Morphological evolution of various fungal species in the presence and absence of aluminum oxide microparticles: Comparative and quantitative insights into microparticle‐enhanced cultivation (MPEC)

Abstract The application of microparticle‐enhanced cultivation (MPEC) is an attractive method to control mycelial morphology, and thus enhance the production of metabolites and enzymes in the submerged cultivations of filamentous fungi. Unfortunately, most literature data deals with the spore‐agglomerating species like aspergilli. Therefore, the detailed quantitative study of the morphological evolution of four different fungal species (Aspergillus terreus, Penicillium rubens, Chaetomium globosum, and Mucor racemosus) based on the digital analysis of microscopic images was presented in this paper. In accordance with the current knowledge, these species exhibit different mechanisms of agglomerates formation. The standard submerged shake flask cultivations (as a reference) and MPEC involving 10 μm aluminum oxide microparticles (6 g·L−1) were performed. The morphological parameters, including mean projected area, elongation, roughness, and morphology number were determined for the mycelial objects within the first 24 hr of growth. It occurred that heretofore observed and widely discussed effect of microparticles on fungi, namely the decrease in pellet size, was not observed for the species whose pellet formation mechanism is different from spore agglomeration. In the MPEC, C. globosum developed core‐shell pellets, and M. racemosus, a nonagglomerative species, formed the relatively larger, compared to standard cultures, pellets with distinct cores.


| INTRODUCTION
Filamentous fungi are of central relevance to biotechnological industry due to their prolific biosynthetic capabilities. Widely exploited for the production of valuable metabolites and enzymes, these organisms constitute the core of important large-scale biomanufacturing processes (Adrio & Demain, 2003;Demain & Martens, 2017). The development of cost-effective cultivation strategies is thus an essential prerequisite for harnessing the enormous biochemical potential of fungal strains. Furthermore, maintaining the economic feasibility of bio-based production is inevitably associated with the development of novel process-related methods that allow for the continuous improvements in terms of titer, yield, and productivity (Nielsen et al., 2017).
Several approaches have been devised to influence one of the key parameters associated with submerged cultivations of filamentous fungi, namely the morphology (Walisko, Moench-Tegeder, Blotenberg, Wucherpfennig, & Krull, 2015). Briefly, depending on a number of factors, the mycelial development in stirred or shaken liquid cultures proceeds in the forms of dispersed hyphae, clumps, or pellets (Papagianni, 2004). The latter can be either micropellets or macropellets whose diameter often exceeds even 1 mm. In filamentous fungi, three distinct mechanisms of pellets formation are distinguished, namely spores agglomeration (e.g., in Aspergillus), mycelial agglomeration (e.g., in Penicillium), and nonagglomerative pellet growth (e.g., in Zygomycetes). Since the optimal production of a particular fungal metabolite or enzyme is typically correlated with a specific morphological form, the biosynthesis of a target molecule can be greatly enhanced by controlling the morphology of the employed strain. It was previously reported that adding mineral microparticles (e.g., talc or aluminum oxide) to cultivation media has a profound effect on mycelial morphology in submerged cultures, and depending on the microbial platform and the molecule of interest, it can lead to the significant increase in product concentration via generating a favorable morphological form (Antecka, Blatkiewicz, Bizukojć, & Ledakowicz, 2016;Coban, Demirci, & Turhan, 2015a, 2015bCoban and Demirci, 2016;Driouch, Hänsch, Wucherpfennig, Krull, & Wittmann, 2012;Driouch, Roth, Dersch, & Wittmann, 2010;Driouch, Sommer, & Wittmann, 2010;Etschmann et al., 2015;Gao, Zeng, Yu, Dong, & Chen, 2014;Gonciarz & Bizukojć, 2014;Kaup, Ehrich, Pescheck, & Schrader, 2008;Yatmaz, Karahalil, Germec, Ilgin, & Turhan, 2016). The effectiveness of this flexible, simple, and relatively inexpensive approach, referred to as the microparticle-enhanced cultivation (MPEC), has been documented by several groups. For instance, the application of talc microparticles resulted in a dispersed morphology of Aspergillus niger SKAn 1015 and, consequently, a eightfold increase in βfructofuranosidase productivity was observed in fed-batch bioreactor cultures (Driouch, Roth, et al. 2010). A similar phenomenon was noted for the fungus Caldariomyces fumago DSM 1256 by Kaup et al. (2008), who recorded a 10-fold improvement in terms of chloroperoxidase formation in the talc-enriched medium. Although the aforementioned studies involved the use of the MPEC for loosening the mycelia and elevating enzyme levels, Gonciarz and Bizukojć (2014) demonstrated the use of talc for decreasing pellet diameter and increasing product titers in the lovastatin-oriented cultivation of Aspergillus terreus ATCC 20542. Other examples of MPEC-related studies were reviewed by , Krull et al. (2013), and Walisko et al. (2015). Currently, the MPEC is regarded as one of the leading modern methods of morphological engineering, a discipline defined by McIntyre, Müller, Dynesen, and Nielsen (2001) as "tailoring morphologies for specific bioprocesses." In a broad perspective, the significant rise of metabolite or enzyme concentration in the MPEC proceeds in parallel with the change in substrate consumption rates, broth viscosity, and oxygen transfer Antecka, Blatkiewicz, et al. 2016).
Despite its advantages and confirmed effectiveness, the mechanism responsible for the interaction between microparticles and the growing mycelia has not been elucidated. Moreover, no quantitative representation of the MPEC-related chain of events has been proposed so far. To the best of our knowledge, the only contribution in this respect was provided by Driouch, Sommer, et al. (2010), who verbally described the influence of talc microparticles on the agglomeration of A. niger SKAn1015 conidiospores in the submerged culture. According to those authors, the agglomeration is disturbed by microparticles in the initial phase of growth and, as a result, the pellets of reduced size or loose mycelia can be formed in the course of the cultivation (Driouch, Sommer, et al. 2010). Importantly, the morphology-related measurements essential for the quantitative considerations were not performed. In the previously published works, the influence of the MPEC on fungal morphology was mainly analyzed in the context of an individually selected species, whereas a more comprehensive perspective is still lacking. Furthermore, the MPEC was mostly applied toward spore-agglomerating aspergilli. Only Kaup et al. (2008) addressed the impact of microparticles on various filamentous microorganisms; however, the presented MPEC-related dataset was rather limited in scope and referred solely to pellet size as the parameter illustrating the effects of the MPEC.
In the current study, the microscopic-level comparison between the microparticle-enriched and standard cultures was conducted for four fungal species, selected on the basis of their growth characteristics, namely Mucor racemosus, Chaetomium globosum, Penicillium rubens, and A. terreus. Mucor racemosus is a dimorphic fungus representing the group of zygomycetes. Depending on culture conditions, it proliferates as budding yeast-like cells or branching hyphae and even if it forms pellets, the pellet formation mechanism is nonagglomerative (Orlowski, 1991). Chaetomium globosum propagation relies on the formation of asci in dark, hairy perithecia of characteristic appearance (Wang et al., 2016). Actually it is difficult to attribute any mechanism of pellet formation to this species, although spore agglomeration seems to be the most probable.
Penicillium rubens and A. terreus exhibit distinct mechanisms of pellet formation, namely the agglomeration of spores and mycelia, respectively (Grimm et al., 2004;Nielsen, 1996).
The efforts presented in this work were based on the size and shape parameters of mycelial objects computed within the framework of digital image analysis. The aim of the study was to provide a quantitative description of the influence of microparticles on fungi displaying different mechanisms of mycelial development in the submerged conditions.

