Evolved α‐factor prepro‐leaders for directed laccase evolution in Saccharomyces cerevisiae

Summary Although the functional expression of fungal laccases in Saccharomyces cerevisiae has proven to be complicated, the replacement of signal peptides appears to be a suitable approach to enhance secretion in directed evolution experiments. In this study, twelve constructs were prepared by fusing native and evolved α‐factor prepro‐leaders from S. cerevisiae to four different laccases with low‐, medium‐ and high‐redox potential (PM1L from basidiomycete PM1; PcL from Pycnoporus cinnabarinus; TspC30L from Trametes sp. strain C30; and MtL from Myceliophthora thermophila). Microcultures of the prepro‐leader:laccase fusions were grown in selective expression medium that used galactose as both the sole carbon source and as the inducer of expression so that the secretion and activity were assessed with low‐ and high‐redox potential mediators in a high‐throughput screening context. With total activity improvements as high as sevenfold over those obtained with the native α‐factor prepro‐leader, the evolved prepro‐leader from PcL (αPcL) most strongly enhanced secretion of the high‐ and medium‐redox potential laccases PcL, PM1L and TspC30L in the microtiter format with an expression pattern driven by prepro‐leaders in the order αPcL > αPM 1L ~ αnative. By contrast, the pattern of the low‐redox potential MtL was αnative > αPcL > αPM 1L. When produced in flask with rich medium, the evolved prepro‐leaders outperformed the αnative signal peptide irrespective of the laccase attached, enhancing secretion over 50‐fold. Together, these results highlight the importance of using evolved α‐factor prepro‐leaders for functional expression of fungal laccases in directed evolution campaigns.


Introduction
Fungal laccases (EC 1.10.3.2, benzenediol:oxygen oxidoreductases) catalyse the oxidation of phenols, aromatic amines and other compounds, with the concomitant reduction of molecular oxygen to water (Solomon et al., 1996;Gianfreda et al., 1999;Alcalde, 2007). The laccase substrate spectrum can be expanded notably through the laccase mediator system, a system based on diffusible electron carriers that become strong oxidizers upon oxidation by laccase to act then on other substratesmostly non-phenolicsthat are otherwise little oxidized by the laccase alone (Morozova et al., 2007;Cañas and Camarero, 2010). Given this broad substrate range and their minimal requirements, fungal laccases belong to the elite of oxidases that can be employed in very distinct areas of biotechnology, from organic synthesis to novel green processes and beyond (Riva, 2006;Kunamneni et al., 2008a,b;Mate and Alcalde, 2017). For decades, these blue multicopper-containing enzymes have attracted much interest and as such, they have been the focus of many attempts to engineer them through directed evolution with a view to adapt them to harsh industrial conditions, making them resistant to high temperature or extreme pH, or functional in the presence of different types of inhibitors or organic solvents, to name but a few (Rodgers et al., 2010;Alcalde, 2015, 2016). Assisted by a strong portfolio of solutions that combine bio-and electro-catalysis, the application of engineered laccases is no longer a pipedream. However, this new age of directed laccase evolution requires tools and library creation methods that can be readily manipulated to help generate superior biocatalysts.
One of the main hurdles when engineering fungal laccases is their poor functional expression in heterologous hosts and limited secretion. Due to its eukaryotic nature and simple fermentation requirements, Saccharomyces cerevisiae is a suitable microorganism to improve recombinant laccases by directed evolution (Gonzalez-Perez et al., 2012). With an efficient DNA recombination apparatus, this yeast allows us to perform a wide array of genetic manipulations, facilitating the generation of molecular diversity. Protein engineering strategies have been used to boost laccase secretion in S. cerevisiae, including (i) the replacement of the native signal peptide with different prepro-leaders, (ii) directed evolution of the mature laccase, (iii) directed evolution of prepro-leaders, and (iv) a combination of these approaches.
