Electronic Journal of Biotechnology

Electronic Journal of Biotechnology 33 (2018) 29–35
Contents lists available at ScienceDirect
Electronic Journal of Biotechnology

Research article
Scaling-up fermentation of Escherichia coli for production of recombinant P64k protein from Neisseria meningitidis
Raúl Espinosa Pérez a,⁎, José García Suárez a, Emilio Narciandi Diaz b, Ricardo Silva Rodríguez c, Evelin Caballero Menéndez d, Héctor Díaz Balaguer a, Alexis Musacchio Lasa e,⁎
a Technological Development Division, Center for Genetic Engineering and Biotechnology, Ave 31, e/ 158 and 190, PO Box 6162, Cubanacán Playa, La Habana 10600, Cuba b Investment Department, Center for Genetic Engineering and Biotechnology, Ave 31, e/158 and 190, PO Box 6162, Cubanacán Playa, La Habana 10600, Cuba c Business Development Group, Center for Genetic Engineering and Biotechnology, Ave 31, e/158 and 190, PO Box 6162, Cubanacán Playa, La Habana 10600, Cuba d Biomedical Research Division, Center for Genetic Engineering and Biotechnology, Ave 31, e/158 and 190, PO Box 6162, Cubanacán Playa, La Habana 10600, Cuba e System Biology Division, Center for Genetic Engineering and Biotechnology, Ave 31, e/158 and 190, PO Box 6162, Cubanacán Playa, La Habana 10600, Cuba

article info
Article history: Received 17 July 2017 Accepted 1 March 2018 Available online 29 March 2018
Keywords: Escherichia coli Fermentation process Fermentation Gram negative diplococcus High molecular weight protein Meningococcal vaccine Neisseria meningitidis Obligate human pathogen Recombinant P64k protein Scale-up Tryptophan promoter

abstract
Background: P64k is a Neisseria meningitidis high molecular weight protein present in meningococcal vaccine preparations. The lpdA gene, which encodes for this protein, was cloned in Escherichia coli and the P64k recombinant protein was expressed in E. coli K12 GC366 cells under the control of a tryptophan promoter. P64k was expressed as an intracellular soluble protein about 28% of the total cellular protein. Several scale-up criteria of fermentation processes were studied to obtain the recombinant P64k protein at the pilot production scale. Results: The best operational conditions at a larger scale production of P64k recombinant protein were studied and compared using the four following criteria: Constant Reynold's number (Re constant), Constant impeller tip speed (n di constant), Constant power consumption per unit liquid volume (P/V constant) and Constant volumetric oxygen transfer coefficients (KLa/k constant). The highest production of the recombinant protein was achieved based on the constant KLa/k scale-up fermentation criterion, calculating the aeration rate (Q) and the impeller agitation speed (n) by iterative process, keeping constant the KLa/k value from bench scale. The P64k protein total production at the 50 l culture scale was 546 mg l-1 in comparison with the 284 mg l-1 obtained at 1.5 l bench scale. Conclusions: The methodology described herein, for the KLa/k scale-up fermentation criterion, allowed us to obtain the P64k protein at 50 l scale. A fermentation process for the production of P64k protein from N. meningitidis was established, a protein to be used in future vaccine formulations in humans.
How to cite: Espinosa R, García J, Narciandi E, et al. Scaling-up fermentation of Escherichia coli for production of recombinant P64k protein from Neisseria meningitidis. Electron J Biotechnol 2018;33:29-35. https://doi.org/10.1016/j. ejbt.2018.03.004.
© 2018 Pontificia Universidad Católica de Valparaíso. Production and hosting by Elsevier B.V. All rights reserved. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
Neisseria meningitidis, a gram negative diplococcus and an obligate human pathogen, causes significant morbidity and mortality in children and young adults worldwide through epidemic or sporadic meningitis and/or septicemia [1].
Immunization is the best and most efficacious preventive approach against meningococcal disease [2]. This is the reason why, first polysaccharide vaccines and later on conjugate vaccines against
⁎ Corresponding authors. E-mail addresses: [email protected] (R. Espinosa Pérez),
[email protected] (A. Musacchio Lasa). Peer review under responsibility of Pontificia Universidad Católica de Valparaíso.

serogroups A, C, Y and W-135 meningococci were developed. However, most work on development of group B meningococcal vaccines has focused upon use of noncapsular surface components (i.e. lipooligosaccharides (LOS) and outer membrane proteins, either individually- or as in proteoliposome vaccine preparations).
Several outer membrane proteins (OMPs) have been examined as vaccine candidates. Besides major OMPs, other surface proteins are under investigation in several laboratories, including highly conserved OMPs that could potentially protect against meningococcal disease in humans [3].
Among proteins present in outer membrane vesicles vaccine preparations, a high-molecular mass protein (P64k) was identified. This protein is common to many meningococcal isolates, and the lpdA gene encoding for this proteins has been isolated and cloned

https://doi.org/10.1016/j.ejbt.2018.03.004 0717-3458/© 2018 Pontificia Universidad Católica de Valparaíso. Production and hosting by Elsevier B.V. All rights reserved. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

