A derivative of Escherichia coli A19 was used in all the fermentations. The strain used had the methionine auxotrophy of A19 reverted to prototrophy in addition to the following deletions: AtnaA AspeA AtonA AendA. The strain was constructed in our laboratory. Where noted, a derivative of this strain was used with the ArelA251::kan and AspoT207::cat alleles (5). The alleles were transferred by PI transduction from the CF1693 strain which was kindly donated by Dr. Michael Cashel, National Institute of Child Health & Human Development.
A rich defined media was formulated to ensure greater consistency between cultures than is possible with the standard complex media used to grow cells for extract preparation. The media was enriched with many amino acids and vitamins to support rapid growth. The inorganic salt and mineral composition was based on a recipe used industrially for high cell density cultures (9). The vitamins were chosen based on chemical analysis sheets obtained from Difco, Inc. which indicated the typical composition of tryptone and yeast extract. All amino acids were supplied except alanine, serine, arginine, cysteine, aspartate, glutamate, and glutamine. More details on the media design will be published elsewhere.
The defined medium used consisted of the following (in g/L): (NH4)2S041.5, NaH2P04H20 3.5, KC12, trisodium citrate dihydrate 0.5. After sterilization, the following vitamins were added (in mg/L): choline chloride 28.6, niacin 25.1, p-aminobenzoic acid (potassium salt) 25.6, pantothenic acid (hemi-calcium salt) 9.4, pyridoxine hydrochloride 1.5, riboflavin 3.9, thiamine hydrochloride 17.7, biotin 0.1, cyanocobalamin 0.01, folic acid dihydrate 0.07. The following minerals were also added after sterilization (in mg/L): FeCl3-6H20 20, Na2Mo04-2H20 3.5, boric acid 1.2, CoC12-6H20 3.4, CuS04-5H20 3.4, MnS04H20 1.9, ZnS04-7H20 3.4. The iron(ffl) chloride and sodium molybdate were prepared in separate stocks with citrate to prevent precipitation. Glucose, magnesium, and amino acids were also added after sterilization (see below).
The glucose feed solution contained 250 g/L glucose and 6.33 g/L MgS04*7H20 and was autoclaved. The amino acid feed solution contained the following (in g/L): Asp-H20 48.9, Gly 70.6, His HC1H20 13.5, lie 35.1, Leu 36.3, Lys-HC1H20 30.8, Met 14.1, Phe 14.1, Pro 46.9, Thr 37.1, Tip 14.3, Tyr 17.3, Val 22.8, betaine 33.1. These concentrations are based on observed amino acid consumption rates. Approximately 180 g/L KOH was necessary to dissolve the amino acids and adjust the pH to 11-11.5. The amino acid solution was filter sterilized prior to use. Fresh glucose/magnesium sulfate and amino acid solutions were prepared for each fermentation.
Several stock vials of each strain were prepared from an overnight culture in LB media. Stocks with 20% (v/v) glycerol were frozen in liquid nitrogen and stored at -80°C until use. For the inoculum, a stock vial was thawed at room temperature and 1 mL was transferred to 5 mL of LB media in a culture tube. The remainder of the stock vial was discarded. The culture was incubated at 37°C with shaking for 4 hours. Then, 1 mL from the tube was used to inoculate 100 mL of defined media with 4 mL of the glucose feed (10 g/L glucose final concentration) and 1.86 mL of the amino acid solution in a 1 L baffled shake flask. The flask was shaken at 37°C overnight for 12-16 hours. Next, 20 mL from the overnight culture was transferred to 400 mL of fresh media with half the amount of glucose and amino acids as the overnight media. Once the culture was in exponential growth and between 2 and 3 OD595 (about 3 hours), the entire culture was used to inoculate the fermentor.
A B. Braun Biostat C fermentor with 10 L working volume was used. The standard baffles were replaced by custom machined 3 cm wide baffles to improve oxygen transfer, especially at high agitation rates. A side feed port for the glucose feed was manufacturing by drilling a hole through a blind stopper and welding a stainless steel tube (1/8 in. ID) into the hole. The outside end of the tubing was connected to a steam supply so the tube could be sterilized by steam injection during the fermentor sterilization cycle.
After sterilization, the dissolved oxygen (D02) probe was calibrated with argon and air. The MFCS/Win software supplied by B. Braun logged data and controlled the operating conditions as follows: aeration 10 L/min, temperature 37°C, pressure 1.5 bar (gauge), pH 7.2. Ammonium hydroxide (-2 M) and 5 N sulfuric acid were used to control pH. Agitation was started at 400 RPM and increased manually to maintain the D02 above 30% air saturation at atmospheric pressure.
