ROP-TB was derived from ESAT-6 and CFP-10. It was designed to include a sequence of 8 peptides consisting of 35 amino acids per peptide. Each peptide overlapped with its adjacent peptides by 10 amino acids. The eight sequential overlapping peptides covered the complete sequences of ESAT-6 and CFP-10. The cathepsin S substrate sequence (LRMK) was interspersed between each peptide. Cathepsin S is known to exist in professional APCs (10). A schematic presentation of ROP-TB is shown in Figure 1A . ROP-TB was expressed by E. coli BL21 cells and then purified by Ni NTA affinity chromatography. ( Figure 1B ). The purity of the protein was confirmed by HPLC to be more than 95% ( Figure 1C ).

To remove any potential endotoxins that may have potentially contaminated the product, ROP-TB was further purified by endotoxin removal chromatography. Subsequent endotoxin levels in ROP-TB before application in cell cultures was tested and were <1 EU/mL. To test the effects of endotoxin on the ROP-TB-based IGRA assay, an enzyme-linked immunosorbent spot (ELISPOT) assay study using ROP-TB spiked with endotoxin was carried out. The doses of endotoxin used (up to 33 EU/mL) did not influence the results of the ROP-TB-related ELISPOT ( Supplementary Figure 1 ). This is consistent with another report showing that only more than 100 EU/mL of endotoxin induces an ELISPOT response (12).

To test the enzymatic effects of cathepsin S on the protein, ROP-TB was incubated in vitro with cathepsin S. After digestion, samples were analyzed using SDS-PAGE. As Figure 1D shows, ROP-TB can be digested in vitro into peptides of different sizes by cathepsin S ( Figure 1D ).

To understand the cross-presentation efficiency of ROP-TB in APCs, it is important to investigate its metabolic rate in the endolysosome. We, therefore, compared ROP-TB with TB peptides and TB wild-type protein (fusion of ESAT-6 and CFP-10) by tracing their intracellular localization at different time points in the mouse dendritic cell line DC2.4 cells. ROP-TB, TB peptides, and TB protein were labeled with a green fluorescent dye (WV=530/30nm) and were incubated with DC2.4 cells for 30 min and 6 h. DC2.4 cells were co-incubated with Lysotracker to emit red fluorescence (WV=630/30nm). The uptake of ROP-TB, TB peptides, or TB protein is shown in green, and endolysosomes are shown in red under a confocal microscope. The co-localization between red and green colors represents the approximate location of the two objects. As shown in Figures 2A–C , ROP-TB, TB peptides, and TB fusion protein are internalized by DC2.4 within 30 min of incubation and are largely co-localized within endolysosomes. This indicates that these antigens are located within the endolysosomal compartment. After 6 hours, the green fluorescence in Figures 2D, E (ROP-TB and TB peptides) but not in Figure 2F (TB fusion protein), disappeared ( Figures 2D–F ), suggesting that ROP-TB and TB peptides have been processed more rapidly ( Figure 2 ).

To test whether ROP-TB can be developed as a diagnostic agent for T cell function, we assessed specific T cell responses induced by synthetic ROP-TB or Mtb peptides ex vivo in a mouse model. We first established the test model by infecting BALB/c mice with Mtb (H37R1) intranasally ( Figure 3 ). Infectivity was validated by acid-fast staining and Hemotoxin-Eosin (H&E) staining of lung tissue ( Figure 3A upper panel). The lungs of infected mice clearly showed infection with H37R1 and presented lesions when compared to non-infected mice ( Figure 3A , lower panel).

Spleen cells from Mtb-infected mice were then isolated and stimulated with different concentrations of synthetic ROP-TB and Mtb peptides. T cell responses to these antigens were measured by the IGRA assay (ELISPOT). As shown in Figure 3B , ROP-TB stimulates stronger specific T cell responses than synthetic peptides, indicating that ROP-TB is a more sensitive antigen in the mouse model than synthetic peptides in detecting T cell responses.