30

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

in Escherichia coli for the further expression of the recombinant protein [4].
An extensive biochemical and biophysical characterization of the P64k protein has been carried out [5]. The protein has been described as a dihydrilopoamide dehydrogenase, based in its high degree of homology with two enzymes of the pyruvate dehydrogenase complex of several species. The first 110 amino acids are similar to those in the lipoyl-binding domain present in the enzyme dihydrolipoamide acetyl-transferase (E2), while the rest of the protein has homology with the enzyme lipoamide dehydrogenase (E3) [6,7]. The gene and protein have been sequenced and its 3D structure determined [8].
From the immunological point of view, the protein was first considered as a vaccine candidate against Neisseria meningitidis infection. The antigenicity of this protein was studied and sera from vaccines and convalescent patients were able to recognize the recombinant protein. At the same time, antibodies elicited in several animal models by the recombinant P64k protein recognized the natural protein, present in the whole cell meningococci and in outer membrane vesicle vaccine preparations [9,10,11]. However, the functional activity of antibodies elicited in several animal models and in human by the recombinant protein alone was not enough in terms of conferring protection against meningococcal disease to consider the protein as a unique vaccine candidate. On the other hand, taking into account that the recombinant protein was highly immunogenic [12,13], its function as an immunological carrier for peptides [14,15] and antigens [16,17] was evaluated. Recently, the P64k protein is employed as a recipient of anti-cancer vaccine candidate [18,19].
In order to satisfy the required quantities of P64k protein for the accomplishment of preclinical and clinical studies first, and subsequently for the use in these conjugate vaccines, it was necessary to scale up the fermentation process of the recombinant E. coli from 1.5 l [20] up to 50 l scale. Based on several scale up fermentation criteria, a comparative analysis was performed and the results are presented herein.
2. Materials and methods
2.1. E. coli strain and expression plasmid
E. coli K12 GC366 [21] strain was transformed with the plasmid pM-152 [10], which contains the entire lpdA gene encoding for the P64k protein under the control of the tryptophan (trp) promoter [22].
2.2. Growth conditions
The recombinant E. coli strain was grown at 37°C in 500 ml shake flasks (New Brunswick Scientific Co., USA) containing 100 ml Luria Bertani medium (LB) [21], supplemented with 10 g glucose l-1, 100 μg tryptophan ml-1 (Merck, Germany) and 50 μg kanamycin ml-1 (Sigma Co. Ltd., USA) for 8 h and 150 rpm.
For growth in fermenters the minimal growth medium (MM) [21] supplemented with 10 g casein hydrolysate l-1, 10 g tryptone l-1, 10 g glucose l-1, 0.015 g CaCl2·2H2O l-1, 0.246 g MgSO4·7H2O l-1 and 50 μg kanamycin ml-1, was used.
The bench scale was carried out in 2.5 l fermenters (B.H Marubishi Co., Ldt., Japan) with 1.5 l working volume and 37°C, 700 rpm, aeration rate 1.5 l min-1, as fermentation parameters. The pH was controlled at 7.0 with 40% (v/v) NaOH or H3PO4 [20].
The scale-up batch process was optimized in 70 l fermenters (B.H. Marubishi Co, Ldt., Japan) with 50 l working volume, at 37°C and pH 7.0, as fermentation parameters. The aeration rate and agitation speed varied depending of the studied scale-up criteria. The fermenter geometries for both working scales are shown in Table 1.

Table 1 Geometry of fermenters employed in the scale-up processes.

Parameters
Total fermentation volume [m3] Working fermentation volume [m3] Vessel diameter (D) [m] Impeller diameter (di) [m] Static height of both (H) [m] Number of impellers (N) Geometric factor (fc) Power number (Np) Density (ρ) [kg m-3]

1.5 l scale
0.0025 0.0015 0.129 0.067 0.115 1 0.61 6 1000

50 l scale
0.070 0.050 0.375 0.150 0.453 2 0.92 6 1000

2.3. Analysis
Four scale-up fermentation criteria were employed: (1) Constant Reynold's number (Re constant), (2) constant impeller tip speed (n di constant), (3) constant power consumption per unit liquid volume (P/V constant) and (4) constant volumetric oxygen transfer coefficients (KLa/k constant).
The cell dry weight, as cell growth correlation parameter, was determined using the Sartorius MA 30 moisture balance (Sartorius, Germany). The total protein concentration was measured by Lowry's method [23]. The recombinant P64k expression levels were analyzed by densitometry, on 12.5% SDS-PAGE of the recombinant E. coli whole cells.
3. Results and discussion
3.1. Bench scale growth, expression and production of P64k recombinant protein
The typical schematic fermentation process at 1.5 l scale, for the production of the P64k recombinant protein is shown in Fig. 1.
During the first 8 h of growth at 1.5 l bench scale, an increase in the optical density (cell dry weight) was observed. The maximum value of the specific growth rate was determined during the first 2 h of growth (μ = 1.2 h-1), being stabilized during the rest of the fermentation process at μ = 0.34 h-1, when an increase in the expression of the recombinant protein was detected. After 8 h of fermentation, the specific growth rate began to decline, reaching values of μ = 0.06 h-1 at 10 h of growth, when the fermentation process finished.
The P64k expression level and the protein production increased through the fermentation process, reaching at 10 h growth the highest productivity level.
At this scale, 2.6 g cell dry wt l-1, 22% of P64k recombinant protein expression levels, 284 mg 1-1 of volumetric production and 28.4 mg l-1 h-1 productivity, were obtained. The biomass yield on initial glucose as substrate (Yx/s) and the product yield on this substrate (Yp/s) values were 0.26 g biomass/g substrate and 0.028 g product/g substrate, respectively.
The P64k expression level is comparable with those values obtained for other recombinant proteins produced in E. coli with a similar expression system, such as p36 (20%) [24] and gag24 (25%) from HIV [25] and Opc from N. meningitidis (32%) [26]. Likewise, the productivity levels, biomass yield on substrate and the product yield on substrate of P64k protein are in the same order as for the above mentioned recombinant proteins.
3.2. Scale-up fermentation process of P64k recombinant protein
The scale-up parameters to determine the best operational conditions at a larger scale production of P64k recombinant protein were studied and compared using the four following criteria:

