Construction of primary chassis cells with efficient protein expression in Thermus thermophilus

Abstract

Background

Thermus thermophilus HB27 is a promising thermophilic chassis for recombinant thermostable protein production, owing to its high optimal growth temperature, which can simplify downstream processing and reduce contamination risks. However, maximizing its potential requires optimized genetic tools and host strains. Key limitations include a shortage of well-characterized strong constitutive promoters and potential degradation of recombinant proteins by proteases. To address these, we established a β-galactosidase reporter system (endogenous TTP0042) to screen for strong constitutive promoters and investigated the impact of deleting specific protease genes on protein expression.

Results

Screening of 13 endogenous promoter regions identified P0984 as exhibiting significantly 13-fold higher activity than the control promoter driving the reporter gene. Constructing a plasmid-free strain (HB27ΔpTT27) successfully minimized 270 kb of the genome; it exhibited auxotrophy for cobalamin (requiring 0.1 µg/ml AdoCbl for growth) and a slightly reduced growth rate compared to the wild-type, while its transformation efficiency remained comparable. Notably, a CRISPR-deficient precursor strain (HB27ΔIII-ABΔI-CΔCRF3) showed a significant (~ 100-fold) increase in transformation efficiency compared to the wild-type, facilitating subsequent genetic manipulations. Systematic knockout of 16 predicted non-essential protease loci was performed. Characterization revealed that deletion of TTC0264 (putative ClpY/HslU) and TTC1905 (putative HhoB) significantly reduced extracellular proteolytic activity. Iterative deletion based on phenotypic analysis led to strain DSP9 (10 protease loci deletions), which maintained robust growth and exhibited enhanced accumulation of the β-galactosidase reporter protein compared to the parental strains.

Conclusions

This study provides foundational advancements for T. thermophilus HB27 chassis development, and genetic tools represent valuable resources for optimizing T. thermophilus as a platform for heterologous thermostable protein production and ideas for antibiotic-free systems.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12934-025-02785-y.

Background

The burgeoning emphasis on renewable energy sources and sustainable industrial practices significantly drives the escalating global demand for thermostable proteins. Thermostable enzymes (e.g., cellulases, xylanases, α-amylases, laccases) are crucial for applications requiring elevated temperatures, such as biofuel production, bioleaching, wastewater treatment, and agricultural biotechnology [1–7]. Utilizing thermostable proteins offers advantages like reduced cooling costs, lower contamination risks, and increased reaction rates [1–7].

However, producing thermostable proteins efficiently often faces challenges in conventional mesophilic hosts like E. coli. Issues include protein misfolding, aggregation, degradation, and lack of necessary post-translational modifications due to differences in cellular machinery and environment [8, 9]. The unique structural features conferring thermostability (e.g., specific hydrophobic interactions, salt bridges) may not be correctly established in mesophilic cytoplasm, potentially requiring specialized chaperone systems or protein engineering approaches [10–14].

Thermus thermophilus, particularly strain HB27, presents a compelling alternative host for producing thermostable proteins [9, 15]. Its optimal growth at high temperatures (65–75 °C) inherently favors the stability and correct folding of many thermostable target proteins. Furthermore, high-temperature cultivation simplifies downstream purification, as many endogenous mesophilic contaminants are eliminated [10, 16, 17]. T. thermophilus HB27 offers additional advantages like rapid growth, high cell density potential, natural competence for transformation, and available genetic tools, including CRISPR-Cas systems for genome editing [4, 5, 18–23]. Consequently, HB27 is an attractive chassis for producing various thermostable enzymes and proteins [4, 24–26]. Developing T. thermophilus as a robust chassis aligns sustainable bioproduction goals by potentially reducing energy inputs for cooling and simplifying processing [27–29]. However, it is important to note that this platform is primarily suited for thermostable proteins; expressing inherently thermolabile proteins (e.g., many mammalian biopharmaceuticals) remains a significant challenge due to the high cultivation temperatures [4, 30].

Despite its potential, optimizing T. thermophilus for industrial-scale production requires addressing key limitations. One is the limited availability of strong, well-characterized constitutive promoters for driving high-level, stable gene expression. While some inducible promoters exist (e.g., P_arg, P_dnaK, P_nar, P_pilA4), they often suffer from drawbacks like complex induction strategies, growth inhibition, or unintended regulation [31–34]. Plasmid-based expression systems can also face instability and metabolic burden [35, 36]. Another significant challenge is the potential degradation of recombinant proteins by endogenous host proteases, which can severely reduce yields [37–39]. While protease activity is essential for cellular health, specific proteases can target heterologous proteins. Previous work in T. thermophilus indicated that deleting certain proteases (e.g., Lon-type) could improve yields, but the roles of many other proteases remain uncharacterized [39].

Therefore, this study aimed to enhance T. thermophilus HB27 as a microbial chassis through a multi-pronged strategy: (1) Screening of 13 endogenous genomic regions to identify strong constitutive promoters. (2) Constructing and characterizing a plasmid-free strain (HB27ΔpTT27) to reduce genome size and explore its potential as an auxotrophy-based selection. (3) Systematically deleted 16 predicted non-essential protease genes to assess their impact on growth, extracellular proteolytic activity, and reporter protein accumulation. We aim to generate characterized strains and genetic tools to facilitate the development of T. thermophilus HB27 for improved recombinant protein production, particularly focusing on thermostable targets.