| Sporulation media
The strains were maintained on slants prepared according to the recommendations of ATCC. For A. terreus, the slant medium contained malt extract (20 g·L −1 ) and casein peptone (5 g·L −1 ). Penicillium rubens spores were grown on the commercially available potato dextrose medium (BTL Ltd., Poland) containing potato extract (4 g·L −1 ) and glucose (20 g·L −1 ). For M. racemosus, the potato dextrose medium was prepared according to the following procedure: 300 g of potatoes was cooked in 500 ml water; potatoes were then discarded and the broth was filled up to 1 L with water; next, glucose (20 g·L −1 ) was added. The slant medium for C. globosum was prepared as follows: the commercially available rabbit food pellets (25 g) were boiled in 1 L of water and the filtrate was collected after 30 min of steeping. All the abovementioned media were solidified with agar. Spores suspended in sterile water were transferred onto new slants. Then, they were cultivated in the thermostated chamber at 26°C. After 10 days, they additionally remained for 3 days at ambient temperature and were later stored at 4°C. The slants for all studied fungal species were renewed every 2 weeks and the slants of the same age were used in each experiment.
For P. rubens ATCC 9178 (originally deposited in ATCC as Penicillium notatum), this medium was modified with regard to the carbon sources and contained lactose (7.5 g l −1 ) and glucose (7.5 g·L −1 ), as glucose is a preferable substrate in the early stages of P. rubens growth (Koffler, Emerson, Perlman, & Burris, 1945). Mucor racemosus ATCC 7924 was cultivated in preprepared Sabouraud medium (BTL Ltd, Poland) containing peptone (5 g·L −1 ), casein peptone (5 g·L −1 ), and glucose (20 g·L −1 ) (Faramarzi, Badiee, Tabatabaei, Amini, & Torshabi, 2008). All media were sterilized at 121°C for 30 min. It must be mentioned that upon the preliminary experiments, whose results are not presented here, it occurred impossible to apply the same medium composition for all tested strains due to the observed growth limitations.
For the MPEC experiments, 6 g·L −1 sterile (separately autoclaved at 121°C for 30 min as dry powder) microparticles of aluminum oxide (Al 2 O 3 ) of mean diameter equal to 10 μm were added to the cultivation media. Upon the previous experiments, this amount proved to be high enough to act on filamentous fungi evolution in the submerged culture (Gonciarz & Bizukojć, 2014). At the same time, it was crucial not to add too many microparticles that would aggravate image processing and analysis.

