Ammonia, ammonium, and fungal growth
Laboratory-based experiments compared the presence or absence of ammonia/ammonium within the phloem medium with the growth performance of L. procerum. Ammonia/ammonium significantly promoted the growth of L. procerum. The spore and mycelium densities of L. procerum revealed that ammonia/ammonium exposure significantly increased the growth of L. procerum (Fig. 1a, b). The fungal growth rate was 109.0%, and 107.0% compared with the control at doses of 1.56 mol/L ammonia and 0.019 mol/L ammonium, respectively (Fig. 1c, d) (df = 35.241, t = −17.330, P < 0.001; df = 38, t = −12.403, P < 0.001 t-test). The activity on fungal growth by ammonia (1.56 mol/L) and ammonium (0.019 mol/L) was determined using the fungal biomass yield method. After 13 days, the fungal biomass yield with ammonia/ammonium (14.58 mg [dry weight] petri−1/27.30 mg [dry weight] petri−1) was significantly greater than that with sterile water (1.73 mg [dry weight] petri−1/4.60 mg [dry weight] petri−1) (Fig. 1e, f) (df = 7.223, t = 14.748, P < 0.001; df = 9.32, t = 13.213, P < 0.001 t-test).
Ammonia, ammonium, and D. valens larval growth
We observed no significant difference in weight change of RTB larvae in phloem media with or without ammonia (1.56 mol/L)/ammonium (0.019 mol/L) (Fig. 2a) (F2, 66 = 2.163, P = 0.124 ANOVA, Dunnett’s test). Bioassays using the phloem medium infused with 1.56 mol/L ammonia or 1 g/L ammonium characterized the effects of ammonia and ammonium on RTB larval growth by L. procerum. RTB larval weight decreased significantly on L. procerum-colonized phloem media compared to fungus-free phloem media. The weight of RTB larvae fed with phloem media colonized by L. procerum with ammonia (1.56 mol/L)/ammonium (0.019 mol/L) presented were significantly higher compared to those with sterile water. In addition, no significant difference was noted between ammonia and ammonium treatments (Fig. 2b) (F3, 80 = 24.865, P < 0.0001 ANOVA, Dunnett’s test).
Ammonium and carbohydrate consumption
Previously, we determined that ammonia can regulate L. procerum consumption sequence of d-pinitol and d-glucose in phloem medium but it did not affect the consumption of other carbohydrates . Here, we verified the same result that ammonium can regulate the L. procerum consumption of d-pinitol and d-glucose using desugared phloem medium (adding 1 g/L of d-glucose and d-pinitol) (Supplementary Fig. S1) . Compared to the untreated control cultures, the marked reduction in d-pinitol began earlier than d-glucose. d-glucose started to decline significantly on the 5th day in plates colonized by L. procerum without the ammonium, and was depleted on the 11th day. However, the content of d-pinitol stayed constant before the 7th day and 38.65% remained on the 11th day (Supplementary Fig. S1a) (d-pinitol: F9, 70 = 462.667, P < 0.001 ANOVA, Tukey’s test; d-glucose: F9, 70 = 1057.377, P < 0.001 ANOVA, Dunnett’s T3 test). In presence of ammonium, d-pinitol in plates colonized by L. procerum started to decline significantly on the 5th day and was used up on the 11th day, and 51.25% d-glucose remained in plates on the 11th day (Supplementary Fig. S1b) (d-pinitol: F9, 70 = 652.753, P < 0.001; d-glucose: F9, 70 = 97.468, P < 0.001 ANOVA, Dunnett’s T3 test). Ammonia and ammonium reduced inhibition of L. procerum on RTB larval growth and increased fungal growth by regulating d-glucose and d-pinitol consumption in the fungus (Figs. 1 and 2). Obviously, d-glucose was a readily available and preferred carbon source for L. procerum compared to d-pinitol . In order to further confirm the relationship between the carbohydrates consumption by the fungus and the fungal growth in the presence of ammonium, we analyzed the carbohydrate composition of the desugared phloem medium (adding 1 g/L, d-glucose and d-pinitol) in L. procerum-colonized area as the fungus grew over time. Without the ammonium, the d-glucose and d-pinitol consumption by the fungus was similar to the whole desugared phloem medium and fungi-colonized area of the desugared phloem medium (Fig. 3a) (d-pinitol: F9, 70 = 628.211, P < 0.001; d-glucose: F9, 70 = 760.623, P < 0.001 ANOVA, Dunnett’s T3 test). With the presence of ammonium, d-pinitol also decreased significantly over time in plates colonized by L. procerum. On the other hand, the content of d-glucose in L. procerum-colonized area initially decreased sharply and then increased to its peak on day 5, and then decreased with time (Fig. 3b) (d-pinitol: F9, 70 = 564.548, P < 0.001 ANOVA, Dunnett’s T3 test; d-glucose: F9, 70 = 103.886, P < 0.001 ANOVA, Tukey’s test). Given these findings, we initially speculated that ammonia and ammonium do not affect d-glucose consumption by L. procerum and d-glucose left in the medium that originated from conversion from other substances.