The evolution of a-factor prepro-leaders from S. cerevisiae is exceptionally relevant, in the hope that they could serve as universal signal peptides in different directed laccase evolution enterprises, an issue that has yet to be addressed. The pioneering work of the Wittrup group indicated that directed evolution of a-factor preproleaders could enhance the expression of different types of proteins in yeast, from full-length antibodies to cellulases (Rakestraw et al., 2009;Dana et al., 2012).
However, when we have tested evolved a-factor preproleaders in different groups of ligninases (e.g. evolved prepro-leaders from laccases to enhance the secretion of unspecific peroxygenases (Molina-Espeja et al., 2014)), the results were not encouraging, suggesting that evolved prepro-leaders may only be successfully exchanged between proteins of similar phylogeny. Conversely, it still remains unclear whether an a-factor prepro-leader that has been evolved to enhance protein expression can be translated to a different enzyme group to achieve similar benefits or can be even effectively transferred between proteins that belong to the same enzyme group. Particularly, the use of evolved prepro-leaders for directed laccase evolution experiments could help enhance secretion levels in highthroughput screening -HTSformat (i.e. cultures in microtiter plates). Should this be the case, the oxidation of high-redox potential mediators that are barely oxidized by laccase might be readily detected during screening such that their oxidation rates could be improved by iterative rounds of directed evolution.
In this study, we combined different native and evolved prepro-leaders from previous directed evolution campaigns with four fungal laccases that display low-, medium-and high-redox potential and a protein sequence identity between 26 and 73%. Twelve a-factor prepro-leader:laccase fusions were constructed and their influence on expression and secretion was assessed in HTS format with low-and high-redox potential mediators (2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid (ABTS) and K 4 [Mo(CN) 8 ] respectively) so that the restricted growth conditions of a directed evolution round in terms of poor cell growth and enzyme secretion were emulated. A secretion pattern driven by the preproleader attached to the laccase was established and discussed within a mutational context.

Results and discussion
The a-factor prepro-leader from S. cerevisiae is classically employed to enhance the secretion of foreign proteins by yeast (Shuster, 1991;Romanos et al., 1992). This secretory leader contains a pre-region of 19 amino acids and a pro-region of 64 amino acids with three Nlinked glycosylation sites, Fig. 1. The canonical pre-leader is implicated in the translocation of the nascent secretory protein, which is removed from the endoplasmic reticulum (ER) membrane by the action of a signal peptidase between residues 19 and 20. At this point a primary oligosaccharide is added, after which the protein is packed into vesicles for transportation to the Golgi where it is further glycosylated by long outer chains of mannose residues. The a-factor pro-leader is thought to display chaperone-like activity, and it is processed in the Golgi compartment through the action of KEX2, STE13 and KEX1 proteases (the latter of which is unnecessary for the heterologous expression of a-factor prepro-leader fusion proteins). Some years ago, our laboratory achieved the heterologous functional expression in S. cerevisiae of two different high-redox potential laccases from basidiomycete PM1 (PM1L) and Pycnoporus cinnabarinus (PcL; Mate et al., 2010;Camarero et al., 2012). After attaching them to the native a-factor prepro-leader, these fusions were subjected to joint rounds of directed evolution to improve secretion. Similarly, we were also involved in the directed evolution of the low-redox potential laccase from the ascomycete Myceliophthora thermophila (MtL). In this case, the laccase, as well as its native prepro-leader and C-terminalwhich was successfully processed after introducing a KEX2 cleavage site, were evolved together (Bulter et al., 2003). In the current work, the native S. cerevisiae a-factor prepro-leader and the evolved afactor prepro-leaders from PM1L and PcL (a native , a PM1L and a PcL respectively) were tested to explore their possible combination with evolved laccase mutants PM1, PcL and MtL, and also with the native laccase isoform LAC3 from the basidiomycete Trametes sp. strain C30 (TspC30L), which has proved to be heterologously expressed by yeast (Klonowska et al., 2005), Fig. 1. The protein sequence identity between these four laccases ranges from 73% to 26%, where three of the four laccases (TspC30, PM1L and PcL) are medium-to highredox potential laccases with a sequence identity window of 69-73% at the amino acid level, Table 1. Accordingly, twelve a-factor prepro-leader:laccase fusions were generated by gene assembly through IVOE (Alcalde, 2010), and secretion was assessed within a HTS context so that conditions found during a directed evolution experiment were rapidly reproduced.