Materials and methods

Bacterial strain and cultivation

E. coli DH5α was used as a host for plasmid engineering and multiplication, cultivated in LB medium (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl) at 37 °C with 200 rpm in a shaker incubator (BLabotery, ZQPL-200). T. thermophilus HB27 (Wildtype), HB27∆TTP0042 (T. thermophilus HB27 with β-galactosidase gene knockout), HB27∆III-AB∆I-C∆CRE3 (T. thermophilus HB27 with I-C, III-A & B CRISPR-Cas system and CRF3 (CRISPR-associated Rossmann Fold) genes knockouts), and T. thermophilus stains were cultivated at 65 °C with 200 rpm shaking in Tt medium (8 g/L tryptone, 4 g/L yeast extract, 3 g/L NaCl, 4.36 mM NaHCO₃,0.26 mM MgSO₄, 0.739 mM MgCl₂0.6 H₂O, 0.012 mM KCl, and 0.249 mM CaCl₂0.2 H₂O) [23, 34]. Petri dishes of T. thermophilus were wrapped in sterilized bags to maintain plate humidity in the static incubator (BLabotery, SPL-250). Strains were listed in the supplementary data Table S1.

Plasmid construction

To construct plasmids for endogenous promoter screening, promoter regions were isolated from the HB27 genome utilizing PCR and primers with about 20 nt at the 5’-end complementary to the plasmid ligation side. Then, the TTP0042 codon (β-glucosidase reporter gene) amplified from the pTT27 plasmid was cloned into pRKP31, which was pre-digested with SalI and NheI. Then, the P31 promoter was replaced with the next screened promoters into the pRKP31-TTP0042 plasmid utilizing 2X CE Mix V3 (Vazyme, C117-01-AA).

To construct the knockout plasmids for replicon repT and the 16 protease loci based on the endogenous CRISPR system type I-B, a spacer that matched the target loci was cloned within two BbsI restriction sites on the pRKP31-AC3 shuttle vector by T4 polynucleotide (BioLabs, M0201S). A spacer fragment was generated by annealing two complementary single-stranded oligo DNA 42 bp in length with 4nt at the 5`-end complementary to the BbsI sticky ends, at 95 °C for 5 min, and slowly cooled down at room temperature to form the double-stranded inserts. The two arms of the homology direct repair (HDR) template were separately amplified using the HB27 genome as template and connected using splicing by 20 bp overhang extension PCR (SOE PCR). Plasmids were constructed in two steps. First, spacers that matched the double-strand break (DSB) site were cloned. Then, the plasmid was linearized by PCR, followed by DpnI (Thermo Scientific, FD1704) enzyme digestion, and the HDR was cloned by 20 bp overhang homology utilizing ABclonal MultiF Seamless Assembly Mix (ABclonal, RK21020) [23]. Fig. S1 illustrates the schematic procedure of endogenous CRISPR-Cas-based gene editing technology.

Plasmids were validated by sequencing (Sangon, Shanghai, China) before being transformed into the HB27 cells. Plasmids used or constructed are listed in Table S3. Primers used for PCR amplification and validation are listed in Table S2. Primers and single-stranded oligo DNA were synthesized by Sangon Co., Shanghai, China. PCR was performed using the Phanta Max Super‐Fidelity DNA polymerase (Vazyme, P505). The amplification was performed at 95 °C for 5 min, 95 °C for 15 s, Tm-5 °C for 15 s, 72 °C for 1 min/1 kb, and 72 °C for 5 min, for 30 cycles. The amplicons and digested fragments were purified with the E.Z.N.A. Cycle-Pure kit (Omega BIO-TEK, D6492- 02). Plasmid extracted utilizing E.Z.N.A. Plasmid Mini Kit I (Omega BIO-TEK, D6943- 02).

Transformation, cell morphology, and growth rate Estimation

The heat shock transformation method was used for DH5α competent cells. 50 µL DH5α was thawed on ice for 5 min, 10 µL recombinant plasmid was gently mixed, and placed on ice for 25 min. Heat shock at 42℃ for 45 s, followed by ice incubation for 2 min, then 900 µL LB medium was pipetted and incubated at 37 °C, 200 rpm for 1 h, and finally spread on LB selection medium [23].

T. thermophilus utilizes a natural competence system to take up exogenous DNA during the mid-log phase. 1 µg plasmid was mixed with 500 µL mid-log phase cells, incubated for 2.5 h at 65℃, 180 rpm, and then spread on a Tt selection medium [23].

The pRKP31 and pRKP31-AC3 Shuttle plasmids were used as backbones for construction with a kanamycin-selectable marker, under selection pressure of 30 µg/mL and 20 µg/mL for E. coli and T. thermophilus, respectively.

For microscopic cell morphology ofT. thermophilus, a fuchsin-dyed smear was prepared and observed using the immersion oil 100X objective lens [40].

For the growth curve, bacteria were acquired by measuring the optical turbidity at an absorbance of 600 nm. Since the optical density of the culture is proportional to the cell density, measuring the turbidity of the culture can be used to estimate the number of bacterial cells. Overnight cultures were diluted 1:100 v/v into 50 mL Tt medium in a shaker incubator at 65℃, 180 r/min. Samples were taken every 2 to 4 h intervals for 24 to 48 h. The strains were cultured in three biological repeats [41].

