Inhibitory activity of aromatic plant extracts against dairy-related Clostridium species and their use to prevent the late blowing defect of cheese

The aim of the present work was the selection of aromatic plant essential oils (EOs) and/or ethanolic extracts (EEs) to prevent the late blowing defect (LBD) of cheese caused by Clostridium spp. EEs resulted more effective than EOs to inhibit dairy-borne Clostridium spp. in vitro . Savory, hyssop, lavender and tarragon EEs, which showed the lowest minimal inhibitory concentration against Clostridium tyrobutyricum , were selected to study the prevention of LBD caused by this bacterium in cheese. Addition of savory and lavender EEs to cheese milk delayed LBD by 2 weeks, but at the end of ripening these cheeses showed similar clostridial vegetative cells counts, spoilage symptoms and propionic, and butyric acids levels than blown control cheese. Tarragon EE, with the highest content in caffeic acid, also delayed LBD by 2 weeks, but it was more effective to inhibit Clostridium , since cheese with tarragon EE showed minor LBD symptoms, lower vegetative cells count and lower concentrations of propionic and butyric acids than the rest of cheeses made with EEs. This fact could be also attributable to the greater number of antimicrobial terpenes (1,8-cineole, 4-terpineol, α -terpineol, isoelemicin, methyl eugenol, and methyl trans -isoeugenol) detected in this cheese. This is the first report on the application of EEs to control C. tyrobutyricum in cheese.


Introduction
A large proportion of food is spoiled along the food supply chain, especially before reaching and with the final consumers.The Food and Agriculture Organization (FAO) reported that one-third of the food produced for human consumption is either spoiled or wasted (Odeyemi et al., 2020), and most spoilage of food is caused by microorganisms.Late blowing defect (LBD) poses a major risk of microbial spoilage in semi-hard and hard cheeses, resulting in texture defects like holes, slits, eyes and cracks, off-flavors, and reduced shelf-life.LBD is mainly caused by Clostridium tyrobutyricum which ferment lactic acid to produce butyric acid, acetic acid, and gasses (CO 2 and H 2 ) that swell the cheeses (Garde et al., 2013), although other species like Clostridium butyricum, Clostridium beijerinckii and Clostridium sporogenes, may also be involved.Clostridium spp.are ubiquitous and spore-forming bacteria, making extremely difficult to control their entrance in the dairy chain and the subsequent spoilage of cheese.
Currently, the control of C. tyrobutyricum in cheese is tackled with different chemical preservatives that, however, are not completely effective, since LBD still remains a problem for many cheesemakers.In some cases, the use of preservatives is limited for health or allergy reasons, and generates a certain rejection in consumers, whose demands of "natural" products require new and much more sustainable approaches for food preservation (Nájera et al., 2021).In this sense, plants essential oils (EOs) are widely used as antimicrobials in foods for the activity of mono-and sesquiterpenes, but aqueous or ethanolic plant extracts, less explored, also exert an antimicrobial activity principally due to its phenolic components (Christaki et al., 2021;Ritota and Manzi, 2020).In EO production, the solid residues obtained after hydrodistillation are rich in phenolic compounds that can be recovered and reused for their antimicrobial properties (Ortiz de Elguea-Culebras et al., 2022).Furthermore, these residues have been deodorized during hydrodistillation, which increases their interest to be used in food.
In cheeses made with milk fortified with lemon balm, basil, thyme, and rosemary EOs, Clostridium inhibition was observed (Licón et al., 2020;Moro et al., 2015), but the effect of these EOs on the LBD was not studied.Moreover, many works have been carried out with collection strains (Bagheri et al., 2020;Bakhtiary et al., 2018;Librán et al., 2013;Licón et al., 2020;Magwa et al., 2006;Shah et al., 2004;Tosun et al., 2022;Vujić et al., 2020), so it would be very interesting to test the efficacy of plant extracts against autochthonous isolates of Clostridium spp. of cheeses with LBD.
Therefore, in the present work we have tested the antimicrobial activity of aromatic plants EOs and EEs (obtained from solid by-products of EO production) against dairy-borne C. tyrobutyricum, C. butyricum, C. beijerinckii and C. sporogenes strains, with the aim of selecting the most active extracts to test their efficacy to prevent LBD caused by Clostridium spp. in cheese.

