafilters were obtained from Millipore. Centricon 10 microconcentration and PM-10 ultrafiltration membranes were from Amicon. Rabbit anti-mouse immunoglobulin G was purchased from Organon Teknika. All chemicals and reagents were analytical grade or purer. A monoclonal antibody designated aHT+N, was prepared in our laboratory and found to be specific for the amino-terminal lobe of transferrin. A complete description of this antibody is given elsewhere (22). This antibody has no reactivity with bovine transferrin. Methods Cells and cell maintenance. Baby hamster kidney cells were transfected with the expression vector pNUT-hTF/2N, which encodes the natural signal sequence and the amino-terminal half-molecule lobe (hTFI2N) of human serum transferrin followed by two stop codons. This expression system has been thoroughly described (21). Frozen stocks of BHK cells transformed by the pNUT-hTF/2N plasmid were stored in liquid nitrogen in 95% fetal calf serum and 5%

HALF-MOLECULE

PRODUCTION

215

dimethyl sulfoxide (DMSO). Cells (~1 X 106) were initially brought up in DMEM-F-12 containing 5% newborn calf serum, penicillin (100 units/ml), and streptomycin (100 pg/ml). Methotrexate (500 PM) was used routinely in all cultures in which cells were harvested for frozen stocks. Methotrexate was absent in ceils that were expanded for mass culture. In a typical experiment, cells were passed at -80% confluency using Versene. Sequential passage of the original cells was to five loo-mm dishes (10 ml volume), then to five T-175 flasks (30 ml volume), and finally to five expanded surface roller bottles (200 ml volume). At the T-175 passage, a serum substitute, Ultraser G, at a level of l%, was used instead of 5% fetal calf serum in DMEM-F-12 lacking phenol red. Cells in the roller bottles were kept in a Bellco cell production roller apparatus set at a speed of 1 rpm. Alginate bead entrapment. Entrapment of the cells in alginate beads was accomplished essentially as described in the Bellco instructions. Briefly, a mixture of 25 ml of cell suspension (from five T-175 flasks harvested at 80% confluency) and 225 ml of sodium alginate solution was slowly dripped from two 60-ml syringes via silicone tubing (3 mm, id.) into a sterile l-liter beaker containing 250 ml of a calcium solution (100 mM CaCl,, 2.5 mM glucose, and 25 mM Hepes, pH 7.2), 250 ml of DMEM-F-12 medium with 1% Ultraser G, and 1 ml of an antifoam solution. This solution was stirred slowly with a magnetic bar. The syringes were driven by a Harvard pump (Model 977) at the lowest setting. After formation, the beads (-400 ml volume) were allowed to settle, the supernatant was decanted, and 900 ml of DMEM-F-12 medium lacking Ultraser G was added. The settling, decantation, and washing steps were repeated three times. The beads were distributed evenly to five roller bottles and the volume of each was adjusted to 300 ml with the DMEM-F-12-Ultraser G medium. All of these operations were carried out under sterile conditions in a laminar flow tissue culture hood. The original isoIsolation of recombinant hTFI2N. lation procedure (21) has been modified to expedite the process and give higher yields of recombinant product more reproducibly. Harvested culture medium was made 0.01% with respect to phenylmethanesulfonyl fluoride (100 mg/liter in 1.5 ml ethanol) and sodium azide to inhibit proteases and bacterial growth, respectively. Sufficient Fe3+ (nitrilotriacetic acid), was added to saturate all the transferrin present and to displace any zinc bound to the transferrin. The medium was then reduced in volume to 50 ml or less using either a 400-ml stirred cell fitted with a PM-10 membrane or eight immersibleCX Ultrafilters. In both cases the medium was kept at 4’C. The culture medium was processed in this manner and frozen until the end of the culture period. After thawing, further reduction, and exchange with MilliQ