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

31

Fig. 1. P64k protein production at 1.5 l bench scale (a) Culture kinetics of E. coli expressing the recombinant protein grown during 10 h at 37°C, 700 rpm, 1.5 l min-1. The cell dry weight (●) is expressed in [g l-1]. The expression level of the recombinant P64k protein (▲) is expressed in percent. The production of the recombinant P64k protein (■) is expressed in mg l-1. (b) SDS-PAGE (12.5%) of samples taken during the fermentation process: lane 1: at 0 h, lane 2: at 2 h, lane 3: at 4 h, lane 4: at 6 h, lane 5: at 8 h and lane 6: at 10 h. Lane 7: purified recombinant P64k protein.

(1) Constant Reynold's number (Re constant), (2) Constant impeller tip speed (n di constant), (3) Constant power consumption per unit liquid volume (P/V constant) and (4) Constant volumetric oxygen transfer coefficients (KLa/k constant).
3.2.1. Scale-up under constant Reynold's number criterion This scale-up procedure was performed using the Reynold's
equation, Re = ρ n di2/μ [27,28,29,30,31], where ρ is the fluid density, μ is the viscosity, n represents the impeller agitation speed, and di the impeller diameter. For this scale-up criterion, then:

n50 l ¼ n1:5 lðdi1:5 l=di50 lÞ2

½Equation 1Š

The aeration rate at the larger scale, considering a constant aeration number (Na = Q/n di3), was calculated as [27,29,32]:

Q 50l ¼ Q 1:5 lðn50 l=n1:5 lÞ ðdi50 l=di1:5 lÞ3

½Equation 2Š

3.2.2. Scale-up under constant impeller tip speed criterion The scale-up procedure using the constant impeller tip speed
criterion [27,28,29,30,31] was performed according [Equation 3]:

n50 l ¼ n1:5 lðdi1:5 l=di50 lÞ

½Equation 3Š

where, n represents impeller agitation speeds and di impeller diameters. The aeration rate at the larger scale was calculated as previously
define in [Equation 2].

3.2.3. Scale-up under constant power consumption per unit liquid volume criterion
This scale-up procedure, using the constant power per unit of volume criteria [27,28,29,30,31], is performed according the following equations:

ðP=VÞ1:5 l ¼ ðP=VÞ50 l

½Equation 4Š

P ¼ Np ρ n3 di5 f c N

½Equation 5Š

where, P corresponds to the non-gassed power, Np the power number,
ρ the fluid density, n the impeller agitation speed, di the impellent diameter, N the number of impellents and fc is the geometric factor. The geometric factor was calculated as fc = ((D H)1/2/(3 di)) and the volume as V = π D2 H/4, where D is the vessel diameter and H is the static height of broth. Substituting [Equation 5] in [Equation 4],

n50 l ¼ n1:5 lðdi1:5 l=di50 lÞ5=3ðD50 l=D1:5 lÞ2=3 ½ðH50 l=H1:5 lÞðN1:5 l=N50 lÞðfc1:5 l=fc50 lފ1=3

½Equation 6Š

The aeration rate at the larger scale was calculated as previously define in [Equation 2].

3.2.4. Scale-up under constant volumetric oxygen transfer coefficients KLa/k criterion
The implementation of constant KLa/k scale-up criterion was performed using the most successful correlations with dimensional equations of the form: KLa = k (Pg/V)α(Vs)β [33], where k, α and β are constants, Pg is the gassed power, V is the liquid volume and Vs is the superficial gas velocity. Taking into account that V = π D2 H/4,
where D is the vessel diameter and H is the static height of broth, and Vs = (4Q /π D2), where Q the is the aeration rate, then.



0:4 

0:2

ðKLa=kÞ ¼ 4 Pg=π D2H 4 Q =π D2

½Equation 7Š

The gassed power factor fp (fp = Pg/P), where P corresponds to the non-gassed power, is determined from the power requirements for agitation in a gassed system [27,28,29,31,32].
The value for the gassed power factor fp is taken from the experimental correlations between Pg and operating variables presented as curves of Pg/P vs aeration number (Na), where Na = (Q/n di3), being Q the aeration rate, n the rotation speed of impeller and di the impeller diameter.
Therefore, from the gassed power factor,

Pg ¼ fpP

½Equation 8Š

32

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

and the non-gassed power calculated as:

P ¼ Nprn3di5Nfc

½Equation 9Š

where Np is the power number, ρ is the fluid density, n is the rotation speed of impeller, N is the number of impellers, and fc is the geometric factor, calculated as fc = (D H)1/2/(3di), and substituting
[Equation 9] and [Equation 8] in [Equation 7],