Bacterial strains and growth conditions
The strains of C. tyrobutyricum (INIA 68 and INIA 69),C. butyricum (INIA 66 and INIA 67),C. beijerinckii (INIA 63 and INIA 65) and C. sporogenes (INIA 70 and INIA 71) were isolated from Manchego cheeses with pronounced LBD and selected because they presented different pulsotypes (represented by several isolates) which produced high amounts of gas and butyric acid in milk and in Bryant and Burkey broth (containing sodium lactate) (Garde et al., 2011(Garde et al., , 2012a  C for 48 h in anaerobic conditions.All anaerobic incubations were carried out in jars with an H 2 plus CO 2 generating kit (AnaeroGen, Oxoid, Basingstoke, UK), with an anaerobiosis indicator.Bacteria were stored at − 80 • C as stock cultures in their corresponding culture media supplemented with 5% glycerol, and sub-cultured twice before their use.
To give rise to LBD in cheese, we prepared a concentrated stock of C. tyrobutyricum INIA 68 spores in skim milk (Difco) following the method described by Ávila et al. (2016).Clostridial spores were obtained by inoculation in modified RCM (without agar and sodium acetate).After incubation at 37 • C for 3 days under anaerobic conditions, the culture was centrifuged (5000×g, 15 min, 20 • C).The cell pellet was washed twice with sterile distilled water, suspended in skimmed milk, and heat-shocked at 80 • C for 20 min to kill off vegetative cells.The spore suspension was aliquoted and maintained at − 40 • C until their use in cheese manufacture.The spore concentration (3.0 × 10 7 spores/mL) was determined on RCM agar after anaerobic incubation at 37 • C for 3 days.
Essential oils (EOs) were obtained by steam distillation, and ethanolic extracts (EEs) from the solid by-products resulting from the production of EOs using a Soxhlet extractor with 60% ethanol (v/v) as described by Muñoz-Tebar et al. (2021).The EEs were vacuum filtered with filter paper, centrifuged (10,000×g, 10 min, 4 • C), and the supernatant filtered (0.22 μm).Plant extracts were light-protected and stored at 4 • C until use.
The chemical composition of the most effective EOs and EEs was determined by gas (GC-FID) and liquid (HPLC-DAD) chromatography, respectively, and was previously reported (Muñoz-Tebar et al., 2021).With respect to the low molecular weight phenolic compounds of the most inhibiting extracts, selected to be tested in cheese, protocatechuic, syringic and caffeic acids, and sinapaldehyde were detected in savory EE, the three acids were also present in hyssop and lavender EEs, and protocatechuic and caffeic acids in tarragon EE (Muñoz-Tebar et al., 2021).

Minimal inhibitory concentration (MIC) of aromatic plant extracts on Clostridium spp.
The MIC of EOs and EEs were determined in RCM and was defined as the lowest concentration that showed a complete inhibition of the growth of the assayed Clostridium strains.Before the microplate assays, EOs were prepared at 20.0 mg/mL in 25% (v/v) Tween 20 (Sigma-Aldrich, St. Louis, United States).Serial two-fold dilutions of EOs and EEs were prepared with distilled water using 96-well microplates with lid to obtain a final concentration range of 10 to 0.020 mg/mL and 500 to 0.24 μL/mL, respectively.Clostridial strains were grown as indicated above for 72 h, diluted to approximately 10 6 CFU/mL in double strength RCM, and then 125 μL were pipetted into the wells containing 125 μL of the diluted extracts, in duplicate.Positive clostridia growth controls were prepared in RCM without any extract.Negative controls were prepared in RCM without clostridia inoculation.In addition, the effects of Tween 20 (25%, v/v) and ethanol (60%, v/v) on Clostridium growth were tested in the same way as aromatic plant extracts.Microplates were incubated at 37 • C for 7 d under anaerobic conditions and bacterial growth was confirmed visually.Those wells showing absence of growth in RCM compared with the negative control wells were interpreted as negative.Experiments were performed in duplicate.