| Cultivation conditions
The shake flask culture of 150 ml working volume run in flatbottomed flasks was used in each experiment. Fungi were cultivated in the thermostated rotary shaker Certomat ® BS-1 at 28°C (B. Braun Biotech International, contemporarily Sartorius Stedim, Germany). The shaking speed was equal to 110 min −1 . The cultivations were started as follows. From each slant, the spores of the individual fungus were washed to the appropriate type (medium composition), and amount (dependent on the number of flasks to be prepared) of liquid medium and the spores were enumerated under a microscope (objective 40×) with the use of a Thoma chamber. In each experiment, the initial number of spores in the cultivation medium was maintained at the level of 10 9 L −1 . If required, the medium was diluted or more spores were added. For the MPEC runs, the appropriate amount of microparticles was added. Next, the prepared suspension was dispensed to shake flasks. The individual experimental run lasted for 24 hr. Samples were taken every 1, 2, or 3 hr depending on the phase of growth from 5 hr until the end of the experiment. Each MPEC run was accompanied by the standard cultivation, which served as the reference. For each fungus, at least three independent experiments were performed.

| Image processing and analysis
The samples from each experiment (standard and MPEC runs conducted in parallel) were subjected to the immediate (no sample storage was applied) microscopic observations. The light microscope OLYMPUS BX53 was used for this purpose (Olympus Corporation, Japan). The microscope was equipped with the high-resolution RGB digital camera OLYMPUS DP27 and controlled by the computer with the image analysis software OLYMPUS cellSens Dimension Desktop 1.16 (Olympus Corporation, Japan). Due to fact that the size of the observed mycelial objects changed by several magnitudes, the variety of objectives: 1.25×, 2.5×, 4×, 10×, 20×, 40×, and 100× had to be used. The viable slides were prepared by dropping approximately 10 μl of the fungal suspension and observed using phase contrast. At least 40 RGB images of resolution 2,448 × 1,920 were snapped to assure the minimum number of mycelial objects for analysis ( Figure S1).
These images were next subjected to the semiautomatic (actions of the operator incidentally required) image processing and analysis procedures, which were programmed in the macrolanguage of the used software. Their steps were as follows: 1. Filtration of the image with the use of a median filter to smooth the edges of the objects. The median filter does not change either the size or the shape of the objects.

2.
Edge detection with the use of a Sobel filter.

3.
Image segmentation based on the enhanced edges of the objects to detect the valid mycelial objects in the green plane of the RGB image. The use of the edge detector made the processing and analysis of all images independent of the colors and shades, which were present in the images because of rapidly changing mycelial objects due to their fast growth and consequent changes of objectives (from 40× to 1.25×) to assure the valid magnification.

Calculation of the morphological parameters (both size and shape)
of the detected mycelial objects: -projected area (A) calculated as the pixel count in the given object multiplied by squared calibration unit.
-mean diameter (D) being the distance of two boundary points on a line through the centre of gravity.
-elongation (E) defined as the squared quotient of longitudinal and transversal deviation of all pixels belonging to the object along the regression line. If E = 1, then the object is ideally circular. The higher the elongation, the more different is the shape of the object from the circle; ultimately, it may become a thin line.
-roughness (R), sometimes called solidity or convexity, calculated as the ratio of projected area to convex area. If R = 1, then the object is ideally smooth and convex without any protrusions on its boundary. The values of roughness decrease, when there are any irregularities (like growing out filaments) on its boundary and the object becomes hairy.
-Morphology number (Wucherpfennig, Hestler, & Krull, 2011) calculated from: where: A -mean projected area, S -solidity (roughness), Eelongation, and D max -maximum object diameter. The value of morphology number close to 1 and at the same time not lower than 0.6 is found for the circular objects like ungerminated spores and ideal pellets (pelleted morphology with few clumps and filaments). If the pellets are irregular or filamentous morphology is evolved, the morphology number is lower than 0.5 (Wucherpfennig et al., 2011).