It is worth noting that cell growth in HTS microculture format is far from ideal (in terms of oxygen availability and stirring limitation), implying severe constrains during the preculture, growth and production phases. Although the use of a rich non-selective medium is preferred in the final stages of larger fermentations, it is not always suitable to produce laccase mutant libraries in an HTS format as it may interfere with the screening of different high-redox potential mediators whose oxidized products could yield responses at the UV/VIS wavelength frontier (unpublished material). Moreover, the secretion of native proteins and ancillary factors by the yeast may also affect the measurements. Therefore, a selective expression medium (SEM) for laccase secretion by S. cerevisiae in HTS format was used to overcome these hurdles. This SEM contained a supplement of copper to favour cofactor uptake by laccases and more importantly, galactose (instead of raffinose or glucose) as the only carbon source to Fig. 1. Native and evolved a-factor prepro-leaders. Processing sites in the pre-and pro-regions are indicated by the blue arrows, the red arrows highlight mutations and the green dotted lines indicate glycosylation sites. a native , native a-factor prepro-leader; a PcL , evolved a-factor prepro-leader from a previous evolution campaign performed on PcL (Camarero et al., 2012); a PM1L , evolved a-factor prepro-leader from a previous evolution campaign performed on PM1L (Mate et al., 2010). The a-factor prepro-leader:laccase fusions were constructed by gene assembly through in vivo overlap extension (IVOE; Alcalde, 2010). All PCR reactions were cleaned, concentrated, loaded in preparative low melting point agarose gels (0.75% w:v) and purified. The constructs were cloned under the control of the GAL1 promoter of the pJRoC30 expression shuttle vector, which was linearized with BamHI and XhoI, and the linear plasmid was concentrated and purified as above. The reaction mixtures contained DNA template (10 ng ll À1 ), 1 mM dNTPs (0.25 mM each), 3% (v/v) dimethylsulfoxide (DMSO) and 0.05 U/of Pfu Ultra DNA polymerase in a final volume of 50 ll, along with the appropriate primers (0.25 lM). The design of the overlapping 40 bp regions between adjacent fragments allowed the homologous recombination machinery of S. cerevisiae to drive the in vivo fusion and cloning of the different genetic elements (protease-deficient S. cerevisiae strain BJ5465). trigger laccase expression under the control of the GAL1/10 promoters (see legend for Fig. 2). SEM allowed laccase activity to be measured at both the near UV and visible wavelengths while providing resistance against plasmid degradation given that selection is exerted during all growth stages. In terms of the screening assays, two different redox mediators were chosen, each with a E°that is pH-independent: ABTS, E°A BTS •+ = +690 versus NHE; K 4 [Mo (CN) 8 ], E°= +780 mV versus NHE. ABTS is a mediator whose radical cation ABTS •+ gives a reliable colorimetric response with a maximum of absorbance~418 nm. This organic molecule is becoming a common substrate for HTS assays in different evolution campaigns involving laccases, peroxidases and peroxygenases (Alcalde, 2015). By contrast, K 4 [Mo(CN) 8 ] is a mediator with a higher redox potential that belongs to the group of transition metal coordination complexes and it can cycle between -4/-3 redox states. As such, K 4 [Mo(CN) 8 ] does not yield a radical product upon oxidation by laccase as the electron exchange is focused on the metallic atom of the complex but it does follow an electron