0:6 

0:4

KLa=k ¼ 4=π D2

Nprdi5N fc fp=H n1:2Q 0:2

½Equation 10Š

According to this scale-up criterion, (KLa/k)1.5l = (KLa/k)50l. The (KLa/k)1.5l value of 8.7 was calculated from experimental conditions having the aeration rate (Q) of 1.5 l min-1 and the impeller agitation speed (n) of 700 rpm at 1.5 l fermentation scale. The n and Q values for the 50 l fermenter conditions, were calculated by iterative process using the [Equation 10], keeping a constant factor of 8.7 for (KLa/k)1.5l. The calculated values for the scale-up processes are shown in Table 2.
Using the calculated fermentation parameter combinations of aeration rate (Q) and the impeller agitation speed (n), three fermentation processes were performed for each condition. The average of P64k recombinant protein production is shown in Table 3. According to these results, the best combination of aeration rate and the rotation speed of impeller parameters to keep a constant KLa/k value of 8.7 was Q of 50 l min-1 and 320 rpm. This set of values was used to perform the comparison among the four scaling up fermentation processes.
The results of total P64k production related to the influence of aeration rate and the impeller agitation speed are shown in Fig. 2. The Fig. 2a represents the row data from the fermentation processes according to Table 3. The highest level of P64k production is obtained in a range of 315–340 rpm of impeller agitation speed and 35–55 l min-1 aeration rates, having the maximum P64k production of 546 mg l-1, with 320 rpm agitation and 50 l min-1 aeration rates. A surface response graph for the P64k production is presented in Fig. 2b. This graph is characterized by a none-linear regression model, with Q (aeration rate) and n (impeller agitation speed) as independent variables [34]. The P64k production is represented in [Equation 11]:

Production ¼ ‐76; 352:6315 þð904:934 QÞ

À

Á

þ ð359:0292 nÞ‐ 2:2763 Q 2 ‐ð2:172 Q nÞ‐ 0:4187 n2

½Equation 11Š

A positive influence of the aeration rate and the impeller speed, as independent variable parameters, over the theoretical total protein production is observed. This means, that at certain value of aeration rate with an increase of the impeller agitation speed, an optimal protein production might be expected. In concordance, at certain

Table 2 Calculated theoretical relation between the aeration rate (Q) and the impeller agitation speed (n) to keep constant a KLa/k value of 8.7, for the studied scale-up fermentation process.

Variant no.
1 2 3 4 5 6 7 8 9 10

Q [l min-1/l s-1/vvm]
10/0.17/0.2 15/0.25/0.3 20/0.33/0.4 25/0.42/0.5 30/0.50/0.6 35/0.58/0.7 40/0.67/0.8 45/0.75/0.9 50/0.83/1.0 55/0.92/1.1

n [rpm/rps]
420/7.00 390/6.50 370/6.17 360/6.00 350/5.83 340/5.67 330/5.50 325/5.42 320/5.33 315/5.25

Pg/V [(W m-3]
5574.87 4463.54 3811.45 3510.68 3226.16 2957.45 2704.09 2583.02 2465.62 2351.84

Vs [m s-1]
0.0015 0.0023 0.003 0.0038 0.0045 0.0053 0.006 0.0068 0.0075 0.0083

KLa/k [W0.4 m-1 s-0.2]
8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7

Table 3 Average of P64k recombinant protein production as a result of triplicate fermentation processes from each variant.

Variant no.
1 2 3 4 5 6 7 8 9 10

Q [l min-1/l s-1/vvm]
10/0.17/0.2 15/0.25/0.3 20/0.33/0.4 25/0.42/0.5 30/0.50/0.6 35/0.58/0.7 40/0.67/0.8 45/0.75/0.9 50/0.83/1.0 55/0.92/1.1

n [rpm/rps]
420/7.00 390/6.50 370/6.17 360/6.00 350/5.83 340/5.67 330/5.50 325/5.42 320/5.33 315/5.25

KLa/k [W0.4 m-1 s-0.2]
8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7 8.7

Production [mg l-1]
277 ± 15 310 ± 23 305 ± 38 293 ± 26 301 ± 30 325 ± 29 398 ± 49 430 ± 48 546 ± 51 405 ± 42

value of impeller speed, when the aeration rate increased a final protein production might also be expected. However, analyzing the combination of these parameters over the final production, a negative influence is observed. This means, that to keep a certain value of production if the impeller agitation speed increased the aeration rate has to decrease. This is characterized by the negative signal in the term (Q n). Nevertheless, the input of this term to the total production value in the equation is negligible. From this equation, the maximum value for protein production of 548 mg l-1 falls in 315 rpm and 48.7 l min-1 aeration speed. These values are similar to those obtained experimentally, and when the KLa/k coefficient was recalculated with these parameters a value of 8.4 was obtained, compared with the 8.7 used for this scale-up criterion.
3.3. Comparison of all scale-up criteria
The scale-up process is one of the key points, not only in fermentation but also in the industry as a whole. In the real practice, several scale-up criteria have been studied and applied, but all of them based in empirical results [35,36].
The constant power input per unit volume of liquid (P/V) and constant volumetric oxygen transfer coefficient (KLa) are among the most scale-up fermentation criteria used in the industry [37,38,39]. No scale-up strategy as such is generally established, so for each process a suitable approach needs to be determined [27,32,40]. In practice, often combinations of methods are used for adequate scale-up of a bioreactor [41,42].
The parameters for the studied scale-up criteria and the influence over the P64k production are shown in Table 4. Scaling up with the constant Reynold's number criterion, very low values for impeller agitation speed (20% regarding 1.5 l scale), P/V (4% regarding 1.5 l scale) and KLa/k (20% regarding 1.5 l scale) were obtained at 50 l scale. This might provoke an inappropriate mixing, resulting in around 50% lower values for cell dry weight and P64k production obtained at 1.5 l scale. This is in agreement that Reynold's scaling up criterion is not a good one to use for such a fermentation process, since the degree of agitation decreases very strongly with the increase of the production scale [42,43,44,45,46].
The impeller tip speed scale-up criterion has some advantages in the case of bioprocesses with shear sensitive microorganisms or the absolute value of the impeller tip speed [42,47,48], but it is also argued that it is not a good enough criterion for scaling [45,49]. In our case, the cell dry weight and the final P64k production increased in 50%, compared to values obtained for Reynold's scale-up criterion. To do so, an increase in the impeller agitation speed and the aeration rate of around 50% had to be implemented.
Maintaining a constant power per unit volume is another popular scale-up method, and has been successfully applied as a primary scale-up parameter for many antibiotic fermentations [40,42].

