Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2026 May 7;17:4105. doi: 10.1038/s41467-026-72617-z

Glucose-dependent acidification strategy by nectar-dwelling bacteria mediates pollen tube burst

Yoshinobu Kato 1,2,✉, Hiroki Miura 1, Seiji Takayama 1, Sota Fujii 1,3,✉
PMCID: PMC13153243  PMID: 42098134

Abstract

Floral nectary gland serves as an ecological hub, including a habitat for microorganisms introduced by pollinators. Although nectar supplies abundant carbon sources, it is limited in nitrogen sources. To address this unbalanced nutritional condition, it is considered that some nectar-dwelling bacteria including those belong to genus Acinetobacter obtain nutrients from pollen protoplasm by inducing pollen tube burst. Here, using gas chromatography-mass spectrometry, we identify gluconic acid in A. nectaris and A. pollinis as the substance responsible for triggering pollen tube burst via acidification of the surrounding environments. Through whole-genome sequencing and functional analysis, we identify a membrane-bound pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenase which produces gluconic acid. Notably, nectar-dwelling Acinetobacter species consistently retain PQQ biosynthetic genes, and expression of the PQQ precursor peptide gene are upregulated in the presence of glucose, suggesting that these bacteria can synthesize PQQ and gluconic acid immediately upon entering the nectary gland. Taken together with previous findings, this glucose-dependent acidification is beneficial for both efficient carbon source utilization and nutrient landscape modification, a bacterial adaptive strategy well-suited to the competitive nectar environment.

Subject terms: Bacterial secretion, Bacterial genomics

Nectar bacteria turn floral sugar into a weapon: Acinetobacter converts glucose to gluconic acid, acidifies nectar, and triggers pollen tube burst, revealing a nectar-adapted strategy coupling glucose oxidation to acidification and nutrient access.

Introduction

Floral nectary gland plays a pivotal role in ecological networks by attracting pollinators such as insects and birds for effective pollen transfer, while offering nutritional rewards to these visitors1. During these interactions, microorganisms are also transferred into floral nectaries, and nectar becomes a habitat for microbial life2. It has long been known that diverse yeast populations inhabit nectaries3. Recently, extensive exploration and isolation of nectar-dwelling bacteria have revealed a more diverse microbial community4,5. Thus, floral nectaries not only support broader ecological systems but also harbor unique microbial communities with their own internal ecological dynamics. Nectar-dwelling microorganisms also alter the chemical composition of floral nectar, producing various effects such as changes in pollinator behaviors2.

Nectar-dwelling microorganisms exhibit a distinctive lifestyle. Floral nectaries are assumed to be initially sterile prior to anthesis, and microorganisms are primarily introduced through pollinator visits2. Once established, these microbes can proliferate rapidly in the nectar environment, often reaching densities of 105–107 cells mL−1 and occasionally as high as 109 cells mL−1 6–8. In floral nectaries, the concept of “priority effects”—where the order and initial abundance of arriving species influence community assembly and interactions—has been proposed2. Therefore, rapid microbial growth is considered critical for successful and widespread colonization of floral nectaries. In this sense, floral nectaries function as highly competitive ecological arenas.

The nectar environment is highly carbon-rich but nitrogen-limited9, raising an intriguing question of how microorganisms acquire essential nitrogen sources for their rapid growth10,11. A previous study reported that pollen grains are frequently transferred into floral nectaries, and that some pollen grains and pollen tubes rupture, releasing their protoplasm11. In another study, co-cultivation experiments with nectar-derived Acinetobacter species and Eschscholzia californica pollen demonstrated that these bacteria can induce pollen germination and bursting of pollen tube tips to promote their own proliferation12. These studies suggest that nectar-dwelling Acinetobacter species gain access to nutrients including nitrogen released from pollen which is protected by a recalcitrant outer shell13.

This study focused on the mechanism by which nectar-dwelling Acinetobacter species trigger pollen tube burst. To clarify this mechanism—which alters the nutrient landscape within floral nectaries—and to address an underexplored aspect of plant-microbe interactions in nectar ecosystems, we aimed to identify the substance secreted by these bacteria that induces pollen tube rupture. We showed that gluconic acid, which is produced by a PQQ-dependent glucose dehydrogenase enzyme, triggers pollen tube burst by acidifying the environment. This rapid glucose-dependent acidification helps bacteria both efficiently use carbon sources and reshape the nutrient landscape, giving them a competitive edge in the nectar environment.

Results

Glucose-dependent acidification by nectar-dwelling Acinetobacter nectaris triggers the pollen tube burst

In this study, we isolated a bacterial strain belonging to the Acinetobacter genus from the nectaries of Abelia grandiflora flowers. Through partial amplification and sequencing of 16S ribosomal RNA followed by a BLASTn search, this bacterium was identified as A. nectaris (Supplementary Fig. 1), a species previously reported not to induce pollen tube burst activity in the previous study12. This assignment was further supported by whole-genome phylogenomic analysis (see below). We tested whether our isolated A. nectaris strain possesses pollen tube burst activity. A. nectaris was incubated in the induction medium containing 1% (w/v) sucrose for 6 h. The filter-sterilized medium after incubation was then mixed with germinating Petunia hybrida pollen. In contrast to the previous study12, the supernatant after incubating A. nectaris isolated in this study showed clear pollen tube burst activity, resulting in the release of protoplasm (Fig. 1a, e). The induction medium without A. nectaris incubation did not cause pollen tube bursting (Fig. 1a, e). We observed that the pH of the induction medium was markedly decreased to ~3.0 after incubation with A. nectaris (Fig. 1f). This acidic condition may have caused the pollen tube burst. To test this possibility, the induction medium after A. nectaris incubation was neutralized to pH 6.9–7.1 with KOH prior to treating the pollen. This neutralization suppressed the bursting activity of pollen tubes in the cultured supernatant (Fig. 1a, e), suggesting that some organic acid produced by A. nectaris triggers the pollen tube burst. Despite the inclusion of 1 mM PIPES buffer (pH 6.5) in the induction medium, the pH drop was observed, suggesting that a sufficient amount of acid was produced to overcome this buffered condition.

Fig. 1. Gluconic acid production depending on glucose in A. nectaris triggers pollen tube burst of Petunia hybrida.

