Plant material

Cardoon (C. cardunculus L.) flowers were collected from an ecotype cultivated at the INIAV campus in Oeiras, Portugal. Only the upper purple portions of the flowers (pistils) were manually separated, cleaned to remove debris and dried at room temperature (20–25°C) for about 30 days in a controlled, shaded environment to preserve enzymatic activity and ensure sample quality.

Preliminary evaluation of NaCl concentration

A preliminary assay was carried out to evaluate how increasing NaCl concentrations affect the enzymatic activity of cardoon flower extracts. Dried pistils (0.04 g/L) were macerated in salt solutions from 0% to 20% (w/v), in 2.5% increments, and all tests were performed in triplicate.

Main trial: preparation of cardoon flower extracts using four different extraction formulations

Based on preliminary results, four extraction solutions were selected: NaCl at 5% (optimal enzymatic performance) and 20% (high-salt condition for potential microbial inhibition), plus two buffer solutions: sodium acetate (pH 5.5), following ISO 23058/IDF 199:2006 (ISO and IDF, 2006), and citrate (pH 3.0), chosen for its acidity and potential antimicrobial effect. Each formulation was prepared in duplicate as independent biological replicates, yielding eight 50 mL extracts.

Dried pistils (2 g) were manually macerated at room temperature with each extraction solution following traditional cheesemaking procedures, then filtered (Whatman No.  40) to a final volume of 50 mL (Gomes et al., Reference Gomes, Belo, Alvarenga, Dias, Lage, Pinheiro, Pinto-cruz, Brás, Duarte and Martins2019, Reference Gomes, Pina, Fernandes, Dias, Reboredo, Martins and Alvarenga2024). Extracts were prepared under controlled laboratory conditions, supplemented with 2% potassium sorbate and stored at 4 °C. Enzymatic and microbiological properties were evaluated at 0, 45, 90 and 135 days, with all assays conducted in duplicate.

Evaluation of technological properties

In the preliminary trial, milk-clotting activity (MCA) was determined according to ISO 23058/IDF 199:2006 (ISO and IDF, 2006) and expressed as international milk-clotting units (IMCU/g), reflecting the activity per gram of dried flower material. In the main trial, MCA was expressed as IMCU/mL to quantify enzymatic activity in the liquid extracts.

Coagulation dynamics were assessed using an Optigraph (Alliance, Frépillon, France), which measures near-infrared signal attenuation during micelle aggregation. This allowed real-time determination of parameters such as flocculation time (R) and curd firmness indices (AR, A2R, A20, A40), following the procedures described by Gomes et al. (Reference Gomes, Belo, Alvarenga, Dias, Lage, Pinheiro, Pinto-cruz, Brás, Duarte and Martins2019, Reference Gomes, Pina, Fernandes, Dias, Reboredo, Martins and Alvarenga2024).

Microbiological analysis

Microbiological characterization during the main trial was performed at each storage time under sterile conditions. The following microbial groups were selected due to their relevance to extract spoilage and food safety: (1) Total mesophilic bacteria were enumerated on Plate Count Agar (PCA) (Merck KGaA, Darmstadt, Germany) at 30°C for 72 h (ISO 4833–1:2013/Amd 1:2022). (2) Following ISO 21527-2:2008, moulds and yeasts on Rose Bengal Chloramphenicol Agar (VWR International BV, Leuven, Belgium) at 25°C for 5–7 days. Results for moulds and yeasts are reported separately in Table 3 for clarity. (3) Enterobacteriaceae on Violet Red Bile Glucose Agar (Biokar Diagnostics, Beauvais, France) plates, at 37°C for 24 h (ISO 21528-2:2017). (4) Psychrotrophic microorganisms on PCA at 6.5°C for 10 days (ISO 17410:2019).

Statistical analysis

Statistical analyses were performed using STATISTICA for Windows (version 7.0, StatSoft Inc., Tulsa, OK, USA). One-way analysis of variance followed by the Scheffé post hoc test was applied to evaluate the effect of NaCl concentration in the preliminary trial (n = 3), with a 95% confidence level (P < 0.05).

Due to the limited number of biological replicates in the main trial (n = 2), results were analysed descriptively and are presented as mean values ± standard deviation.

