A Streptomyces lividans SipY deficient strain as a 1 host for protein production : standardization of 2 operational alternatives for model proteins . 3

This work was supported by the Spanish Ministry of Economy and Competitiveness (project number CTQ2011-28398-CO2-01), the Spanish Ministry of the Environment and Rural and Marine affairs (Grant No. EGO22008) and Grant PIE201220E003 from the CSIC. The authors from UAB are members of the Research group 2014SGR452 and of the Biochemical Engineering Unit of the Reference Network in Biotechnology (XRB), Generalitat de Catalunya.


Introduction 20
Actinomycetes are ubiquitous soil bacteria containing a high G+C genome 1 , and the Streptomyces genus 21 belongs to the actinomycetes family and encompasses aerobic mycelia-forming soil bacteria, which are 22 natural producers of many antibiotics and other biologically active molecules 2 . A particularly appealing 23 property of the Streptomyces species is its natural ability to produce and secrete extracellular proteins 3 .

24
This, together with the knowledge acquired over the years on the genetic manipulation of these 25 microorganisms 1 and the accumulated scale-up cultivation experience, have rendered the streptomycetes 26 as a group of interesting bacteria to be exploited in Biotechnology, expressing and secreting extracellular 27 proteins of industrial application 4-6 . Streptomyces lividans is the most utilised species for protein 28 production due to its limited restriction-modification system and relative low endogenous protease 29 3 activity 7 . In the last ten years, several proteins have been expressed in this strain with variable active 1 protein concentrations (1-10 mg.L -1 in some cases, up to 300 mg.L -1 in others) 4,5, 8-10 . 2 Secretory proteins are synthesised by means of a signal peptide at their amino end that helps to trigger the 3 protein to the translocation machinery present at the cell membrane. Type I signal peptidases (Type I 4 SPases) are responsible for cleaving the signal peptide that allows the release of the translocated proteins 5 to the culture broth. Four type I signal peptidases are identified in Streptomyces lividans: Sip W, SipX, 6 SipY and Sip Z 11 . SipY appears to be the major SPase, since the secretome of a SipY defective S.lividans 7 strain is severely affected in this strain 12 . Despite this handicap, the SipY deficient strain was shown to 8 be very useful for extracellular protein production in S. lividans, because the absence of the major signal 9 peptidase was partially compensated by the three remaining active signal peptidases. This permitted the 10 efficient overproduction of extracellular enzymes when their genes were propagated in multicopy as has 11 been revealed for agarase 13 in the absence of extracellular proteases. The diminished extracellular 12 proteolytic activity should favour the extracellular accumulation of the overproduced protein, potentially 13 improving its stability and the diminished presence of the otherwise naturally-produced extracellular 14 proteins 12 should considerably simplify the downstream processing of the secreted overproduced protein.

15
Bacterial proteins using the secondary Tat (Twin Arginine Translocation) secretory pathway are naturally 16 released to the medium properly folded and fully functional 14,15 . Therefore, the overproduction of a Tat-17 secreted protein by an S. lividans SipY deficient strain should potentially render a considerable amount of 18 correctly folded, highly stable extracellular product. The practical implementation of protein production 19 processes requires the development of operational procedures to maximise protein productivity. The first 20 objective of this work is to improve process development using a Tat-dependent protein, agarase, as a 21 model to study its overproduction by the SipY mutant strain, with special emphasis on the bioreactor 22 operation mode.

23
Additionally, the Sip Y deficient strain has been used for the production of Streptomyces lividans laccase 24 (Small laccase, SLAC), predicted to be secreted via the Tat pathway. The Streptomyces coelicolor laccase 25 was characterized in 2004 16 as a two-domain protein and showed high thermal stability, detergent 26 resistance and an optimum pH in the basic range (around 8). This last fact is remarkable because, 27 although some laccases of fungal 17 and algal origin 18 have been reported to exert the maximum activity at 28 neutral pH, most of the commercially available laccases-generally from fungal origin-have their optimum 29 at acidic pH. Hence, SLAC is a promising enzyme for pollutants' degradation in liquid effluents near 30 4 neutral pH like urban wastewater. In addition the Streptomyces SLACs could also be useful as a 1 biocatalyst for biotransformations performed in the basic pH range; i.e. in single pot multienzymatic 2 reactions, coupling oxidation and aldol addition 19 . In consequence, the possibility of efficient extracellular 3 production of SLAC by the S. lividans SipY deficient strain is worth investigating, and the bioreactor 4 operation alternatives developed for agarase were extended to SLAC synthesised in the SipY deficient