Fig. 1

Open in a new tab

a–d Representative images of pollen tube burst assay. Supernatants from A. nectaris cultured with sucrose (a), no sugar (b), fructose (c), or glucose (d) were added to germinating Petunia hybrida pollen. Asterisks indicate ruptured pollen grains. Scale bars = 200 µm. e The rates of ruptured pollen tubes per germinated pollen. Data are mean ± SD. For each supernatant (n = 3), at least 30 pollen tubes were counted in each replicate. S, -, F, and G indicate sucrose, no sugar, fructose, and glucose, respectively. P-values determined by two-sided Dunnett’s test in each carbon source were 1.39 × 10-5, 1.64 × 10-5, 0.63, 0.25, 7.35;× 10-6, and 1.02 × 10-5, from left to right. Statistical significance is indicated by asterisks: n.s., P ≥ 0.05; ***P < 0.001 (f) pH of the induction medium containing sucrose, fructose, glucose, or no sugar after A. nectaris incubation. pH of the induction medium without incubation was also shown. Data are mean ± SD. All groups consisted of n = 6. S, G, F, and - indicate sucrose, glucose, fructose, and no sugar, respectively. Different letters indicate significant difference (P < 0.001) compared with the other conditions after A. nectaris incubation, as assessed by two-sided Tukey’s HSD test. g GC-MS analysis of the induction medium containing no sugar, sucrose, fructose, or glucose after A. nectaris incubation. Total ion current chromatogram is shown. The numbered peaks were confirmed with reference standards according to their retention time and mass spectrum patterns in Supplementary Figs. 2–5. I.S. indicates the internal standard. Experiments were repeated three times with similar results, and the figure presents a representative dataset. Note that the retention time in no sugar was slightly different from the other samples because it was analyzed on a different date. Arrowheads at 17.44 and 19.24 min indicate estimated peaks originating from derivatives of palmitic acid and stearic acid, respectively, which were likely contaminants introduced during the experimental procedure. Source data are provided as a Source Data file.

Among the components of the induction medium, we considered that 1% (w/v) sucrose (29 mM) is the most likely source of organic acid production. To verify this hypothesis, we tested pollen tube burst activity using induction media containing other sugars. Whereas no sugar and fructose did not induce pollen tube burst activity or pH drop (Fig. 1b, c, e, f), the medium containing glucose induced both (Fig. 1d, e, f). Furthermore, the decrease of pH was more severe in the glucose medium than in the sucrose medium (Fig. 1f), suggesting that glucose is a precursor for this organic acid. To identify the organic acid, we analyzed the induction medium containing sucrose, glucose, or fructose after A. nectaris incubation using gas chromatography-mass spectrometry (GC-MS) (Fig. 1g). The numbered peaks were confirmed by comparing the retention time and the mass spectrum pattern with the standard reference (Supplementary Figs. 2–5). Peaks at 17.44 min and 19.24 min were estimated as derivatives of palmitic acid and stearic acid, respectively, by NIST library search (Fig. 1g). The same peaks were detected even in the sucrose reference standard (Supplementary Fig. 3), indicating that they were likely contaminants introduced during the experimental procedure in GC-MS analysis. In sucrose- or glucose-containing medium, peaks originated from gluconic acid was detected (Fig. 1g: peaks 3, 4, and 6). It is noted that gluconic acid readily dehydrates and converts into the cyclic ester glucono-delta-lactone and glucono-gamma-lactone14, and these derivatives were also detected as peaks 3 and 4. The retention time and the mass spectrum pattern of peaks 3, 4, and 6 matched those of the reference standard for gluconic acid (Supplementary Fig. 2). Although the retention time of the peak 4 was also similar to that of the reference standard for glucose, their MS spectrum patterns were different (Supplementary Fig. 5). In the no-sugar or fructose-containing medium, peaks 3, 4, and 6 were not detected (Fig. 1g), consistent with the absence of pollen tube burst activity and the pH drop (Fig. 1b, c, e, f). To further confirm the effect of gluconic acid on the pollen tube burst, gluconic acid was added exogenously to germinated pollen. A dose-dependent effect on pollen tube burst became significant at pH values below 3.8 (Supplementary Fig. 6a, b). These results indicate that the substance causing the pollen tube burst is gluconic acid.

We also tested whether the pollen tube burst could be induced in other plant species. In Abelia grandiflora, the species from which A. nectaris was isolated in this study, and in Arabidopsis thaliana, the pollen tube burst was more significantly induced in the glucose-containing induction medium compared to the fructose-containing induction medium, in which gluconic acid was not produced (Supplementary Fig. 7a–d). This result suggests that the acidic condition produced by A. nectaris can induce pollen tube burst across a broad range of plant species.

Glucose-dependent acidification is conserved in other Acinetobacter species

In the previous study, the pollen tube burst activity was well confirmed in A. pollinis SCC47712. We also tested whether this species also generates gluconic acid from glucose. Similar to A. nectaris isolated in this study, pollen tube burst activity and a pH drop were confirmed in the sucrose- or glucose-containing medium but not in the no-sugar or fructose-containing medium (Fig. 2a–c). The peaks originated from gluconic acid (peaks 3, 4, and 6) were also detected in GC-MS analysis of the glucose- or sucrose-containing medium after incubation (Fig. 2d). These results suggest that the same strategy for inducing pollen tube burst is conserved in other nectar-dwelling Acinetobacter species.

Fig. 2. Gluconic acid production depending on glucose in A. pollinis SCC477 triggers pollen tube burst of Petunia hybrida.

Fig. 2

Open in a new tab

a Representative images of pollen tube burst assay. Supernatants from A. pollinis SCC477 cultured with sucrose (a), no sugar (b), fructose (c), or glucose (d) were added to germinating Petunia hybrida pollen. Asterisks indicate ruptured pollen grains. Scale bars = 200 µm. b The rates of ruptured pollen tubes per germinated pollen. Data are mean ± SD. For each culture supernatant (n = 3), at least 30 germinated pollen tubes were counted in each replicate. S, -, F, and G indicate sucrose, no sugar, fructose, and glucose, respectively. P-values determined by two-sided Dunnett’s test in each carbon source were 5.62 × 10-3, 6.58 × 10-3, 0.26, 0.61, 1.93 × 10-4, and 3.88 × 10-4, from left to right. Statistical significance is indicated by asterisks: n.s., P ≥ 0.05; **P < 0.01; ***P < 0.001 (c) pH of the induction medium containing sucrose, fructose, glucose, or no sugar after incubation. Data are mean ± SD. All groups consisted of n = 6. S, G, F, and - indicate sucrose, glucose, fructose, and no sugar, respectively. Different letters indicate significant difference (P < 0.001) compared with the other conditions after A. pollinis incubation, as assessed by two-sided Tukey’s HSD test. d GC-MS analysis of the medium containing no sugar, sucrose, fructose, or glucose after A. pollinis incubation. Total ion current chromatogram is shown. The numbered peaks were confirmed with reference standards according to their retention time and mass spectrum patterns in Supplementary Figs. 2–5. I.S. indicates the internal standard. Experiments were repeated three times with similar results, and a representative dataset are presented. Note that the retention time in the no sugar sample was slightly different from the other samples because it was analyzed on a different date. Arrowheads at 17.44 and 19.24 min indicate peaks originating from derivatives of palmitic acid and stearic acid, respectively, as estimated by NIST library search; these peaks are likely contaminants introduced during the experimental procedure. Source data are provided as a Source Data file.