5.
Removal of any debris and invalid objects with the use of various size and shape filters: in the images with spores, the removal of noncircular objects upon the value of circularity and in the ones with large pellets the removal too small objects (upon projected area) were required. Some debris had to be manually removed by the operator in this step too. In some images, the areas of interest (AOIs) were manually established to improve the detection of valid mycelial objects and avoid debris originating mainly from microparticles.

6.
Transfer of image analysis data into a spreadsheet file.
Additionally, such parameters as the diameter of pellet core or the ratio of core diameter to pellet diameter were measured manually in the selected images.
The parameters of the analyzed mycelial objects were subjected to statistical analysis. The mean values, standard deviation, and confidence band upon t-Student distribution (α = 0.05) were determined.
Specifically, in some cases upon mean projected area, two classes of the mycelial objects (smaller and larger ones, regarding their mean projected area) were distinguished by introducing a threshold value ( Figure S2). Details are provided in Section 3.

| RESULTS
The quantitative description of spore-to-mycelium evolution of four fungal species was presented. The analysis was performed on the basis of microscopic images, examples of which are shown in Figures S3-S6 for the standard cultivation and in Figures S7-S10 for the MPEC. After 8 hr of cultivation, the mean projected area of small agglomerates and large agglomerates reached 1.3 × 10 3 μm 2 and 4.2 × 10 4 μm 2 , respectively ( Figure 1).

| Growth of Aspergillus terreus in the standard and microparticle-enhanced cultivations
First germ tubes emerged from the external regions (edges) of the agglomerates not earlier than at 10 hr of the standard cultivation, initiating the next stage of fungal growth, that is, the germination of spores ( Figure S3f). The emergence of germ tubes from the agglomerated Development of Aspergillus terreus in the standard cultivation (left panels) and MPEC (right panels) quantified by morphological parameters spores clearly confirmed the agglomerative (aggregative) mechanism of pellet formation in A. terreus. The morphology number dropped to 0.37, the elongation was equal to 1.78, and roughness decreased to 0.77 (Figure 1), so the shape of mycelial objects differed from the ideal circle ( Figure S3f).
Within the next hours of A. terreus mycelial evolution in the standard cultivation, the germ tubes got longer and their transformation into hyphae took place. This was reflected by the decrease in morphology number to the lowest observed level of 0.35 in 12 hr of cultivation ( Figure 1). Evolved hyphae covered the cores of the agglomerates and radially grew toward the cultivation medium ( Figure S3g). Then, all the agglomerates achieved more regular (more circular) shapes. Almost all objects began to resemble circular or ellipsoid pellets after 15 hr of cultivation ( Figure S3h). Due to the change in shape, the elongation decreased to 1.94. At this time, the morphology number increased to 0.46. The mean projected area of small agglomerates and large agglomerates was equal to 1.1 × 10 3 μm 2 and 2.1 × 10 5 μm 2 , respectively ( Figure 1).
In 17 hr of the growth in the standard cultivation, A. terreus ultimately formed regular hairy pellets with cores of circular or ellipsoid shape and the morphology number of 0.56. Roughness had then its smallest value equal to 0.67, which showed that many filaments grew out of the agglomerates forming a hairy-like structure around the core of the pellets ( Figure S3i). Ultimately, hairy pellets with small cores of regular round shapes and similar size were achieved at the end of the standard cultivation after 24 hr of growth ( Figure S3j). The elongation took the value of 1.07 (close to a number corresponding to a circle) and thus returned to the level recorded at the beginning of the experiment. Roughness increased to 0.81 (still quite hairy objects were observed) and the morphology number reached 0.67. The mean projected area was equal to 2.4 × 10 6 μm 2 and the two classes of objects of different sizes were not observed any more ( Figure 1).
In the case of A. terreus growth in the MPEC, many differences were seen from 8 hr of the run compared to the standard cultivation, which is proved by the microscopic images ( Figure S7a-j) and the values of morphological parameters (Figure 1). Similarly as in the standard cultivation, two classes of mycelial objects: the ones larger than 10 4 μm 2 (large objects) and the ones smaller than 10 4 μm 2 (small objects) were also dis- and was much larger than in the standard cultivation (rarely exceeding 1 × 10 3 μm 2 ). On the other hand, the mean projected area of large objects in the MPEC was not higher than 1.6 × 10 5 μm 2 , while in the standard cultivation it exceeded 7 × 10 5 μm 2 (compare Figures S3f and S7f).
In 10 hr of the MPEC, the mean projected area (calculated for all objects) reached 4.6 × 10 3 μm 2 , and this value was significantly lower (almost by two magnitudes) than the mean projected area in the standard cultivation, which was equal to 1.3 × 10 5 μm 2 (Figure 1).
Between 8 and 12 hr, the objects in the MPEC were more circular in comparison with those in the standard cultivation ( Figure S7f). This was indicated by the values of elongation (much lower than 2.00) and morphology number (higher than 0.40) in the MPEC (Figure 1 where the pellets were far more hairy and with longer filaments growing out of the pellet core. The ratio of filaments to diameter was 0.33, while in the standard cultivation it was only 0.21. Most microparticles were embedded in the structure of the pellets ( Figure S7j); however, the core-shell pellets were not observed. In contrast, the pellets having from 1 to 4 cores (multicore pellets) were observed in the MPEC ( Figure S7j). It was another important difference with respect to the standard cultivation. This may indicate that, due to the presence of microparticles, the evenly evolved pellets (the ones with the core formed) had the tendency to loosely agglomerate with each other. This agglomeration was probably taking place with the contribution of long filaments protruding from the pellets. In sum, the mean projected area of MPEC pellets was smaller (even in the case of multicore pellets) in comparison with the standard cultivation. In 24 hr of the run, the elongation parameter, after certain fluctuations, reached 1.35. This value was higher than the one recorded in the standard cultivation, what demonstrates that MPEC pellets were also less circular ( Figure S7j).
The morphology number was equal to 0.44 and was lower than in the standard cultivation. The addition of Al 2 O 3 microparticles in the MPEC also changed the structure of A. terreus pellets, which was reflected by the value of roughness (0.64). Pellets in the standard cultivation were smoother, as their roughness was at the level of 0.81 (Figure 1).
In addition, few dispersed mycelia, namely branched and unbranched hyphae and clumps, were visible ( Figure S7j).