Quantitative and qualitative β-galactosidase assay

Quantitative assessment of β-galactosidase activity was performed based on the established O-Nitrophenyl-β-D-galactopyranoside (ONPG) assay as described [42, 43]. Reporter plasmids were introduced into the T. thermophilus HB27 ΔTTP0042. Three distinct single colonies harboring the respective reporter plasmid were randomly selected for each independent experimental replicate. These transformed strains were cultivated in Tt medium under the optimum conditions. Upon reaching the log growth phase, as determined by an optical density at (OD₆₀₀ ~ 1.0), bacterial cells were harvested via centrifugation. The resulting cellular pellets were resuspended in 10 mM Tris-HCl buffer (pH 8.0). To obtain crude cellular lysates, the resuspended cells underwent controlled sonication. Following lysis, cellular debris was eliminated through centrifugation at 13,000 × g for 30 min at 20 °C to obtain clarified supernatants. The protein concentration within these clarified cellular extracts was precisely quantified using the Bradford reagent (Tiangen, PA102). A 50 µL aliquot of the clarified supernatant was combined with 450 µL of reaction buffer (2.8 mM ONPG in 50 mM sodium phosphate, pH 6.5). Samples were incubated at 65 °C for a precisely controlled duration of 20 min, after which the enzymatic reaction was terminated by adding an equal volume of 1 M Na₂CO₃. The absorbance of the o-nitrophenol was spectroscopically estimated at 420 nm using a microplate reader (Victor Nivo 3 S; PerkinElmer). One unit of specific β-galactosidase activity is operationally defined as 1 nmol ρ-nitrophenol produced per min per mg total protein.

In the qualitative plate assays, 5 µL of mid-log phase (OD₆₀₀ ~ 1.0) screened strains were spotted on Tt plates supplemented with 100 mg/L X-gal. Plates were incubated at 65℃ for 2 days to observe the color change [44].

Extracellular protease assay

The protease mutations were screened for extracellular protease activity by skim milk plate assay [45]. 2.5 µL of mid-log phase (OD₆₀₀ ~ 1.0) mutants were spotted on Tt plates supplemented with 10% (v/v) skimmed milk. The clear zone diameter was measured 3 days after incubation at 65 °C by flooding the plates with 10% trichloroacetic acid (TCA). The relative enzyme activity was calculated using the following formula:

Where: REA is relative enzyme activity, CZD is clear zone diameter, and CD is colony diameter.

Statistical analysis

All histogram drawings and statistical analysis were completed using the GraphPad Prism 8 software (GraphPad Software Inc.). Data are expressed as the mean ± standard deviation (SD). The significance of differences between groups’ mean values was analyzed using Duncan’s Multiple Range Test (DMRT) at a P < 0.05 significance level.

Results

Construction of promoter screening reporter vector

T. thermophilus HB27 β-galactosidase is involved in lactose metabolism and acts as a prevalent reporter enzyme used in studying the activities of promoters [32, 34]. T. thermophilus HB27 β-galactosidase transcript by three loci (TTP0042, TTP0220, TTP0222) located in the pTT27 plasmid [19]. The TTP0042 locus was detected as the main source of β-galactosidase activity in T. thermophilus HB27. Thus, the β-galactosidase-deletion strain (HB27ΔTTP0042) was used as a competent cell for promoter screening. Thirteen promoter regions were inserted into pRKP-TTP0042 by homology ligation. Promoters named as follows: Pslp (206 bp), Pago (300 bp), Pnqo (414 bp), P31 (216 bp), P031 (214 bp), P43 (195 bp), P214 (160 bp), P215 (184 bp), P0416 (853 bp), P0984 (858 bp), P1578 (842 bp), P1706 (849 bp), P1798 (861 bp). The pRKP-TTP0042 screening vector contains the promoter-less endogenous TTP0042 β-galactosidase, heat–stable Kanamycin resistance selectable marker derived with Pslp promoter, E. coli replicon puc-ori, and T. thermophilus replicon repA ori (Fig. 1A). Transformants grown up on Tt selection plates were picked up for verification using specific primers (Fig. S2 and Table S2).

Fig. 1.

Construction, Quantitative, and Qualitative β-galactosidase for screened promoters. (A) Map of reporter gene expression vectors shows the integration space of screened promoters “P”, to drive “TTP0042- purple color” β-galactosidase reporter gene as promoter expression indicator. The red color represents “KanR” kanamycin selectable marker gene with Pslp-promoter-driven and gray color represents “repA ori, puc-ori” replicon region of T. thermophilus HB27 and E. coli, respectively. (B) Colony visual validation the screened promoter constructions, wild type HB27 strain, and the β-galactosidase defect strain as a control (HB27ΔTTP0042); (C) Comparison of β-glucosidase activities of the monitored promoters constructions based on o-nitrophenyl β-D-galactoside (ONPG) method. One unit of specific β-galactosidase activity is operationally defined as 1 nmol ρ-nitrophenol produced per min per mg total protein. Bars represent ± SD of three replicates. Different letters indicate significant differences according to Duncan’s multiple range tests at P < 0.05

Effect of different promoters on β-galactosidase expression

To analyze the potential transcription level of the 13 promoters. The HB27ΔTTP0042 transformants were examined qualitatively and quantitatively by X-Gal plates and ONPG assay (Fig. 1B, C). Transformants, Control (HB27ΔTTP0042), and the HB27 (wild-type) were grown in Tt medium under optimal growth conditions to mid-log phase, and then the β-glucosidase activities were examined. The Pago, P031, P43, P0416, and P1578 promoters have a significantly lower activity as appears in yellow colonies on the X-Gal plates (Fig. 1B). Interestingly, the quality and quantity assay show that Pslp, Pnqo, P31, P214, P215, P0984, P1706, and P1798 promoters were significantly higher than the wild-type and control as shows in Fig. 1B, C. Among the 13 promoters, the P0984 promoter shows significantly 13-fold higher activity, which recommends the utilization of the P0984 promoter for the subsequent thermostable protein production.