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

33

Fig. 2. Schematic representation of results from the scale up variants using KLa/k criteria. (a) Influence of aeration rate and agitation speed on the final production of P64k protein. The production of the recombinant P64k protein (■) is expressed in [mg l-1], the aeration (●) is expressed in [l min-1] and the agitation speed (▲) is expressed in [rpm], (b) Surface
response graph from this analysis (n: agitation speed, Q: aeration rate).

When scaled up with constant P/V criterion, similar values of cell dry weight and P64k production were obtained, as compared to those for the constant impeller tip speed criterion. This was observed in spite of the 30% increase of the impeller agitation speed and the aeration rates used for the constant impeller tip speed scale up criterion.
For instance, scale-up based on constant P/V resulted in an increase of the shear rate (calculated as: n di) of 30%, as compared with the constant impeller tip speed criterion and the KLa/k scale up criterion. This was the highest value among all the studied criteria. With this criterion, a P/V value of 5.2 kW m-3 was obtained, slightly higher than the range of values reported as a general rule (1 to 3 kW m-3) [27,43]. This increase in the shear stress may correlate with an increase of the possible cell damage and therefore with the decrease of the final production of the protein of interest [45,50].
The volumetric mass transfer coefficient is the most often applied physical scale-up variable. It includes relevant parameters that influence oxygen supply such as agitation and aeration [36,40,51,52]. Therefore, a strong emphasis has been put on maintaining a constant KLa in scale-up of process to ensure the same oxygen supply rate and

to satisfy the oxygen demand of the desired cultivation. The same KLa values can be achieved in different sized vessels by adjusting the operating conditions, i.e. stirrer speed and aeration rates on the larger scale to match the optimum KLa conditions determined in the smallscale studies [53]. Several correlations estimating KLa have been developed over the years for a range of different vessel sizes [33,37,54, 55] and successfully scaled-up of processes from bench scale to pilot scale based on this criterion or in conjunction with another scale-up parameter have been established [56,57,57,58].
In this study, when the KLa/k scale up criterion was used, assuming a constant value of 8.7 (calculated for the 1.5 l scale), the aeration rate increased in 5–6 times (50 l min-1) as compared with values for the impeller tip speed (7.5 l min-1) and P/V (9.8 l min-1) criteria. However, a similar impeller agitation speed value was obtained when compared theses three scale-up criteria. Using the parameters calculated for the KLa/k scale up criterion, no shear increase occurred and the highest cell dry weight and P64k production was achieved.
Using the estimated parameters for the KLa/k scale up criterion (n = 320 rpm, Q = 50 l min-1), a total of 10 fermentation processes were

Table 4 Parameters for comparison of scale-up fermentation criteria.

Parameters

1.5 l scale

50 l scale Rea

n [rpm/rps]
P [W]
Re
n di P/V [W m-3] Pg/V [W m-3] Q [l min-1/l s-1/vvm] Vs [m s-1] KLa/k [W0.4 m-1 s-0.2] Dry weight [g l-1] P64k production [mg l-1]

700/11.67 7.85 5.20E + 04 46.9 5223.34 5064.01 1.5/0.03/1.00 0.0019 8.7 2.6 ± 0.11 284 ± 12

139.7/2.33 10.57 5.20E + 04 20.95 211.4 204.97 3.4/0.06/0.07 0.0005 1.8 1.4 ± 0.06 168 ± 7.0

a Constant Reynold's number scale-up criterion. b Constant impeller tip speed scale-up criterion. c Constant power consumption per unit liquid volume scale-up criterion. d Constant volumetric oxygen transfer coefficients scale-up criterion.

n dib
312.7/5.21 118.64 1.10E + 05 46.95 2372.4 2300.02 7.5/0.13/0.15 0.0011 5.7 2.7 ± 0.13 338 ± 15

P/Vc
406.7/6.78 261 1.50E + 05 61 5223.34 5060.01 9.8/0.16/0.20 0.0015 8.2 3.0 ± 0.15 383 ± 19

KLa/kd
320/5.33 127.18 1.20E + 05 48 2543.6 2465.62 50.0/0.83/1.00 0.0076 8.7 3.9 ± 0.19 546 ± 27

34

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

Fig. 3. P64k protein production at 50 l production scale (a) Culture kinetics of E. coli expressing the recombinant protein grown during 10 h at 37°C, 320 rpm, 50 l min-1. The cell dry weight (●) is expressed in [g l-1]. The expression level of the recombinant P64k protein (▲) is expressed in percent. The production of the recombinant P64k protein (■) is expressed in mg l-1. (b) SDS-PAGE
(12.5%) of samples taken during the fermentation process: lane 2: at 0 h, lane 3: at 2 h, lane 4: at 4 h, lane 5: at 6 h, lane 6: at 8 h and lane 7: at 10 h. Lane 1: purified recombinant P64k protein.