216

MASON

H,O, the medium was centrifuged at 10K rpm for 15 min in an SS-34 rotor in a Sorvall RCBB refrigerated centrifuge. The concentrated medium was then loaded onto a column of DEAE-Sephacel (2.5 X 40 cm) equilibrated with 10 mM Tris-HCl, 0.02% NaN,, pH 8.0. The column was developed with a single step consisting of 500 ml of 150 mM Tris-HCl, pH 8.0. All colored fractions from this elution were pooled, concentrated to ~10 ml and loaded onto a Sephacryl S-100HR column (5 X 80 cm) equilibrated and run in 0.1 M NH,HCO,. The column was pumped from the bottom with a Pharmacia P-l pump set to give a flow rate of -60 ml/h. Fractions of 140 drops/tube were collected. In most cases the recombinant hTF/SN was pure at this point. Purity was assessed by determining the A,,/A,,, and A,,,IA,, ratios, as well as by gel electrophoresis (see below). Final purification using a Pharmacia FPLC and a column of Polyanion SI (1 X 10 cm) was sometimes required. This step has been described in detail (21). Protein concentration was determined by absorbance measurement using the following coefficient and molecular weight for Fe-hTF/2N: E,, nm,i% = 12.6 and 37,132 (21). SDS-PAGE was carried out using 5-15% gradient gels and a 4% stacker (23). Radioimmunoassay of recombinant hTFI2N. A competitive solid-phase immunoassay was used to determine the concentration of recombinant hTF/2N in the culture medium and at various stages of the purification (21). Briefly, 1 pg of rabbit anti-mouse immunoglobulin G in 100 ~1 of 14 mM Na,CO,, 34 nM NaHCO, was added to Removawells and incubated overnight at 4°C. The wells were washed three times with 200 ~1 of assay buffer (50 mM Tris-HCl, pH 7.4, containing 0.1 M NaCl, 0.02% NaN,, and 0.1% bovine serum albumin). An appropriate dilution (1:8300 in assay buffer) of the monoclonal antibody, designated crHT+N,, in a volume of 200 ~1 was added to all wells except those used to determine background. Following incubation for l-l.5 h at 37”C, each well was washed three times with 200 ~1 of assay buffer. Pure recombinant hTF/2N radioiodinated with Iodogen (21) was added at a level of 2.5 ng/200 ~1 in the presence or absence of unlabeled standards and samples. A standard curve was generated by competition of Fe-lz51-hTF/2N with 8-200 rig/well of unlabeled recombinant hTF/BN. After incubation for l-2 h at 37”C, the wells were washed as above, separated, placed into Omnivials, and assayed for radioactivity. RESULTS

While our protocols for production and isolation of the recombinant human amino-terminal half-molecule domain from BHK cells were being fine-tuned, several observations which have become the foundation of the current report were made. First, it was noted that the BHK cells containing the plasmid grew better in

ET

AL.

DMEM-F-12 than in DMEM alone. The presence of 15 mM Hepes buffer in this medium appears to stabilize the cells, especially at the roller bottle stage. The presence of zinc sulfate at a level of 432 pg/liter may contribute to the increased production observed. Second, whereas in the initial experiments with DMEM or DMEM-F-12 the roller bottles were gassed twice a day with 5% CO,95% air it was noted that this had the opposite effect on pH than was desired. Gassing was therefore discontinued. Third, although the growth of the cells and production of recombinant hTFl2N was adequate in either fetal or newborn calf serum at a level of 5%, isolation of the recombinant protein from the medium containing serum was tedious and time consuming. The use of Ultraser G, a serum substitute, led to equal or better production of the recombinant protein and greatly simplified the isolation procedure. BHK cells containing the plasmid expressing the recombinant hTF/BN were passed into five expanded surface roller bottles as described under Methods. The pH of the medium was monitored daily. It had been noted that production of hTF/2N correlated, as might be expected, with cell density and also with a drop in pH. During the first 5 days in the roller bottles, as the cells are growing to confluency, the pH was: Day 1,7.6; Day 2, 7.2; Day 3, 7.1; Day 4,6.9; and Day 5,6.8. After a change of medium, the pH on the subsequent 3 days was 7.3, 7.1, and 6.9. Following another medium change the pH was 7.2, 6.9, and 6.7. For the remainder of the experiment the pH dropped to 6.7 or below by the second day after a medium change. As indicated in Fig. 1A the expression of hTF/BN initially climbed from -50 pg/ml on Day 6 to a maximum of 120 pg/ml on Day 22. The production rate then gradually fell from -80 pg/ml to a low of -20 pg/ml in the next 40 days in culture. As shown in Fig. lB, 1500 mg of recombinant hTF/2N was produced by the cells and secreted into the medium over the 60-day time course. The main reason for the fall in the production rate of recombinant hTF/2N appeared to be the loss of cell mass from the roller bottle surface. Thus, as the culture period progressed, patches of BHK cells sloughed off the surface into the medium. A possible solution to this problem was to use alginate beads to encapsulate the cells. Following the harvesting and pooling of BHK cells from 10 T-150 flasks, half of the cells were transferred directly into five expanded surface roller bottles while the other half were encapsulated in alginate beads as described under Methods. As shown in Fig. 2A the production rate (pg/ml) of hTF/2N encapsulated in the beads lagged behind that observed from the cells adhering to the surface of the roller bottle. Eventually, however, the production rate reached a similar level and maintained that level for the duration of the particular experiment. The data shown in Fig. 2A give the standard deviations from analysis of the culture media done