PQQ-dependent dehydrogenase in A. nectaris can produce gluconic acid from glucose

To identify the biosynthetic genes involved in gluconic acid production, whole-genome sequencing of this A. nectaris strain was conducted using a Nanopore sequencer. A complete and circular chromosome and two plasmids were assembled. The genome size and the number of the annotated genes were approximately half of those of Escherichia coli (Supplementary Table 1). The phylogenomic tree was reconstructed using this genome (Supplementary Fig. 8). Consistent with the BLAST search using the 16S ribosomal RNA sequence (Supplementary Fig. 1), the strain isolated in this study belong to the Acinetobacter nectaris clade. The whole-genome average nucleotide identity (ANI) values against other A. nectaris strains were ~95% (94.7–95.2%) (Supplementary Table 2), meeting the generally accepted threshold (≥ 95%) for species delineation. Hence, this bacterium was designated as A. nectaris UT841 strain, and two plasmids were designated as pAnec1 and pAnec2, respectively.

In the previous study, another A. nectaris strain (BB362) caused little or no pollen tube burst12, leading us to speculate that one or both plasmids in our strain encode gene(s) for gluconic acid production. To test this possibility, we tried to eliminate these plasmids using plasmid incompatibility with ori sites predicted with Ori-Finder 202215 (Supplementary Fig. 9a–c; see also Methods). pAnec1 was no longer amplified in PCR, indicating that it was successfully eliminated in ∆pAnec (Supplementary Fig. 9c). During the elimination procedure of the pAnec1 plasmid, pAnec2 was also spontaneously lost in ∆pAnec (Supplementary Fig. 9c). These bacteria, in which the endogenous plasmids were eliminated, still showed pollen tube burst activity, a pH drop, and gluconic acid production (detected as peaks 3, 4, and 6 in GC-MS) in response to glucose, suggesting that neither pAnec1 nor pAnec2 did not encode genes for gluconic acid production (Supplementary Fig. 9d–g).

Next, we focused on the chromosome and found the genes encoding pyrroloquinoline quinone (PQQ)-dependent dehydrogenase. Some PQQ-dependent dehydrogenases can catalyze glucose oxidation in many bacteria, including acetic acid bacteria and some Acinetobacter species16–18. E. coli also possesses the gcd gene, which codes for a membrane-bound PQQ-dependent glucose dehydrogenase19. In the A. nectaris UT841 genome, two genes (AnecUT841_18170 and AnecUT841_19130) encoding membrane-bound PQQ-dependent dehydrogenases were annotated by DFAST20. To assess whether these two genes encode enzymes with catalytic activity for glucose oxidation, a complementation test was performed in E. coli. In this assay, BW25113 (a parent strain of the Keio collection) as the wild type (WT) and JW0120-KC from the Keio collection served as the gcd mutant21. We subcloned the coding sequences of AnecUT841_18170 or AnecUT841_19130, fused with the native gcd promoter19, into pUC18 and introduced into the gcd mutant. Since E. coli does not have the ability to produce PQQ on its own22, 1 µM PQQ was supplied exogenously in this assay. Consistent with a previous study22, the WT strain exhibited a significant PQQ-dependent pH decrease and the peaks originated from gluconic acid (peak 6) in GC-MS, whereas the gcd mutant did not show these phenomena (Fig. 3a, b). While introduction of AnecUT841_19130 did not rescue the gcd mutant, the strain transformed with the plasmid containing AnecUT841_18170 exhibited a pH decrease and gluconic acid production in the presence of 1 µM PQQ, similar to the WT (Fig. 3a, b). These results suggest that A. nectaris possesses a functional PQQ-dependent glucose dehydrogenase (Anec-GCD).

Fig. 3. AnecUT841_18170 encodes a PQQ-dependent glucose dehydrogenase.

Fig. 3

Open in a new tab

a pH of the induction medium containing glucose after E. coli incubation under the presence or absence of 1 µM PQQ. Data are mean ± SD. All groups consisted of n = 9. P-values determined by two-sided paired-sample t-tests with Holm–Bonferroni correction were 5.68 × 10-9 in WT, 0.976 in gcd, 1.34 × 10-5 in gcd + pUC18/Pgcd-AnecUT841_18170, and 0.976 in gcd + pUC18/Pgcd-AnecUT841_19130. Statistical significance is indicated by asterisks: n.s., P ≥ 0.05; ***P < 0.001. b GC-MS analysis of the induction medium containing glucose after E. coli incubation. Total ion current chromatogram is shown. The numbered peaks were confirmed with reference standards according to their retention time and mass spectrum patterns in Supplementary Figs. 2–5. I.S. indicates the internal standard. Source data are provided as a Source Data file.

PQQ biosynthesis factors are strictly conserved in nectar-dwelling Acinetobacter species

To investigate the evolution of the PQQ-dependent gluconic acid production system in the nectar-dwelling bacteria, a reciprocal BLASTp search was performed. In the reciprocal BLASTp search for Anec-GCD (AnecUT841_18170), most Acinetobacter species showed their best hits to this protein, and each best hit was then matched back to Anec-GCD in a second BLASTp search against the protein database of A. nectaris UT841 (Fig. 4a), suggesting that orthologous genes of Anec-GCD are widely conserved in the Acinetobacter genus. This implies that nectar-dwelling bacteria use the conventional glucose dehydrogenase to rupture pollen tubes. Consistent with this, gluconic acid production has also been reported in other Acinetobacter species which are not nectar-dwelling bacteria17,18.

Fig. 4. PQQ biosynthetic factors are strictly conserved in nectar-dwelling Acinetobacter species.

Fig. 4

Open in a new tab

a Summary of the reciprocal BLASTp results. GenBank IDs of the reciprocally best hits are listed in Supplementary Data 1. The topology of the phylogenomic tree on the right side is based on Supplementary Fig. 8. b Schematic illustration of the gene cluster encoding PQQ biosynthetic factors in A. nectaris UT841 genome.

Next, we focused on the genes related to the PQQ biosynthesis. Whereas some Acinetobacter species requires PQQ supplementation to produce gluconic acid in previous studies17,18, we did not supply any exogenous PQQ to A. nectaris UT841 or A. pollinis SCC477, suggesting that they can produce PQQ on their own. In the A. nectaris UT841 genome, a gene cluster of PQQ biosynthetic genes was identified (pqqA–E; Fig. 4b). These PQQ biosynthetic factors were analyzed using a reciprocal BLASTp search. While PqqE was widely conserved across the Acinetobacter genus, non-nectar-dwelling species possessing PqqA–D were limited (Fig. 4a). However, all nectar-dwelling Acinetobacter possessed PqqA–E. It is noted that PqqA, a precursor peptide for PQQ23, was not detected in A. boissieri ANC4422, A. apis ANC5114, and A. nectaris CIP110549, despite the presence of PqqB–E. This is likely because pqqA was not annotated in their genome database due to its small size (~30 amino acids). However, amino acid sequences showing high similarity (>80%) were found through a tblastn search using the PqqA sequence from A. nectaris UT841 as a query (Supplementary Fig. 10 and Fig. 4a). These findings suggest that almost all nectar-dwelling bacteria possess the PQQ biosynthetic pathway, supporting the importance of PQQ-dependent gluconic acid production in nectaries.