Delayed-type hypersensitivity (DTH) reflects T cell responses to antigen stimulation in vivo. To investigate whether ROP-TB can be processed and presented in vivo to specific T cells, guinea pigs were infected with Mtb H37R1, and DTH-related skin reactions to ROP-TB, ESAT-6/CFP-10, and PPD (purified protein derivative) were investigated. Both ROP-TB and ESAT-6/CFP-10 trigger positive responses to DTH, as well as PPD ( Figure 4A ). DTH elicited by ROP-TB was stronger than that elicited by ESAT-6/CFP-10 and PPD in 100 μg and 10 μg doses (p<0.01) ( Figure 4A ).

Bovine TB (M. Bovis) epidemics have been reported in many parts of the world. This has posed a risk for humans drinking infected, unpasteurized milk. Practically all cattle on a farm need to be slaughtered even if only one head of cattle is found to be infected with M. bovis. This causes great economic loss. Thus, as a preventative measure, a tuberculin-based skin test using a PPD is usually performed to screen for infected cattle (13). BCG may also be used as a vaccine for M. bovis infection (14) if there is another available diagnostic reagent which will not be interfered by the vaccination.

Since M. bovis and Mtb share the same ESAT-6 and CFP-10 sequences (15), we wondered whether our ROP-TB reagent could be used as another choice of M. bovis skin test reagent. Thus, the application ROP-TB as a diagnostic reagent for M. bovis infection was tested in cattle from two farms (Farm 1, n=12 and Farm 2, n=40, respectively) recruited for ROP-TB and PPD- based skin tests. As shown in Figure 4B , both ROP-TB and PPD stimulates skin reactions in the cattle from Farm 1. Among the 12 cattle, 7 show positive skin reactions to ROP-TB stimulation ( Figure 4C , left panel) and 7 to PPD stimulation ( Figure 4C , right panel). Figure 4D shows the interrelationship of ROP-TB-stimulated and PPD-stimulated skin reactions in these cattle. None of the cattle in Farm 2 produced a positive skin reaction to antigens. Table 1 summarizes the skin reactions of both farms. The sensitivity and specificity of ROP-TB with PPD are 85.7% and 97.7%, respectively ( Table 1 ).

To observe whether ROP-TB can promiscuously stimulate the memory T cell response in humans, we tested ROP-TB for a specific T cell immune response measured by ELISPOT in a population of diverse HLA phenotypes. We recruited 18 patients with tuberculosis who were diagnosed by sputum test, acid-fast staining, and lung computed tomography (CT) scan. The demographic data for these patients are summarized in Table 2 . Their HLA phenotypes are shown in Table 3 . We then compared the ELISPOT response of the PBMC of the 18 patients with the following antigens: 8 individual overlapping synthetic peptides derived from ESAT-6 and CFP-10, pooled peptides of the 8 overlapping synthetic peptides derived from ESAT-6 and CFP-10, and ROP-TB (which is composed of linked and chained peptides of the 8 peptide sequences). As shown in Table 4 , the 18 patients having different HLA types ( Table 3 ) respond to different individual peptides. However, most patients, 16/18 in ROP-TB and 15/18 in pooled peptides, respond to the pooled peptides and ROP-TB ( Table 4 ). This experiment shows that ROP-TB, like the pooled peptides, is able to stimulate memory T cell responses in different HLA types of TB patients.

The cDNA genes encoding native ESAT-6-CFP-10 and the ROP-TB respectively were designed and synthesized as previously described (6).

For purification, the corresponding plasmids were transformed into E. coil BL21 (Solarbio, China). A single colony was harvested in LB medium (50 μg/mL Kanamycin) and cultured overnight. Bacteria from the overnight culture were then inoculated into the new LB medium at a 1:100 ratio and cultured until the OD600 reached 0.6. Isopropyl β- d-1-thiogalactopyranoside (IPTG, Solarbio, China) was added at a final concentration of 0.5 mM and the bacteria were collected by centrifugation after 16 h of incubation. Lysis buffer (20mM Tris-HCl, 0.5 M NaCl, 2 mM EDTA, pH 8.0) was used to resuspend the bacteria, which were then lysed by sonication. The insoluble fractions were separated by centrifugation at 15,000 g for 1 hour.