25
Batch and fed-batch cultures were performed in NMMP medium 1   2 Mycelium stored at -80ºC was used to inoculate 3mL of R5 medium with thiostrepton and/or kanamycin 3 as required and set to grow in 50-mL falcon flasks. After 24h of incubation at 30ºC and 150 rpm, the cells 4 were transferred to 500mL or 1L baffled flasks containing 50mL or 250mL of the same medium. These 5 cultures were grown for 24h at 30ºC and 150 rpm and the corresponding biomass harvested by 6 centrifugation; biomass at an initial concentration of 0.1g·L -1 of wet weight was used to inoculate a 3L 7 bioreactor or a 7 L bioreactor.

10
Quantity One (Bio-rad) was used to quantify the band observed after the reaction with the antibodies 11 against DagA.

14
The agarase production by the SipY deficient strain, S. lividans TK21Y62(pAGAs5) 13 , was studied in 15 comparison to S.lividans TK21(pAGAs5) (from now the wild type strain). Bacterial cell cultures were 16 performed in a 3L bioreactor using either mannitol or glucose as carbon source. The average results

17
concerning biomass growth and agarase production in mannitol for both strains are shown in Table 1.

18
Biomass and total extracellular protein concentration were higher for the wild type than for the SipY 19 deficient strain, while total extracellular agarase and specific agarase activities per gram of biomass and 20 per mg of protein were higher for the deficient strain; this was expected since the accumulation of 21 extracellular proteins is severely diminished in the SipY deficient strain compared to the wild type 12,13 .

22
Cultures grown in the presence of glucose as carbon source (data not shown) produced 30% less agarase 23 than those grown in the presence of mannitol. The time course of a representative mutant strain culture in 24 the presence of mannitol as carbon source is shown in Figure 1. Biomass growth starts before mannitol 25 consumption, presumably using casamino acids as a carbon source 28 , with a specific growth rate of 26 µ=0.12 h -1 at the exponential phase. During extracellular agarase accumulation the protease levels in the 27 medium wer lower than the detection limit.

8
Additionally, the mutant exhibited better rheological properties compared to the wild type. The SipY 1 deficient strain grows in a disperse suspension considerably diminishing cellular adhesion to the reactor 2 wall, a very significant drawback of the wild type strain which grows in clump aggregates. This 3 remarkable disperse growth characteristic may simplify the standardization and scale up of production 4 due to its better rheological behaviour.

5
All the above results suggest that S. lividans TK21Y62 can provide advantages for the overproduction of 6 Tat secretory proteins. Further improvement of this production was attempted by extending the bioreactor 7 operation to fed-batch mode. Fed-batch operational mode was assayed as a method to simultaneously increase agarase concentration 10 and productivity (expressed as U.L -1 .h -1 ). The strategy consisted of mannitol pulse addition, recovering its 11 initial concentration in the culture medium. All the other nutrients were calculated to be in excess of the 12 necessary amount to support the biomass concentration achieved in the process. Pulse addition was 13 performed avoiding total mannitol depletion. As can be seen in Figure 1, when mannitol was totally 14 depleted, biomass concentration decreased due to cell lysis and a total extracellular protein increment was 15 detected in the supernatant, without increasing the agarase activity.

15
The bacterium was cultured in a 5L batch bioreactor using an NMMP modified medium as indicated in 16 the Materials and Methods section with 10g·L -1 of either mannitol or glucose as carbon sources. In both 17 conditions sustained growth and laccase synthesis and secretion were obtained. Batch growth on mannitol 18 is depicted in Figure 3A.