13C6-labeled glucose treatment
The mass spectra of d-glucose and d-glucose-13C6 standards were shown in Supplementary Fig. S2. On the basis of the mass spectra for d-glucose and d-glucose-13C6, we decided to consider the ions m/z 319 and m/z 323, which have large mass numbers and relatively high intensities, as monitor ions. Without ammonium, there was no difference in the concentration of natural glucose between fungus-free and the fungus-colonized medium, and the concentration of labeled glucose significantly decreased in fungus-colonized medium when compared to the fungus-free medium (Fig. 4a, b) (m/z 319: df = 14, t = 1.110, P = 0.293; m/z 323: df = 5.686, t = 22.951, P < 0.001 t-test). In the presence of ammonium, the concentration of natural glucose increased and the labeled glucose was depleted on the fungus-colonized medium compared to the fungus-free medium (Fig. 4b, c) (m/z 319: df = 11, t = −8.309, P < 0.001; m/z 323: df = 7.035, t = 114.821, P < 0.001 t-test). Compared to the control, labeled glucose was also consumed faster by the fungus in the presence of ammonium (Fig. 4). Even with desugared phloem medium only supplied with d-pinitol (1 g/L), the production of glucose was detected in L. procerum-colonized phloem media in the presence of ammonium (Supplementary Fig. S3a, b). However, when ammonium was absent, no glucose was found in L. procerum-colonized phloem media (Supplementary Fig. S3a, b). Furthermore, in the L. procerum-colonized minimal medium (glucose-free medium with only 1 g/L d-pinitol added), no glucose was produced in either the presence or absence of ammonium (Supplementary Fig. S3c).
The preceding results indicated that ammonium exposure caused the conversion of phloem substances to glucose in L. procerum. Thus, we subsequently analyzed the content of glucogenesis-related substances in the phloem. It is well known that pine phloem is rich in cellulose . Other carbohydrates related to glucogenesis in Pinus tabuliformis phloem were extracted and analyzed. Starch and sucrose were the only carbohydrates related to glucogenesis detected in the phloem (Table 1). All sample preparations from the phloem (311.87 ± 17.08 mg/g, DW), the phloem medium (236.72 ± 7.03 mg/g, DW) and the desugared phloem (258.61 ± 4.46 mg/g, DW) contained relatively high percentages of starch. Very small amounts of sucrose were present in the sample preparations from the phloem (1.65 ± 0.06 mg/g, DW) and the phloem medium (2.20 ± 0.13 mg/g, DW), and it was not detected in the preparations from the desugared phloem.
No glucose was produced in the L. procerum-colonized minimal medium (adding 1 g/L d-pinitol) despite the presence or absence of ammonium (Supplementary Fig. S3c). However, in the minimal medium supplemented with 1 g/L d-pinitol and 20 g/L starch, glucose was detected only in the L. procerum-colonized minimal medium with the addition of starch in the presence of ammonium (Fig. 5a, b) (d-glucose: df = 6, t = −18.630, P < 0.001 t-test). In addition, after introducing two other glucogenesis-related carbohydrates (cellulose and sucrose) to the minimal medium (adding 1 g/L d-pinitol), there was still no glucose detected in the minimal medium in the presence of ammonium (Supplementary Fig. S4). Further investigation revealed that L. procerum and RTB larvae performed better on minimal media with glucose than that on minimal media with only starch (Supplementary Fig. S5) (df = 42, t = 2.957, P = 0.008 t-test). All of these results indicated that starch is the only substance in phloem available for the production of glucose, and L. procerum is able to convert the nutrient-poor but abundant carbohydrate (starch) into a highly nutritional carbon source (glucose).
The bacterial volatile ammonia, as a nitrogen source, induces the L. procerum starch metabolism pathway to convert starch to high nutritional-value glucose
As evidenced that starch is responsible for the production of glucose by L. procerum, we then used next generation sequencing to determine which gene was involved in this process (Fig. 6a). Genes for hydrolyzing starch to glucose/digestion of starch, including α-amylase (AMYA3), which is involved in the breakdown of long chain carbohydrates, amyloglucanase (AMYG) and α-glucosidase (α-glu1 and α-glu5), which are involved in breaking down starch to the monosaccharide glucose, were upregulated by the exposure of ammonium (Fig. 6b). Other genes related to cellulose and sucrose metabolism showed no response to ammonium exposure (Supplementary Fig. S6). qRT-PCR analyses also found that exposure to ammonium significantly induced the expression of AMYA3, AMYG and α-glu5 (Fig. 6c) (AMYG: df = 4, t = −8.402, P < 0.001; AMYA3: df = 4, t = −4.693, P < 0.001; α-glu5: df = 4, t = −5.618, P = 0.005; α-glu6: df = 4, t = 0.288, P = 0.787; α-glu1: df = 4, t = −1.247, P = 0.280 t-test).