performed. At the 50 l scale, during the first 6 h of growth an increase in the optical density (and the corresponding total cell dry weight) was observed. The maximum value of the specific growth rate was obtained during the first 2 h of growth (μ = 1.4 h-1), being stabilized during the rest of the fermentation process at μ = 0.33 h-1. After 8 h of fermentation the specific growth rate began to decline, reaching values of μ = 0.03 h-1 at 10 h of growth.
The P64k expression level and the protein production increased through the fermentation process, reaching at 10 h growth the highest productivity level. As an average, at 50 l scale (using the KLa/k scale up criterion), 3.9 g cell dry wt l-1, 28% level of P64k carrier protein expression, 546 mg l-1 volumetric production and 54.6 mg l-1 h-1 productivity were obtained. The biomass yield on initial glucose as substrate and the product yield on this substrate values were 0.39 g biomass/g substrate and 0.054 g product/g substrate, respectively. A behavior of these fermentation processes is shown in Fig. 3.
The total P64k protein production increased 1.9 folds, when data from both scale were analyzed. This augment is influenced by the 6% slightly increase in P64k protein expression and the 1.5 fold increase in the final cell growth. These values are comparable with those obtained for other recombinant proteins produced in E. coli under the same expression system such as gag24, gp-41 and p36 from HIV [24,59,60], and the Opc from N. meningitidis [26], where the proteins in study become “toxic” for the cell.
In conclusion, the methodology employed, using the KLa/k scale-up fermentation criterion and calculating the aeration rate (Q) and the impeller agitation speed (n) by iterative process, keeping constant the KLa/k value from bench scale, allowed us to obtain the highest P64k protein production at 50 l scale, in comparison with the other studied scale-up fermentation criteria. As a result, we have established a fermentation process for the production of the P64k recombinant protein from N. meningitidis, a protein to be used in future vaccine formulations in humans.
References
[1] Pizza M, Rappuoli R. Neisseria meningitidis: Pathogenesis and immunity. Curr Opin Microbiol 2015;23:68–72. https://doi.org/10.1016/j.mib.2014.11.006.
[2] Crum-Cianflone N, Sullivan E. Meningococcal vaccinations. Infect Dis Ther 2016;5: 89–112. https://doi.org/10.1007/s40121-016-0107-0.

[3] Frasch CE, Bash MC. Neisseria meningitidis vaccines. In: Madame Curie Bioscience Database [Internet]. Landes Bioscience: Austin (TX); 2000-2013 Available from: https://www.ncbi.nlm.nih.gov/books/NBK6030/.
[4] Silva R, Menendez T, Alonso LM, Iglesias E, Musacchio A, de Jesus Leal M, et al. Characterisation of the lpdA gene from Neisseria meningitidis by polymerase chain reaction, restriction fragment length polymorphism and sequencing. FEMS Microbiol Lett 1999;174:191–9. https://doi.org/10.1111/j.1574-6968.1999.tb13568.x.
[5] Gomez R, Madrazo J, Gonzalez LJ, Chinea G, Musacchio A, Rodriguez A, et al. Structural and functional characterization of P64k protein from Neisseria meningitidis. Biotecnología Aplicada 1999;16:83–7.
[6] Martin A. The P64k Neisseria meningitidis dihydrolipoamide dehydrogenase participates in catalysis of the pyruvate dehydrogenase multienzyme complex. Biotecnología Aplicada 2003;20:102–6.
[7] Bringas R, Fernandez J. A lipoamide dehydrogenase from Neisseria meningitidis has a lipoyl domain. Proteins 1995;21:303–6. https://doi.org/10.1002/prot.340210404.
[8] de la Sierra IL, Prange T, Fourme R, Padron G, Fuentes P, Musacchio A, et al. Crystallization and preliminary X-ray investigation of a recombinant outer membrane protein from Neisseria meningitidis. J Mol Biol 1994;235:1154–5. https://doi.org/10.1006/jmbi.1994.1067.
[9] Guillen G, Alvarez A, Silva R, Morera V, Gonzalez S, Musacchio A, et al. Expression in Escherichia coli of the lpdA gene, protein sequence analysis and immunological characterization of the P64k protein from Neisseria meningitidis. Biotechnol Appl Biochem 1998;27:189–96. https://doi.org/10.1111/j.1470-8744.1998.tb00494.x.
[10] Rodriguez RS, Houssein SMS, Nieto GG, Fernandez JR, Novoa LI, Grillo JM, et al. Nucleotide sequence coding for an outer membrane protein from Neisseria meningitidis and use of said protein in vaccine preparations. Patent No EP 1992;0474313:A2.
[11] Gonzalez S, Nazabal C, Vina L, Caballero E. Influence of several adjuvants on the immune response against a recombinant meningococcal high molecular mass antigen. In: Brown F, Haaheim LR, editors. Modulation of the immune response to vaccine antigens. Dev Biol Stand, Karger: Basel; 1998. p. 269–76.
[12] Guillen G, Silva R, Coizeau E, Novoa LI, Selman M, Morales J, et al. Cloning and expression in Escherichia coli of a high molecular weight outer membrane protein (PM6) from Neisseria meningitidis. In: Conde-Glez CJ, Morse S, Rice P, Sparling F, Calderón E, editors. Pathobiology and immunology of Neisseriaeceae. Chapter Vaccines and Immunity, INSP: Mexico; 1995. p. 834–40.
[13] Perez AE, Dickinson FO, Banderas F, Serrano T, Llanes R, Guzman D, et al. Safety and preliminary immunogenicity of MenC/P64k, a meningococcal serogroup C conjugate vaccine with a new recombinant carrier. FEMS Immunol Med Microbiol 2006;46: 386–92. https://doi.org/10.1111/j.1574-695X.2006.00047.x.
[14] Gonzalez S, Alvarez A, Caballero E, Vina L, Guillen G, Silva R. P64k meningococcal protein as immunological carrier for weak immunogens. Scand J Immunol 2000; 52:113–6. https://doi.org/10.1046/j.1365-3083.2000.00788.x.
[15] Borroto D, Perez Alea M, Fernandez W, Bello D. Vectors and P64k gene targeting for tandem affinity purification in Neisseria meningitidis. J Microbiol Methods 2006;65: 187–93. https://doi.org/10.1016/j.mimet.2005.07.013.
[16] Guillen G, Alvarez A, Gonzalez S, Musacchio A, Niebla O, Coizeau E, et al. Expression in Escherichia coli and immunological characterization of a hybrid class 1-64 kDa protein from Neisseria meningitidis. Biotecnología Aplicada 1996;13:271–5.
[17] Gonzalez S, Musacchio A, Silva R, Guillen G, del Valle JA, Herrera L. Enhancement of the immunogenicity of recombinant PM6 and HBsAg by double experimental vaccine. In: Conde-Glez CJ, Morse S, Rice P, Sparling F, Calderón E, editors.