transfer route, as ABTS (Rochefort et al., 2004). While its reaction product gives reliable response at 388 nm, K 4 [Mo(CN) 8 ], like other high-redox potential mediators, is hardly oxidized by low-redox potential laccases. Therefore,  Under these premises, the secretion of a-factor prepro-leader:laccase fusions grown in SEM/HTS format was evaluated using ABTS and K 4 [Mo(CN) 8 ]. Notably, when PM1L was fused to the evolved a-factor preproleader from PcL (a PcL , Fig. 1), secretion augmented~7fold irrespective of the redox mediator tested, Fig. 2A. Similar results were obtained with PcL fusions, although the total activity detected in the microculture broth was less than that of the PM1L fusions due to their weaker expression (2 and 8 mg l À1 for PcL and PM1L mutants respectively; Camarero et al., 2012;Mate et al., 2010), Fig. 2B. Thus, evolved a-factor prepro-leaders conferred a similar pattern of secretion to the high-redox potential laccases PM1L and PcL, in the order a PcL > a PM1L > a native . The strongest secretion of the medium-redox potential TspC30L was also achieved when fused to a PcL (with a production of~500 ABTS U l À1 and a secretion pattern a PcL > a native > a PM1L ) despite the fact that this prepro-leader was originally evolved for PcL. By contrast, secretion of the low-redox potential MtL was similar for both the a native and a PcL constructions, Fig. 2C and D. Thus, the strong correlation between protein sequence identity and secretion driven by the different  ; a native , native a-factor prepro-leader from S. cerevisae; a PcL , evolved a-factor prepro-leader from a previous evolution campaign performed on PcL (Camarero et al., 2012); a PM1L , evolved a-factor prepro-leader from a previous evolution campaign performed on PM1L (Mate et al., 2010). Measurements were made in triplicate on supernatants from three independent fermentations, and they are expressed as the mean including standard deviation. A single S. cerevisiae colony was picked from the SC dropout plate for each laccase construct, inoculated in minimal SC medium (20 ml) and incubated for 48 h at 30°C and 220 rpm (Minitron-INFORS, Biogen Spain). An aliquot of cells was used to inoculate minimal SC medium (20 ml) in a 100 ml flask (optical density at 600 nm [OD 600 ] 0.25), the cells were allowed to complete two growth phases (6 to 8 h; OD 600 = 1) and 2 ml of the culture was them added to the laccase expression medium (18 ml) in a 100 ml flask. After incubation for 72 h at 30°C and 220 rpm, the cells were harvested by centrifugation at 4500 rpm and 4°C (Eppendorf 5810R centrifuge, Germany) and supernatants assayed for ABTS activity as described previously. Minimal SC medium contained 100 ml of 6.7% (w/v) sterile yeast nitrogen base, 100 ml of a 19.2 g l À1 sterile yeast synthetic dropout medium supplement without uracil, 100 ml of sterile 20% (w/v) raffinose, 700 ml of sddH 2 O and 1 ml of chloramphenicol (25 g l À1 ). YP medium contained 10 g of yeast extract, 20 g of peptone and sddH 2 O to 650 ml. Laccase expression medium contained 144 ml of 1.55xYP, 13.4 ml of 1 M KH 2 PO 4 (pH 6.0) buffer, 22.2 ml of 2% (w/v) galactose, 0.4 ml CuSO 4 (1M), 0.200 ml of chloramphenicol (25 g l À1 ) and sddH 2 O to 200 ml. prepro-leaders indicates that while the secretion of medium-and high-redox potential laccases (with a sequence identity in the range 69-73%) can be improved by attaching them to a-factor prepro-leaders evolved for their functional expression, MtLwhich shares 26-30% sequence identity with its laccase counterpartsdoes not follow the same rules, at least within a HTS format (see below).