MICs of aromatic plant extracts on cheese starter
The MICs of EOs and EEs against cheese starter MA 16 (Choozit™ MA 16, Danisco, Laboratorios Arroyo, Santander, Spain), composed of Lactococcus lactis subsp.lactis and Lactococcus lactis subsp.cremoris, were also determined in GM17 broth (Difco) as described above.Starter MA 16 was grown in GM17 broth at 30 • C under aerobic conditions for 18 h previous inoculation of double strength GM17 broth with approximately 10 7 CFU/mL.Microplates were incubated at 30 • C for 2 d and bacterial growth was confirmed visually.Experiments were performed in duplicate.

Cheese making with C. tyrobutyricum spores and ethanolic extracts
To study the inhibition of C. tyrobutyricum by selected EEs and their effect on cheese LBD, we used C. tyrobutyricum INIA 68, a wild potent cheese spoiler able to induce the appearance of LBD during cheese ripening (Gómez-Torres et al., 2015).Semi-hard-type cheeses were manufactured from pasteurized (72 • C for 15 s) ewe milk as described by Gómez-Torres et al. (2015).In brief, six vats were made and inoculated with: vat (1) starter MA 16 and ethanol (control cheese without LBD); vat (2) starter, clostridial spores and ethanol (positive control of LBD); vat (3) starter, spores, and savory EE; vat (4) starter, spores and hyssop EE; vat (5) starter, spores and lavender EE; and vat (6) starter, spores and tarragon EE.Starter MA 16 was inoculated to vats' milk at ~7.0 log CFU/mL, and C. tyrobutyricum INIA 68 at ~3.4 log spores/mL.Ethanol (60%, v/v) was added to milk in vats 1-2 at a final concentration of 1.5%, and EEs and were added to milk in vats 3-6 at a final concentration of 25 μL/mL (which implied 1.5% ethanol).The cheeses were vacuum-sealed and ripened at 12 • C during 60 d.

Detection of late blowing defect
During cheese ripening, we regularly monitored the appearance of spoilage symptoms due to Clostridium metabolism: swollen package, cracks/splits in cheese and rancid odor due to butyric acid production.

Microbiological analysis of cheeses
At each sampling time, curd (5 g) or cheese (5 g) samples were aseptically taken.Samples were stomached in 45 mL of warm sterile 2% (w/v) sodium citrate solution.Cheese homogenates were diluted in sterile 0.1% (w/v) peptone solution (Difco).A Thoma counting chamber was used for clostridial vegetative cells counting under phase contrast microscopy, RCM agar for clostridial spore counting (3 d at 37 • C in anaerobiosis) after heat shock of cheese dilutions at 80 • C for 20 min, and PCA (Difco) with 0.1% skim milk, for starter lactococci counting (24 h at 30 • C).Clostridial spores and lactococci counts were expressed as log CFU/g of cheese, and clostridial vegetative cells as log cells/g.Detection limits were 0.40 and 1.40 log CFU/g, and 3.91 log cells/g for clostridial spores and lactococci counts, and clostridial vegetative cells, respectively.

Physicochemical analysis of cheeses
Cheese pH was measured with a Crison pH meter (model GPL 22, Crison Instruments, Barcelona, Spain) using a Crison penetration electrode (model 52-3.2).Dry matter content was determined after drying to constant weight in a vacuum oven at 100 • C. Lactic, propionic, and butyric acids were determined by HPLC as described by Ávila et al. (2016).Color parameters at the cheese surface were measured using a CM-700 spectrocolorimeter with SpectraMagic NX VA.9 software (Minolta Camera Co., Osaka, Japan) as described by Gómez-Torres et al. (2014).Volatile compounds of cheeses were extracted by automated solid-phase microextraction (SPME) and analyzed by gas chromatography-mass spectrometry (GC-MS) as described by Gómez-Torres et al. (2016).All determinations were carried out in duplicate.

Statistical analysis
Cheeses were manufactured in duplicate experiments on different days, and all determinations were done in duplicate.Statistical treatment of data was performed by SPSS software (IBM SPSS Statistics version 22, IBM Corp., Armonk, NY, USA).Data were analyzed by ANOVA using a general linear model and comparison of means was carried out by Tukey's test.