The protein was further purified by nickel affinity chromatography using Ni-NTA (GE, USA). To detect expressed soluble TB protein (ESAT-6-CFP-10), Ni-NTA resin was added to the soluble fractions for 1 h. The protein was washed with lysis buffer containing 10 mM imidazole (washing buffer) and eluted with lysis buffer containing 300 mM imidazole (elution buffer). The insoluble ROP-TB in the pellet was resuspended in 8 M urea buffer and centrifuged at 15,000 g for 1 h. The rest procedures were similar to those of ESAT-6-CFP-10 except that the washing and elution buffer contained 8 M urea. For refolding, the eluted proteins (ESAT-6-CFP-10 and ROP-TB) were exchanged by PBS for dialysis.

Ni-NTA purified ROP-TB was further purified by endotoxin removal chromatography (BIOENDO, China). Briefly, the column (1.5 mL) was washed with balanced buffer, then ROP-TB was flowed through the column at 0.5 mL/min and the effluent collected.

The endotoxin present in ROP-TB was detected using the Limulus amebocyte lysate assay (BIOENDO, China) following the manufacturer’s instructions. Briefly, ROP-TB was diluted with endotoxin-free water by serial two-fold dilutions, then diluted ROP-TB was added, respectively, to the Limulus amebocyte lysates, which were dissolved in 100 μL endotoxin-free water. The mixtures were incubated in ampoule bottles at 37°C for 1 hour. If a gelatinous substance formed at a certain dilution, it was considered positive: Endotoxin concentration was calculated as maximum positive × 0.125 EU/mL, which indicated the sensitivity of Limulus amebocyte lysate.

Purified protein (ROP-TB) in PBS and 4mM DTT was incubated with Cathepsin S (SinoBiological, China) at 37°C for 0.5, 1, 2, 4, 8, and 22 h. The mass ratios of enzyme and protein in digestion were 1:50 and 1:100, respectively. Samples with different digestion times were run through SDS-PAGE for measuring enzyme activity.

To observe antigen uptake, 1 μM ROP-TB, TB peptides, and TB protein (ESAT-6-CFP-10) were incubated with DC2.4 cells (a gift from K Rock of Massachusetts University, United States) in culture plates (3x10⁵ per dish) at 37°C for 30 minutes or 6 hours. ROP-TB, TB peptides, and TB protein marked with the LIVE/DEAD cell stain kit (Invitrogen, USA). To stain endolysosomes, 10 μL of 1:200 diluted (red) Lysotracker (Invitrogen, USA) was added to the cell culture one hour before incubation. After incubation, cells were washed and fixed with 4% paraformaldehyde for 10 min. The samples were further stained with DAPI (Solarbio, USA) prior to observation under a confocal microscope (ZEISS, LSM800).

The mouse experiment was outsourced to the Shanghai Public Health Center (Shanghai, China). All procedures were approved by the Committee of the Shanghai Public Health Center and followed the national guidelines for the use of experimental animals. Briefly, all mice (BALA/c, age 6–8 weeks, SPF) were infected intranasally with Mtb (H37R1) at 5x10⁵ CFU in 40 μL. After infection, the animals were housed in cages kept in laminar flow safety enclosures in a Level III biosafety facility.

To verify the establishment of the Mtb model, the lung tissues of infected and uninfected mice were sectioned for fast acid staining. The sections were placed in a staining dish and stained with Carbol Fuschin. After heating to steam and incubation for 5 min, the lung sections were washed with slow running water, decolorized with acid alcohol, counterstained with Malachite green for 1 min, and then air-dried. The final sections were positive when red and negative otherwise. The procedure of H&E staining (Sigma)for the evaluation of lymphocyte infiltration is the following: samples were first fixed overnight with 4% paraformaldehyde. After embedding in paraffin, the samples were cut into 2-μm thick sections and subjected to H&E staining. Histopathological analysis was performed using a Zeiss microscope.