R. Espinosa Pérez et al. / Electronic Journal of Biotechnology 33 (2018) 29–35

35

Pathobiology and immunology of Neisseriaeceae. Chapter Vaccines and Immunity, INSP: Mexico; 1995. p. 825–9. [18] Gonzalez G, Crombet T, Torres F, Catala M, Alfonso L, Osorio M, et al. Epidermal growth factor-based cancer vaccine for non-small-cell lung cancer therapy. Ann Oncol 2003;14:461–6https://doi.org/10.1093/annonc/mdg102. [19] Crombet T, Rodriguez P, Neninger E, Garcia B, Lage A. CIMAvax EGF (EGF-P64K) vaccine for the treatment of non-small-cell lung cancer. Expert Rev Vaccines 2015;14: 1303–11. https://doi.org/10.1586/14760584.2015.1079488. [20] Espinosa R, Caballero E, Musacchio A, Silva R. Production of a recombinant, immunogenic protein, P64k, of Neisseria meningitidis in Escherichia coli in fed-batch fermenters. Biotechnol Lett 2002;24:343–6. https://doi.org/10.1023/A%3A1014525222704. [21] Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: A laboratory manual. 2nd ed. New York: Cold Spring Harbor Laboratory; 1989. [22] Novoa LI, Machado JA, Fernandez JR, Benitez JV, Narciandi E, Rodriquez JL et al. Method for the expression of heterologous proteins produced in fused form in Escherichia. coli, use thereof, expression vectors and recombinant strains. Patent No. EP0416673A1, 1991. [23] Lowry O, Rosebrough N, Farr A, Randall R. Protein measurement with the Folin phenol reagent. J Biol Chem 1951;193:265–75. [24] Narciandi E, Delgado A. Production of the recombinant p36 protein from HIV-2 in Escherichia coli using the trp promoter. Biotechnol Lett 1992;14:1103–8. http://dx.doi/org/10.1007/BF01027010. [25] Narciandi E, Motolongo J, Perez L. High level production of the recombinant gag24 protein from HIV-1 in Escherichia coli. Afr J Biotechnol 2006;5:103–7. https://doi.org/10.5897/AJB2006.000-5015. [26] Espinosa R, Musacchio A, Silva R, Caballero E, Garcia JA, Diaz H. Scale-up of recombinant Opc protein production in Escherichia coli for a meningococcal vaccine. J Biotechnol 2006;127:109–14. https://doi.org/10.1016/j.jbiotec.2006.06.005. [27] Aiba S, Humphrey AE, Millis NF. Biochemical engineering. 1st ed. New York: Academic Press Inc.; 1964; 141–76. [28] Asenjo JA, Merchuk JC. Bioreactor system design. New York: Marcel Dekker Inc.; 1994 [p. 219-524]. [29] Doran PM. Bioprocess engineering principles. 2nd ed. New York: Academic Press Inc.; 2012. [30] Galindo E. Fronteras en biotecnología y bioingeniería. México: Sociedad Mexicana de Biotecnología y Bioingeniería; 1996. [31] Quintero R. Ingeniería bioquímica: Teoría y aplicaciones. 1st ed. Alhambra Mexicana: Mexicoam; 1981 [p. 100-275]. [32] Wang DIC, Cooney CL, Demain AL, Dunnill P, Humphrey AE, Lilly MD, et al. Fermentation and enzyme technology. Preparative biochemistry. 1st ed. New York: WileyInterscience; 1979. [33] van't Riet K. Review of measuring methods and results in nonviscous gas-liquid mass transfer in stirred vessels. Ind Eng Chem Process Des Dev 1979;18:357–64. [34] Myers RH, Montgomery DC, Anderson-Cook CM. Response surface methodology: Process and product optimization using designed experiments. fourth ed. New York: John Wiley and Sons; 2016. [35] Margaritis A, Zajic JE. Mixing, mass transfer, and scale-up of polysaccharide fermentations. Biotechnol Bioeng 1978;20:939–1001. https://doi.org/10.1002/bit.260200702. [36] Marks DM. Equipment design considerations for large scale cell culture. Cytotechnology 2003;42:21–33. https://doi.org/10.1023/A:1026103405618. [37] Vilaca PR, Badino J, Facciotti MCR, Schmidell W. Determination of power consumption and volumetric oxygen transfer coefficient in bioreactors. Bioprocess Eng 2000; 22:261–5. https://doi.org/10.1007/s004490050730. [38] Shukla VB, Parasu Veera U, Kulkarni PR, Pandit AB. Scale-up of biotransformation process in stirred tank reactor using dual impeller bioreactor. Biochem Eng J 2001; 8:19–29. https://doi.org/10.1016/S1369-703X(00)00130-3. [39] de Figueiredo MML, Calderbank PH. The scale-up of aerated mixing vessels for specified oxygen dissolution rates. Chem Eng Sci 1979;34:1333–8. https://doi.org/10.1016/0009-2509(79)80025-1.