Inhibitory activity of essential oils against Clostridium spp. in the microplate assay
The MICs of EOs against the different Clostridium species are shown in Table 1.The sensitivity of Clostridium to EOs was strain dependent.In general, EOs did not inhibit many Clostridium strains.Oregano EO, with carvacrol and thymol as the major compounds (Muñoz-Tebar et al., 2021), inhibited all C. tyrobutyricum and C. beijerinckii strains, and one C. butyricum strain.Savory EO showed lower antimicrobial spectrum than oregano EO, inhibiting only the growth of all C. beijerinckii strains, one C. tyrobutyricum strain and one C. butyricum strain.The analysis of savory EO revealed that its main components were carvacrol and p-cymene (Muñoz-Tebar et al., 2021).The anticlostridial spectrum of tarragon EO was limited to all C. beijerinckii strains, and the one of hyssop, lavender, santolina and marjoram EOs to two C. beijerinckii strains.The major components of tarragon EO were methyl eugenol, elimicin and (z)-iso-elemicin, those of hyssop EO were 1,8-cineole and β-pinene + mircene, and those of lavender EO were linalool, 1,8-cineole, α-pinene, and camphor (Muñoz-Tebar et al., 2021).The marjoram EO contained mainly 1,8-cineole, linalool and β-pinene + mircene (data not shown).In the case of santolina EO, numerous chemical compounds were identified, being 1,8-cineole and 8-methylene-3-oxatricyclo [5.2.0.0 (2,4)]nonane the major ones (Ortiz de Elguea-Culebras et al., 2017).On the other hand, C. sporogenes strains resulted resistant to all EOs, and C. beijerinckii were the most sensitive to EOs with MICs from 0.04 to 8.75 mg/mL.Tween 20 did not inhibit the growth of any of the Clostridium strains at 12.5%, the maximum concentration tested.
In our work, oregano EO was the most effective EO in the inhibition of Clostridium spp.The higher content of carvacrol and thymol, and the presence of trans-caryophyllene in oregano EO with respect to the rest of EOs (Muñoz-Tebar et al., 2021), could be related to its greater efficiency in clostridial inhibition.Antimicrobial properties have been described for these compounds (Angane et al., 2022;Magwa et al., 2006;Muñoz-Tebar et al., 2021;Rao et al., 2019).
The difference found among studies could be due to the different assays used to investigate Clostridium sensibility and/or to different concentration of the EOs bioactive compounds exerting antimicrobial activity, determined by plants genetics, and affected by different factors such as soil composition, climate, plant management, and phenological stage (Ritota and Manzi, 2020).Another factor that could explain this variability is the different sensitivity of the strains to EOs.

Inhibitory activity of ethanolic extracts against Clostridium spp. in the microplate assay
We found that all clostridial strains were sensitive to EEs at least at M. Ávila et al. one of the tested concentrations (Table 2).EEs sensitivity was strain dependent.MICs ranged from 11.7 to 437.5 μL/mL for C. tyrobutyricum, from 15.6 to 500 μL/mL for C. butyricum, from 0.9 to 250 μL/mL for C. beijerinckii and from 7.8 to 312.5 μL/mL for C. sporogenes.Hyssop, lavender and tarragon EEs showed lower MICs (<40 μL/mL) than the other EEs for all Clostridium species.In addition, oregano, savory, and marjoram EEs showed MICs <40 μL/mL for one C. beijerinckii strain, two C. tyrobutyricum and one C. butyricum strains, and all C. sporogenes strains, respectively.Ethanol did not inhibit the growth of any of the Clostridium strains at 30%, the maximum concentration tested.Low molecular weight phenolic compounds detected in the oregano, savory, tarragon, hyssop and lavender EEs were protocatechuic, syringic and caffeic acids, and sinapaldehyde (Muñoz-Tebar et al., 2021).Caffeic acid was one of the major compounds in all the samples, with the highest amount obtained in the tarragon EE.The highest concentration of protocatechuic acid was found in the oregano EE, sinapaldehyde was only present in the savory and oregano EEs and syringic acid in hyssop and lavender EEs.In the case of the phenol composition of santolina and marjoram EEs, salvianolic and rosmarinic acid were predominant, respectively (Oalđe et al., 2020;Ortiz de Elguea-Culebras et al., 2017).Antimicrobial activity for all these compounds has been described in the literature (Oulahal and Degraeve, 2022;Panyo et al., 2016).
In spite of the proven antibacterial effect of EOs for cheese preservation, several factors limit their application: intense flavor, potential interactions with food components may be a hindrance to their efficiency, their price and their potential toxicity at high doses despite their GRAS status (Dupas et al., 2020).For these reasons, recent studies on dairy product preservation have focused on the effects of plant extracts other than EOs.Nevertheless, only a few studies concerning the inhibition of Clostridium spp.by plant EEs are available.EEs of Eremophila duttonii and Helichrysum plicatum showed antimicrobial activity against C. sporogenes (Shah et al., 2004;Vujić et al., 2020).The growth of Clostridium perfringens was inhibited by caper, Eremophila duttonii and eucalyptus EEs (Hematian et al., 2020;Shah et al., 2004;Ullah et al., 2021).