PQQ biosynthetic pathway is likely upregulated in the presence of glucose

To investigate whether the PQQ-dependent gluconic acid production system is upregulated in response to external sugar conditions, RNA-seq analysis was performed after incubating A. nectaris UT841 in media containing fructose, sucrose, or glucose, as well as in a sugar-free medium. The expression levels of Anec-gcd and pqqB–E did not show drastic changes among the different sugar conditions (Fig. 5b–f and Supplementary Table 3). However, the expression of pqqA, which encodes the precursor peptide for PQQ biosynthesis, was markedly upregulated in the sucrose- and glucose-containing induction buffer, with the strongest upregulation observed in the presence of glucose (Fig. 5a and Supplementary Table 3). These results suggest that PQQ production is likely enhanced in response to glucose by increasing the supply of the PQQ precursor.

Fig. 5. pqqA expression is upregulated in the presence of sucrose and glucose.

Fig. 5

Open in a new tab

a–f Gene expression levels of PQQ biosynthetic factors and a PQQ-dependent glucose dehydrogenase, normalized by transcripts per million (TPM). The detail values are shown in Supplementary Table 3. g Proposed working model illustrating the mechanism of pollen tube burst mediated by nectar-dwelling Acinetobacter species.

Discussion

This study shows that A. nectaris and A. pollinis acidify their exogenous environment by producing gluconic acid from glucose, which leads to the pollen tube burst (Fig. 5g). Although bursting occurred at the tips of pollen tubes, the mechanism depending on gluconic acid is likely distinct from sperm release during fertilization, which is triggered by RALF peptides24,25. Taken together with previous observations11,12, this mechanism appears to alter various environmental factors, including pH and nutrient availability, thereby shaping the ecology of floral nectaries. In the previous study, other nectar-dwelling Acinetobacter species was also shown to mediate rupture of pollen tubes12, suggesting that the mechanism revealed in this study is likely conserved in those species. Moreover, a nectar-dwelling Gluconobacter species was reported to consume glucose efficiently and decreased nectar pH to around 2 in floral nectaries26, suggesting that nectary acidification may be a general strategy beyond the Acinetobacter genus. Pollen tube burst was observed at pH values below 3.8 (Supplementary Fig. 6), and such acidic conditions can occur in floral nectary glands as a result of activity by nectar-dwelling bacteria26. To achieve a pH of 3.8 in our condition, the required concentration of gluconic acid was 0.012% (w/v). This value is likely an overestimate for natural nectaries because the assay was performed under PIPES-buffered conditions. We induced the pollen tube burst activity in the presence of 1% (w/v) sucrose or glucose. Generally, concentration of sucrose and/or glucose in nectar—the substrates for gluconic-acid production—exceeds 10% (w/w) across a broad range of plant species27, suggesting that nectar typically contains sufficient substrate to generate gluconic acid and produce an acidic environment. In addition, this physicochemical and non-specific strategy should act on pollen from a broad range of plant species, including Eschscholzia californica (Papaveraceae)12, Petunia hybrida (Solanaceae; Fig. 1), Abelia grandiflora (Caprifoliaceae; Supplementary Fig. 7a, b), and Arabidopsis thaliana (Brassicaceae; Supplementary Fig. 7c, d), which aligns well with the ecology of nectar-dwelling bacteria whose dispersal is passive and dependent on pollinators. Given the non-specific nature of this pH-dependent mechanism, other microorganisms might also employ a similar strategy using different organic acids.

Gluconic acid is the most abundant organic acid in honey28, and it is widely accepted that its production is primarily driven by glucose oxidases of honeybees29,30. This study shows that nectar-dwelling bacteria can also produce gluconic acid from sucrose and glucose. Whereas honeybees secret glucose oxidases that depend on flavin adenine dinucleotide (FAD) as a cofactor31, nectar-dwelling Acinetobacter species utilize a different type of enzyme, PQQ-dependent glucose dehydrogenase. The microbial community present in nectar may also influence the chemical composition of honey in a distinct way.

Although A. nectaris showed clear pollen tube burst activity in this study, it caused little or no pollen tube burst in the previous study12. This is probably because pollen germination scarcely occurred during co-cultivation with A. nectaris, and gluconic acid could not reach the pollen tubes in the previous experiment12. In contrast, pollen was pre-germinated and pollen tubes were exposed to an acidic environment, resulting in pollen tube burst in this study. It may be possible that promotion of pollen germination and pollen tube burst by nectar-dwelling Acinetobacter are independent mechanisms.

In our comparative genomic analysis, many Acinetobacter species were found to lack the PQQ biosynthesis pathway (Fig. 4a). A previous study shows that the requirement of exogenous PQQ supplementation for gluconic acid production varies even within the same species18, indicating that loss of the PQQ biosynthesis pathway occurs easily and frequently during the Acinetobacter evolution. Nonetheless, nectar-dwelling Acinetobacter species—despite having undergone genome size reduction during their evolution32—showed strict conservation of the genes encoding PQQ biosynthesis factors (Fig. 4a). This may reflect the niche and the survival strategy of the nectar-dwelling Acinetobacter species. Floral nectar is assumed to be initially sterile, and microorganisms transmitted by pollinators must grow rapidly to gain the advantage of “priority effects” and become the dominant species in the floral nectar33. Therefore, they are unlikely to rely on exogenous supplementation of PQQ from other microorganisms due to the rapid onset of growth-related metabolic demands. Given this situation, it seems that there is a selective pressure for nectar-dwelling bacteria to retain the PQQ biosynthetic pathway for rapid production of gluconic acid. Consistent with this, the expression of pqqA, which encodes the precursor peptide of PQQ, was strongly upregulated by sucrose and glucose (Fig. 5a). The upregulation of pqqA expression was not observed in Pseudomonas34, and this feature might be a unique in nectar-dwelling Acinetobacter species.

Anec-GCD (AcecUT841_18170) belongs to the membrane-bound PQQ-dependent glucose dehydrogenase family, which couples to the respiratory chain and can transfer electrons derived from the glucose oxidation to ubiquinone35. While AnecUT841_19130 did not rescue the gcd mutant (Fig. 3), this observation does not exclude the possibility that AnecUT841_19130 has glucose dehydrogenase activity, since lack of complementation may reflect limitations of the E. coli expression system (e.g., improper folding). After incubation of A. nectaris UT841 or A. pollinis SCC477, glucose was almost not detected in the GC-MS analysis (Supplementary Fig. 5), whereas sucrose and fructose were still detected (Fig. 1g and Supplementary Figs. 3 and 4). This suggests that these bacteria can efficiently consume glucose and drive both the respiratory chain and gluconic acid production. Gluconic acid production by membrane-bound PQQ-dependent glucose dehydrogenases is beneficial for both efficient carbon source utilization and nitrogen source acquisition from pollen. In addition, the pH decrease caused by gluconic acid exhibits as antimicrobial properties36, contributing to demonstrate “inhibitory priority effect”37. From these perspectives, the acidification strategy based on the gluconic acid production is highly advantageous and well-suited to the competitive nectar environment and the unique lifestyle of nectar-dwelling bacteria.

