Determination of a Dynamic Feeding Strategy for Recombinant Pichia pastoris Strains

Abstract

The knowledge of certain strain specific parameters of recombinant P. pastoris strains is required to be able to set up a feeding regime for fed-batch cultivations. To date, these parameters are commonly determined either by time-consuming and labor-intensive continuous cultivations or by several, consecutive fed-batch cultivations. Here, we describe a fast method based on batch experiments with methanol pulses to extract certain strain characteristic parameters, which are required to set up a dynamic feeding strategy for P. pastoris strains based on specific substrate uptake rate (q_s). We further describe in detail the course of actions which have to be taken to obtain the desired dynamics during feeding.

1 Introduction

Recombinant protein production with the methylotrophic yeast Pichia pastoris is a key process not only in academic research but also in the biopharmaceutical industry. To date, several of the implemented fermentation strategies for P. pastoris are based on the Invitrogen protocol suggesting constant feeding profiles for fed-batch cultivations (http://tools.invitrogen.com). Different strategies, like a feed forward regime based on a constant specific growth rate (μ; e.g., [1, 2]), are based on this protocol. However, these strategies do not aim at minimizing substrate consumption or at improving and optimizing the specific productivity.

One of the major goals of each recombinant protein production process is a maximum productivity. Currently, the outcomes regarding a possible interdependency between the specific productivity (q_p) and the specific growth rate (μ) of P. pastoris are inconsistent as some studies show that q_p does not relate to μ [1, 3, 4], whereas another study demonstrates growth association [5]. Due to these controversial findings, another parameter than μ, namely, the specific substrate uptake rate (q_s), has been analyzed for its possible correlation with q_p recently [6-8].

However, regardless of which control parameter is chosen to set up a feeding profile for P. pastoris, specific parameters of the respective strain have to be determined. This can either be done by very time-consuming and labor-intensive continuous cultivations [9] or by several, consecutive fed-batch cultivations [10, 11]. However, to meet industrial demands, a fast and easy-to-do characterization of recombinant P. pastoris strains to extract bioprocess-relevant strain characteristic parameters for the subsequent set-up of production processes is essential.

Here, we describe a novel, fast method based on batch experiments with methanol pulses to extract a minimal set of strain characteristic parameters, which are required to set up a dynamic feeding strategy for P. pastoris strains based on q_s.

2 Materials

Prepare all media and solutions with analytical grade reagents and deionized water.

2.1 Medium for Preculture

Yeast nitrogen base (YNB) medium per L: 1.0 M potassium phosphate buffer (dissolve 118.1 g KH₂PO₄ and 23.0 g K₂HPO₄ in 1,000 mL distilled water, pH 6.0), 3.4 g YNB w/o Amino acids and Ammonium Sulfate, 10 g (NH₄)₂SO₄, 400 mg biotin, 20 g glucose. Weigh YNB w/o amino acids and ammonium sulfate (NH₄)₂SO₄, biotin, and glucose in a beaker. Add 100 mL of 1.0 M potassium phosphate buffer (pH 6.0), dissolve by stirring and set with water to 1 L. Filter-sterilize through a 0.2 μm cutoff filter into a sterile flask. If the strain of interest carries a resistance gene, add the respective antibiotics according to the specific resistance marker present in the strain (e.g., Zeocin, Kanamycin) to an appropriate concentration (e.g., 100 μg/mL medium). Store at 4 °C.

2.2 Medium for Batch and Fed-Batch Culture

1. Basal salt medium (BSM) per L: 26.7 mL of 85 % (v/v) phosphoric acid; 1.17 g CaSO₄ · 2H₂O, 18.2 g K₂SO₄, 14.9 g MgSO₄ · 7H₂O, 4.13 g KOH, 44 g C₆H₁₂O₆ · H₂O, 0.3 mL Antifoam. Weigh the chemicals in a beaker, dissolve in around 600 mL of water, and then fill up to 725 mL in a measuring cylinder. Fill the bioreactor with this medium and autoclave.

2. C-source per L: weigh 220 g glucose monohydrate or 200 g glycerol (depending on which C-source you prefer) and fill with water to 1 L. Autoclave and store at 4 °C.

3. Trace metal solution (PTM1) per L: 6.0 g CuSO₄ · 5H₂O, 0.08 g NaI, 3.0 g MnSO₄ · H₂O, 0.2 g Na₂MoO₄ · 2H₂O, 0.02 g H₃BO₃, 0.5 g CoCl₂, 20.0 g ZnCl₂, 65.0 g FeSO₄ · 7H₂O, 0.2 g biotin, 5 mL H₂SO₄. Weigh the chemicals in a beaker and fill with water to 1 L. Filter-sterilize through a 0.2 μm cutoff filter into a sterile flask and store at room temperature.