Inhibitory activity of aromatic plant extracts against cheese starter in the microplate assay
Differences in the antimicrobial potential of aromatic plant extracts against starter lactococci, depending on plant source, were detected.Starter lactococci were resistant to hyssop, santolina, marjoram and tarragon EOs, while oregano, savory and lavender EOs showed MICs of 1.72, 4.38, and 5.00 mg/mL, respectively (Table 3).Lactococci were resistant to Tween 20 at 12.5%, but ethanol, unlike clostridial strains, inhibit lactococci growth at a concentration ≥7.5%.However, hyssop,

Viability of C. tyrobutyricum and symptoms of late blowing defect in cheeses made with ethanolic extracts
In view of the microplate assays results, EEs of savory, hyssop, lavender and tarragon were selected to study the inhibition of C. tyrobutyricum INIA 68 (a potent cheese spoiler isolated from a blown Manchego cheese) in cheese and prevent LBD.
Control cheese (made only with starter) showed an even surface, without slits, and a lactic odor.Blowing control cheese (made with starter and C. tyrobutyricum spores), showed puffy package, slits, and rancid odor on day 14 of ripening, matching with a decrease in spore counts due to spore germination and the subsequent appearance of vegetative cells, responsible for the butyric acid fermentation (Table 4).In cheese with hyssop EE, symptoms of LBD manifested also on day 14, along with a decrease in spore counts and the emergence of vegetative cells (Table 4).However, the addition savory, lavender and tarragon EEs to cheese slowed down the appearance of vegetative cells and associated LBD by 2 weeks.In all cheeses made with C. tyrobutyricum, spore counts declined up to non-detectable levels during ripening, indicating absence of sporulation.Such lack of sporulation of this strain has been also reported in ewes' milk cheese ( Ávila et al., 2016, 2017).
The presence of EEs in cheeses seemed to promote the germination of clostridial spores, with spore counts 0.21-0.55units lower (P < 0.01) in cheeses with EEs than in blown control cheese at day 1.Furthermore, no clostridial spores were detected at 14 d of ripening in cheeses with hyssop, lavender, and tarragon EEs, whereas in blown control cheese spores become undetectable after 28 d.However, despite the increased spore germination, Clostridium vegetative cell counts were lower (P < 0.01) in cheese with hyssop EE than in blown control cheese at 14 and 28 d, and emerged belatedly in cheeses with savory, lavender, and tarragon EEs.Interestingly, cheese with tarragon EE, with the highest content in caffeic acid (Muñoz-Tebar et al., 2021), showed lower (P < 0.01) counts of clostridial vegetative cells and minor symptoms of LBD than blown control cheese at the end of ripening (Fig. 1).These results might indicate that EEs both foster Clostridium spore germination and inhibit the vegetative cells responsible for LBD, delaying the butyric acid fermentation.In this sense, licorice EE appeared to have germination-inducing properties that increased the sensitivity of the spores of Clostridium botulinum to thermal treatment (Cui et al., 2011).Despite the tested concentration of EEs in our cheeses was similar or even higher that their MICs against C. tyrobutyricum INIA 68 in vitro (Table 2), it was not effective to control completely the growth of Clostridium.Other authors have also observed that higher concentration of natural compounds was needed to achieve the same antimicrobial effect in cheese than in vitro (Ritota and Manzi, 2020).In cheese, due to the complexity of the food matrix, there are intrinsic and extrinsic factors that can interfere with the activity of these compounds, they may be lost during cheese processing or become unstable after light, temperature, oxygen, and pH exposure (Gouvea et al., 2017).