Methods

Isolation of A. nectaris UT841

A. nectaris UT841 was isolated from nectaries of Abelia grandiflora flowers at Yayoi-Campus, the University of Tokyo. Floral nectaries were excised and homogenized in sterile water. To remove debris, the homogenate was subjected to brief centrifugation (1000 × g for 30 sec), and the supernatant was then spread onto a Luria-Bertani (LB) agar plate. The plate was incubated for three days at room temperature, resulting in the emergence of colonies including A. nectaris. To ensure purity, single-colony isolation was performed three times on fresh LB plates.

Bacterial Strains

All bacterial strains used in this study are listed in Supplementary Table 4.

Plasmid construction and transformation

All oligonucleotides used in this study are listed in Supplementary Data 2. The ampicillin resistance gene in pUC18 was replaced with the kanamycin resistance gene derived from pCambia1300. The sequence of pUC18, excluding the ampicillin resistance gene, was amplified by inverse PCR. The kanamycin resistance gene from pCambia1300 was amplified and subcloned into this sequence using the In-Fusion cloning system (Takara-Bio, Kusatsu, Japan). The resulting construct was designated as pUC18_KanR. The predicted ori site in pAnec1, ranging from 20,000 to 25,000 bp, was amplified from the genomic DNA of A. nectaris UT841 and subcloned into pUC18_KanR, which had been digested with SalI and EcoRI, using In-Fusion cloning system (Takara-Bio). The resulting construct was designated as pUC18_KanR/pAnec1ori. The entire sequence was further amplified by inverse PCR, and the amplified SacB gene from pACRISPR38 was subcloned into this sequence using In-Fusion cloning system (Takara-Bio). pACRISPR was a gift from Q. Ji (Addgene plasmid 113348). The resulting construct was designated as pUC18_KanR_SacB/pAnec1ori.

The native gcd promoter region and the coding sequences of AnecUT841_18170 and AnecUT841_19130 were amplified from the genomic DNA of E. coli BW25113 and A. nectaris UT841, respectively. Each coding sequence was fused to the native gcd promoter by second PCR and subcloned into pUC18, which had been digested with SalI and EcoRI, using In-Fusion cloning system (Takara-Bio).

A. nectaris UT841 was cultured in a medium containing 1.5% (w/v) tryptone, 0.5% (w/v) soytone, 0.5% (w/v) NaCl, and 5% (w/v) fructose at 28˚C. E. coli BW25113 and JW0120-KC (the gcd mutant) were cultured in Terrific Broth at 37˚C. When OD600 reached 0.5–1.0, the cells were centrifuged at 10,000 × g for 1 min and resuspended in sterile 10% (v/v) glycerol. This washing procedure was repeated five times, and resuspended cells were flash-frozen in liquid nitrogen and stored at −80˚C. Plasmid DNA was mixed with the competent cells and transformed by electroporation at 2000 V using a ELEPO21 electroporator with 1 mm gap cuvettes (Nepa Gene, Ichikawa, Japan).

Elimination of plasmids using plasmid incompatibility

Using pCambia1300, a binary vector for E. coli and Agrobacterium tumefaciens, A. nectaris was transformed by electroporation, confirming that the kanamycin resistance gene in pCambia1300 was functional in A. nectaris (Supplementary Fig. 9b). On the other hand, pUC18_KanR, in which the ampicillin resistance gene was replaced with the kanamycin resistance gene originated from pCambia1300, could not be introduced, indicating that the ColE1-type ori in pUC18 was not functional in A. nectaris (Supplementary Fig. 9b). We then subcloned the predicted ori of pAnec1 into the pUC18_KanR backbone. This plasmid was successfully introduced into A. nectaris, suggesting that the predicted ori of pAnec1 was functional in A. nectaris (Supplementary Fig. 9b). We also subcloned the SacB gene for plasmid curing and introduced it into A. nectaris again. After repeating streaking and single-colony isolation 20 times, pAnec1 was not amplified in PCR, indicating that it was successfully eliminated (Supplementary Fig. 9c). After further streaking on kanamycin-free plates 30 times, the introduced plasmid was also removed, as the kanamycin resistance gene was not amplified (Supplementary Fig. 9c). During the elimination procedure of the pAnec1 plasmid, pAnec2 was also spontaneously lost (Supplementary Fig. 9c).

Testing pollen tube burst activity and GC-MS analysis

A. pollinis SCC4774 was derived from American Type Culture Collection (ATCC) under the type strain deposit number TSD-214. A. nectaris UT841, A. nectaris SCC477, or ∆pAnec was cultured in a medium containing 1.5% (w/v) tryptone, 0.5% (w/v) soytone, 0.5% (w/v) NaCl, and 5% (w/v) fructose for 14 h at 28 °C. The cells were pelleted by centrifugation at 10,000 × g for 1 min and resuspended in induction medium (1 mM PIPES-KOH buffer (pH 6.5) containing 1 mM CaCl₂, 1 mM Ca(NO₃)₂, 1 mM MgSO₄, and 0.01% (w/v) H3BO3) modified from previously reported pollen germination media39,40. The cells were pelleted again by centrifugation at 10,000 × g for 1 min and resuspended in the same buffer, except that it contained 1% (w/v) glucose, sucrose, or fructose. The cells were incubated for 6 h at room temperature with gentle shaking, and the supernatant was collected after centrifugation at 10,000 × g for 1 min, followed by filtration through a 0.22 µm filter.

Pollen grains from Petunia hybrida were suspended in pollen germination medium (1 mM PIPES-KOH buffer (pH 6.1) containing 1 mM CaCl₂, 1 mM Ca(NO₃)₂, 1 mM MgSO₄, 0.01% (w/v) H3BO3, and 18% (w/v) sucrose), modified from previously studies39,40, and incubated for 2 h at 23 ˚C in the dark. For germination of pollen from Abelia grandiflora and Arabidopsis thaliana (Col-0), we used the same medium without PIPES to increase the pollen germination efficiency39. 10 μM epibrassinolide was added for Col-0 pollen41. The pollen concentrations were ~40–50, 15–30, and 60–80 grains µL−1 in P. hybrida, A. grandiflora, and A. thaliana, respectively. The percentages of pollen germination were 90% for P. hybrida, 15–30% for A. grandiflora, and 30–40% in for A. thaliana. Germinated pollen was mixed with 1/10 volume of the induction medium before or after incubation with bacteria. Pollen grains and tubes were observed using LSM880 confocal microscope (Carl Zeiss, Oberkochen, Germany). Differential interference contrast image acquisition was controlled using Zen v2.3 SP1 software (Carl Zeiss). The rates of pollen tube burst per number of germinated pollen were calculated using at least thirty germinated pollen for P. hybrida and A. thaliana and fifteen germinated pollen for A. grandiflora.