Preliminary trial with increasing NaCl concentrations

The preliminary assay assessed the effect of increasing NaCl concentrations, revealing that moderate salt levels, particularly 5% NaCl, provided favourable coagulation performance, with higher MCA and stronger curd formation. Intermediate NaCl concentrations improved enzymatic performance relative to the unsalted control, whereas both low and high salt levels were associated with delayed flocculation and weaker curd formation, indicating reduced efficiency (Table 1). These results reinforce earlier findings showing that elevated NaCl levels can inhibit the proteolytic system of C. cardunculus L. flowers (Sousa and Malcata, Reference Sousa and Malcata1996). The 5% NaCl formulation showed a favourable balance between milk-clotting performance and curd structure and is also expected to limit microbial growth, supporting extract stability during storage, in agreement with Conceição et al. (Reference Conceição, Martins, Alvarenga, Dias, Lamy, Garrido, Gomes, Freitas, Belo, Brás, Paulino, Duarte and Karaman2018).

Table 1.

Mean values (standard deviations) of the technological properties of the cardoon flower extracts with 2.5% increments of NaCl concentrations

Table 1 Long description

NaCl concentration from 0% to 20% is compared for cardoon flower extract milk-clotting performance and curd firmness, reported as mean with standard deviation. Milk-clotting activity rises from 36.02 at 0% to a maximum of 85.46 at 5%, then declines to about 38 to 47 at 15% to 20%. Coagulation time drops from 1632 seconds at 0% to the fastest value, 840 seconds at 5%, then increases again, reaching 1597 seconds at 17.5%. Early firmness measures AR and A2R are highest at 0% (2.57 and 3.95) and are lower at 5% (1.76 and 2.66), with intermediate values at most other salt levels. Firmness at 20 minutes is zero at 0%, 15%, 17.5%, and 20%, but is highest at 5% (0.94) and remains low to moderate at 2.5% to 12.5%. Firmness at 40 minutes is highest around 5% to 7.5% (about 2.56 to 2.58) and lowest at 17.5% (1.25). Superscript letters indicate statistically significant differences within each column, so small numeric gaps may not reflect meaningful differences.

MCA, milk-clotting activity; R, coagulation time; AR and A2R, curd firmness after two and three times the coagulation time (R) after coagulant addition. respectively, A20 and A40, curd firmness at 20 and 40 min, respectively, after coagulant addition; IMCU/g, International Milk-Clotting Units; s, seconds; V, volts. Different superscript letters in each column indicate significant differences (P < 0.05, n = 3, Scheffé test).

Technological properties of the extracts prepared with four extraction formulations

Table 2 summarizes the technological performance of the extracts obtained with the four formulations. The 20% NaCl extract consistently showed the lowest MCA values across all storage times, suggesting an inhibitory effect of high salt levels. As expected, MCA trends followed those of the flocculation time (R), which increased over storage, particularly in the 20% NaCl formulation (Dias et al., Reference Dias, Gomes, Silvério, Freitas, Fernandes, Martins, Jasnau Caeiro, Lageiro and Alvarenga2025). Initial MCA values were consistent with the preliminary trial after unit conversion, indicating consistency between the preliminary and main trials.

Table 2.

Mean values (standard deviations) of the technological properties of the extracts during the storage period

Table 2 Long description

Mean values with standard deviations are reported for multiple technological properties across four extraction formulations (20% NaCl, 5% NaCl, citrate at pH 3.0, acetate at pH 5.5) at 0, 45, 90, and 135 days of storage. Milk-clotting activity is consistently lowest with 20% NaCl (about 1.4 to 2.15 IMCU per mL) and higher with citrate and acetate, peaking around day 90 at about 4.30 and 4.25 IMCU per mL, respectively; 5% NaCl stays near 2.90 to 3.55. R time is shortest at day 0 for citrate (about 641 seconds) and longest overall for 20% NaCl, rising to about 1460 seconds at day 90 before dropping to about 1271 at day 135. AR and A2R are generally highest for 5% NaCl, reaching about 2.07 volts (AR) and 3.30 volts (A2R) at day 90, while citrate and acetate remain lower and 20% NaCl is lowest at baseline. A20 is zero for 20% NaCl from day 45 onward, but increases over time for the other formulations, with 5% NaCl reaching about 1.44 volts at day 135 and citrate and acetate reaching about 1.26 and 1.02. A40 is highest for 5% NaCl at later storage (about 3.37 volts at day 135), while citrate and acetate rise more moderately and 20% NaCl stays near 1.0 to 1.3. Comparisons should consider the reported variability, which is larger for some measures and time points, especially early clotting activity and R time in several formulations.