Organic acids content of cheeses
In control cheese, the concentration of lactic acid did not diminish during ripening, and propionic and butyric acid were not detected   9.12 ± 0.11 c 9.02 ± 0.06 ab 9.05 ± 0.14 bc 9.14 ± 0.04 c 8.92 ± 0.06 a 9.05 ± 0.17 bc a Results are mean (n = 4) ± SD of duplicate determinations in two cheese trials.Means in the same row followed by different letters differ significantly (P < 0.01).
(Table 5).Blowing control cheese and blown cheeses with savory, hyssop, and lavender EEs, showed a similar decrease in lactic acid content (Table 5) and the formation of butyric acid was concomitant with the appearance of blowing symptoms, consequence of butyric acid fermentation by vegetative cells (Tables 4 and 5).Propionic acid was also detected in these cheeses.Nonetheless, cheese made with tarragon EE and clostridial spores showed, in general, a slower (P < 0.01) consumption of lactic acid and lower (P < 0.01) levels of propionic and butyric acids than the rest of blown cheeses made with EEs, consistent with its minor spoilage symptoms (Fig. 1).We have previously observed that, in cheeses with LBD, the concentration of butyric acid increase and the concentration of lactic acid decrease with the degree of LBD (Garde et al., 2012a(Garde et al., , 2020)).

Cheese pH, dry matter content and lactococci counts
Table 6 shows the values of pH, dry matter and starter lactococci counts for the cheeses made with C. tyrobutyricum INIA 68 and aromatic plants EEs during ripening.Cheeses pH values increased during ripening (P < 0.01), and dry matter content rose an average 2.4% in cheeses from day 1-14 (P < 0.01), due to the loss of moisture, and remained practically constant until 60 d.The addition of aromatic plants EEs and/or C. tyrobutyricum spores to milk did not alter the pH values of cheeses, but dry matter was generally lower in the cheeses made with the EEs than in the control cheese.Some studies have demonstrated that certain plant extracts do not significantly affect cheese pH or rather assist in maintaining a suitable pH (Gouvea et al., 2017).The addition of EEs of rosemary powder or of Echinophora platyloba to cheese did not affect pH values (Ghaderi et al., 2021;Himed-Idir et al., 2021).Fortification of UF-soft cheese with ginger EE reduced pH value but total solids were not affected (El-Aziz et al., 2012).Cream cheese supplemented with Moringa oleifera EE showed increased total solids (Mohamed et al., 2018).
Lactococci counts decreased (P < 0.01) from 1 to 60 d of ripening in all cheeses.In general, the addition of savory, hyssop, lavender, and tarragon EEs to cheese milk did not affect (P > 0.01) starter lactococci counts, registering only lower (P < 0.01) counts in the cheese with lavender EE than in control cheese at 60 d.These results are in accordance with the lactococci resistance observed in vitro to EEs and ethanol at their concentration in milk (25 μL/mL EEs, implying 1.5% ethanol) (Table 3).The lactic acid bacteria reported in cheese have demonstrated their resistance to plant extracts at concentrations that inhibit the growth of pathogenic microorganisms (Gouvea et al., 2017).During refrigerated storage, the growth of L. lactis ssp.lactis and L. lactis ssp.cremoris in UF-soft cheese fortified with ginger EE was enhanced compared with control cheese (El-Aziz et al., 2012).