The oxidation of K 4 [Mo(CN) 8 ] was followed readily in the HTS context for both PM1L and PcL, the latter displaying lower responses due to its more limited secretion. As expected, the medium-redox potential TspC30L also gave a reliable response with this substrate, albeit to a much lesser extent than its high-redox potential laccase counterparts, Fig. 2A-C. Finally, no activity was recorded with the low-redox potential MtL, irrespective of the fusion tested, Fig. 2D. These results highlight the benefits of combining K 4 [Mo(CN) 8 ] with SEM for evolving and/or searching high-redox potential laccases.
Given that the growth conditions in the HTS/SEM experiments were restricted (i.e. OD 600 < 1), to fully analyse the effects of evolved prepro-leaders on secretion while circumventing possible metabolic/culture burdens, the ensemble of laccase fusions were tested in flask fermentations with rich medium (with OD 600~3 5-40), Fig. 3. Under these conditions, evolved preproleaders outperformed the secretion achieved by a native , no matter the laccase attached. This was especially conspicuous for high-redox potential laccases, the secretion of which increased up to~50-fold when they were associated to a PM1L or a PcL , Fig. 3A. By contrast, laccase cultures with SEM in flask followed a similar secretion pattern as that obtained in HTS/SEM experiments but they were precluded for larger scale production due to the limited growth of yeast in SEM (with OD 600 < 15). Thus, the composition of the medium, the format and the culture conditions become key drivers when assessing laccase activity, such that the secretion observed in HTS format within a directed evolution experiment cannot always be extrapolated to large fermentations.
Both the a PcL and a PM1L evolved prepro-leaders are derived from several rounds of directed evolution to enhance the secretion of PcL and PM1L, and they share common features. First, a similar mutation was introduced independently in the canonical pre-leader of each signal peptide (A9D and V10D for a PcL and a PM1L , respectively, see Fig. 1). These mutations are located in the hydrophobic domain of the pre-region that is involved in ER targeting. In our previous studies, we showed that, individually, these mutations improve the secretion of their fused laccase sequences by reducing markedly hydrophobicity during the extrusion of the polypeptide laccase chain into the bilayer of the ER, while their combination did not benefit secretion (Mate et al., 2010;Camarero et al., 2012). In addition to the V10D mutation, a PM1L contains a mutation in one of the three sites for N-linked glycosylation of the pro-leader (N23K, within the Asn-X-Ser/ Thr recognition motif). Similarly, a PcL carries the S58G mutation located in the second N-glycosylation site and although in this case the glycosylation site was not lost, it seems plausible that its affinity for sugar anchoring might have changed. The effect of such substitutions on secretion remains uncertain; however, a similar change at the third glycosylation site (N57D) was also reported in the best evolved a-factor preproleader appS4 that improved antibody secretion, reflecting the possible role that these three glycosylation sites could have on exocytosis (Rakestraw et al., 2009). The F48S of a PcL is another mutation located at the pro-leader. A similar substitution (F48/S/V) was again observed in four leaders evolved for antibodies secretion, which highlights how this mutation enhances the secretion of a variety of proteins, even those from quite distant families. Finally, the mutations E86G and A87T respectively found in a PcL and in a PM1L modify the STE13 processing site which, in turn, could affect the performance of KEX2 in the Golgi compartment during the final maturation stages.

Conclusions
We describe here the use of evolved a-factor prepro-leaders for the functional expression in S. cerevisiae of fungal laccases with different redox potentials to perform directed evolution experiments. When we tested such prepro-leaders within a HTS context, assaying different redox mediators, their secretion was mainly related to the laccase sequences from which they were evolved. By contrast, in flask fermentations with rich medium the evolved signal sequences improved secretion regardless of the laccase attached, taking one step closer to their 'universality' at least within the laccase enzyme group. These evolved leaders share certain similarities with other a-factor prepro-leaders evolved to express proteins from different sources, which opens a new avenue to engineer universal signal peptides for expression in yeast.