| Growth of Penicillium rubens in the standard and microparticle-enhanced cultivations
In contrast to A. terreus, the studied cultures of P. rubens were initiated from slightly elongated spores. Their mean projected area was Roughness was almost the same in both cultivations, being equal to about 0.80 in both runs. The morphology number was equal to 0.57 and was lower in the MPEC than in the standard cultivation.
In 5 hr of the standard cultivation, this parameter was equal to 0.49 ( Figure 2).
Similarly as in the standard run, two classes of objects, namely the ones with area exceeding 10 4 μm 2 (large objects) and the ones with area lower than 10 4 μm 2 (small objects) were distinguished ( Figure S2).
However, in the MPEC, they were found 1 hr earlier (in 7 hr).
The first regular circular agglomerates were observed in 10 hr The morphology number in the MPEC was not as high as in the standard cultivation (Figure 2).
The first oval hairy pellets with the distinct cores were observed in 15 hr of the MPEC, 2 hr earlier than in the standard cultivation ( Figure   S8h). Moreover, at this stage, many branched, unbranched hyphae, and clumps were seen ( Figure S8h). Accordingly, the highest elonga- In 24 hr of the MPEC, besides dispersed hyphae, hairy pellets of nonideal ellipsoid shape were developed ( Figure S8i). The morphology number took the value of 0.18. Furthermore, pellets formed in the MPEC were similar in structure and shape to the pellets grown without microparticles. With regard to size, it is impossible to unequivocally state which of them were bigger. Taking mean projected area of objects of both classes into account, the objects from the MPEC were astonishingly bigger (A = 1.9 × 10 5 μm 2 ) that the ones from the standard culture (A = 1.7 × 10 5 μm 2 ). However, if large objects (pellets) were taken into account (Figure 2), they were smaller in the MPEC (A = 3.7 × 10 6 μm 2 vs. A = 2.3 × 10 6 μm 2 ). Small objects from the standard culture were smaller (A = 1.8 × 10 3 μm 2 ) than those from the MPEC (2.5 × 10 3 μm 2 ). So, it seemed that the action of microparticles depended on the size of objects. It led to decreased size of larger objects and increased size of smaller ones. The similar effect was also noticed for A. terreus (see Section 3.2.1). Microparticles somewhat stabilized the agglomerates in P. rubens. In 24 hr of the MPEC, the roughness parameter reached 0.61 and was lower than in the standard cultivation (0.7), which meant that the objects were more hairy and probably of more loose structure ( Figure S8i).