Cheese color
The color parameters of cheeses are shown in Table 7. Addition of aromatic plants EEs resulted in homogeneous cheese color.All cheeses showed negative values for a* parameter (greenish direction), whereas b* values were positive (yellowish direction) similar to those reported for pasteurized ewe milk cheeses (Gómez-Torres et al., 2015).As ripening advanced, the greenish-yellowish degree of cheeses increased (P < 0.01), while lightness (L*) decreased.No significant differences were found for L* and a* parameter (except at 60 d) between control cheese and blown control cheese with Clostridium spores, although b* values were lower (P < 0.01) in the blown cheese.
The color parameters of the cheeses were affected by EEs addition, probably due to the inherent greenish yellow to brownish color of EEs (Fig. 1).The cheeses made with EEs presented lower (P < 0.01) lightness than control cheese throughout ripening.Regarding a* parameter, cheeses made with savory and tarragon EEs presented higher (P < 0.01) values than control cheese, and those made with hyssop and lavender EEs had lower (P < 0.01) values (Table 7).Cheeses made with EEs presented higher (P < 0.01) values of b* than control cheese at 1 day.In the case of savory and hyssop EEs, these differences were minimized along ripening.However, differences remained significant (P < 0.01) in the case of lavender and tarragon EEs (Table 7).Despite this, in general cheeses made with EEs presented an appropriate color for cheese, characterized by lower lightness and higher yellowness, which gave them the aspect of a more mature cheese (Fig. 1).Some of the low molecular weight phenolic compounds (caffeic, syringic and protocatechuic acids, and sinapaldehyde) identified in our aromatic plants EEs (Muñoz-Tebar et al., 2021) resulted in brownish yellow colors when added to gelatin or chitosan films (Erge and Eren, 2021;Liu et al., 2017;Yang et al., 2019).Plant extracts have a distinctive effect on cheese color.In general, UF-soft cheeses fortified with ginger EE exhibited a slight gradual increase in whiteness and greenish degree at week 1 and 6 of storage, and yellowish degree of cheese was more affected by fortification level with ginger extract (El-Aziz et al., 2012).Lemon grass and cress EEs improved the color attributes of supplemented processed cheese, whereas the cinnamon, rosemary, cloves, oregano, and sage EEs obtained lower panelists' scores than control cheese (Tayel et al., 2015).

Volatile profile of cheeses
A total of fifty-eight compounds were identified by SPME/GC-MS in the volatile fraction of 60-d-old cheeses, including 4 sulfur compounds, 4 hydrocarbons, 2 aldehydes, 5 ketones, 6 alcohols, 8 esters, 8 acids, 10 terpenes and terpenoids, and 11 benzene compounds.Table 8 shows volatile compounds significantly (P < 0.01) affected by Clostridium and/ or EEs addition to cheese milk.Cheese milk contamination with  et al., 2015).In general, cheeses made with spores and EEs showed lower levels of these compounds than the blown control cheese (Table 8) according to the spoilage delay in these cheeses.In accordance with the HPLC results, the relative abundance of butyric acid was lower (P < 0.01) in the cheese with tarragon EE than in the rest of cheeses with EEs.On the contrary, levels of ethyl acetate, hexanoate, lactate and octanoate were lower in cheeses with LBD than in control cheese (Table 8).The same trend was observed by Gómez-Torres et al. (2015) for some of these compounds in cheeses with LBD.
Clostridium inhibition was observed in cheeses made with aromatic plants EOs (Licón et al., 2020;Moro et al., 2015) but the effect of these EOs on the LBD was not studied.β-Pinene, D-limonene, 1,8-cineole, α-terpineol and methyl eugenol, among other compounds, were detected in the volatile fraction of these cheeses.
Levels of 1-hexanol were higher in cheeses made with hyssop and lavender EEs (Table 8).Higher concentrations of 1-hexanol have been reported in cheeses made with herbs than in control cheeses (Kavaz et al., 2013;Sulejmani et al., 2020).Ethylbenzene was only detected in cheeses made with hyssop and tarragon EEs, and the levels of styrene, 2-ethyl-1,4-dimethylbenzene and phenylethanol were higher in cheeses made with hyssop, with savory, hyssop, or lavender, and with hyssop, lavender or tarragon EEs, respectively (Table 8).Higher levels of styrene were detected in cheeses made with saffron (Librán et al., 2014), and of styrene and phenylethanol in cheese made with garlic (Sulejmani et al., 2020).
The presence of antimicrobials such as terpenes and terpenoids in cheeses made with EEs may explain the delay in the appearance of LBD in these cheeses by inhibiting C. tyrobutyricum, especially in cheese with tarragon EE that showed 6 of these compounds in its volatile fraction (Table 8).However, caffeic, syringic and protocatechuic acids, and sinapaldehyde identified in our aromatic plants EEs (Muñoz-Tebar et al., 2021) may have also been able to contribute to the inhibition of C. tyrobutyricum, since they showed antimicrobial activity (Oulahal and Degraeve, 2022;Panyo et al., 2016).The antibacterial activity from plant extracts could be due to their diverse content in bioactive substances which may combine to perform their antimicrobial action (Tayel et al., 2015).