Two µL of the induction medium after incubation with bacteria and neutralization was mixed with 1% (w/v) sorbitol as the internal standard and used for GC-MS analysis. Samples were dried overnight in a vacuum desiccator over anhydrous CaSO4. 250 µL of methoxylamine hydrochloride (in pyridine) was added and incubated at 60˚C for 90 min. After cooling to room temperature, 250 µL of N,O-Bis(trimethylsilyl)trifluoroacetamide with 1% trimethylchlorosilane was added and incubated at 70˚C for 60 min. After cooling to room temperature, 1 µL of the derivatized sample was analyzed using a GCMS-QP2020NX (SHIMADZU, Kyoto, Japan) equipped with an SH-I-5Sil MS capillary column (30 m, 0.25 mm inner diameter, 0.25 µm film thickness; SHIMADZU). Samples were injected using split injection (split ratio 1:10) under helium carrier gas flow set at 1.0 mL per minute. The injector temperature was set to 280˚C. The column temperature was maintained at 80 ˚C for 2 min, increased at a rate of 10 ˚C per minute to 325 ˚C, and held at 325 ˚C for 3.5 min. The spectra were obtained by electron ionization and recorded with a scanning range of 50–550 m/z. Three biological replicates were analyzed in each experiment. The detailed GC-MS results are reported in Supplementary Data 3.

Complementation assay of the gcd mutant in E. coli

BW25113, gcd, pUC18/Pgcd-AnecUT841_18170 in gcd, or pUC18/Pgcd-AnecUT841_19130 in gcd was cultured in Terrific Broth for 14 h at 37 ˚C. The cells were pelleted by centrifugation at 10,000 × g for 1 min and resuspended in induction medium (1 mM PIPES-KOH buffer (pH 6.5) containing 1 mM CaCl₂, 1 mM Ca(NO₃)₂, 1 mM MgSO₄, and 0.01% (w/v) H3BO3). This washing procedure was repeated, and the cells were resuspended in the same buffer, except that it contained 1% (w/v) glucose with or without 1 µM PQQ. The cells were incubated for 18 h at 37 ˚C with gentle shaking. The supernatant was collected after centrifugation at 10,000 × g for 1 min, followed by filtration through a 0.22 µm filter. GC-MS analysis was performed as described above.

Statistical analyses

All of the statistical analysis and plotting was done using R42.

DNA preparation and Nanopore sequencing

A. nectaris UT841 was cultured in a medium containing 1.5% (w/v) tryptone, 0.5% (w/v) soytone, 0.5% (w/v) NaCl, and 5% (w/v) fructose for 14 h at 28˚C. 2 mL of the cultured medium was used for extraction of genomic DNA using the DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany), following the manufacturer’s protocol. Two columns were used, and DNA was eluted in Buffer AE (50 µL each). 60 µL of extracted DNA (100 ng/µL) was treated with the Short Read Eliminator Kit XS (PacBio, California, USA) to remove fragmented DNA, following the manufacturer’s protocol. The DNA was resuspended in 55 µL of buffer EB.

For library preparation, 47 µL of DNA was used with the SQK-LSK114 kit (Oxford Nanopore Technologies, Oxford, UK) according to the manufacturer’s protocol. 11 µL of the resultant library (68.4 ng/µL) was used for nanopore sequencing (FLO-MIN114, R10.4.1) using MinKnow software (version 22.10.7) for 15 h. The read speed was set to 260 bp/sec.

The simplex basecalling was executed using Guppy_basecaller (version 6.4.2 + 97a7f06), employing the ‘dna_r10.4.1_e8.2_260bps_sup.cfg’ configuration file and the ‘--do_read_splitting’ option. Duplex pairs were identified using sub-commands ‘pairs_from_summary’ and ‘filter_pairs’ from Duplex Tools (version 0.2.19). The duplex basecalling was performed using Guppy_basecaller_duplex (version 6.4.2 + 97a7f06) with the ‘dna_r10.4.1_e8.2_260bps_sup.cfg’ configuration file and the ‘--duplex_pairing_mode’ option.

Lambda DNA sequences were removed from the duplex-called FASTQ files using NanoLyse43 (version 1.2.1). Subsequently, the reads were trimmed by 50 bp at both the head and tail ends, and the reads shorter than 500 bp were eliminated using NanoFilt43 (version 2.8.0). The resulting FASTQ files were subjected to de novo assembly using Flye44 (version 2.9.1-b1780) with the ‘--nano-hq’ and ‘--scaffold’ options. Finally, the assembled genome was annotated with DFAST20 (version 1.2.18). The pqqA gene was manually annotated. The ANI values were calculated using FastANI45 (version 1.34).

Phylogenomic analysis

Genomic sequences of Acinetobacter species were acquired from the NCBI database (https://www.ncbi.nlm.nih.gov/data-hub/genome/). The corresponding RefSeq or GenBank accession numbers are listed in Supplementary Data 4. Both the retrieved and the A. nectaris sequences assembled in this study were annotated using Prokka46 (version 1.14.6). Subsequently, the annotated protein sequences were analyzed by PhyloPhlAn 3.047 (version 3.0.67), which identified 365 shared proteins in the PhyloPhlAn database. These shared proteins concatenated, aligned, and trimmed by PhyloPhlAn were used to reconstruct a phylogenomic tree with IQ-TREE 248 (version 2.2.0 COVID-edition). The “Q.insect+F + R4” substitution model was selected by ModelFinder49, and a consensus tree was constructed based on 1000 bootstrap replicates for both SH-aLRT and Ultrafast Bootstrap analyses.

RNA-seq analysis

A. nectaris UT841 was cultured in a medium containing 1.5% (w/v) tryptone, 0.5% (w/v) soytone, 0.5% (w/v) NaCl, and 5% (w/v) fructose for 14 h at 28˚C. The cells were pelleted by centrifugation at 10,000 × g for 1 min and resuspended in induction medium (1 mM PIPES-KOH buffer (pH 6.1) containing 1 mM CaCl₂, 1 mM Ca(NO₃)₂, 1 mM MgSO₄, and 0.01% (w/v) H3BO3). The cells were pelleted again by centrifugation at 10,000 × g for 1 min and resuspended in the same buffer containing 1% (w/v) glucose, sucrose, or fructose, or without any sugars. The cells were incubated for 6 h at room temperature with gentle shaking. After incubation, the incubated medium was mixed with two volumes of RNAprotect Bacteria Reagent (QIAGEN) and incubated for 15 min at room temperature. RNA was then extracted using the RNeasy kit (QIAGEN) following the manufacturer’s protocol. RNA-seq libraries were constructed using the SQK-PCB109 kit (Oxford Nanopore Technologies) according to the manufacturer’s protocol. The resulting libraries were used for nanopore sequencing (FLO-MIN106, R9.4.1) using MinKnow software (version 22.10.10). Base calling was performed using the Guppy basecaller (version 6.4.6+ae70e8f). Reads were mapped to the A. nectaris UT841 genome using the minimap250 (version 2.24-r1122). Mapped reads were counted using featureCounts51 (version 2.0.3). Read counts were normalized by calculating transcripts per million (TPM)52. Two independent biological replicates were analyzed for each condition.