| Growth of Chaetomium globosum in the standard and microparticle-enhanced cultivations
The standard and MPECs of C. globosum were started from large, elongated, lemon-shaped spores ( Figure S5a) Figure S2). In 12 hr of the standard cultivation of C. globosum, the first regular pellets of circular shape were observed ( Figure S5f).
Their mean projected area reached 3.9 × 10 5 μm 2 , the elongation was equal to 1.29, and the morphology number raised to 0.36 (Figure 3).
Roughness increased steadily up to 0.58. Within the next 4 hr of the cultivation, the mean projected area was successively rising due to hyphal agglomeration and proliferation. The pellets grew and acquired a more and more circular shape, what is seen in Figure S5g Pellets that were ultimately formed in 24 hr of C. globosum MPEC were described as the core-shell pellets covered with long filaments ( Figure S9j). The corresponding morphology number was equal to 0.45 ( Figure 3). Interestingly, this value was almost identical for the standard cultivation. It indicated that the main effect of microparticles was associated with the formation of the mineral core in the pellets. The cores of the pellets from the standard cultivation were less distinct ( Figure S5j). The hyphae forming the pellet were uniformly distributed. In contrast, the pellets of A. terreus had visibly denser hyphae in their centers. (Figure S3j). C. globosum pellets from the MPEC were almost of the same size (A = 2.6 × 10 6 μm 2 ) as in the standard cultivation (A = 2.9 × 10 6 μm 2 ). The elongation in the MPEC reached the value of 1.25 and was higher than in the standard cultivation (1.18).
Roughness in the MPEC was ultimately equal to 0.67 and this value did not significantly differ from the one determined for the standard culture ( Figure 3).

| Growth of Mucor racemosus in standard and microparticle-enhanced cultivations
The spores of M. racemosus, from which the standard and microparticles-enhanced cultivations started, were ellipsoid (Figures F I G U R E 4 Development of Mucor racemosus in the standard cultivation (left panels) and MPEC (right panels) quantified by morphological parameters S6a and S10a) and their sizes varied from 2 to 6 μm 2 (the mean diameter equal to 2 μm). The morphology number, equal to 0.71, was lower than the one observed for A. terreus and P. rubens. The elongation parameter reached the value 1.36, which was higher compared to A. terreus and P. rubens. Roughness was equal to 0.97 (Figure 4).
The first stage of M. racemosus growth in the standard cultivation comprised the rapid process of spores swelling leading to the change in spore shape into almost ideal circles with the values of morphology number, elongation, and roughness equal to 0.89, 1.06, and 0.97, respectively. Within 6.5 hr of the experiment, the spore size increased by two orders of magnitude exceeding 200 μm 2 ( Figure S6b).
After 6.5 hr of the standard cultivation, the next stage of M. racemosus growth began, namely the germination of spores ( Figure S6c).  Figure S10i). It was the most significant outcome of microparticles addition.
The first stage of M. racemosus proliferation, namely spore swelling, was slightly slower in the MPEC in comparison with the standard cultivation. Within first 6.5 hr of the MPEC, the mean projected area of spores reached only 190 μm 2 , while in the standard cultivation it exceeded 200 μm 2 (Figure 4). In the MPEC, the elongation reached its minimal value equal to 1.17, which corresponded to the most spherical shape observed for M. racemosus objects. The elongation in the MPEC was also lower than in the standard cultivation (the minimum was achieved earlier in 5 hr). At this point, the spores in the standard culture already began to elongate ( Figure S6b).
The formation of M. racemosus agglomerates in the MPEC started in 10 hr of growth. Two classes of objects were distinguished ( Figure   S2). The sizes of these classes were completely different from the ones observed for the standard culture. It was the only case out of all studied cultivations that the threshold value was at the level of 10 5 μm 2 , not 10 4 μm 2 ( Figure S2). The mean projected area for the large agglomerates was equal to 7.5 × 10 4 μm 2 in the standard cultivation and 2.9 × 10 5 μm 2 in the MPEC (Figure 4). Numerous pellets with distinct cores, which were observed in 24 hr of M. racemosus MPEC, had rather irregular shapes ( Figure S10i).
This was reflected by the value of morphology number equal to 0.53. It is worth mentioning that the morphology number in the standard cultivation was much lower than 0.5, indicating the presence of dispersed mycelium rather than the pelleted morphology. The elongation in the MPEC of M. racemosus reached 2.04 and was higher than these values observed for A. terreus, C. globosum or P. rubens. The mean projected area of large agglomerates in the MPEC cultivation was equal to 9 × 10 5 μm 2 , whereas in the standard cultivation this parameter had a lower value of 2 × 10 5 μm 2 (Figure 4). Roughness in the MPEC cultivation was also higher than in the standard cultivation and reached 0.83.
Again, the observed effect of microparticles was far from typical in this case, as one would normally expect the size of agglomerates and pellets to be lower in the MPEC than in the standard submerged cultures.