Abbreviations for technological properties follow those defined in Table 1.

Table 3.

Mean values (standard deviations) of microbial counts (log CFU/mL) for cardoon flower extracts during storage (0–135 days), under different extraction formulations

Table 3 Long description

Microbial counts in cardoon flower extracts are reported as mean log CFU per mL with standard deviations for five microbial groups across 0, 45, 90, and 135 days, comparing four extraction formulations. Total mesophilic bacteria are the only group consistently detected beyond day 0, staying low in all formulations. Citrate at pH 3.0 starts at 0.40 at day 0 and is not detected at 45, 90, and 135 days, while 20 percent NaCl and 5 percent NaCl remain around 0.89 to 1.27 from day 45 to 135. Acetate at pH 5.5 shows low mesophilic counts that generally decline over time, from 1.56 at day 0 to 0.54 at day 135. Molds are present at day 0 in all formulations, highest with acetate at pH 5.5 at 1.33, then drop to not detected by day 135 in every formulation. Yeasts appear only at day 0 in 5 percent NaCl, citrate pH 3.0, and acetate pH 5.5, and are not detected at later time points. Enterobacteriacea and psychrotrophic microorganisms are not detected at any time in any formulation. Values marked not detected indicate results below the method’s detection limit, so small differences near zero should be interpreted cautiously.

n.d., not detected (below the limit of detection of the method, corresponding to 0 log CFU/mL, according to ISO standards).

Gel firmness measured at twice the coagulation time (AR) decreased during storage due to increasing R values, with lower firmness at the end of the trial (Fox et al., Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017a, Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017b). The 5% NaCl extract maintained higher curd firmness at later storage times compared with the other extraction formulations. Overall, storage time did not affect this parameter. A similar pattern was observed for A2R, which reflects structural development beyond initial coagulation and is relevant for firmer cheese types. The 5% NaCl formulation showed higher A2R values at 90 days, although no clear differences were observed across storage times for this parameter.

For the A20 parameter, storage time affected only the 5% NaCl extract, which showed higher firmness at later stages due to its lower R value, allowing curd development within the 20-min interval. Formulation type also influenced A20: at the initial time, the citrate buffer produced the lowest R and highest firmness, while the 20% NaCl extract did not initiate coagulation within 20 min at any storage point, explaining the absence of measurable firmness. The A40 parameter, less dependent on R, did not vary over storage time but showed differences among formulations.

Several extraction systems, including water, saline solutions and buffered systems, have been reported for obtaining milk-clotting enzymes from C. cardunculus flowers, with extraction conditions strongly influencing coagulation properties and curd firmness (Ben Amira et al., Reference Ben Amira, Mokni, Yaich, Chaabouni, Besbes, Blecker and Attia2017). In this study, 5% NaCl appeared to be a promising formulation, providing higher MCA and improved gel firmness, while higher salt concentrations reduced activity, consistent with previous reports on the inhibitory effects of excessive NaCl on the proteolytic system of C. cardunculus L. enzymes (Sousa and Malcata, Reference Sousa and Malcata1996; Conceição et al., Reference Conceição, Martins, Alvarenga, Dias, Lamy, Garrido, Gomes, Freitas, Belo, Brás, Paulino, Duarte and Karaman2018), which is consistent with the reduced activity observed in the 20% NaCl extracts.

Although the NaCl introduced through cardoon extracts is negligible due to the low dosages used in cheesemaking, extraction conditions may still indirectly influence cheese quality by affecting coagulation kinetics and curd structure, with the 5% NaCl extract showing favourable technological characteristics under the conditions tested (Fox et al., Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017a, Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017b; Alvarenga et al., Reference Alvarenga, Fernandes, Gomes, Baltazar, Fiates, G, Santos, Conceição and Dias2024).