4. Base solution to set the pH: 2–3 M NH₄OH.

2.3 Methanol Solution for Pulses

For methanol pulses during batch cultivations use pure methanol supplemented with PTM1. For this purpose add 12 mL PTM1 solution into 1 L pure methanol, filter-sterilize this solution into a sterile bottle and store this solution at 4 °C.

2.4 Feeding Medium

Depending on the goal of the fed-batch cultivation (either biomass formation or induction of protein expression) different feeding media can be prepared. For P. pastoris glucose and glycerol are prominent C-sources for biomass formation, whereas methanol is used for the induction of protein expression.

1. Glucose feed per L: 275 g glucose monohydrate, 12 mL PTM1, 0.3 mL antifoam.

2. Glycerol feed per L: 250 g glycerol, 12 mL PTM1, 0.3 mL antifoam.

3. Methanol feed per L: 300 g methanol (use a balance), 4 mL PTM1, 0.3 mL antifoam.

The glucose and the glycerol feed can be sterilized via autoclavation; the methanol feed is sterile-filtered through a 0.2 μm cutoff filter into a sterile flask in order to avoid methanol evaporation.

2.5 Equipment

For a standard fed-batch experiment the following equipment is at least required:

1. Bioreactor (e.g., 5 L working volume glass bioreactor; Infors, Switzerland).

2. pH and pO₂ probe.

3. Air and oxygen lines.

4. Offgas analyzer (e.g., infrared cell for CO₂ and a zirconium dioxide sensor for O₂ concentration; DasGip, Germany).

5. Pumps and tubings for base and feed.

6. Balances (reactor balance, feed balance, base balance)— connected to the process information management system.

7. Process information management system (PIMS; e.g., Lucullus, SecureCell, Switzerland).

8. Spectrophotometer, centrifuge and dry oven for sample preparation.

9. HPLC for exact determination of methanol concentrations (e.g., Agilent Technologies, USA) equipped with a Supelco guard column, a SUPELCOGEL C-610H ion-exclusion column (Sigma-Aldrich, USA) and a refractive index detector (Agilent Technologies, USA).

3 Methods

3.1 Preculture of Pichia pastoris

Start a pre-culture of the P. pastoris strain of interest in 100 mL of YNB medium in 1 L baffled shaking flasks at 220 rpm and 28 °C for maximum 24 h (to guarantee good aeration only 1/10 of the total volume of the flask is filled with medium). The preculture is inoculated with 1 mL of frozen glycerol stock (see Notes 1-3).

3.2 Batch Cultivation in Bioreactors

After autoclaving the BSM in a bioreactor vessel, aseptically add the C-source (e.g., 40 g/L glucose or glycerol as final concentration in the vessel). Then adjust the temperature and the stirring speed to the desired values, before the pH in the bioreactor is adjusted to pH 5.0 using 25 % (v/v) ammonia solution (NH₄OH) by manually pumping the solution into the bioreactor. Aseptically transfer sterile PTM1 solution to the bioreactor (4.5 mL/L BSM; see Note 4). Aseptically transfer the preculture into a sterile inoculation flask (i.e., a vessel providing a connection to the bioreactor). The inoculum should be around 10 % of the final volume in the bioreactor.

Dissolved oxygen (dO₂) is measured with a sterilizable polarographic dissolved oxygen electrode. The pH is measured on-line with a sterilizable electrode and maintained constant with a step controller using 2–3 M NH₄OH which also represents the N-source during cultivation. The exact concentration of NH₄OH in the base bottle is determined by titration with 0.25 M potassium hydrogen phthalate (KHP; see Note 5). Base consumption is determined gravimetrically by putting the base bottle on a balance and recording the loss in weight over time. Set the cultivation temperature (e.g., 28 °C) and fix the agitation to the highest possible setpoint (e.g., 1,495 rpm) to guarantee good aeration. Aerate the culture with 2.0 vvm dried air (i.e., volume per volume per minute; in 1 L cultivation volume 2.0 vvm correspond to 2 L of dry air per minute). Measure the offgas of the culture by using an infrared cell for CO₂ and a paramagnetic cell for O₂ concentration. Temperature, pH, dO₂, agitation in the vessel, as well as CO₂ and O₂ in the offgas are measured on-line and logged in a process information management system.