| DISCUSSION
All the experiments carried out in this study focused only on the mycelial evolution, and interactions of microparticles and mycelia of various fungal species. As the most important events of fungal evolution took place during the first 24 hr of growth in the preculture, fungal products, either secondary metabolites or enzymes, were not detectable in the broth in the end of each experiment. It is the reason why product concentrations were not determined in this study, despite the well-known fact that the MPEC is used to increase fungal productivity.
With regard to the behavior of the tested fungal species in the studied submerged cultures, only the results obtained for A. terreus were clearly in accordance with the mechanism of pellet formation by spore agglomeration described in literature (Bizukojc & Ledakowicz, 2010;Metz & Kossen, 1977). Two stages of pellets formation were distinguished. In the first stage, two size classes of agglomerates (area above and under 10 4 μm 2 ) were observed. In the second stage, all objects agglomerated, creating round shaped, hairy pellets. A similar two-stage agglomeration process was previously recorded for A. niger.
In the first stage of the corresponding study, 10% of spores present in the cultivation medium agglomerated. Afterwards, the spores germinated and entered the second stage of agglomeration. As a result, over 90% of the A. niger spores agglomerated into pellets (Grimm et al., 2004).
The agglomeration started after germination of spores and resulted in hardly spherical pellets, much different from those formed by C. globosum or A. terreus (Nielsen, Johansen, Jacobsen, Krabben, & Villadsen, 1995). Two classes of objects were distinguished.
However, unlike for A. terreus, these two classes were observed in the broth until the end of the 24-hr cultivation period. Ultimately, the dispersed hyphae dominated over the pellets. This type of morphology was somewhat a mixture of large pellets and small hyphal elements.
Although C. globosum formed fairly elegant circular hairy pellets in the end of the experiment, its morphological evolution more resembled the hyphal agglomeration scheme typical for penicilli.
Nongerminated spores hardly agglomerated with each other but had the affinity to the remnants of hairy perithecia. In fact, they often germinated before agglomeration. Some of them germinated inside the perithecia. Furthermore, their germ tubes emerged much earlier that the ones in A. terreus. The germinated spores with elongating hyphae formed the hairy pellets. Their shape was similar to the one observed for A. terreus, but completely different from the shape of P. rubens pellets. It is difficult to discuss these observations with literature as no references concerning the early stages of C. globosum growth have been published so far.
Last but not least, a nonagglomerative lower fungus M. racemosus ultimately formed a limited number of pellets; however, they were much smaller and, above all, looser than those of A. terreus, C. globosum and P. rubens. These pellets were also formed from a much lower number of germinated spores (hyphal elements) than those of the spore-or hyphae-agglomerative fungi. The mechanism of M. racemosus pellet formation was similar to the one described for an actinobacterium Streptomyces tendae. According to Vecht-Lifshitz, Magdassi, and Braun (1990), S. tendae exhibited a nonagglomerative mechanism of pellets formation. The agglomeration of mycelium occurred if the number of spores in the inoculum exceeded 10 6 L −1 . As a result, pellets containing not more than 300 germinated spores were formed and this number is at least by two magnitudes lower than for agglomerative fungal species.
In previous literature reports, the effect of mineral microparticles used in the MPEC has actually been attributed to the decrease in the size of agglomerates (pellets) and change in their structure toward the looser one (Krull et al., 2013). However, the majority of experiments involved the spore-agglomerating species (mostly Aspergillus) and any discussions or data regarding other species are hardly found (Kaup et al., 2008).  (Gao et al., 2014;Coban and Demirci, 2016). For Mortierella isabellina, unlike M. racemosus, the formation of small pellets and the decrease in their size due to the presence of microparticles was observed (Gao et al., 2014). The case of Rhizopus oryzae was more interesting and complicated. This nonagglomerating fungus grew as bulk, dispersed mycelium, but in the certain range (around 10 g·L −1 ) of talc or aluminum oxide microparticles the formation of pellets took place (similarly as for M. racemosus tested here). Interestingly, with the further increase in microparticles concentration the culture returned to a more dispersed morphology (Coban and Demirci, 2016). Nevertheless, in light of previous literature findings, the stabilizing effect of microparticles on pellets is plausible.
For example, this mechanism was previously observed for agglomerative A. terreus growing under high shear stress in the bioreactor culture (Gonciarz et al., 2016).
In sum, when applying microparticles one must be aware that their action toward filamentous fungi is species-dependent. It is also crucial to understand the consequences of morphological changes with regard to aeration, mixing and mass transfer conditions and resulting increase (or undesired decrease) of the productivity of the culture. The tests presented here were performed solely for precultures and it was clearly demonstrated that microparticles contributed to the development of a particular morphological form. No matter whether the agglomerates are formed or not or if their size increases or decreases, the application of microparticles is fully justified as long as the resulting morphological form is favorable in terms of bioprocess performance.
The correlation between morphological form and the production of metabolites and enzymes needs to be studied individually for each bioproduct and species. This individual approach to the MPEC was also previously emphasized by Etschmann et al. (2015).
Despite the need for the aforementioned individual approach and a lack of products formed in the precultures studied here, one may speculate on the optimum fungal morphology of tested species and plausible positive outcomes of the use of these precultures as the inocula. Generally, the fact that the pellets are smaller and looser makes them more productive with regard to secondary metabolites and enzymes. It is the case of lovastatin producer A. terreus (smaller and less dense pellets) and cellulolytic enzymes producer C. globosum (smaller core-shell pellets with thin mycelial layer around the core). Diffusion of oxygen in such pellets is more effective, as it was many times shown in literature (increase in lovastatin production, Gonciarz et al., 2016), (core-shell pellets favoring β-fructofuranosidase production, Driouch et al., 2012). The formation of secondary metabolites and hydrolytic enzymes is catabolism-dependent and requires significant supplies of energy and thus oxygen. With regard to the penicillin-producer P. rubens, the structure of pellets was stabilized, although the size of the pellets was slightly altered. More stable pellets might decrease the viscosity of broths and increase the convective oxygen transfer.
Formation of micropellets (not macropellets) of little diffusion resistance by M. racemosus may be favorable for the overall run of the growth process, as the convective oxygen transfer for highly dispersed and, at the same time, highly viscous fungal suspensions is weak. Less viscous micropellets suspension will increase convective oxygen transfer in the bioreactor. Mucor racemosus is not a profound producer of metabolites (it was tested here as the representative of lower fungi); however, it does biotransform steroids in the oxygen-dependent reactions (Faramarzi et al., 2008) and the changes in morphology observed in the present work might be advantageous in this context.
Ultimately, the findings from the presented experiments may prove valuable for those who work with the same fungal genera, leaving alone the fact that the precultures thoroughly studied here are going to be used in our future experiments as the inocula for the larger scale bioreactor processes.