The effect of NaCl on MCA and coagulation dynamics is likely linked to changes in ionic strength affecting the stability and activity of cardoon aspartic proteases. Moderate salt concentrations may support enzyme activity, while higher levels can inhibit proteolysis, explaining the reduced MCA and delayed flocculation observed at 20% NaCl (Tsouli, Reference Tsouli1974; Sousa and Malcata, Reference Sousa and Malcata2002). Similar sensitivity of cardoon proteases to extraction conditions has also been reported in more recent studies (Alavi and Momen, Reference Alavi and Momen2020).

The results also show that enzymatic performance and microbial stability depend on the combined effects of pH and ionic environment. Cardoon proteases remain active under moderately acidic conditions, preserving MCA in citrate-based extracts, while citrate and low pH contribute to microbial inhibition (Alavi and Momen, Reference Alavi and Momen2020; Coban, Reference Coban2020; Lambré et al., Reference Lambré, Barat Baviera, Bolognesi, Cocconcelli, Crebelli, Gott, Grob, Lampi, Mengelers, Mortensen, Rivière, Steffensen, Tlustos, Van Loveren, Vernis, Zorn, Roos, Andryszkiewicz, Liu and Chesson2022). Overall, the technological performance and preservation of cardoon flower extracts appear to depend on the interplay between enzyme activity, ionic strength and pH, factors that may ultimately influence cheese texture and quality (Fox et al., Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017a, Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017b; Alvarenga et al., Reference Alvarenga, Fernandes, Gomes, Baltazar, Fiates, G, Santos, Conceição and Dias2024).

Microbiology evaluation

The citrate buffer was associated with the lowest mesophilic bacterial counts, with no detectable mesophilic bacteria observed during refrigerated storage (Table 3). In contrast, both saline extract formulations maintained detectable mesophilic counts throughout storage, indicating that salt alone was not sufficient to suppress growth under the tested conditions. These results align with Coban (Reference Coban2020), who reported that organic acids exert antimicrobial activity by disrupting membrane integrity and interfering with cellular processes, supporting the enhanced inhibitory effect observed in the citrate formulation.

Fungal growth was detected at the beginning of storage but was absent in all formulations by the end of the trial, with moulds contributing more to early fungal counts than yeasts. Yeasts appeared only at the initial evaluation, and notably, the 20% NaCl extract showed no yeast presence at any time. Enterobacteriaceae and psychrotrophic microorganisms were undetectable across all formulations and storage periods. The addition of 2% potassium sorbate likely contributed to suppressing microbial growth, although the persistent presence of mesophilic bacteria indicates that inhibition was not complete.

Given that cardoon extract is typically used at very low dosages in cheesemaking (0.2–0.3 g/L) (Alvarenga et al., Reference Alvarenga, Martins, Caeiro, Garcia, Pássaro, Coelho, Santos, Lampreia, Martins and Dias2021, Reference Alvarenga, Fernandes, Gomes, Baltazar, Fiates, G, Santos, Conceição and Dias2024), the resulting potassium sorbate concentration in cheese would be around 0.1 g/kg, well within regulatory safety limits (Regulation [EC] No. 1332/2008). Overall, microbial loads remained low, though not fully eliminated, contrasting with reports on lyophilized extracts, which show greater stability during storage (Tejada et al., Reference Tejada, Vioque, Gómez and Fernández-Salguero2008).

In addition to formulation and preservation strategies, the microbiological safety of cardoon flower extracts strongly depends on good hygiene practices throughout flower handling and extract preparation. Proper cleaning of raw plant material, controlled drying conditions, sanitary maceration and filtration procedures and the use of food-grade water and additives are essential to minimize initial microbial contamination. Several authors have emphasized that traditional preparation methods may lead to variable microbial loads if hygienic conditions are not adequately controlled (Tejada et al., Reference Tejada, Vioque, Gómez and Fernández-Salguero2008; Conceição et al., Reference Conceição, Martins, Alvarenga, Dias, Lamy, Garrido, Gomes, Freitas, Belo, Brás, Paulino, Duarte and Karaman2018).