Conclusions
In the present work, aromatic plants EEs resulted more effective than EOs in the in vitro inhibition of Clostridium growth.In cheese, EEs promoted germination of C. tyrobutyricum spores, restrained vegetative cells growth and retarded the consequent appearance of LBD, without hardly altering or harming cheese characteristics.Nevertheless, considering the lower levels of clostridial spores usually found in naturally contaminated milk, a greater inhibition may be expected in real cheese production.Tarragon EE, with the highest content of caffeic acid, was the most effective extract, since the cheese made with it showed less LBD symptoms.This fact could be also attributable to the greater number of terpenes detected in the volatile fraction of this cheese.
As the number of works concerning the use of EEs in food is very limited, further studies are necessary to establish their role in cheese preservation.With respect to LBD prevention, it would be worth studying various EEs concentrations, other plants EEs, and combinations of different EEs and with other antimicrobials.The effect of promising EEs on the organoleptic characteristics of cheese should also be determined.

Table 1
Minimal inhibitory concentration a of aromatic plants essential oils (EOs) against Clostridium spp., in RCM at 37 • C after 7 d under anaerobic conditions.Values represent EOs concentration which caused a complete inhibition of clostridial growth in 4 replicates.Range of concentration tested for EOs: 0.020-10 mg/ mL.NO: no growth inhibition at the highest concentration tested. a

Table 2
(Christaki et al., 2021;de Campos et al., 2022;de Carvalho et al., 2015;Licón et al., 2020)tridium spp., in RCM at 37 • C after 7 d under anaerobic conditions.Although starter lactic acid bacteria are considered relatively resistant to the antimicrobial effects of EOs/extracts and usually their growth is not significantly affected, similarly to our results some of studies have shown the opposite results for extracts such as EOs from oregano, lemon balm and thyme(Christaki et al., 2021;de Campos et al., 2022;de Carvalho et al., 2015;Licón et al., 2020).To our knowledge, there is no information regarding the effect of EEs from aromatic plants on cheese starter lactococci.In our study lower concentrations of EEs were necessary to inhibit the majority of tested Clostridium spp.than for starter lactococci, making EEs interesting to be applied in cheese to control Clostridium spp.spoilage without negatively affecting the starter.
a Values represent EEs concentration which caused a complete inhibition of clostridial growth in 4 replicates.Range of concentration tested for EEs: 0.24-500 μL/ mL.M.Ávila et al.

Table 3
Minimal inhibitory concentration a of aromatic plants extracts against starter lactococci in GM17 at 30 • C after 2 d under aerobic conditions.Values represent EOs and EEs concentration which caused a complete inhibition of clostridial growth in 4 replicates.Range of concentration tested for EOs: 0.020-10 mg/mL and EEs: 0.24-500 μL/mL.NO: no growth inhibition at the highest concentration tested.Appearance of late blowing defect, counts a of clostridial spores and vegetative cells of cheeses made with C. tyrobutyricum INIA 68 and aromatic plants ethanolic extracts (EEs).
a a Results are mean (n = 4) ± SD of duplicate determinations in two cheese trials.Means in the same row followed by different letters differ significantly (P < 0.01).ND, not detected.M.Ávila et al.

Table 5 Concentration
a of lactic, propionic and butyric acids of cheeses made with C. tyrobutyricum INIA 68 and aromatic plants ethanolic extracts (EEs).Cheese starter Spores of C. tyrobutyricum INIA 68a Results are mean (n = 4) ± SD of duplicate determinations in two cheese trials.Means in the same row followed by different letters differ significantly (P < 0.01).ND, not detected.

Table 6
Values a of pH, dry matter and starter lactococci counts of cheeses made with C. tyrobutyricum INIA 68 and aromatic plants ethanolic extracts (EEs).

Table 7
Color a of cheeses made with C. tyrobutyricum INIA 68 and aromatic plants ethanolic extracts (EEs).
a Results are mean (n = 4) ± SD of duplicate determinations in two cheese trials.Means in the same row followed by different letters differ significantly (P < 0.01).

Table 8
Volatile compounds a significantly affected by late blowing defect or EEs addition in 60-day-old cheeses made with C. tyrobutyricum INIA 68 and aromatic plants ethanolic extracts (EEs).Results are mean (n = 4) ± SD of duplicate determinations in two cheese trials, expressed as relative abundances to internal standard cyclohexanone.Means in the same row followed by different letters differ significantly (P< 0.01).ND, not detected.