Fluorescein isothiocyanate (FITC) was purchased from Solorbio (Shanghai, China). Lipofectamine CRISPMAX was purchased from Thermo Fisher (China). The serum mediums were purchased from Natocor (Wuhan, China). All the other chemical reagents were purchased from Sigma-Aldrich (Mainland, China). E. coli DH5α was used as the host for DNA manipulations, while E. coli BL21 (DE3) was the host for recombinant expression of CL7-tagged eGFP, Cas9, cellulases, and CBM domains. The P. pastoris strain GS115 and the vector pPICZαA were obtained from Invitrogen (Carlsbad, CA, United States). The vectors pET23a-T and pET23a-CL7 were constructed and stored in our laboratory previously (Li et al., 2019; Dong et al., 2020). In addition, the plasmids to express CL7-tagged eGFP and Cas9 (Qiao et al., 2019) were constructed before. The genes encoding exo-mode cellobiohydrolases (CBH) from Yarrowia lipolytica, endoglucanase (EG) from Clostridium thermocellum DSM1237, glucose-tolerant β-glucosidase (BGL) from Thermoanaerobacterium thermosaccharolyticum DSM 571, and CBM from Thermobifida fusca were synthesized by Sangon Biotech (Shanghai, China). The plasmids of the above-mentioned cellulases and CBM were all constructed before (Dong et al., 2020).

The E. coli strains were grown in LB medium (1% tryptone, 0.5% yeast extract, 1% NaCl) supplied with 100 μg/m ampicillin. P. pastoris yeasts were grown firstly in YPD plates (1% yeast extract, 2% peptide, and 2% glucose) supplemented with 100 µg/ml of Zeocin and then cultured in BMGY/BMMY medium base (20.0 g/L peptone, 10.0 g/L yeast extract, 100 mmol PBS broth, pH 6.0).

The plasmids of IM7-displaying yeast were digested with Pme1 and transformed into GS115 competent cells. Transformants were firstly isolated by incubation at 28°C for 48 h on YPD plates supplemented with 100 µg/ml of Zeocin. Then, five to ten single colonies of transformants were inoculated in 20 ml of BMGY in 250 ml flasks and cultivated at 28°C under 200 rpm. After 24 h, the cells were centrifuged at 5,000 × g for 5 min, resuspended in 20 ml of BMMY medium containing 1% (v/v) methanol, and continued to grow at 28°C and 200 rpm for 24 h.

To express the CL7-tagged proteins in E. coli, 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) was added to the cells when the cells were grown to an OD₆₀₀ of 0.6. Then, the cells were grown at 18°C for 12 h. The E. coli cells collected were resuspended in PBS buffer containing 200 mm NaCl and 10 mm CaCl₂ (pH 7.4) and then sonicated on ice for 20 min. The cell lysates were either purified by Ni-NTA affinity columns or directly incubated with the Ca-alginate–encapsulated yeast strains.

The yeast strains were harvested and washed twice by ice-cold water, resuspended, and blocked in 1 ml PBS buffer (200 mM NaCl, pH 7.4) with 1 mg/ml BSA for 1 h at 4°C with rotation. Then, 1 µL of mouse anti-HA tag monoclonal antibodies was added to the cell suspension of 1,000 µL and then incubated at room temperature with rotation for 2 h. In the next step, the cells were washed three times with PBS and resuspended in 200 µL of PBS with the addition of 1 µL of FITC-conjugated goat anti-mouse IgG (H + L) antibodies, followed by incubation of them at room temperature for 1 h with rotation. Finally, the cells were washed three times with PBS, resuspended in 1 ml of PBS, and examined by fluorescence microscopy (IX73, Olympus, Tokyo, Japan). Flow cytometric analysis (FACS) was analyzed with a flow cytometer (CytoFLEX, Beckman Coulter, Suzhou, China) to estimate the percentage of the fluorescence-positive IM7-displaying yeast cells.

Ca-alginate yeast-encapsulated beads were prepared by dropping a mixture of sodium alginate solution and yeast cells (10 g/L) into a saturated boric acid containing 4% (w/w) CaCl₂ solution under slow magnetic stirring at 25°C. The resulting beads were 3 ± 0.2 mm in diameter and stirred in this solution for 30 min. They were then collected by filtration, washed three times with sterile and deionized water, and stored in Tris-HCl buffer (pH 7.0) at 4°C.

Samples of SA@yeast beads were coated under vacuum with a thin layer of gold and examined by scanning electron microscopy (SEM) (JSM 6510LV, JEOL, Japan) at 10–20 kV. Electron microscopy showed that yeast cells were attached to the Ca-alginate beads in the inner region.

The target plasmid DNA was synthesized and inserted into a pET23a(+) vector. For a typical cleavage assay, the Cas9 RNP (300 ng) was diluted into the cleavage buffer (25 mM Tris-HCl, pH 8.0, 150 mM KCl, 5 mM MgCl₂, 5 mM DTT, 5% glycerol). The plasmids (200 ng) were cleaved in the cleavage buffer. The mixture was incubated at 37°C for 30 min, and then the temperature was elevated to 85°C for 5 min to stop the reaction. The cleavage products were analyzed by agarose gels and viewed by GoldView staining.