Construction of HB27 plasmid-free strain

To step forward chassis cell construction and minimize the T. thermophilus HB27 genome. We deleted the pTT27 plasmid (~ 270 kb) [19]. We utilized the CRISPR-deficient strain HB27ΔIII-ABΔI-CΔCRF3, previously constructed by our lab as a competent cell due to its enhanced transformability, lack CRISPR type III subtype A&B, type I subtype C, and CRISPR-associated Rossmannfold (64 kbp deletion) as mapped in Fig. 2A. The pTT27 plasmid replication origin (repT) was identified, 42nts were selected after PAM sequence (5′-TTC-3′) in the middle of the repT locus as a specific region for DSB, and an approximately 500 bp region flanking repT locus were cloned in the pRKP31-AC3 vector for HDR (Fig S1). Colonies grown up in the Tt selection medium were screened by colony PCR with three pairs of primers flanking the target deletion region (Table S2, Fig. S3A). PCR results confirmed the DSB sandwiched by the HDR; control strain was 3319 bp, 2214 bp, and 3322 bp, while those of the plasmid-free strain showed no bands, suggesting a successful deletion was established (Fig. S3B). As pTT27 encodes key carotenoid biosynthesis genes (crtI, crtB) [19], the resulting HB27ΔpTT27 colonies were white, in contrast to the wild-type strain (Fig. S3C).

Fig. 2.

Construction and Characterization of HB27 giant plasmid (pTT27)-free strain. (A) Schematic diagram of the HB27ΔIII-ABΔI-CΔCRF3 and giant plasmid deletion strains construction; (B) Growth rate of strains grown up in Tt medium; (C) Comparison of transformation efficiency of the strains, the efficiency was calculated based on Colony-Forming-Unit per microgram of introduced plasmid DNA. (D) Cell morphology of the strains visualized with an oil immersion objective (1000X). Bars represent ± SD of three replicates

Characterization of HB27 plasmid-free strain

To evaluate the plasmid-free strain (HB27ΔpTT27) as a potential chassis, its phenotypic properties (growth rate, morphology, transformation efficiency) were compared with the control (HB27ΔIII-ABΔI-CΔCRF3) and the wild-type HB27.

Growth rate: As shown in Fig. 2B, the growth of HB27ΔIII-ABΔI-CΔCRF3 was comparable to wild-type HB27, reaching stationary phase (OD₆₀₀ ~ 5) around 30 h. However, HB27ΔpTT27 failed to grow significantly in the standard Tt medium. Since pTT27 harbors the cobalamin (Vitamin B12) biosynthesis gene cluster (TTP0001-TTP0023) [19, 46]. We hypothesized that this deletion caused auxotrophy. Supplementation with 0.1 µg/mL AdoCbl rescued growth, although the growth rate remained slightly slower, and the final OD was lower than wild-type and competent strains. This confirms that pTT27 encodes essential genes for cobalamin synthesis and potentially other factors affecting optimal growth under these conditions. This auxotrophy could potentially be applied to antibiotic-free selection systems in future applications [47].

Transformation efficiency: We assessed natural transformation efficiency using the pRKP0984-TTP0042 plasmid. Wild-type HB27 exhibited a baseline efficiency of approximately 2.02 × 10⁴ CFU/µg. The plasmid-free HB27ΔpTT27 strain showed a similar efficiency (2.13 × 10⁴ CFU/µg), indicating that removal of the pTT27 plasmid did not significantly impair DNA uptake under these conditions. In contrast, the competent CRISPR-deficient strain HB27ΔIII-ABΔI-CΔCRF3 displayed significantly enhanced efficiency (1.78 × 10⁶ CFU/µg), nearly a 100-fold increase over the wild-type (Fig. 2C). This pronounced increase is likely due to the removal of CRISPR-Cas systems that target foreign DNA. Such enhancement is consistent with observations in other bacteria, where inactivation of defense systems like restriction-modification (R-M), StySA, or BREX systems improves transformation efficiencies [48–52]. This high transformability proved advantageous for the subsequent protease gene deletions.

Cell morphology: Microscopic examination revealed that HB27ΔIII-ABΔI-CΔCRF3 morphology was indistinguishable from wild-type bacilli. However, the plasmid-free HB27ΔpTT27 strain consistently displayed a slightly elongated cell shape compared to the wild-type (Fig. 2D). This subtle morphological change might be linked to the altered growth kinetics or potential cell envelope modifications resulting from the large deletion [53], warranting further investigation.