Initially, the fermentor contained 10 L of the defined media with 93 mL of the amino acid solution added. The glucose/magnesium sulfate solution was added to a final glucose concentration of 5 g/L (200 mL). BASF Mazu® DF-204 antifoam was added (1-2 mL) to control foaming.
A laptop computer (Toshiba Portege 7200CTe) with a National Instruments DAQCard-6024E and MATLAB 6.1 software with the Data Acquisition Toolbox (The Mathworks, Inc. Natick, MA) was used to control glucose and amino acid feeding. The analog recorder output of the Biostat C was connected to the input of the data acquisition card and the card's analog outputs were used to control two MasterFlex 77521-40 peristaltic pumps with EasyLoad II pump heads (Cole-Parmer, Vernon Hills, IL). The D02 signal from the Biostat C (1 Hz sampling rate) was analyzed by a MATLAB program to detect sharp increases or drops. The point-to-point differences in the most recent 5 samples were averaged and compared to the averaged slope for the previous 85 samples (5-89 seconds ago). If the difference in the two slopes was greater than a threshold value, a sudden rise or drop in D02 was signaled.
Glucose feeding was started 3.5 hrs after inoculation at a constant 0.6 mL/min to avoid a period of complete glucose exhaustion at the transition from batch to feeding mode. Glucose limitation was detected by a sharp spike in the D02 which usually occurred between 4 and 4.5 hours of fermentation time. Amino acid feeding was started at 4.25 hrs or when glucose became limiting, whichever came first. Amino acids were fed at 27% of the glucose feed rate, except during glucose pulses when the amino acids feed was maintained on the previous smooth exponential feed profile.
Once glucose limitation was detected, glucose pulse testing started. The glucose feed rate was increased temporarily to supply excess glucose and then reduced until a rise in D02 was detected. The glucose feed changes were scaled to the current feed rate as follows: feed below 2.5 mL/min, up 75%, down 50%; feed between 2.5 and 5 mL/min, up 50%, down 25%; feed above 5 mL/min, up 30%, down 25%. If glucose was limiting, the pulse of glucose would cause the D02 to drop as metabolism sped up. Once die excess glucose was depleted, metabolism would slow and D02 would rise. If a pulse failed to drop the D02, glucose was assumed to be in excess. The reduced glucose feed rate was maintained until a rise in D02 indicated glucose limitation. The pulse test was then repeated. The total glucose added during the pulse test was divided by the time from pulse start to D02 rise to determine the glucose utilization rate. Glucose feeding was then resumed at a feed rate set to 90-100% of the calculated uptake rate. The feed rate was continuously increased exponentially between pulse tests with a rate constant between 0.3 and 1.0 /hr depending on the fermentation (see Results for details). Pulse tests were initiated every 15-20 minutes to adjust the feed rate as necessary.
Samples were withdrawn from the side sampling port on the fermentor and passed through a cooling coil in an ice-water bath before being collected in ice cold 15 mL disposable tubes. Approximately 40 mL of culture was flushed through the sampling system before a sample of about 10 mL was collected. A 1 mL portion of the sample was immediately centrifuged at 20,000 g, 4°C for 5 min. Two 0.4 mL portions of the supernatant were frozen in liquid nitrogen and stored at -80°C until later analysis. The cell density was measured at 595 nm by diluting the sample as necessary to get a reading between 0.1 and 0.25 AU.
Glucose and acetate in the sample supernatants were quantitated using an Aminex HPX-87H HPLC column (Bio-Rad, Hercules, CA) on an Agilent 1100 HPLC with a refractive index detector. An isocratic elution was performed with 5 mM H2S04 at 0.4 mL/min and 35°C for 40 min. Injections were 20 ^L of undiluted sample supernatant. A calibration curve was prepared with standards from 0.1 to 100 mM. The minimum detectable concentrations were about 0.1 mM for acetate and 0.05 mM for glucose.
Inorganic phosphate in the sample supernatants was assayed using the procedure of Saheki et al (70). Samples were diluted to be in the range of standards (0.05-2 mM) and assayed in triplicate.
Extract preparation and protein synthesis of chloramphenicol acetyl transferase (CAT) were done as previously described (77). The extracts designated "2YTPG" in Figure 9 were prepared from cells grown in the traditional manner (9) except that 18 g/L glucose, 3.0 g/L KH2P04, and 7.0 g/L K2HP04 were added to the growth media as described in (12). We have seen no effect of the glucose and phosphate addition on CAT production in this system (unpublished data).
Was this article helpful?