[40] Schmidt FR. Optimization and scale up of industrial fermentation processes. Appl Microbiol Biotechnol 2005;68:425–35. https://doi.org/10.1007/s00253-005-0003-0.
[41] Sweere APJ, Luyben KC, Kossen NWF. Regime analysis and scale-down: Tools to investigate the performance of bioreactors. Enzyme Microb Technol 1987;9:386–98. https://doi.org/10.1016/0141-0229(87)90133-5.
[42] Ju LK, Chase GG. Improved scale-up strategies of bioreactors. Bioprocess Eng 1992;8: 49–53. https://doi.org/10.1007/BF00369263.
[43] Einsele A. Scaling-up of bioreactors. Process Biochem 1978;13:13–4. [44] Ettler P. Problems and possibilities in fermentation scale-up procedure. In: Malik VS,
Sridhar P, editors. Industrial biotechnology. Oxford and IBH: New Delhi; 1992. p. 297–305. [45] Junker BH. Scale-up methodologies for Escherichia coli and yeast fermentation processes. J Biosci Bioeng 2004;97:347–64. https://doi.org/10.1016/S1389-1723(04)70218-2. [46] Bisio A, Kabel RL. Scaleup of chemical processes: Conversion from laboratory scale tests to successful commercial size design. Inc.: John Wiley & Sons; 1985 [p. 406]. [47] Steel R, Maxon WD. Dissolved oxygen measurements in pilot- and production-scale novobiocin fermentations. Biotechnol Bioeng 1966;8:97–108. https://doi.org/10.1002/bit.260080109. [48] Maranga L, Cunha A, Clemente J, Cruz P, Carrondo M. Scale-up of virus-like particles production: Effects of sparging, agitation and bioreactor scale on cell growth, infection kinetics and productivity. J Biotechnol 2004;107:55–64. https://doi.org/10.1016/j.jbiotec.2003.09.012. [49] Doble M, Kruthiventi AK, Gaikar VG. Biotransformations and bioprocesses. New York: Marcel Dekker Inc; 2004. [50] Charles M, Wilson J, Fermenter design, in: Flickinger, M. C., Drew, S.W. (Eds.), Bioprocess technology: fermentation, biocatalysis, and bioseparation, vol. 2. John Wiley and Sons, New York; 1999, p. 1157–1189. [51] Alam Z, Razali F. Scale-up of stirred and aerated bioengineering bioreactor based on constant mass transfer coefficient. Jurnal Teknologi 2005;43:95–110. [52] Yawalkar AA, Heesink ABM, Versteeg GF, Pangarkar VG. Gas-liquid mass transfer coefficient in stirred tank reactors. Can J Chem Eng 2002;80:840–8. https://doi.org/10.1002/cjce.5450800507. [53] Stanbury PF, Whitaker A, Hall SJ. Chapter 9 — Aeration and agitation. Principles of fermentation technology. second ed. Oxford: Melbourne: Butterworth-Heinemann; 2003. p. 537–618. [54] Linek V, Kordac M, Fujasova M, Moucha T. Gas–liquid mass transfer coefficient in stirred tanks interpreted through models of idealized eddy structure of turbulence in the bubble vicinity. Chemical Engineering and Processing: Process Intensification 2004;43:1511–7. https://doi.org/10.1016/j.cep.2004.02.009. [55] Zhu Y, Bandopadhayay PC, Wu JI. Measurement of gas-liquid mass transfer in an agitated vessel — A comparison between different impellers. J Chem Eng Jpn 2001;34: 579–84. [56] Wong I, Hernandez A, Garcia MA, Segura R, Rodriguez I. Fermentation scale up for recombinant K99 antigen production cloned in Escherichia coli MC1061. Process Biochem 2002;37:1195–9. https://doi.org/10.1016/S0032-9592(01)00299-0. [57] Shin CS, Hong MS, Lee J. Oxygen transfer correlation in high cell density culture of recombinant E. coli. Biotechnol Tech 1996;10:679–82. https://doi.org/10.1007/BF00168479. [58] Flores ER, Perez F, de la Torre M. Scale-up of Bacillus thuringiensis fermentation based on oxygen transfer. J Ferment Bioeng 1997;83:561–4. https://doi.org/10.1016/S0922-338X(97)81137-3. [59] Narciandi E, Garcia J, Motolongo J, Machado JA, Benitez JV, Novoa LI, et al. Produccion, aislamiento y semipurificacion de la proteina transmembranica gp-41 sintetizada en Escherichia coli. Biotecnología Aplicada 1993;10:36–40. [60] Narciandi E. High-level production of p36 from HIV-2 in fed-batch culture of recombinant Escherichia coli. J Ferment Bioeng 1996;81:360–2. https://doi.org/10.1016/0922-338X(96)80593-9.