Construction of protease-encoding loci knockout strains

Bioinformatic analysis identified 16 predicted nonessential protease-encoding genes within the T. thermophilus HB27 genome distributed across the chromosomal loci as listed in Table 1. We knocked out these 16 loci to optimize the highly transformable HB27ΔIII-ABΔI-CΔCRF3 strain. To ensure precise locus deletion without cluster gene interruption, we targeted the protein-coding sequences while preserving promoter elements and regulatory regions (Fig. 3A). Utilizing the pRKP31-AC3 plasmid-mediated HDR recombination system based on the endogenous CRISPR-Cas9 Type I-B machinery. We successfully constructed 16 individual protease knockout strains. In vivo transcript of the designed crRNA generates ribonucleoprotein (RNP) complexes. These RNPs effectively cleaved the protease locus, with precise repair mediated by homologous recombination utilizing an HDR template, thus introducing the designed null mutations. As illustrated in Fig. 3A, B, Table S2, the PCR-based validation strategy confirmed the homogeneity of the mutant population. Validated strains were then subjected to iterative rounds of subculturing in an antibiotic-free Tt medium to facilitate plasmid loss, yielding marker-free protease deletion strains for future editing and ensuring a complete gene editing mutant. As illustrated in Fig. 3A ‘exam-genome-F/R’ primer pair confirmed the target deletion by amplifying across the modified region, yielding amplicons of reduced size depending on the target locus compared to wild-type allele (Fig. 3B), ‘exam-sp-F/exam-genome-R’ primer combination detect the presence of any residual wild-type alleles, ‘exam-plasmid-F/R’ primer confirm the subsequent loss of the editing plasmid. This multi-faceted screening approach generates 16 individual protease deletion loci, designated DP1 through DP16, each carrying a precise deletion in the intended protease-encoding locus.

Table 1.

Selected proteases identified in the genome sequence of T. thermophilus HB27

Locus	Annotation	General Function
TTC0035	ATP-dependent zinc metalloprotease FtsH, EC:3.4.24.- (DP1)	Acts as a processive, ATP-dependent zinc metallopeptidase for both cytoplasmic and membrane proteins. Plays a role in the quality control of integral membrane proteins.
TTC0174	ATP-dependent clp protease ATP-binding subunit clpA / Negative regulator of genetic competence Clpc/MecB (DP2)	Chaperone and an integral component of the ATP-dependent ClpAP protease, participates in regulating protein degradation and the dissolution and degradation of protein aggregates.
TTC0251	ATP-dependent Clp protease ATP-binding subunit ClpX (DP3)	ATP-dependent specificity component of the Clp protease. It directs the protease to specific substrates. Can perform chaperone functions in the absence of ClpP.
TTC0264	ATP-dependent hsl protease ATP-binding subunit hslU (DP4)	The HslU subunit of the HslU-HslV complex functions as an ATP-dependent ‘unfoldase’. The binding of ATP and its subsequent hydrolysis by HslU are essential for the unfolding of protein substrates, subsequently hydrolysed by HslV [54]. HslU recognizes the N-terminal part of its protein substrates and unfolds these before they are guided to HslV for hydrolysis [55]. In Peptidase Family T1.
TTC0265	ATP-dependent protease hslV, EC:3.4.25.- (DP5)
TTC0417	S1C family serine proteases (DP6)	Catalysis of the hydrolysis of internal, alpha-peptide bonds in a polypeptide chain by a catalytic mechanism that involves a catalytic triad consisting of a serine nucleophile that is activated by a proton relay involving an acidic residue (e.g., aspartate or glutamate) and a basic residue (usually histidine).
TTC0481	Membrane metalloprotease (DP7)	Catalysis of the hydrolysis of peptide bonds by a mechanism in which water acts as a nucleophile, one or two metal ions hold the water molecule in place, and charged amino acid side chains are ligands for the metal ions.
TTC0492	CPBP family of intramembrane metalloproteinases, Abortive infection protein (DP8)	Catalysis of the hydrolysis of internal, alpha-peptide bonds in a polypeptide chain by a mechanism in which water acts as a nucleophile, one or two metal ions hold the water molecule in place, and charged amino acid side chains are ligands for the metal ions.
TTC0663	M66 family metalloproteinases (DP9)	This family of metallopeptidases contains StcE, a virulence factor found in Shiga-toxigenic Escherichia coli organisms. StcE peptidase cleaves C1 esterase inhibitor [56].
TTC0687	ClpP/crotonase-like domain superfamily (DP10)	ClpP is an ATP-dependent protease that cleaves several proteins, such as casein and albumin. It exists as a heterodimer of ATP-binding regulatory A and catalytic P subunits, both of which are required for effective levels of protease activity in the presence of ATP [57].
TTC0950	M50B family Zinc metalloprotease (DP11)	Catalysis of the hydrolysis of peptide bonds by a mechanism in which water acts as a nucleophile, one or two metal ions hold the water molecule in place, and charged amino acid side chains are ligands for the metal ions.
TTC0974	PrcB C-terminal domain-containing protein (DP12)	This domain is found at the C terminus of Treponema denticola PrcB. PrcB interacts with the PrtP protease (dentilisin) and is required for the stability of the protease complex [58].
TTC1110	Rhomboid family intramembrane serine protease (DP13)	Catalysis of the hydrolysis of internal, alpha-peptide bonds in a polypeptide chain by a catalytic mechanism that involves a catalytic triad consisting of a serine nucleophile that is activated by a proton relay involving an acidic residue (e.g., aspartate or glutamate) and a basic residue (usually histidine).
TTC1111	Periplasmic serine protease (DP14)	Catalysis of the hydrolysis of a peptide bond. A peptide bond is a covalent bond formed when the carbon atom from the carboxyl group of one amino acid shares electrons with the nitrogen atom from the amino group of a second amino acid.
TTC1128	ATP-dependent zinc metalloprotease FtsH, EC:3.4.24.- (DP15)	Homologous to TTC0035.
TTC1905	S1C family serine protease (DP16)	Homologous to TTC0417.