The IGRA-based T cell function assay was performed using ELISpot kits (Mebtech, Sweden) (6). Briefly, spleen cells were separated and incubated at 2.5x10⁵ cells/well with 5 μg/well ROP-TB or peptides at 37°C for 24 h. Then, 96-well culture plates were precoated with IFN-γ antibodies. After the incubation, the cells were discarded, and the plates were washed with PBS followed by incubation with biotinylated anti-IFN-γ antibodies for 2 h at room temperature. The plates were washed with PBS and an enzyme-labeled anti-biotin antibody was added for 1 h. Finally, BCIP/NBT was used for color development, and the reaction was stopped by washing plates with tap water. Spots were counted with an Elispot reader (Autoimmun Diagnostike, Strasburg, Germany).

The human experiments were reviewed and approved by the Ethics Committee of Changzhou No. 3 People’s Hospital. Eighteen patients with tuberculosis infection were recruited from the hospital and provided informed written consent prior to the collection of their detailed clinical data ( Table 2 ). HLA typing of the patients was performed by Weihe Biotechnology INC (Jiangsu, China).

IGRA-based T cell testing was performed using ELISpot kits (Mebtech, Sweden). Briefly, 2.5x10⁵/150 μL of PBMCs were stimulated overnight with 5 μg/well of ROP-TB, either pooled or as singular peptides in anti-IFN-γ-Ab precoated plates (Millipore, Bedford, MA). Cells were discarded and biotinylated anti-IFN-γ antibodies were added for a 2-h incubation at room temperature followed by another 1-h incubation at room temperature with an enzyme-labeled anti-biotin antibody. After the color development, the reaction was stopped by washing the plates with tap water, and the plates were air dried. Spots were counted with an Elispot reader (Autoimmun Diagnostike, Strasburg, Germany).

The experiment was outsourced to the Shanghai Public Health Center (Shanghai, China). All animals were infected with Mtb (H37R1) of 0.5 mL (5x10⁴ CFU) through nasal drop. After infection, the animals were housed in cages contained within laminar flow safety enclosures in a Level III biosecurity facility.

The DTH-based skin test was performed as follows: guinea pigs were divided into six groups of five animals per group. They were injected intradermally on the flank with 0.1 mL of ESAT-6-CFP-10 (1 mg/mL, 0.1 mg/mL), ROP-TB (1 mg/mL, 0.1 mg/mL), national standard and negative control (physiological saline). The diameters of the erythema were read after 24–72 hours.

The experiments were approved by the Ethics Committee of Southwest University (Ethics approval number: IACUC-20210320-03). Cattle from two farms (n=12 and n=40, respectively) were recruited. To evaluate the potential of a DTH-based skin test with ROP-TB, we selected commercially available PPD, which is the standard reagent for the diagnosis of M. bovis TB, to be compared with ROP-TB in the experiments. The PPD skin test was performed following the standard Chinese diagnostic technique for tuberculosis in animals. ROP-TB antigen (30 μg/animal) was injected into a volume of 0.1 mL on one side of the cow neck, while PPD (2500 IU/animal) on the other side. Skin thicknesses at injection sites were measured between 48 and 72 h of injection. For the readout of the skin test, if the skin thickness of the cattle was ≥ 4 mm, the animal was considered skin test-positive and negative if the thickness was <2 mm. In the case that the difference was nearly or equal to 2 mm, the result was considered inconclusive.

Statistical analysis was performed with Student’s t test. P values <0.05 were considered statistically significant. Figure 4A was generated and analyzed using GraphPad Prism version 8 (GraphPad Software Inc.).