Imaging analysis of A. oryzae penetration into steamed rice.
Yamada-nishiki (sake rice) and Chiyo-nishiki (eating rice) were polished to 50% or 90% (removed 10% outside), and used to make koji, the steamed rice with A. oryzae mycelia (see Materials and methods), referred as Y90, C90 and C50, respectively (Fig. 1A). A. oryzae mycelia grew on and in the koji pellets. We observed surfaces and cross sections of koji pellets by a zoom microscopy (Fig. 1A, B). To evaluate the degree of invasive growth into the steamed rice, called ‘haze-komi’, we used the A. oryzae strain, in which histone H2B are labeled with GFP, to make koji. The koji was sliced in 30 mm sections using a cryomicrotome. The sections were observed by a fluorescent microscopy. High intensity of GFP signals covered the periphery of rice in the Y90, C90 and C50 koji (Fig. 1C). Moreover, we could detect each hypha with GFP signal at cellular level in the rice. We quantified the haze-komi by measuring how far hyphae penetrated from the koji surface in the Y90, C90 and C50 (Fig. 1D, Supplemental figure 1A). The hyphal lengths from the koji surface in C90 and Y90 were comparable, 371 + 60 and 311 + 38 mm, respectively (n = 20 hyphae in 3 independent koji). Notably, the hyphae in Y50 penetrated more deeply 1.4 - 1.6 times, 501 + 66 mm (n = 20 hyphae in 3 independent koji). There was no clear difference in the GFP intensities on the surface between the different types of rice.
A. oryzae penetration inter-rice cells and intra-rice cells.
In the koji sections, rice endosperm cells were observed in bright field and UV light irradiation as well due to the autofluorescence (Fig. 2A). GFP signal from A. oryzae often indicated similar patterns with rice cells, suggesting that hyphae often grow between rice cells, which is consistent with the previous report (32). In addition, higher magnification images showed hyphal growth inside of rice cells (Fig. 2A arrows, Supplemental figure 1B). Confocal microscopy imaging of the koji sections confirmed that hyphae grew inside of rice cells and frequent co-localization of fungal signal on outlines of rice cells (Fig. 2B, Movie 1 and 2, Supplemental figure 1C). The hyphal growth inter-rice cells and intra-rice cells were observed similarly in the Y90, C90 and C50.
We tested the chemical reagent TOMEI (see Materials and methods), that turns plant tissues transparent, for the koji. Confocal imaging visualized network-like GFP signal from A. oryzae co-localized with the arrangement of rice cells (Fig. 2C).
SEM analysis of A. oryzae penetration intra-rice cells.
We observed the cross sections of koji by fluorescent microscopy and found that some hyphae grew through surrounding space like a furrow in the rice (Fig. 3A, arrow and dotted line). Scanning Electron Microscopy (SEM) also indicated the hyphae in furrows on the rice cross sections (Fig. 3B, white arrows), although the shapes of rice cells were not clearly observed. The SEM imaging showed that some hyphae came from or went into a hole on the rice cross sections (Fig. 3B, yellow arrows). The holes and furrows appeared to be tunnels formed by the sugar degradation during the hyphal growth in rice cells, which is in agreement with the previous report (33).
Time-lapse imaging of A. oryzae penetration into steamed rice.
To monitor the time course of A. oryzae growth into the steamed rice, we applied fluorescent live imaging for the cross sections of koji Y50 and Y90 (Fig. 4). The conidia were inoculated on the surface of rice, then the koji was incubated for 7 hours. Z-stack images of the cross sections were taken every 10 or 20 min for 14 hours. The Z-stack merged images were shown by time-lapse movies (Movie 3, 4). We could visualize A. oryzae hyphae (green) penetration into the steamed rice (red), which is termed haze-komi, by live imaging (Fig. 4). In Y50, hyphae grew from the surface of rice towards the center of rice (Fig. 4A, B). While some hyphae grew through the rice cell shape, others changed the growth direction when they bumped against rice cells (arrows). The hyphae appeared to grow on the outside rice surface following the growth into the rice. In Y90, hyphal signal increased under the surface of rice (Fig. 4C, D). In contrast to Y50, most hyphae did not continue to penetrate towards the center of rice but grew close to the surface, mainly ~300 mm, with more branching than in Y50.
X-ray CT analysis of A. oryzae penetration into steamed rice.