Fig. 3.

Construction and Characterization of T. thermophilus HB27 strains; (A) The exam-genome-F/R primer pair confirmed the successful deletion of the target gene by amplifying across the modified region, yielding amplicons of reduced size compared to the wild-type allele. The exam-sp-F/exam-genome-R primer combination was designed to detect the presence of any residual wild-type alleles (HB27ΔIII-ABΔI-CΔCRF3), confirming the homogeneity of the mutant population, the exam-plasmid-F/R primer pair was used to confirm the subsequent loss of the editing plasmid; (B) Colony PCR amplicons of protease gene deletion verification image (M: 5 kb DNA Marker (5000, 3000, 2000, 1500, 1000, 750, 500, 250, 100); CK: HB27ΔIII-ABΔI-CΔCRF3, and DP1-16: protease deletion strains); (C-E) Growth rate of strains grown up in Tt medium; (F) Extracellular protease activity based on skimmed milk plate assay; (G) β-galactosidase expression level monitored based on o-nitrophenyl β-D-galactosidase (ONPG) method as indicator for protease activity impacts. One unit of specific β-galactosidase activity is operationally defined as 1 nmol ρ-nitrophenol produced per min per mg total protein. Bars represent ± SD of three replicates. Different letters indicate significant differences according to Duncan’s multiple range tests at P < 0.05

Characterization of protease-encoding loci knockout strains

Growth rate: The 16 individual protease deletion strains (DP1-DP16) were monitored in a liquid Tt medium, revealing distinct growth phenotypes among strains. DP6-DP11, exhibited growth profiles indistinguishable from the wild-type strain, reaching stationary phase by 30 h of incubation with a comparable OD₆₀₀ ~ 5 (Fig. 3D). DP1-DP4, DP12-DP16 displayed subtle variations in growth rates relative to the wild-type, while reached the stationary phase by 30 h, with slightly lower OD than the wild-type (Fig. 3C, E). DP5 (TTC0265 deletion, putative FtsH) manifested a significantly attenuated growth rate with extended incubation period of 40 h to reach the stationary phase (Fig. 3C). Notably, DP1 (TTC0035 deletion), DP5 (TTC0265 deletion), and DP16 (TTC1905 deletion) consistently demonstrated a reduction in maximal biomass accumulation.

Extracellular proteolytic activity was assessed using Tt skim milk plates incubated for 72 h at 65 °C (Fig. 3F) to compare the resulting hydrolysis zones of the strains (Table 2). Notably, DP2, DP3, DP9, DP11-DP15, exhibited hydrolytic zones with no statistically significant difference compared to HB27 and HB27ΔIII-ABΔI-CΔCRF3, indicating unimpaired extracellular protease function. In contrast, DP1, DP5-DP8, and DP10 generated significantly smaller zones of hydrolysis, suggesting a reduction in secreted protease activity. Remarkably, DP4 (TTC0264 deletion, putative ClpY/HslU) and DP16 (TTC1905 deletion, putative HhoB) displayed severely attenuated hydrolytic activity (1.61 ± 0.58 mm in diameter). These observations strongly implicate the proteases encoded by TTC0264 and TTC1905 as key contributors to overall secreted proteolytic activity under these conditions.

Table 2.

Relative extracellular protease activity for T. thermophilus HB27 and deletion strains

Strain	Relative Enzyme Activity	Strain	Relative Enzyme Activity
HB27	2.57 ± 0.50a	HB27ΔIII-ABΔI-CΔCRF3	2.5 ± 0.87ab
DP1	2.19 ± 0.29de	DP2	2.5 ± 0.50ab
DP3	2.52 ± 0.29a	DP4	1.64 ± 0.50f
DP5	1.95 ± 1.26e	DP6	2.26 ± 1.15cd
DP7	2.23 ± 0.76cd	DP8	2.23 ± 0.29cd
DP9	2.47 ± 0.76ab	DP10	2.14 ± 0.50de
DP11	2.40 ± 0.58abc	DP12	2.43 ± 1.32abc
DP13	2.31 ± 0.58bcd	DP14	2.57 ± 0.50a
DP15	2.52 ± 0.29a	DP16	1.61 ± 0.58f

Bars represent ± SD of three replicates. Different letters indicate significant differences according to Duncan’s multiple range tests at P < 0.05

Reporter protein accumulation assesses the impact of protease deletions on heterologous protein expression capabilities, using the pRKP0984-TTP0042 reporter plasmid. DP1, DP2, DP3, DP5, and DP8 exhibited levels statistically indistinguishable from the competent strain HB27ΔIII-ABΔI-CΔCRF3 (Fig. 3G). Conversely, the remaining protease deletion strains demonstrated a statistically significant reduction in β-galactosidase activity.