| CONCLUSIONS
Upon the performed experiments, three general conclusions can be drawn: • The action of microparticles in the MPEC seems to be dependent on the size of mycelial objects.
• Microparticles may accelerate and favor agglomeration of mycelium, as the spores and small mycelial objects (like free-branched and unbranched hyphae) have the affinity to microparticles, what ultimately results in a kind of scaffold for the construction of the agglomerates.
• Unlike in the case of small mycelial objects, microparticles make the agglomeration of the larger mycelial objects more difficult, sometimes destroying the agglomerates, what leads to the decrease in pellet size.
Regarding the individual species one can conclude that: • As A. terreus is a typical spore-agglomerative fungal species, microparticles addition caused the decrease in pellets size and changed the internal pellet structure. It also contributed to prolonging the agglomeration of small pellets, resulting in the formation of multicore pellets.
• In the case of P. rubens, a typical hyphal agglomerative species, microparticles addition accelerated the agglomeration of hyphae but hardly influenced the size of the evolved pellets.
• In spite of the fact that the mechanism of pellet formation in C. globosum cannot be categorized as spore-agglomerative or hyphal agglomerative, the effect of microparticles was in this case similar to the one observed for spore-agglomerative A. terreus. Microparticles addition accelerated the agglomeration stage, decreased the size of the pellets, and also led to the development of core-shell pellets.
• The significant effect of microparticles on M. racemosus was observed. This nonagglomerative species formed the noticeable number of pellets with distinct cores due to the presence of microparticles.