Besides the expression of these genes, we also measured the enzymatic activity of α-amylase, amyloglucanase, and α-glucosidase by double antibody sandwich methods. Compared to the control, ammonium-treated L. procerum showed higher enzymatic activities of α-amylase (df = 4, t = −45.369, P < 0.001 t-test), amyloglucanase (df = 4, t = −30.845, P < 0.001 t-test) and α-glucosidase (df = 4, t = −285.833, P < 0.001 t-test) (Fig. 6d). Taken together, these outcomes strongly indicated that the bacterial volatile ammonia mediated the conversion of starch to glucose via the activation of starch hydrolyzes-related enzymes.
Deletion of amylolytic enzyme genes causes a growth defect of L. procerum and RTB larvae
Given these findings, we subsequently generated ΔAMYG and ΔAMYA3 mutants of L. procerum by homologous recombination. Carbohydrate composition in the desugared phloem medium (adding 1 g/L, d-glucose and d-pinitol) showed that, when exposed to ammonium, the glucose left in the culture medium colonized by ΔAMYG mutant was significantly reduced compared to wild-type (WT), while deletion of AMYA3 was not (Fig. 7a) (AMYG: df = 7.426, t = 10.572, P < 0.001; AMYA3: df = 14, t = 4.248, P = 0.180 t-test). We found that loss of AMYG led to growth defects of L. procerum (Fig. 7b, c) (b: df = 14, t = 12.648, P < 0.001; c: df = 14, t = 4.075, P = 0.003 t-test), and weight loss of RTB larvae (Fig. 7d) (df = 43, t = 3.681, P = 0.001 t-test). The morphology of ΔAMYG mutant is also different from WT (Supplementary Fig. S7).
The inhibition of glucose sensing/metabolism and starch metabolism pathway caused by ΔSUC1 mutant
Having confirmed the conversion of starch to glucose, we then investigated which factors regulate this conversion. A transcription factor deletion of COL-26 and its homology gene has been reported to have a critical role in utilization of starch and integration of carbon and nitrogen metabolism in Neurospora crassa and Trichoderma reesei . In this work, the homology gene of COL-26 was identified and named as SUC1 in L. procerum (Supplementary Fig. S8). When compared to WT strains, ΔSUC1 mutants showed growth defects in the presence of ammonium and desugared phloem medium, which also had an extra 1 g/L of d-glucose-13C6 and d-pinitol added (Supplementary Fig. S8a, b). To investigate the functions of SUC1 in this process, we analyzed the carbohydrate composition in the plates and evaluated transcriptional changes in the ΔSUC1 mutant when switched to desugared phloem medium containing 1 g/L of d-glucose-13C6 and d-pinitol under identical conditions as with the WT parent strain (see above). We found that loss of SUC1 has no effect on the content of pinitol and glucose in desugared phloem medium (adding 1 g/L of d-glucose-13C6 and d-pinitol) following fungus consumption in the presence of ammonium (Fig. 8a) (d-pinitol: df = 14, t = 1.87, P = 0.086; d-glucose: df = 14, t = −1.277, P = 0.226 t-test). Furthermore, glucose left in plates with ΔSUC1 mutant was mainly labeled glucose. In contrast, glucose left in plates with WT mainly was natural glucose (Fig. 8b and Supplementary S9) (m/z 319: F2, 21 = 228.376, P < 0.001; m/z 323: F2, 21 = 162.146, P < 0.001 ANOVA, Tukey’s test). Based on the transcription analysis, we found that the main genes in the glucose regulon gene set and starch regulon gene set were downregulated in the ΔSUC1 mutant (Fig. 8c) (PK: df = 4, t = 5.745, P = 0.005; PFK: df = 4, t = 6.450, P = 0.003; PGD: df = 4, t = 4.752, P = 0.009; G6PD: df = 4, t = 3.076, P = 0.037; AMYA3: df = 4, t = 14.602, P < 0.001; AMYG: df = 4, t = 4.983, P = 0.008; α-glu1: df = 4, t = 3.315, P = 0.030; α-glu5: df = 4, t = 0.085, P = 0.936 t-test), including two key genes of the glycolysis (6-phosphogluconate dehydrogenase: PGD and glucose 6‐phosphate dehydrogenase: G6PD), two key genes of the pentose phosphate pathway (pyruvate kinase: PK and phosphofructokinase: PFK), and three amylolytic genes (AMYA3, AMYG and α-glu1). In addition, deletion of SUC1 also impaired the growth of L. procerum (Fig. 8d, e) (d: df = 14, t = 10.25, P < 0.001; e: df = 14, t = 5.785, P < 0.001 t-test) and RTB larvae (Fig. 8f) (df = 33.878, t = 3.276, P = 0.002 t-test) significantly. The morphology of ΔSUC1 mutant was different from WT as well (Supplementary Fig. S7). These data indicated that deletion of the SUC1 repressed the glucose sensing/metabolism and starch metabolism pathway in L. procerum with ammonium, which therefore disrupted the beneficial association of L. procerum and RTB larvae.