Based on integrated analysis (growth, protease activity, reporter expression), a subset of 10 loci (TTC0174, TTC0251, TTC0264, TTC0417, TTC0481, TTC0492, TTC0663, TTC0687, TTC0950, TTC0974) were selected for multiple knockout construction, starting from DP6 (ΔTTC0417) which showed wild-type-like growth and slightly reduced reporter expression.

Construction of multiple protease-encoding loci knockout strains

To develop T. thermophilus HB27 chassis cells with enhanced heterologous protein expression capabilities, we generated a series of the above-mentioned knockout loci using the DP6 strain (HB27ΔIII-ABΔI-CΔCRF3ΔTTC0417) as a competent cell with subsequent rounds of gene editing. We constructed nine multiple protease loci knockout strains, named DSP1-DSP9 (Fig. S4), and confirmed colony PCR (Fig. S5).

The successful construction of these iterative knockout strains (DSP1-DSP9) represents a step toward optimizing T. thermophilus HB27. However, knockout efficiency progressively decreased with each successive deletion, ultimately limiting further gene editing using the I-B system, and underscores the need for alternative strategies to further refine this chassis.

Characterization of multiple protease-encoding loci knockout strains

Growth rate: All multiple-deletion strains (DSP1-DSP9) exhibited growth profiles generally comparable to the wild-type strain, reaching log phase around 24 h, though final biomass was slightly reduced (OD₆₀₀ 3.57–3.88 vs. ~ 4.2) (Fig. 4A). DSP9 showed the least growth impairment among the multiple-deletion strains.

Fig. 4.

Characterization of T. thermophilus HB27 accumulated protease-encoding gene defect strains (WT HB27, Ctrl.: HB27ΔIII-ABΔI-CΔCRF3 and DSP1-DSP9). (A) Growth rate of accumulated protease defect strains was grown up in Tt medium. (B) β-galactosidase expression level monitored based on o-nitrophenyl β-D-galactosidase (ONPG) method as indicator for protease activity impacts. One unit of specific β-galactosidase activity is operationally defined as 1 nmol ρ-nitrophenol produced per min per mg total protein. Bars represent ± SD of three replicates. Different letters indicate significant differences according to Duncan’s multiple range tests at P < 0.05

Extracellular protease activity: DSP1-DSP3 showed markedly smaller hydrolysis zones compared to the parental strains, consistent with the accumulation of deletions including TTC0264 (putative ClpY/HslU). Other strains (DSP4-DSP6, DSP8-DSP9) showed activity closer to the parental strain, suggesting compensatory effects or that the deleted proteases in these combinations contribute less to casein hydrolysis under these conditions (Table 3).

Table 3.

Relative extracellular protease activity for T. thermophilus HB27 and multiple deletion strains

Strain	Relative Enzyme Activity	Strain	Relative Enzyme Activity
HB27	3.83 ± 0.76bcd	HB27ΔIII-ABΔI-CΔCRF3	4.83 ± 1.04b
DSP1	1.00 ± 0.50e	DSP2	1.50 ± 0.50e
DSP3	1.17 ± 0.76e	DSP4	4.33 ± 0.76bc
DSP5	3.17 ± 0.76cd	DSP6	4.33 ± 0.29bc
DSP7	6.83 ± 0.29a	DSP8	3.00 ± 0.50d
DSP9	4.83 ± 0.58b

Bars represent ± SD of three replicates. Different letters indicate significant differences according to Duncan’s multiple range tests at P < 0.05

Reporter gene expression: β-galactosidase activity varied among the multiple deletion strains (Fig. 4B). DSP1-DSP2 and DSP4-DSP5 showed activity similar to the parental strains. DSP3 showed reduced activity. Significantly, strains DSP6 through DSP9 exhibited enhanced β-galactosidase activity compared to the starting strain, with DSP6 showing the highest (11,480 U/mg) and DSP9 also showing strong activity (10,844 U/mg). This suggests that the combined deletion of these specific proteases reduces degradation or turnover of the intracellular β-galactosidase reporter.

Based on this characterization, DSP9 emerged as a promising first-generation chassis candidate. Its genome size is reduced compared to wild-type (~ 3.5% deletion from chromosome relative to WT), it maintains robust growth kinetics, and shows enhanced reporter protein accumulation (approximately twofold higher activity than the wild-type strain under these conditions).

Discussion

This study aimed to develop improved chassis strain and genetic tools for T. thermophilus HB27, focusing on enhancing its utility for heterologous protein production, particularly thermostable proteins. We successfully identified a strong native constitutive promoter (P0984), generated and characterized a plasmid-free strain (HB27ΔpTT27), demonstrated significantly improved transformation efficiency by deleting specific CRISPR-Cas loci, and systematically evaluated the impact of single and multiple protease loci deletions.

The identification of strong constitutive promoters as Pslp, Pnqo, P31, P214, P215, P0984, P1706, and P1798 expands the limited promoter toolbox for T. thermophilus, providing valuable elements for achieving high-level gene expression without complex induction strategies, contrasting with previously characterized inducible or regulated promoters [31–34].

The construction of the plasmid-free HB27ΔpTT27 strain represents a significant genome reduction (~ 11% of total genome size). While this strain exhibited cobalamin auxotrophy and slightly impaired growth, necessitating supplementation, it offers potential advantages. These include potentially increased genetic stability in further recombinant thermostable protein production. Its comparable transformation efficiency suggests basic competence machinery is unaffected. The auxotrophy presents an opportunity for developing antibiotic-free selection systems, a desirable trait for industrial applications [47], which is currently under investigation. However, the limitations (auxotrophy, slower growth) must be considered, as adding a specific supplement quantity is incomparable to a naturally occurring genetic system.