RECOMBINANT

TRANSFERRIN

HALF-MOLECULE

I

n4 0

10

20

30

40

50

60

B

/

217

PRODUCTION

born or fetal calf serum. It was found that five to six batches of culture medium (- 1000 ml each) could be processed simultaneously using the protocol described under Methods. In our previous work, only 1000 ml was processed at a time. The chromatograms from the DEAE-Sephacel, Sephacyl S-lOOHR, and Polyanion SI columns are shown in Figs. 4A, 4C, and 4D. Samples of each peak from the various columns were run on an SDS gel (Fig. 4B) to monitor the isolation procedure. Recovery data are presented in Table 2. Approximately half of the Azm units in the original medium were eliminated in the reduction and exchange step which proceeds the DEAE-Sephacel column. Approximately 25% of the A,, units placed onto this column was recovered in the single-step elution. For the recombinant hTF/PN there was an -80% recovery following the DEAE-Sephacryl column. Virtually all of the hTF/BN placed onto the Sephacyl S-100HR column was recovered. The overall recovery of hTFI2N was -80%. Similar results have been obtained from many different preparations.

60A 60 -10

20 Days

30

in roller

40

50

60

bottles

FIG. 1. Production rate, expressed as micrograms per milliliter (A) and as total cumulative milligrams collected (B), of hTF/ZN produced by BHK cells in continuous production in five expanded surface roller bottles. On the days indicated -1000 ml of DMEM-F-121% Ultraser G medium was collected and assayed for recombinant product. One bottle was lost to contamination on Day 48 and a second on Day 52.

on two separate days with two separate standard curves. A typical standard curve is shown in Fig. 3. In terms of overall production for the 25 days during which media were collected, the roller culture cells produced -430 mg vs 290 mg for the BHK cells in alginate beads. Thus the cells adhering to the roller bottle produced -30% more recombinant protein than the encapsulated cells over this period. In the course of setting up BHK cells containing the plasmid expressing hTF/2N, a range of production levels was found. The data presented in Table 1 show the maximum product expression in a given experiment (counting the day the cells were placed in the roller bottles as Day 1) and the culture conditions including serum vs serum substitute and adherent cells vs alginate beads. Having achieved a reasonable level of production of hTF/BN, it was necessary to develop a more efficient isolation procedure. The use of Ultraser G aided in this endeavor by reducing the protein content of the culture supernatant -4.5-fold relative to that of new-

0

5

10

15

Days

in roller

20

25

-I 30

bottles

FIG. 2. Production rate, expressed as micrograms per milliliter (A) and as total cumulative milligrams collected (B), of hTF/PN produced by BHK cells in continuous production. The cells were either adhering to five expanded surface roller bottles (m) or entrapped in alginate beads in five expanded surface roller bottles (A). On the days indicated -1000 ml of DMEM-F-12-1% Ultraser G medium was collected and assayed for recombinant product. Error bars in A indicate standard deviation from two complete sets of determinations, each done in duplicate.

218

MASON

l.O--

0 s .u .Fl F c3

-l.O--2.0

--

-3.0

--

-4.0

4 0.5

1.0 Log

1.5 rig/well

2.0

2.5

FIG. 3. Standard curve for the competitive radioimmunoassay used to determine the amount of recombinant hTF/ZN in the culture medium. Monoclonal antibody (uHT+N, (specific to the amino-terminal domain of human serum transferrin) was incubated overnight at 4°C on rabbit anti-mouse IgG-coated Removawells. Radioiodinated hTFI2N preincubated with unlabeled competitor (3-200 ng) was added to the washed Removawells and the binding was measured after incubation for 1.5 h at 37°C and washing. Each point is the mean of duplicate determinations. Logit B/B, = In [(B/&)/l-B/B,,)], where B is the amount of radiolabeled sample bound in the presence of competitor and B, is the amount of radiolabeled hTFI2N bound in the absence of unlabeled competitor. Standard curves are highly reproducible and typically have correlation coefficients of 0.99 or better.