Reciprocal BLASTp analysis and tblastn analysis

Reciprocal BLASTp analysis and tblastn analysis were performed using the BLAST+ package (version 2.13.0+). Protein sequences of Acinetobacter species were acquired from the NCBI database (https://www.ncbi.nlm.nih.gov/data-hub/genome/). The corresponding RefSeq or GenBank accession numbers are listed in Supplementary Data 5.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary information (5.4MB, pdf)
Supplementary Data 1 (13.8KB, xlsx)
Supplementary Data 2 (10.4KB, xlsx)
Supplementary Data 3 (28.4KB, xlsx)
Supplementary Data 4 (11.1KB, xlsx)
Supplementary Data 5 (11.1KB, xlsx)
Reporting Summary (93.2KB, pdf)
Transparent Peer Review file (350KB, pdf)

Source data

Source Data (22.6KB, xlsx)

Acknowledgements

We thank Maki Nidome, Mizuki Saito, Aya Yoshida, and Sachiyo Soneda for their technical assistance. This work was supported in part by Grants-in-Aid for Transformative Research Areas (22H05174 to S.F.), Grants-in-Aid for Scientific Research (21H05030 to S.T.; 24K01692 to S.F.), Grant-in-Aid for Challenging Exploratory Research (23K17987 to S.F.), Grant-in-Aid for Early-Career Scientists (23K14207 to Y.K.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT), Japan Science and Technology Agency (JST) PRESTO program (JPMJPR21D4 to Y.K.), and the Suntory Rising Stars Encouragement Program in Life Sciences (to S.F.).

Author contributions

Y.K., S.T. and S.F. designed the research. Y.K. and H.M. isolated A. nectaris UT841. Y.K. performed almost all the experiments. Y.K., S.T., and S.F. prepared the manuscript with input from H.M.

Peer review

Peer review information

Nature Communications thanks Ling Ding and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The annotated genome sequence, as well as the the raw data files for whole-genome sequencing and RNA-seq, have been deposited in the DNA Data Bank of Japan (DDBJ) and the DDBJ Sequence Read Archive under BioProject ID PRJDB20387. The source data underlying Figs. 1–3 and Supplementary Figs. 6, 7, and 9 are provided as a Source Data file. Additional details on datasets that support the findings of this study will be made available by the corresponding author upon request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Yoshinobu Kato, Email: a-ykato@g.ecc.u-tokyo.ac.jp.

Sota Fujii, Email: a-fujii@g.ecc.u-tokyo.ac.jp.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-72617-z.