The enzyme activity of cellulosome or free cellulases against microcrystalline cellulose (Avicel) or phosphoric acid–swollen cellulose (PASC) or carboxymethyl cellulose (CMC) was detected by 3,5-dinitrosalicylic acid (DNS) assay (Bailey et al., 1992). The PASC (86.2) was prepared from Avicel (Sangon Biotech, Shanghai, China) as described previously (Zhang et al., 2006). Minicellulosomes or free cellulases were incubated with 0.1% cellulose substrate in 50 mM citrate buffer (pH 4.8) with 10 mM CaCl₂ at 50°C for 30 min. After addition of DNS and boiling for 10 min, the reducing sugars were detected at 540 nm. One unit of the enzyme activity was defined as the amount of enzyme that released 1 micromol of product from the cellulose substrate at 50°C in 1 min.

Firstly, we tested the possibility to directly adopt the IM7-displaying yeast cells (Supplementary Figure S1) to pack chromatograph columns for purification of the target proteins. Unfortunately, the elution efficiency of protein is very low in this way since the yeast cells are possibly packed too tightly to block the column. To solve the issue, we decided to encapsulate the engineered IM7-displaying yeast cells in calcium alginate beads (Figure 1A) and then use them for protein purification (Figure 1B). After many assays (Supplementary Table S1), we found that the best coating solution is mixing the IM7-displaying yeast cells with the solutions containing 2.5% (w/w) sodium alginate (SA) (Fan et al., 2017; Kim et al., 2017; Benucci et al., 2019), 0.5% (w/w) gelatin, and 2% (w/w) polyvinyl alcohol (PVA). Meanwhile, the saturated boric acid containing 4% CaCl₂ was employed as the cross-linking reagent. As a result, the Ca-alginate yeast-encapsulated beads (hereafter named “SA@yeast beads”) were constructed with an average diameter of 3 mm (Figure 2A). The SEM image (Figure 2B) shows the interiors of these beads, clearly indicating that the yeast cells were well encapsulated. Importantly, there are plenty of pores inside the beads, allowing the free proteins to move in and out efficiently. This is the foundation for our method.

To verify the versatility of our method, we harnessed the SA@yeast beads to purify various proteins with different sizes, including eGFP (enhanced GFP) (∼27 kDa) (Zhang et al., 2017), Streptococcus pyogenes Cas9 (spCas9, ∼170 kDa) (Jinek et al., 2012), and Yarrowia lipolytica exo-mode cellobiohydrolase (Wei et al., 2014) (CBH, ∼100 KDa). Following a standard procedure, the Escherichia coli cell lysates were incubated with the beads and washed by a high-salt buffer containing 500 mM NaCl. To remove the N-terminal CL7 tag of the target protein, an engineered human HRV 3C proteinase with a CL7 tag was added for on-column cleavage followed by collecting the elution. As shown in Figure 2C, the eGFP proteins were purified in a single step with a purity of over 95%. Similar results were observed for both Cas9 and CBH (Supplementary Figure S2), demonstrating that our method can be used as a universal tool for protein purification. The protein loading capacity of this CL7/IM7 yeast affinity system is ∼10–15 mg per 1 ml SA@yeast beads (Table 1). Compared with the Ni-NTA affinity purification approach, the CL7/IM7 affinity system significantly increases the purification level of the end protein products (Supplementary Figure S3). More importantly, it maintains or improves the activity of the protein products (Vassylyeva et al., 2017). For instance, the purified Cas9 enzymes made by this method show excellent endonuclease activity (Figure 3).

To extend the scope of the method, we adopted the system to coat the multienzyme complex for biocatalysis. The cellulosome, a complicated multienzyme machine produced by cellulolytic microorganisms, was chosen to perform the proof-of-concept experiment. To achieve efficient carboxymethyl cellulose hydrolysis, three cellulases including an endoglucanase (EG) from Clostridium thermocellum DSM1237 (Leis et al., 2017), a glucose-tolerant β-glucosidase (BGL) from Thermoanaerobacterium thermosaccharolyticum DSM 571 (Zhao et al., 2013), and a CBH from Yarrowia lipolytica (Wei et al., 2014), as well as a carbohydrate-binding module (CBM) from Thermobifida fusca (Forsberg et al., 2014), were fused with an N-terminal CL7 tag and over-expressed in E. coli. Then, the IM7-displaying yeast cells were in vitro incubated with the E. coli lysates containing three cellulases and CBM at a ratio of 1:1:1:1 (Dong et al., 2020), leading to assembly of minicellulosomes on the surface of these yeast cells. Next, those yeast cells were encapsulated in Ca-alginate beads, and their cellulase activity against Avicel, phosphoric acid–swollen cellulose [PASC (86.2)], and carboxymethyl cellulose (CMC) was determined by measuring the reducing sugars within the first 30 min. Compared with free cellulases in solution, the coated multienzyme complex exhibited significantly enhanced catalytic activity (Figure 4A). More importantly, these SA@yeast beads can be repeatedly used for at least six times with very slight activity decreases (Figure 4B), demonstrating great potential in the production of cellulosic ethanol in the future.