DISCUSSION

In the original report of the expression of hTF/BN using the expression vector pNUT-hTF/2N, levels of lo-15 pg/ml of recombinant protein were realized. In the present study, conditions have been changed to achieve levels ranging from 55 to 120 pg/ml of hTF/2N. In our work we have sought to increase expression levels on a relatively small scale suitable to manipulation in an academic environment with limited resources. It should be clear from the work described that we have not exhaustively examined every variable but have developed a system that allows production of one to several

TABLE Maximum

Experiment No.

Rate

of Production

of hTF/2N, of BHK

Maximum hTF/ZN (bdml) 43 35 54 119 60 96 56 68

Note.

Eight

separate

production

runs

AL.

hundred milligrams of recombinant product in a reasonable amount of time. Likewise in the isolation procedure we have sought to maximize recovery and minimize the number of steps and time required. We believe that both of these goals have been realized. The current protocol for production and isolation can easily be handled by one person. Major improvements over our previous protocol include the substitution of DMEM-Ham’s F-12 containing 15 mM Hepes for the DMEM used in the original protocol. The BHK cells grow better and have consistently higher production rates with this richer medium with a minimum of handling, i.e., no gassing of bottles. The use of Ultraser G, a serum substitute, led to equal or better production of the recombinant protein and greatly facilitated the subsequent isolation procedure. A potential problem with using Ultraser G is that it contains human holo-transferrin, thereby compromising the radioimmunoassay used to measure the amount of recombinant hTF/2N in the medium. It was found, however, that the level of holo-hTF is sufficiently low (1.09 ? 0.31 pglml, n = 7) relative to the recombinant protein such that the presence of holo-hTF has little effect. Thus the 5000 ml of culture medium which was the starting material in Table 1 contains -5.5 mg of holohTF compared to -270 mg of recombinant hTF/BN. Furthermore, the holo-hTF is completely separated from the hTF/2N on the Sephacyl S-100HR column. The continued use of expanded surface roller bottles rather than bioreactors seems warranted because the bottles are disposable and allow several different cell lines to be grown simultaneously. The principal goal of our research is to produce recombinant hTFI2N and various mutants of this protein for use in NMR studies. Such work requires -1.0 ml of a l-2 mM protein solution. To simplify the interpretation of the NMR spectrum, substituted amino acids are often used in the culture medium. Using the system described above we have successfully incorporated [c-13C]methionine and fully or

1

Expressed as Micrograms per Cells in Continuous Culture

Days

in culture 14 12 11 22 19 19 19 22

are shown.

ET

Milliliter

of Culture

Supernatant,

Conditions 5% 5% 1% 1% 1% 1% 1% 1%

Fetal calf Newborn Ultraser Ultraser Ultraser Ultraser Ultraser Ultraser

serum calf serum G G G G G G

Cells Cells Cells Cells Cells Beads Beads Beads

in in in in in

roller roller roller roller roller in roller in roller in spin

bottle bottle bottle bottle bottle bottle bottle flask

RECOMBINANT

TRANSFERRIN

HALF-MOLECULE

219

PRODUCTION 123456789

2.0

A 1.5

g

c”

10 . 0.5

Tube

p ;~I~ 60

70

I

number

60

90

100

110

0

10

20

30 Time

40

50

60

(min)

FIG. 4. Chromatogram of culture supernatant containing recombinant hTF/BN on a DEAE-Sephacel column (A). Chromatogram of peak a from the DEAE column on a Sephacryl S-100HR column (C). Chromatogram of peak e from the S-100HR column on a Polyanion SI column (D). Electrophoresis of fractions from each column on SDS-PAGE (5-X% gradient gel) under reducing conditions visualized with Coomassie blue (B). Samples are: lanes 1 and 9, Bio-Rad low-molecular-weight standards with molecular weights, top to bottom, of 97,406,66,200,45,000, 31,000,21,500, and 14,400; lane 2, original culture supernatant post reduction but prior to DEAE column; lanes 3-8, fractions from the columns as indicated by the lowercase letters at the bottom. The arrowhead to the right of lane 9 indicates the position of the recombinant hTF/ZN. Approximately 30 pg of protein was loaded in lanes 2 and 3. Approximately 15 pg of sample was loaded in lanes 4-8.