To complement fluorescence microscopy results and obtain more accurate information on mycelial penetration into the steamed rice, we performed a X-ray CT (Computed Tomography) scan analysis (see methods). The intact C90, Y90 and Y50 incubated for 48 hours were set in the X-ray CT device, respectively. The X-ray CT scan produces cross-sectional tomographic images by use of computer-processed combinations of many X-ray measurements taken from different angles, allowing to observe the inside of objects without cutting. The 3D section images were shown by sequence images (Fig. 5A). The fungal signals were determined by the different peak found in CT value (X-ray absorption) line profiles between the koji (rice + A. oryzae) and the rice without the fungus (Fig. 5B). The rice and fungal mycelia are shown in white and yellow, respectively (Fig. 5A, Movie 5-7). In the C90 and Y90, the fungal signals were detected mainly close to the surface of rice. In the Y50, in contrast, the signals were detected both close to the surface and inside of rice.
The rice and fungal volumes were calculated from the 3D data. The ratios of fungal volume per koji, rice + fungus, volumes were indicated in C90, Y90 and Y50 (Fig. 5C). The fungal ratios in the C90 and Y90 were 0.44 + 0.02, 0.32 + 0.03, respectively, while the fungal ratio in the Y50 was 0.67 + 0.1 and significantly higher than those in the C90 and Y90, (n = 3, p < 0.001). The X-ray CT scan analysis also supports deeper invasive growth in the Y50 than the C90 and Y90.
Transcriptome analysis
We compared the transcriptome profiles of A. oryzae in Y50, Y90 and C90 by RNA-seq analysis. The effects of the variety of rice and the polishing rate on the growth of A. oryzae, enzyme production, and metabolism production in sake-koji have been investigated previously (34). One of the most important roles of A. oryzae in rice koji is the supply of enzymes, vitamins, nutrition, such as glucose, amino acids and peptides, that are necessary for sake brewing. From the viewpoint, we compared the expression of genes (Supplemental Table 1). Especially, the comparison between Y50 and Y90 was summarized in Table 1. Digestion of starch and supply of glucose by amylases are the basis of alcohol fermentation. The expression of genes for a-amylases (amyA, amyB and amyC) and glucoamylase (glaA) was 5.2 and 1.7 times higher in Y50 than in Y90, whereas the expression level of maltases, which hydrolyse maltose to glucose, was comparable. The metabolic genes in glycolysis, TCA cycle and electron transport chain were compared by a heatmap (Supplemental figure 2).
Acid proteases are involved in the digestion of the main protein of rice glutelin, also called oryzanin. The enzyme breaks down the protein body containing glutelin, resulting in disruption of the rice structure. Carboxypeptidases degrade peptides and supply amino acids. The expression of major acid protease gene, pepA, was 13 times lower in Y50 than that in Y90, whereas the expression level of carboxypeptidase genes was almost unchanged.
The outer surface of rice, aleurone layer, is rich in lipids and fatty acids, and thier contents decreases as the polishing rate decreases (35). When the polishing rate decreases, the ratio of saturated fatty acids and unsaturated fatty acids changes (35). In sake brewing, fatty acids are important in the production of yeast-derived aroma components (ginjo aroma, especially ethyl caproate). When the unsaturated fatty acid content increases, the production of ethyl caproate in yeast is suppressed. The secreted lipase is necessary for supplying lipids and fatty acids to yeast. The expression of fatty acid synthase genes, fasA and fasB, was 5-7 times higher in Y50 than those in Y90. Additionally, the expression of sdeA and sdeB genes for delta-9-stearic acid desaturase, which converts palmitic acid and stearic acid to palmitoleic acid and oleic acid, respectively, was 13-20 times higher in Y50 than Y90, whereas the expression level of lipase genes was almost unchanged.
Supply of phosphate affects the following yeast fermentation (36). Phytic acid is a preserved state of phosphate in plants, and its content decreases as the rice polishing rate decreases (35). Phytases function to release phosphate from phytic acid (37). The expression of phytases, acid phosphatases and alkaline phosphatases tend to increase in Y50 compared to Y90.
Most of sake yeasts lack some of the genes related to vitamin biosynthesis. In addition, enzymes involved in fermentation require vitamins as cofactor. Supply of vitamins from koji is essential to proceed with fermentation (38). Vitamins are abundant in the outer surface layer and germ of rice, and their amounts decrease as the rice polishing rate decreases (39). The expression of synthesis genes for thiamine, pantothenate and biotin (Vitamin B1, 5 and 7, respectively) increased in Y50 compared to Y90.