19
Growth was maintained until the total depletion of the carbon source occurred, with a specific growth rate 20 of 0.09 h -1 at the exponential phase. The behaviour of the mutant strain was as expected, producing a low 21 amount of secretory proteins due to the major signal peptidase mutation. Laccase activity accumulated 22 extracellularly until a maximum of 5.8 U.L -1 , following a temporal secretion pattern compatible with a 23 protein secreted via the Tat pathway 30 . As in the case of agarase, extracellular protease levels were not 24 detectable and a decrease in biomass concentration and laccase activity took place after total mannitol 25 consumption, which was attributed to cell lysis and possible laccase degradation by intracellular released 26 proteases. A similar result was obtained in glucose batch cultures (Fig 3B), with a higher specific growth 27 rate, (0.15 h -1 ), reaching a maximum biomass concentration of 7.5 g.L -1 , The biomass concentration also 28 decreased at the end of culture, probably due to cell lysis. Extracellular laccase accumulation followed the 29 expected pattern although the secreted amount was lower, 3.1 U.L -1 , and no activity decay parallel to the 1 cell lysis was observed.

2
In summary, maximum values of laccase production, under comparable conditions, are presented in Table   3 2. As can be seen, laccase activity is considerably higher for mannitol than for glucose, in terms of total 4 production, specific production per gram of biomass and specific production per mg of extracellular 5 protein.

6
Fed-batch cultures for laccase production 7 Agarase production by the SipY deficient strain increased with one single addition of mannitol and 8 further additions of carbon or nitrogen sources did not improve the process (Fig 2). Therefore, this 9 operational strategy has been applied for laccase production with mannitol or glucose fed-batch additions, 10 restoring the initial 10 g.L -1 concentration of the carbon sources. The time profiles for both cultures are 11 shown in Figure 4. As can be seen, mannitol addition led to a moderate increase in laccase extracellular 12 activity while glucose addition exerted a drastic effect by almost doubling the enzyme activity.

13
In the mannitol fed-batch culture (Fig. 4A), maximum biomass concentration and laccase activity were

25
Mannitol was a better carbon source for the agarase production by the SipY deficient strain grown in 26 batch cultures; the use of glucose as carbon source revealed a lower agarase production, as previously 27 described 21 .

28
The fact that bacterial growth starts before mannitol consumption, presumably using casamino acids 1 present in the medium as carbon source (Figure 1 promoter. As expression is constitutive, it cannot be decoupled from growth, competing for nutrient 10 resources, and, as a result, when the cellular growth stops, gene expression stops as well, which results in 11 no increase of agarase synthesis.

12
A comparison of the two operation modes: batch, and fed-batch was carried out for agarase production 13 (Table 3). Higher productivity was obtained using the fed-batch process, which was feasible for practical 14 application to produce the model protein agarase in a bench-scale bioreactor. The fed-batch operation 15 with two mannitol additions resulted in the production of a more concentrated agarase, facilitating a much 16 easier downstream processing of the extracellular enzyme.As expected, the estimated specific agarase 17 activity per mg of protein in batch and fed-batch operation is nearly the same (Table 3), since in both 18 cases agarase itself was the main contributor to the extracellular protein concentration.

19
Concerning laccase, a comparison of fed batch culture production in the presence of mannitol or glucose 20 as carbon source is presented in Table 4, comprising the data obtained at periods reflecting the effect of 21 each pulse (see Figure 5). A similar production of laccase (5.35 U.L -1 ) is obtained after 55 hours' growth,

22
corresponding to either a batch mannitol culture or to two additions in fed-batch glucose culture.

23
Consequently, the operation in fed-batch mode with mannitol or glucose additions led to similar 24 productivities yet with three fold more substrate consumption in the case of the glucose. Nevertheless, for 25 process optimization, it must be considered that mannitol price per Kg is more than six fold higher than 26 glucose and the proposed fed-batch operation with glucose as carbon source seems to afford more 27 economic incentives.

28
In conclusion, the above results confirm the potentiality of the SipY defective strain as a producer of