partially deuterated phenylalanine, tyrosine, tryptophan, and histidine. We have found that with the cells adhering to the roller bottle surface, 100% incorporation of the substituted amino acids is achieved with a single medium change. In addition, we have incorporated enough fluorotryptophan to obtain a strong NMR signal, although this compound has a deleterious effect on the BHK cells over time. The use of alginate beads to encapsulate the BHK cells containing the plasmid appeared to be effective in prolonged production levels, at least over 25 days (Fig. 2). The decline in production (Fig. 1) by adherent cells appeared to be related to a loss of cell mass from the roller bottle surface. Use of the beads may be reasonable for the cells in which adherence is a problem or in which long-term culturing is necessary due to low production levels. As a routine procedure, however, the beads offer no advantage and in fact have some disadvantages. These include the additional manipulations involved in encapsulating the cells, the additional culture medium required to wash the newly made beads, and the longer

wash in/out times required to introduce synthetic amino acids. Finally, as shown in Fig. 2, there is a lag associated with the beads which led to -30% less recombinant protein over the 25-day period than was found for the cells adhering to the surface of the roller bottles. Our previous work (21) showed that the recombinant hTFI2N produced by the BHK cells was monodisperse on SDS-PAGE and competent in terms of iron binding as shown by urea gels and by spectral data. The ability of the recombinant protein to bind iron reversibly implies that the eight disulfide linkages are correctly formed such that the three-dimensional structure required for iron binding is intact. Using our previous isolation protocol, a “minor” form of recombinant hTF/BN was isolated after chromatography on the Polyanion SI column. The minor form, which represented 14% of the total recombinant protein, had the same mobility on SDS-PAGE and the same amino acid sequence at the N-terminus. It differed in its elution from the Polyanion SI column and in its behavior on urea gels. In addition a mixture of both apo and iron-containing major and

220

MASON TABLE

Recovery BHK

of Cells

2

Recombinant hTF/ZN from Containing

Original culture medium Original culture medium Post reduction/Pre DEAE column Post DEAE-Sephacel Post S-100 HR Fraction 1 Fraction 2 Fraction 3 Fraction 4

the

the Supernatant Expression Vector

hTF/ZN

of

Volume (ml)

Total 4,

hTF/BN (md

Yield (So)

5000

7950

268

100

R

150 230

3611 927

282 227

105 85

R R

80 80 122 120

85 296 164 287

0 2 10 228

4 85

R R R” S

minor recombinant hTFI2N was observed. In our current protocol we do not observe the minor form and obtain all of the recombinant half-molecule in its iron form. Therefore we suggest that the minor form may have been an artifact of the isolation procedure, which was more time consuming and involved more handling of the sample. In conclusion, expression levels of the amino-terminal lobe of human transferrin in transformed BHK cells have been increased 5- lo-fold by some simple changes in the culturing protocol. In addition isolation of the recombinant protein from the culture medium has been simplified and accelerated such that recoveries of -80% are routinely achieved. The system as described is ideal for the production of the recombinant hTF/BN and various mutants of this protein for use in NMR studies. ACKNOWLEDGMENT work

was supported

by Grant

DK21739

from

the USPHS.

REFERENCES 1. Aisen, P., and proteins. Annu.

AL. 7. Anderson, B. F., Baker, H. M., Dodson, ball, S. V., Waters, J. M., and Baker, human lactoferrin at 3.2 A resolution. 84, 1769-1773.

E. J., Norris, G. E., RumE. N. (1987) Structure of Proc. Natl. Acad. Sci. USA

8. Anderson, B. F., Baker, H. M., Norris, G. E., Rumball, S. V., and Baker, E. N. (1990) Apolactoferrin structure demonstrates ligand-induced conformational change in transferrins. Nature 344, 784-787. 9. Anderson, B. F., Baker, H. M., Norris, G. E., Rice, D. W., and Baker, E. N. (1989) Structure of human lactoferrin: Crystallographic structure analysis and refinement at 2.8 A resolution. J. Mol. Biol. 209, 711-734. 10.

Note. R, the amount of recombinant hTF/BN was determined by competitive radioimmunoassay. S, amount of recombinant hTF/PN was determined by spectral analysis. a The DMEM-F-12-1% Ultraser G medium contains -5.5 mg of holo-TF, most of which elutes in this fraction.