References

  • 1.Roy, R., Schmitt, A. J., Thomas, J. B. & Carter, C. J. Review: Nectar biology: from molecules to ecosystems. Plant Sci262, 148–164 (2017). [DOI] [PubMed] [Google Scholar]
  • 2.Quevedo-Caraballo, S., de Vega, C., Lievens, B., Fukami, T. & Álvarez-Pérez, S. Tiny but mighty? Overview of a decade of research on nectar bacteria. New Phytol245, 1897–1910 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Sandhu, D. K. & Waraich, M. K. Yeasts associated with pollinating bees and flower nectar. Microb. Ecol.11, 51–58 (1985). [DOI] [PubMed] [Google Scholar]
  • 4.Alvarez-Perez, S. et al. Acinetobacter pollinis sp. nov., Acinetobacter baretiae sp. nov. and Acinetobacter rathckeae sp. nov., isolated from floral nectar and honey bees. Int. J. Syst. Evol. Microbiol. 71, 004783 (2021). [DOI] [PubMed]
  • 5.Álvarez-Pérez, S., Lievens, B., Jacquemyn, H. & Herrera, C. M. Acinetobacter nectaris sp. nov. and Acinetobacter boissieri sp. nov., isolated from floral nectar of wild Mediterranean insect-pollinated plants. Int. J. Syst. Evol. Microbiol.63, 1532–1539 (2013). [DOI] [PubMed] [Google Scholar]
  • 6.Herrera, C. M., García, I. M. & Pérez, R. Invisible floral larcenies: microbial communities degrade floral nectar of bumble bee-pollinated plants. Ecology89, 2369–2376 (2008). [DOI] [PubMed] [Google Scholar]
  • 7.Herrera, C. M., de Vega, C., Canto, A. & Pozo, M. I. Yeasts in floral nectar: a quantitative survey. Ann. Bot.103, 1415–1423 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Vannette, R. L. & Fukami, T. Contrasting effects of yeasts and bacteria on floral nectar traits. Ann. Bot.121, 1343–1349 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Nicolson, S. W. & Thornburg, R. W. Nectar chemistry. in Nectaries and Nectar 215–264 (Springer Netherlands, 2007).
  • 10.Pozo, M. I. & Jacquemyn, H. Addition of pollen increases growth of nectar-living yeasts. FEMS Microbiol. Lett.366, 1–5 (2019). [DOI] [PubMed] [Google Scholar]
  • 11.Herrera, C. M. Scavengers that fit beneath a microscope lens. Ecology98, 2725–2726 (2017). [DOI] [PubMed] [Google Scholar]
  • 12.Christensen, S. M., Munkres, I. & Vannette, R. L. Nectar bacteria stimulate pollen germination and bursting to enhance microbial fitness. Curr. Biol.31, 1–8 (2021). [DOI] [PubMed] [Google Scholar]
  • 13.Li, F.-S., Phyo, P., Jacobowitz, J., Hong, M. & Weng, J.-K. The molecular structure of plant sporopollenin. Nat. Plants5, 41–46 (2018). [DOI] [PubMed] [Google Scholar]
  • 14.Sawyer, D. T. & Bagger, J. B. The lactone-acid-salt equilibria for D-glucono-δ-lactone and the hydrolysis kinetics for this lactone. J. Am. Chem. Soc.81, 5302–5306 (1959). [Google Scholar]
  • 15.Dong, M.-J., Luo, H. & Gao, F. Ori-Finder 2022: a comprehensive web server for prediction and analysis of bacterial replication origins. Genomics. Proteomics Bioinformatics20, 1207–1213 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Matsutani, M. & Yakushi, T. Pyrroloquinoline quinone-dependent dehydrogenases of acetic acid bacteria. Appl. Microbiol. Biotechnol.102, 9531–9540 (2018). [DOI] [PubMed] [Google Scholar]
  • 17.Van Schie, B. J., De Mooy, O. H., Linton, J. D., Van Dijken, J. P. & Kuenen, J. G. PQQ-Dependent Production of Gluconic Acid by Acinetobacter, Agrobacterium and Rhizobium Species. Microbiology133, 867–875 (1987). [Google Scholar]
  • 18.Bouvet, P. J. M. & Bouvet, O. M. M. Glucose dehydrogenase activity in Acinetobacter species. Res. Microbiol.140, 531–540 (1989). [DOI] [PubMed] [Google Scholar]
  • 19.Yamada, M., Asaoka, S., Saier, M. H. & Yamada, Y. Characterization of the gcd gene from Escherichia coli K-12 W3110 and regulation of its expression. J. Bacteriol.175, 568–571 (1993). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Tanizawa, Y., Fujisawa, T. & Nakamura, Y. DFAST: a flexible prokaryotic genome annotation pipeline for faster genome publication. Bioinformatics34, 1037–1039 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Waligora, E. A. et al. Role of intracellular carbon metabolism pathways in Shigella flexneri Virulence. Infect. Immun.82, 2746–2755 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Matsushita, K. et al. Escherichia coli is unable to produce pyrroloquinoline quinone (PQQ). Microbiology143, 3149–3156 (1997). [DOI] [PubMed] [Google Scholar]
  • 23.Puehringer, S., Metlitzky, M. & Schwarzenbacher, R. The pyrroloquinoline quinone biosynthesis pathway revisited: A structural approach. BMC Biochem9, 8 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ge, Z. et al. Arabidopsis pollen tube integrity and sperm release are regulated by RALF-mediated signaling. Science (80-.)358, 1596–1600 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Mecchia, M. A. et al. RALF4/19 peptides interact with LRX proteins to control pollen tube growth in Arabidopsis. Science (80-.)358, 1600–1603 (2017). [DOI] [PubMed] [Google Scholar]
  • 26.Vannette, R. L., Gauthier, M.-P. L. & Fukami, T. Nectar bacteria, but not yeast, weaken a plant–pollinator mutualism. Proc. R. Soc. B Biol. Sci.280, 20122601 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Chalcoff, V. R., Aizen, M. A. & Galetto, L. Nectar concentration and composition of 26 species from the temperate forest of South America. Ann. Bot.97, 413–421 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.White, J. W. & Subers, M. H. Studies on honey inhibine. 2. a chemical assay. J. Apic. Res.2, 93–100 (1963). [Google Scholar]
  • 29.Bucekova, M. et al. Honeybee glucose oxidase—its expression in honeybee workers and comparative analyses of its content and H2O2-mediated antibacterial activity in natural honeys. Naturwissenschaften101, 661–670 (2014). [DOI] [PubMed] [Google Scholar]
  • 30.White, J. W. Jr., Subers, M. H. & Schepartz, A. I. The identification of inhibine, the antibacterial factor in honey, as hydrogen peroxide and its origin in a honey glucose-oxidase system. Biochim. Biophys. Acta73, 57–70 (1963). [DOI] [PubMed] [Google Scholar]
  • 31.Brudzynski, K. A current perspective on hydrogen peroxide production in honey. A review. Food Chem332, 127229 (2020). [DOI] [PubMed] [Google Scholar]
  • 32.Sanchez, V. A., Renner, T., Baker, L. J. & Hendry, T. A. Genome evolution following an ecological shift in nectar-dwelling Acinetobacter. mSphere10, e0101024 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Jacquemyn, H., Pozo, M. I., Álvarez-Pérez, S., Lievens, B. & Fukami, T. Yeast–nectar interactions: metacommunities and effects on pollinators. Curr. Opin. Insect Sci.44, 35–40 (2021). [DOI] [PubMed] [Google Scholar]
  • 34.An, R. & Moe, L. A. Regulation of pyrroloquinoline quinone-dependent glucose dehydrogenase activity in the model rhizosphere-dwelling bacterium Pseudomonas putida KT2440. Appl. Environ. Microbiol.82, 4955–4964 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Matsushita, K., Shinagawa, E., Adachi, O. & Ameyama, M. Quinoprotein D-glucose dehydrogenase of the Acinetobacter calcoaceticus respiratory chain: membrane-bound and soluble forms are different molecular species. Biochemistry28, 6276–6280 (1989). [DOI] [PubMed] [Google Scholar]
  • 36.Nieto-Peñalver, C. G., Savino, M. J., Bertini, E. V., Sánchez, L. A. & de Figueroa, L. I. C. Gluconic acid produced by Gluconacetobacter diazotrophicus Pal5 possesses antimicrobial properties. Res. Microbiol.165, 549–558 (2014). [DOI] [PubMed] [Google Scholar]
  • 37.Chappell, C. R. et al. Wide-ranging consequences of priority effects governed by an overarching factor. Elife11, 1–33 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Chen, W. et al. CRISPR/Cas9-based genome editing in Pseudomonas aeruginosa and cytidine deaminase-mediated base editing in Pseudomonas species. iScience6, 222–231 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Boavida, L. C. & McCormick, S. TECHNICAL ADVANCE: Temperature as a determinant factor for increased and reproducible in vitro pollen germination in Arabidopsis thaliana. Plant J52, 570–582 (2007). [DOI] [PubMed] [Google Scholar]
  • 40.Meng, J.-G. et al. Central-cell-produced attractants control fertilization recovery. Cell186, 3593–3605.e12 (2023). [DOI] [PubMed] [Google Scholar]
  • 41.Vogler, F., Schmalzl, C., Englhart, M., Bircheneder, M. & Sprunck, S. Brassinosteroids promote Arabidopsis pollen germination and growth. Plant Reprod27, 153–167 (2014). [DOI] [PubMed] [Google Scholar]
  • 42.R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2016. (2017).
  • 43.De Coster, W., D’Hert, S., Schultz, D. T., Cruts, M. & Van Broeckhoven, C. NanoPack: visualizing and processing long-read sequencing data. Bioinformatics34, 2666–2669 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kolmogorov, M., Yuan, J., Lin, Y. & Pevzner, P. A. Assembly of long, error-prone reads using repeat graphs. Nat. Biotechnol.37, 540–546 (2019). [DOI] [PubMed] [Google Scholar]
  • 45.Jain, C., Rodriguez-R, L. M., Phillippy, A. M., Konstantinidis, K. T. & Aluru, S. High throughput ANI analysis of 90K prokaryotic genomes reveals clear species boundaries. Nat. Commun.9, 5114 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Seemann, T. Prokka: rapid prokaryotic genome annotation. Bioinformatics30, 2068–2069 (2014). [DOI] [PubMed] [Google Scholar]
  • 47.Asnicar, F. et al. Precise phylogenetic analysis of microbial isolates and genomes from metagenomes using PhyloPhlAn 3.0. Nat. Commun.11, 2500 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol.37, 1530–1534 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K. F., von Haeseler, A. & Jermiin, L. S. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods14, 587–589 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Li, H. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics34, 3094–3100 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics30, 923–930 (2014). [DOI] [PubMed] [Google Scholar]
  • 52.Li, B., Ruotti, V., Stewart, R. M., Thomson, J. A. & Dewey, C. N. RNA-Seq gene expression estimation with read mapping uncertainty. Bioinformatics26, 493–500 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary information (5.4MB, pdf)
Supplementary Data 1 (13.8KB, xlsx)
Supplementary Data 2 (10.4KB, xlsx)
Supplementary Data 3 (28.4KB, xlsx)
Supplementary Data 4 (11.1KB, xlsx)
Supplementary Data 5 (11.1KB, xlsx)
Reporting Summary (93.2KB, pdf)
Transparent Peer Review file (350KB, pdf)
Source Data (22.6KB, xlsx)

Data Availability Statement

The annotated genome sequence, as well as the the raw data files for whole-genome sequencing and RNA-seq, have been deposited in the DNA Data Bank of Japan (DDBJ) and the DDBJ Sequence Read Archive under BioProject ID PRJDB20387. The source data underlying Figs. 1–3 and Supplementary Figs. 6, 7, and 9 are provided as a Source Data file. Additional details on datasets that support the findings of this study will be made available by the corresponding author upon request. Source data are provided with this paper.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES