Materials and Methods

In the months of December 2012, February 2013, and May 2013, and based on a random sampling (^{Steel et al., 1997}), 49 samples in total of dried calyxes of a creole variety were collected, of 30 g each, from warehouses located in Ayutla, Ayutla, and Las Mesas, San Marcos, in the state of Guerrero.

Identification of fungi associated with stored roselle calyxes. For the isolation, purification and characterization of colonies, five calyxes were taken at random for each sample and were cut into pieces of approximately 1 cm2; they were disinfected with 1 % sodium hypochlorite for 1.5 to 2 minutes and rinsed three times with sterile distilled water; they were dried with sterile paper towels; subsequently four pieces were sown in Petri dishes with PDA culture medium (^{Crous et al., 2009}) brand Bioxon(r). The dishes were incubated at 25-26 ºC for 10 days or until the development of structures. Likewise, five pieces were placed in a humidity chamber and were left in natural light-dark conditions until the development of structures (^{Crous et al., 2009}). After this time, the appearance of fungal microorganisms was verified and the necessary sub-cultures were made in PDA culture medium in order to initiate the purification process with the hyphae point technique and monoconidial sowing (^{Leslie and Summerell, 2006}; ^{Crous et al., 2009}). With each of the pure isolates, an identification was made at the genus level based on macroscopic morphological characteristics observed through a Zeiss stereoscopic microscope, and microscopic characteristics (according to the Saccardo System) based on the measurement of 100 vegetative and reproductive structures of the cultures (^{Crous et al., 2009}), observed through a Nikon Eclipse Ci compound microscope with a series AmScope MU 1000 microscope camera. The identification was made based on the taxonomic keys of ^{Ellis (1971)}, ^{Sutton (1980)}, ^{Hanlin (1997)}, ^{Barnett and Hunter (1998)} and Leslie and Summerell (2009).

In order for molecular identification, DNA was extracted from the mycelia of the cultures grown in PDA following the AP method (Sambrook and Russell, 2001). In a Techne(r) thermal cycler, model TC-512, the PCR analysis was carried out in order to amplify the internal regions in the ribosomal genes (rDNA) located between subunits 18S-5.8S-28S. The universal indicators ITS 4 (TCCTCCGCTTATTGATATG) and ITS 5 (GGAAGTAAAAGTCGTAACAAGG) were used, which were synthesized in the Instituto de Biotecnología of the UNAM. Each reaction mix contained 2.5 µL of buffer of 10X reaction, 1.25 µL of MgCl, 0.5 µL of dNTP's (Promega(r)), 1.0 µL of ITS 4, 1.0 µL of ITS 5, 0.5 of Taq DNA polymerase (Promega 5 µ/µL), 2.0 µL of DNA and 16.25 µL of injectable water to obtain a final volume of 25 µL (^{Innis et al., 1990}; ^{Crous et al., 2009}). The amplification program was: initial denaturing temperature of 95 ºC for 5 minutes, followed by 30 cycles at 95 ºC for 45 seconds for the denaturing, 57 ºC for 45 seconds for the hybridization and 72 ºC for 1 minute for the extension and a final cycle of 72 ºC for 5 minutes for the final extension and a final refrigeration temperature of 10 ºC (Innis et al., 1990; Crous et al., 2009). The quality of the products of DNA extraction and PCR amplification was verified with an electrophoresis in 1% agarose gel (Agarose LE axygen(r)); for the PCR products a 100 pb molecular weight marker was used (DNA Ladder Progema(r)). The electrophoresis was carried out in a First Light Illuminator UV at 90 V for 30 min. (^{Innis et al., 1990}; ^{Crous et al., 2009}).

The product obtained from the amplification was sent to be sequenced at Macrogen Inc. (Seoul, Korea). The sequences were edited with the program Gene Runner 5.0.33 Beta and the homology of the sequencing in question was determined based on the data of the NCBI (National Center for Biotechnology Information).

Determining the incidence of diseased calyxes. The incidence of diseased calyxes was evaluated using the method proposed by ^{Campbell and Madden (1990)} and it was determined in a nominal manner (diseased or healthy), without considering the severity of the disease. The symptoms of calyxes considered diseased could show up as spots on the calyx or smut at the tips.

Evaluation of storage conditions of the calyxes. The humidity of the stored calyxes was determined according to method 14.004 of the AOAC (1984). Likewise, monthly averages were obtained for the following variables: temperature, luminosity and relative humidity from each of the sampled warehouses. The evaluated period was from December 2012 to May 2013, registering the values of each variable, every four hours with the aid of a Hobo Data Loggers U12.

Data analysis. With the data obtained from all the variables under study, analyses of variance were carried out with the SAS System version 9 software, and a comparison of measures was done using the Least Significant Difference (LSD) method with a significance level of 5 % (^{Steel et al., 1997}). In addition, a correlation analysis was done (^{Steel et al., 1997}) between the following variables: temperature, relative humidity, luminosity, humidity of the calyxes, incidence of diseased calyxes, isolated number of genera and their frequency with the data observed from both warehouses.

Results and Discussion

Even while the norm NMX-FF-115-SCFI-2010 classifies a sole quality level designated as dried roselle flower (calyx), three quality levels were identified from greatest to least: Extra, Supreme and Commercial. Such classification was established by the collectors and is in function of the visual appearance of the product, therefore its economic value varies in the market (Figure 1).

Figura 1 Commercial quality of national roselle calyxes A) Comercial B)Suprema C)Extra, collected in Ayutla and Las mesas, December 2012 - May 3013. 

40.8% of the collected samples corresponded to the Commercial quality, which presented broken calyxes, a dark red color, and showed spots and/ or smut at the tips; 24.5% were calyxes of Extra quality which were observed whole, with an intense red color and visually clean; the Supreme quality is an intermediate level and corresponded to 34.7%.

Identification of fungi associated with stored roselle calyxes. According to the cultural and morphological characterization, 16 genera of fungi were isolated and identified as associated with dried and stored roselle calyxes, their identity and accumulated frequency are shown in Figure 2. No significant difference was found in the frequency between Aspergillus when compared to Alternaria and Nodulisporium, though there was a significant difference between Aspergillus and the rest of the isolates.

Figura 2 Accumulated frequency and average of the 16 genera of fungi isolated in the different periods. 

^{Crous et al. (2009)} mentioned that the most common genera that could be found dispersed in the air were Alternaria, Aureobasidium, Aspergillus, Cladosporium, Eurotium, Fusarium, Mucor, Penicillium, Phoma, Rhizopus, Scopulariopsis, Stachybotrys, Ulocladium and Wallemia. On the other hand, in a study carried out by ^{Essien et al. (2013)}, the following were frequently isolated from the spore atmosphere: Curvularia, Drechslera, Nigrospora, Pithomyces and Stemphylium. The aforementioned supposes that some of the isolated genera in this investigation could be regularly found in the area of study. Nevertheless, some of the genera have also been isolated from fresh calyxes yet to be harvested, which supports the theory that some of the fungi are carried from the field.

The sequencing of the amplicons confirmed the identification of the isolated organisms through a BLAST that produced a ≥97% similarity with the studies done by other authors and reported in the GenBank, with the exception of Neosartorya which was 88%. Through the characteristics of easy recognition, Rhyzopus and Penicillium were identified only at the genus level based on their cultural and morphological characteristics. Table 1 shows the identity percentage with the BLAST of 14 isolated genera of dried and stored roselle calyxes.

Table 1 Location in the GenBank of 14 genera of fungi obtained from dried calyxes stored in the collection centers in Ayutla and Las Mesas. 

Two species of Chaetonium and two species of Fusarium were identified. It is important to mention that phylogenetic studies show that Drechslera has been accepted within the species Curvalaria and Bipolaris, therefore a synonym of C. australiensis and B. australiensis, respectively (^{Manamgoda et al., 2012}; ^{Deng et al., 2014}).

The most frequently isolated genus was Aspergillus and according to ^{Hedayati et al. (2007)} and ^{Adebayo-tayo and Samuel (2009)}, the spores of Aspergillus could easily contaminate the product during the drying process of the calyxes, with relative humidity being the most important variable, this is attributed to its high frequency and again the post-harvest management plays an important role in the contamination of the product. Various authors have reported the presence of A. flavus, A. niger, A. terreus, A. glaucus, A. candidus, A. fumigatus, A. ochraceus, A. rubrum, A. wentii and yeast in beverages based on roselle extract (^{Owusu and Odamtten, 1999}; ^{Adebayo-tayo and Samuel, 2009}), therefore it has not been ruled out that its presence could stem from storage practices.

Some organisms isolated also reported by ^{Owusu and Odamtten (1999)} in evaluating the microflora of tea of Ghanaian herbs, composed of dried leafs of Cinnamon, Hibiscus and Citronella, where 16 fungi species isolated in Hibiscus belonging to the genera Aspergillus, Cladosporium, Eurotium, Manoascus, Paecilomyces, Penicillium, Septodochium, Rhizopus and Syncephalastrum. They also determined samples in the field infected with A. candidus, A. fumigatus, A. ochraceus, vgtC. cladosporioides, C. herbario, Penicillium digitatum, Syncephalastrum spp. with the capacity to infect samples of dried leaves in storage. Furthermore, ^{Omemu et al. (2006)} found a total count of 3.2x104 UFC/ml of fungi on dried roselle calyxes, isolating Saccharomyces cerevisiae, Candida krusei, Rhyzopus oligosporum, Mucor spp., A. flavus and Penicillium citrinum. Other authors have isolated A. niger, A. fumigatus, A. flavus, Trichoderma sp., Rhizopus stolonifera, Penicilliumcitrinum, Saccharomycescerevisiae and Geotrichum in beverages based on roselle calyxes (^{Ojokoh et al., 2002}; Doughari and Elmahmood, 2008; ^{Nwafor and Ikenebomeh, 2009}). The most notable consequence of Aspergillus is the possible production and accumulation of mycotoxins. The presence of aflatoxins produced by A. flavus and A. parsiticus and ochratoxins produced by A. niger, A. ochraceus and A. carbonarius are frequently found in agricultural products (^{Perrone et al., 2007}). ^{Battilani et al. (2003)} reported A. japonicas as the producer of ochratoxin A (OA) in the cultivation of grapes; however, the aforementioned was not corroborated in the studies carried out by ^{Parenicová et al. (2001)}, ^{Samson et al. (2004)} and ^{Perrone et al. (2007)}. In this work, only the identification of the fungi associated with dried and stored roselle calyxes was determined, therefore the presence of any toxin in the product cannot be asserted as a consequence of the presence of the isolated fungi.

Determination of the incidence of diseased calyxes. In comparing the average incidence of diseased calyxes between warehouses, no significant difference was found between Ayutla (60.5 %) and Las Mesas (48.6 %) (DMS=14.99); however, when compared by date of collection, in the month of May 2013, a greater average value (71.11 %) was obtained compared to the months of December 2012 (42.04%) and February 2013 (53.79 %) (DMS=16.5).

Meanwhile in Ayutla, no significant difference was found comparing the three dates of collection in the same warehouse; in Las Mesas a greater incidence was observed in May 2013 when compared to December 2012. When comparing the two warehouses, in December 2012 a greater incidence was observed in Ayutla. The aforementioned is attributed to the fact that in the first months of harvest, in the warehouse in Ayutla, roselle was stored in its different quality grades, meanwhile, in Las Mesas only product with Extra and Supreme quality was received where no calyxes with necrosing were allowed. However, for the months of April onward, in which there was a product shortage, Las Mesas also received Commercial quality, and consequently there was the presence of calyxes with spots, matching the incidence between the two warehouses.

The incidence of diseased calyxes was determined in a nominal manner (diseased or healthy), therefore a calyx of Extra quality could be considered as diseased in the same manner as a Commercial quality calyx, even when the former barely shows a small spot. The aforementioned is unreliable at the time of determining the quality of the final product, given that the collector only bases on the visual appearance and the selection criteria varies between collectors. The Commercial grade showed a greater incidence of diseased calyxes (67.66%) compared to the quality of greater economical price, Extra (44.02%) (DMS=17.276). The aforementioned is attributed to the number of samples collected by quality grade, that is to say, it was more frequent to find calyxes with spots and/ or smut on the tips (40.8%) than healthy calyxes (24.5%). The average number of genera of fungi isolated in calyxes with different quality grades did not show a significant difference, this is related to the visual criteria of the collectors at the time of classifying the Commercial qualities of roselle and the aforementioned consideration.

The statistical analysis, with a 5% level of significance, did not indicate differences between warehouses, in the number of isolated genera (11.6 for both warehouses, DMS=4.719) and their frequency (9.9 % and 5.8 % for Ayutla and Las Mesas, respectively, DMS=4.7405); however, by increasing the level of significance to 10 % a significant difference was found in the frequency of the isolated genera between Ayutla and Las Mesas (DMS=3.6399). This difference is attributed to the quality grades received in each warehouse and to the post-harvest management between the same. Given that in Ayutla it was common to observe the reception of Commercial quality calyxes, storage was carried out for a greater amount of time in bulk hosts, sharing the same space for the different quality grades. In conjunction with this, the humidification and the constant shoveling of the product favors that the spores released are easily distributed in the air, and according to ^{Omemu et al. (2005)} and ^{Adebayo-tayo and Samuel (2009)}, that could fall on the product to damage it afterwards.

Evaluation of the storage conditions. Regarding the percentage of the average humidity content of calyxes, no significant difference was determined between the samples collected in Las Mesas (15.0 %) and Ayutla (14.5 %) (DMS=2.0656). Figure 3 shows the humidity determined for the 49 processed samples and their location regarding the limits established by the NMX-FF- 115-SCF-2010. 81.63% of the samples reported a humidity content greater than 12%, a condition that favors the development of fungi during storage (^{Augustburger, 2000}); in addition, ^{Gadinaru et al. (2003)} and ^{Galicia et al. (2008)} mention that this factor also affects the stability of the color of the calyxes, given that it increases the reaction velocities of the degradation of Anthocyanins, a component responsible for the coloration of the calyxes.

Figure 3 Humidity of the stored roselle calyxes; the sequence of the samples corresponds to the collection order. The dotted lines correspond to the values, minimum, medium and maximum established by the NMX-FF-115-SCFI-2010. 

When comparing the humidity content of the calyxes by date of collection, it was significantly determined that in both warehouses the more humid calyxes were collected in the months of February and May 2013 (Figure 4).

Figure 4 Average humidity of the calyxes in different sampling periods. ¹Averages with the same letter inside the line and within the bars are statistically the same (DMS between warehouses, P<0.05=3.3361; DMS average, P<0.05=2.2185). 

The variations in the percentage of humidity in the product between the different sampling dates are due to the post-harvest practices that are carried out in each warehouse; generally, the collection of the calyxes is carried out in the months of November and December, shortly after the producer has dried and packaged them. Meanwhile, in the hotter months the product is already in storage and sprayed with water in order to avoid it breaking and losing weight, which causes the calyxes to reabsorb humidity (^{Juliani et al., 2009}). ^{Vallecillo and Gómez (2004)} mention that the storage conditions and the management of the product depend on the roselle not reabsorbing humidity from the ambience and favoring the growth of fungi, which results in it being necessary to design and evaluate practices that inhibit this condition. During the material gathering, the average temperature that was registered in Las Mesas was greater (27.6 ºC) than that of Ayutla (27.1 ºC) (DMS=0.162), while the relative humidity in Ayutla was greater (58.4 %) compared to Las Mesas (52.4 %) (DMS=0.8993). The luminosity of Las Mesas was greater (38.3 lux) than the luminosity of Ayutla (16.8 lux) (DMS=4.5752). The collection center of Ayutla is located at an altitude of 391 m above sea level and the warehouse of Las Mesas at 432 m above sea level.

^{Augustburger et al. (2000)} recommends storing the product packaged in polyethylene or polypropylene bags or cardboard boxes in spaces protected from the sun at maximum temperatures of 15-20 ºC and a maximum ambient humidity of 60% and under these conditions the product could be preserved for up to 12 to 18 months.