Therefore, although the present study suggests that formulation, particularly the use of citrate buffer or moderate NaCl concentrations, can contribute to improved microbial stability during storage, these strategies should be considered complementary to hygienic processing practices rather than stand-alone safety measures. This integrated approach is particularly relevant when scaling up artisanal cardoon extract production for broader industrial application, as also highlighted in safety assessments of C. cardunculus L. enzymes for food use (Lambré et al., Reference Lambré, Barat Baviera, Bolognesi, Cocconcelli, Crebelli, Gott, Grob, Lampi, Mengelers, Mortensen, Rivière, Steffensen, Tlustos, Van Loveren, Vernis, Zorn, Roos, Andryszkiewicz, Liu and Chesson2022).

The use of vegetable coagulants in cheesemaking has traditionally depended on empirical knowledge and artisanal practices, which contribute to the typical texture and sensory identity of these cheeses (Roseiro et al., Reference Roseiro, Barbosa, Ames and Wilbey2003; Conceição et al., Reference Conceição, Martins, Alvarenga, Dias, Lamy, Garrido, Gomes, Freitas, Belo, Brás, Paulino, Duarte and Karaman2018). Thus, the variability associated with plant-based coagulants should be seen not only as a technological challenge but also as an intrinsic feature of traditional cheesemaking.

Although this study focuses on C. cardunculus L., the species traditionally used in Iberian PDO cheesemaking, other Cynara species, such as Cynara humilis L., are also used as vegetable coagulants in Mediterranean regions. Differences in species, ecotype and floral composition are known to influence enzyme profiles, MCA and proteolytic behaviour within Cynara spp. (Roseiro et al., Reference Roseiro, Barbosa, Ames and Wilbey2003; Amira et al., Reference Amira, Besbes, Attia and Blecker2017; Alavi and Momen, Reference Alavi and Momen2020).

C. cardunculus L. was selected for this study due to its technological relevance, mandatory use in several PDO cheeses and the fact that its enzymes have been evaluated and approved for food use by the European Food Safety Authority (Lambré et al., Reference Lambré, Barat Baviera, Bolognesi, Cocconcelli, Crebelli, Gott, Grob, Lampi, Mengelers, Mortensen, Rivière, Steffensen, Tlustos, Van Loveren, Vernis, Zorn, Roos, Andryszkiewicz, Liu and Chesson2022). Therefore, while the effects of pH and ionic strength observed here may be broadly relevant to plant-based coagulants, caution is required when extrapolating these findings to other Cynara species or botanical sources, and further comparative studies are needed to assess their wider applicability.

A limitation of the present study is the lack of validation in actual cheesemaking systems, as all results were obtained under controlled laboratory conditions. In real cheese manufacture, factors such as milk composition, processing parameters and production scale are known to significantly influence coagulation behaviour, curd structure and final cheese quality, particularly when vegetable coagulants are used (Roseiro et al., Reference Roseiro, Barbosa, Ames and Wilbey2003; Galán et al., Reference Galán, Cabezas and Fernández-salguero2012; Fox et al., Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017a, Reference Fox, Guinee, Cogan, McSweeney, Fox, Guinee, Cogan and McSweeney2017b). Therefore, while the proposed extraction strategies show promising effects at the extract level, their practical relevance should be confirmed through pilot or industrial-scale cheesemaking trials (Alvarenga et al., Reference Alvarenga, Fernandes, Gomes, Baltazar, Fiates, G, Santos, Conceição and Dias2024).

In conclusion, the preliminary trial confirmed that cardoon flower extracts prepared with 5% NaCl achieved the highest MCA and the best coagulation performance, while very low and high salt concentrations impaired enzymatic efficiency and delayed curd formation. Across the four extraction formulations tested in the main trial, the 5% NaCl solution consistently delivered the most favourable technological properties, including higher gel firmness and more efficient coagulation dynamics. The citrate buffer also performed well, particularly in terms of MCA, whereas the acetate buffer showed intermediate behaviour. As expected, the 20% NaCl extract exhibited the lowest activity and slowest coagulation, confirming its limited suitability for cheesemaking applications.

These results provide practical guidance for improving the consistency and shelf-life of cardoon flower extracts, supporting their use in both traditional and standardized cheesemaking. Optimizing extraction formulations may reduce variability associated with artisanal preparation while maintaining the properties required for high-quality cheese production. Future studies should validate these strategies at pilot or industrial scale and assess their applicability to other Cynara species and plant-based coagulants.