Beside the genes related to brewing and fermentation, genes for conidiophore development, flbA-D and brlA (40), were up-regulated in Y50 compared to Y90, whereas abaA, which is required for phialide differentiation, was unchanged. In A. oryzae, flbC was reported to regulate the expression of genes specifically under solid-state cultivation conditions, possibly independent of the conidiation regulatory network (41).
Increase number of nuclei in A. oryzae hyphae
We observed nuclei labeled with GFP of A. oryzae in the koji and found that the number of nuclei often varied in each hypha (Supplemental figure 3). Since the increase of nuclei in A. oryzae was predicted to be correlated to the high secretion capacity of several enzymes, we focused on the phenotype. The nuclear distribution in A. oryzae has been analyzed in hyphae and especially in conidia (42, 43), which indicated multi-nuclear conidia; the number of nuclei in each conidium varied from 1 to 7 in A. oryzae strains used in sake brewing. We investigated the nuclear distribution in A. oryzae hyphae grown in detail by using the minimal medium but not the rice koji. Some of hyphae contained less than 20 nuclei in the tip compartments, the hyphal cell from the tip to the first septum (Fig. 6A, upper). Other hyphae contained more than 200 nuclei in the tip compartments (Fig. 6A, lower). The hypha containing such a high number of nuclei was imaged by the Z-stack confocal microscopy and shown in 3D imaging (Fig. 6B, Movie 8).
We classified the pattern of nuclear distribution into three types as follows. Class I; nuclei distribute at a constant interval without overlapping. Class II; nuclei align but sometimes overlap. Class III; nuclei scatter over through hyphae but not align. We counted the ratios of class I-III in the time course at 24-, 48- and 72-hours growth (Fig. 6C). At 24 hours, class II was large, while class I and III were approximately 20%. At 48 and 72 hours, class III increased to 60% and more than 70%, respectively. The class III hyphae were usually thicker than those of class I. We measured the hyphal width at tip compartments at 24-, 48- and 72-hours of growth (Fig. 6D). At 24 hours, the hyphal width was 3 to 6 mm. At 48 hours, the ratio of 7 to 10-mm hyphae increased, then at 72 hours, the hyphal width ranged 3 to 12 mm.
As a comparison, we investigated the nuclear distribution in the model fungus Aspergillus nidulans as well in the same way. The ratios of class I and II did not vary approximately 40% and 60%, respectively, in the time course at 24-, 48- and 72-hours (Fig. 6C). The hyphal widths were comparable at 24 and 48 hours, then the peak shifted by 2 mm wider at 72 hours (Fig. 6D). These results indicate that A. oryzae increases the nuclear number and hyphal width in the time course of 1-3 days, which may correlate with the high secretory capacity of several enzymes.
Synchronous mitosis in A. oryzae
To analyze the mechanism of the increase in nuclear number, we investigated the nuclear distribution in mitosis. Synchronous nuclear division in a hyphal compartment has been known in A. nidulans (44-46). We used the A. nidulans strain expressing NLS of the TF StuA tagged with GFP (45,47). The GFP protein localizes in nuclei in interphase, while they move out from the nuclei to cytoplasm due to partial disassembly of nuclear pore complex during closed mitosis (Fig. 7A, Movie 9) (46, 48). The nuclear membrane envelope is intact but permeable, known as partially open mitosis. After mitosis, GFP signals moved back in the two-fold number of nuclei.
We investigated the nuclear distribution in A. oryzae mitosis, where histone H2B labeled with GFP remained in nuclei during mitosis. Time-lapse imaging revealed the synchronized mitosis within 5 min in the tip compartment (Fig. 7B, Movie 10). Even in the class III hypha, a lot of nuclei divided within 5 min in the tip compartment (Fig. 7C, Movie 11). We performed Z-stack and time-lapse of the class III hyphal mitosis and revealed the synchronized mitosis, although the time and space resolutions were not sufficient to demonstrate all nuclei enter mitosis at same time, Movie 11 showing that the H2B-GFP signals undergo condensation simultaneously suggests a synchronous nuclear division (49). The nuclei moved a lot after mitosis in A. oryzae which is consistent with that in A. nidulans (48). Since the size of the H2B-GFP signal did not show significant difference from 24 to 72 hours, 2.5 + 0.3 and 2.5 + 0.2 mm (n=10 from Movie 10 and 11), the increase in nuclear number is not due to fragmentation of nuclei.
The nuclear distribution in class III hypha of A. oryzae resembles that of the another model fungus Neurospora crassa (Fig. 7D, Movie 12, 13), whose mitosis is not clear to be synchronous or not (50).