The dramatic (~ 100-fold) increase in transformation efficiency observed in the CRISPR-deficient strain (HB27ΔIII-ABΔI-CΔCRF3) is a key finding. This highlights the significant barrier that native CRISPR-Cas systems pose to the introduction of foreign DNA in this strain. This result aligns with findings in other bacteria where disabling defense systems (CRISPR-Cas, R-M, BREX) enhances genetic accessibility [48–52]. This improved transformability was instrumental in enabling the subsequent efficient construction of multiple protease knockout strains.

Our systematic protease knockout strategy provided insights into their roles in HB27. Deletion of TTC0264 (putative ClpY/HslU) and TTC1905 (putative HhoB) markedly reduced extracellular caseinolytic activity, identifying them as major secreted proteases under these conditions. The varied effects of single deletions on growth (e.g., significant impairment in ΔTTC0265/DP5) and reporter expression highlight the complex roles and potential redundancy of proteases in cellular physiology [59].

A notable observation was the enhanced accumulation of the intracellular β-galactosidase reporter in strains DSP6-DSP9, which harbor multiple protease deletions, including TTC0264 (putative ClpY/HslU). This occurred despite TTC0264 being implicated primarily in extracellular activity based on the skim milk plate assay. TTC0264 and TTC1905 are predicted extracellular proteases; their knockout indirectly stabilizes intracellular proteins by reducing stress-induced misfolding, which involves significant metabolic reorganization [60–62]. Thus, we emphasize that extracellular protease knockouts can indirectly stabilize intracellular proteins by reducing stress responses that lead to misfolding. As well as the deletion of major proteases might trigger complex regulatory responses (e.g., stress responses, altered expression of other intracellular proteases) that indirectly affect β-galactosidase stability or synthesis [63, 64]. While the exact mechanism requires further study, the empirical result of increased reporter accumulation in these multiple knockout strains (DSP6-DSP9) supports their potential benefit for improving yields of certain intracellular proteins.

The deletions performed target homologs of key protease systems known to be involved in protein quality control (e.g., Clp system components like ClpY/HslU, FtsH-like proteases, HhoB serine protease). In other organisms, these systems degrade misfolded or damaged proteins, regulate protein turnover, and are crucial for stress responses [60, 61, 65–71]. Our results, such as the growth defect in DP5 (ΔTTC0265/FtsH-like) and the enhanced reporter stability in DSP6-DSP9 (lacking TTC0264/ClpY/HslU among others), are broadly consistent with these general roles. However, this study focused on the phenotypic outcomes (growth, extracellular activity, reporter expression) rather than a detailed mechanistic analysis of protein quality control pathways. The observed effects likely stem from altering the balance of protein synthesis, folding, and degradation, but the specific substrates and regulatory consequences within T. thermophilus require further investigation.

Limitations and Future Directions: This study provides foundational tools and strains. A key limitation, as highlighted by the reviewers, is the reliance on β-galactosidase as the primary reporter. While useful for comparative purposes within this study (e.g., promoter strength, effect of protease knockouts), demonstrating the practical utility of these strains requires testing the expression of industrially relevant target proteins, particularly thermostable enzymes [5]. Direct quantitative comparisons of protein yields between these optimized T. thermophilus strains and established mesophilic hosts (E. coli, Pichia sp) were beyond the scope of this initial work but represent an important future direction, especially for targets where Thermus offers intrinsic advantages. Furthermore, the observed drawbacks of the plasmid-free strain (growth defect, auxotrophy) need to be addressed or accounted for in specific applications. Evaluating the performance of the engineered strains (especially DSP9) under realistic fermentation conditions is crucial. Future work could also involve proteomic analysis to confirm protease depletion and investigate global cellular responses, as well as exploring combinatorial effects of the developed tools (e.g., using promoter P0984 in strain DSP9).

Conclusions

This study successfully generated and characterized valuable resources for engineering T. thermophilus HB27 as a host for heterologous protein production. Key contributions include the identification of the strong constitutive promoter P0984, the construction and characterization of a plasmid-free strain (HB27ΔpTT27) revealing its auxotrophic nature and potential for genome streamlining, the demonstration that deleting specific CRISPR-Cas loci dramatically enhances transformation efficiency, and the development of protease deletion strains (DP series and multi-deletion DSP series). Strain DSP9, harboring 10 targeted protease gene deletions built upon the CRISPR-deficient background, emerged as a promising first-generation chassis, exhibiting robust growth and enhanced intracellular accumulation of reporter protein. While this work establishes a strong foundation, further validation using industrially relevant thermostable proteins and optimization under bioprocess conditions are necessary next steps.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Y.Q.L. and M.M.: Conceptualization, Conducted experiment, Data curation, Formal analysis, Visualization, Investigation, Writing– original draft, review & editing; J.W.W. and X.Y.B. & Y.S.: Resources, Methodology; Y.J.L.: Conceptualization, Supervision, Proofreading, Funding, Project administration, Resources.

Funding

This study was supported by the National Key Research and Development Program of China (2022YFA0912200), National Natural Science Foundation of China (32170096), and Key Research and Development Project of Hubei Province (2023BBB025).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.