3.3 Analysis of Growth-Parameters During Cultivations

Harvest 5 mL of culture broth by centrifugation in 10 mL glass tubes (4,300 × g, 4 °C, 10 min), wash the pellet twice with 5 mL deionized water and determine the dry cell weight (DCW) after drying at 105 °C to a constant weight in an oven (approximately 2–3 days; see Note 5). Optical density of the culture broth throughout the process is measured using a spectrophotometer at a wavelength of 600 nm (OD₆₀₀). Dry cell weight measurement and OD₆₀₀ have to be correlated to be able to use the measured OD₆₀₀ values for q_s adaptation in subsequent fed-batch cultivations (see Notes 6 and 7).

3.4 Batch Cultivation with Methanol Pulses

After the complete consumption of the C-source (e.g., glucose or glycerol), which is indicated by an increase of dissolved oxygen and a drop in offgas activity, perform the first methanol pulse of a final concentration in the bioreactor of 0.5 % (v/v). Following pulses are performed one after the other with 1 % (v/v) methanol as soon as methanol in the bioreactor is depleted (evident in the offgas analysis; Fig. 1).

Fig. 1.

Experimental strategy for the fast determination of strain specific parameters of P. pastoris using a batch experiment with methanol pulses of 0.5 and 1 % (v/v). continuous line, carbon dioxide emission rate (CER); circle, calculated specific substrate uptake rate (q_s). Figure adapted with permission from [14]

To obtain the specific rates for substrate uptake during each pulse, a minimum of two samples has to be taken, one directly after methanol addition and the other before complete methanol depletion. The samples are used to determine the concentration of residual substrate, OD₆₀₀ and the dry cell weight. Determined values at the beginning and the end of the respective pulse are used to calculate an average value of the specific substrate uptake rate (q_s) according to Eq. 1 (in short: take a sample—pulse methanol— when the offgas signal drops, indicating depletion of methanol, take another sample—measure the exact methanol concentration in these two samples by HPLC—calculate the volumetric methanol uptake rate and relate it to the total biomass content at the latter sample point).

$$q_s(\mathrm{mmol}∕\mathrm{g}∕\mathrm{h})=\frac{\frac{\mathit{\Delta }\mathrm{MeOH}\left(\mathrm{g}\right)}{\frac{\mathit{\Delta }\text{time}\left(\mathrm{h}\right)}{\text{biomass}\left(\mathrm{g}\right)}}}{M_{\left(\mathrm{MeOH}\right)}\times 1,000}$$ (1)

M_(MeOH) = 32.04 (g/mol), biomass = amount of biomass in gram at the time point of the latter sampling

The parameters which can be determined by this strategy are: (1) adaptation time of P. pastoris the methanol, (2) the q_s during the adaptation pulse and (3) a maximum q_s for methanol (q_{s max}) (see Note 8). We recommend performing at least four methanol pulses to obtain statistically significant parameters (see Note 9).

3.5 Fed-Batch

During fed-batch the dissolved oxygen (dO₂) signal is used to adjust air-in flow to keep levels >30 % dO₂ at all time points. In case air flow is not sufficient to keep this dO₂ level, pure oxygen is added. After the complete consumption of the substrate at the end of the batch phase (see Notes 10 and 11), which is indicated by an increase of dissolved oxygen and a drop in offgas activity, the feeding phase of the fed-batch feed is started. The feed rate is measured and controlled using a gravimetrically based PID flow controller (see Note 12). The culture is first adapted to methanol: the adaptation is performed at a q_s setpoint of 0.5 mmol/g/h (Fig. 2). As soon as the offgas signal (CO₂) gets constant, the q_s setpoint can be stepwise increased up to the predetermined q_{s max} of the respective strain (Fig. 2). To be able to analyze the physiology and the productivity of the P. pastoris strain at each q_s step, it is required to allow the cells to adapt to the respective q_s step before increasing it (see Notes 13 and 14).

Fig. 2.

Fed-batch cultivation of a P. pastoris strains on methanol with a stepwise increase of q_s to q_{s max}. straight line, set point for q_s; black dot, calculated q_s values; black triangle, methanol concentration in the supernatant. The q_{s max} of the respective P. pastoris strain was determined with 1.94 mmol/g/h before. When this level is exceeded in fed-batch cultivations, methanol accumulates in the cultivation broth. Figure adapted with permission from [15]

3.6 Substrate Concentrations

Samples are centrifuged (20,000 × g, 15 min) and then concentrations of methanol are determined in cell free samples by HPLC. The mobile phase is 0.1 % H₃PO₄ with a constant flow rate of 0.5 mL/min and the system is run isocratic. Calibration is done by measuring standard points in the range of 0.1–10 g/L methanol.