The storage conditions used in the zone of study differ from those recommended by ^{Augustburger et al. (2000)}; given that the product is mostly stored in bulk hosts at ambient temperature, which had an average registered value of up to 28.6 ºC, while the relative humidity registered values of up to 66.4 % in Ayutla. According to ^{Arauz (1998)}, the majority of the pathogenic fungi sporulate better in high relative humidity than at low levels. The variation in the luminosity was related to the conditions of the very facilities, the collection volume and the post- harvest practices that are used in each warehouse; however, the light influence on the development of the diseases, in particular in natural conditions, has a lesser importance than that of the temperature or the humidity.

A correlation analysis was carried out between the variables: temperature, relative humidity, luminosity, humidity of the calyxes, incidence of diseased calyxes, number of isolated genera and frequency of the isolated genera. The correlation coefficients of the storage variables of both warehouses are shown in Table 2.

Table 2 Correlation Coefficients between the storage variables evaluated in Ayutla and Las Mesas. 

The correlations in italic show a significant difference of P<0.05. Temp=Temperature, HR=Relative humidity, Lum=Luminosity, Hum-cal=Humidity of the calyxes, Inc=Incidence of diseased calyxes, Ngén=Number of isolated genera, Frec=Frequency of the isolated genera.

A rather high negative correlation was determined between the temperature and relative humidity. Similarly, ^{Arauz (1998)} mentions that an indirect influence of the temperature on the development of pathogens is its effect on humidity; as the temperature increases, the relative humidity decreases and the foliar surfaces dry faster.

On the other hand, the storage temperature and the humidity of the calyxes showed a low positive correlation, by increasing the temperature the humidity of the stored calyxes also increased, this is due to it decreasing the relative humidity, the calyxes continue to dry and the collectors spray the product with water during its stay at the collection centers in order to facilitate their management and avoid weight loss, therefore, as it has already been mentioned, the calyxes can reabsorb humidity (^{Juliani et al., 2009}).

The low positive correlation between the temperature and the incidence of diseased calyxes is attributed to the post-harvest practices, where it must be avoided for the product to reabsorb ambient humidity, which would favor the growth of fungi (^{Vallecillo and Gómez, 2004}) and therefore the incidence of diseased calyxes would increase.

By increasing the temperature, the luminosity of the storage increases. This moderate positive correlation depends on the design and/or post- harvest practices of each warehouse; however, the luminosity is a factor of lesser importance for the development of diseases.

The number of isolated genera showed a low negative correlation with its accumulated frequency, this is attributed to the visual classification of the qualities, which is carried out by the collectors at the time of receiving the roselle, and which is not based on scientific criteria; given that an apparently healthy calyx could be received, of Extra quality, and this same calyx could be superficially contaminated with spores found in the air.

Roselle is a non-perishable product, which requires certain instructions to conserve its natural state after harvest. The drying process of the calyxes must be fast in order to avoid mildew, and in this fashion the product is stored at a humidity below 12%, with adequate ventilation.

Conclusions

The average humidity content of the analyzed roselle samples was 14.5% above what is technically recommended by various authors, including what is established in the NMX-FF-115- SCF-2010; this favored the development of fungi. 81.63% of the samples exceeded the upper limit established by the accepted humidity content for its commercialization.

The incidence of diseased calyxes was greater in the month of May 2013 (71.11 %) compared to the months of December 2012 (42.04 %) and February 2013 (53.79 %). No significant difference was found in the incidence of diseased calyxes collected in the warehouses in Ayutla (60.5 %) and Las Mesas (48.6 %). Likewise, the Commercial quality calyxes showed greater incidence than the calyxes of Extra quality.

16 genera of fungi associated with the dry and stored roselle calyxes were isolated; the 5 most frequently isolated genera were Aspergillus, Alternaria, Nodulisporium, Chaetomium and Thielavia.

Considering α=0.1, a significant difference in the accumulated frequency of this fungi was determined between warehouses. The main cause of the difference is due to the management practices used in each warehouse.

The obtained results in this study allowed knowledge of the conditions under which the harvested and dried roselle are stored in Guerrero, which must be corrected to avoid favoring the presence of the identified organisms.