This

ET

Listowsky, I. (1980) Iron transport Rev. Biochem. 49, 357-393.

and

storage

2. Brock, J. H. (1985) Transferrins, in “Metalloproteins, Part II. Metal Proteins with Non-redox Roles” (Harrison, P., Ed.), pp. 183-262, Macmillan & Co., London. 3. Huebers, H. A., and Finch, C. A. (1987) The physiology of transferrin and transferrin receptors. Physiol. Rev. 67, 520-582. 4. Harris, D. C., and Aisen, P. (1989) Physical biochemistry of the transferrins, in “Iron Carriers and Iron Proteins” (Loehr, T. M., Ed.), pp. 239-351, VCH Publishers, New York. 5. Aisen, P. (1989) Physical biochemistry of the transferrins: Update, 1984-1988, in “Iron Carriers and Iron Proteins” (Loehr, T. M., Ed.), pp. 353-371, VCH Publishers, New York. 6. Thorstensen, K., and Romslo, I. (1990) The role of transferrin in the mechanism of cellular iron uptake. B&hem. J. 271, l-10.

Bailey, S., Evans, R. W., Garratt, R. C., Gorinsky, B., Hasnaint, S., Horsburgh, C., Jhoti, H., Lindley, P. F., Mydin, A., Sarra, R., and Watson, J. L. (1988) Molecular structure of serum transferrin at 3.3-A resolution. Biochemistry 27, 5804-5812. R. C., Butcher, N. D., Brown, S. A., and Brown-Ma11. Woodworth, son, A. (1987) ‘H NMR study of effects of synergistic anion and metal ion binding on pH titration of the histidinyl side-chain residues of the half-molecules of ovotransferrin. Biochemistry 26, 3115-3120. R. C. (1986) Effects of synergistic anions on the pro12. Woodworth, ton magnetic resonance spectra and pH titration of histidinyl side chains in the C-terminal half-molecule of ovotransferrin. J. Znorg. Biochem. 28,245-251. 13.

Lineback-Zins, J., and Brew K. (1980) Preparation andcharacterization of an N-terminal fragment of human serum transferrin containing a single iron-binding site. J. Biol. Chem. 256, 708713. 14. Zak, O., Leibman, A., and Aisen, P. (1983) Metal-binding properties of a single-sited transferrin fragment. Biochim. Biophys. Acta 742,490-495. 15. Valcour, A. A., and Woodworth, R. C. (1987) Proton magnetic resonance spectroscopy of human transferrin N-terminal halfmolecule: Titration and hydrogen-deuterium exchange. B&hemistry 26,3120-3125. P. (1985) Preparation and properties of a 16. Zak, O., and Aisen, single-sited fragment from the C-terminal domain of human transferrin. Biochim. Biophys. Acta 829, 348-353. 17.

Oe, H., Doi, E., and Hirose, boxyl-terminal half-molecules a novel procedure and their 1072.

18. Williams, molecule

J., and fragments

M.

(1988) Amino-terminal and carof ovotransferrin: Preparation by interactions. J. B&hem. 103,1066-

Moreton, K. (1988) The of transferrin. Biochem.

19. MacGillivray, R. T. A., Mendez, E., Sinha, Lineback-Zins, J., and Brew, K. (1982) The sequence of human serum transferrin. Proc. 79,2504-2508.

dimerization of halfJ. 251,849-855. S. K., Sutton, M. R., complete amino acid Natl. Acad. Sci. USA

20.

MacGillivray, R. T. A., Mendez, E., Shewale, J. G., Sinha, S. K., Lineback-Zins, J., and Brew, K. (1983) The primary structure of human serum transferrin: The structures of seven cyanogen bromide fragments and the assembly of the complete structure. J. Biol. Chem. 258,3543-3553.

21.

Funk, W. D., MacGillivray, R. T. A., Mason, A. B., Brown, S. A., and Woodworth, R. C. (1990) Expression of the amino-terminal half-molecule of human serum transferrin in cultured cells and characterization of the recombinant protein. Biochemistry 29, 1654-1660.

22.

Mason, A. B., and Woodworth, R. C. (1991) Monoclonal antibodies to the aminoand carboxyl-terminal domains of human transferrin. Hybridoma 10. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227,&W-685.

23.