Nectar peroxide: assessing variation among plant species, microbial tolerance, and effects on microbial community assembly

Introduction

Many flowering plants offer energy-dense nectar to attract pollinators. While nectar is an essential attractant for pollinators, it also presents a liability (González-Teuber & Heil, 2009; Heil, 2011). In addition to transporting pollen, pollinators act as vectors for dispersal-limited bacteria and fungi (Antonovics, 2005; Herrera et al., 2009; Harper et al., 2010; Morris et al., 2020). Some are plant pathogens and enter plants via floral tissues (González-Teuber & Heil, 2009; Sasu et al., 2010; McArt et al., 2014) while other microbes primarily reside within floral nectar, a chemically diverse habitat rich in nutrients (Pozo et al., 2014; Chappell & Fukami, 2018; Adler et al., 2021). Microbial growth can modify nectar characteristics including pH, floral scent, sugar ratios and total concentration, amino acids, volume, and temperature, which have been shown to alter pollinator foraging behavior (Herrera et al., 2009; Álvarez-Pérez et al., 2012; Pozo et al., 2014; Aizenberg-Gershtein et al., 2015; Schaeffer et al., 2017; Rering et al., 2018; Vannette & Fukami, 2018; de Vega et al., 2022). However, many plant traits are hypothesized to protect nectar against potential disadvantageous changes induced by microbes (Adler, 2000; Carter & Thornburg, 2004). High sugar concentrations reduce the number of bumble bee-vectored yeast species able to survive within the nectar of Helleborus foetidus (Herrera et al., 2009). Undetermined chemical properties of nectar also reduce the growth of bacterial wilt in cucumber (Sasu et al., 2010).

These findings suggest the potential adaptive value of nectar antimicrobial mechanisms (Adler, 2000; Herrera et al., 2009). However, much remains to be understood about the identity of and the mechanisms by which floral nectar components are responsible for reducing microbial growth, and whether strategies are conserved across phylogenetically diverse plant species. Floral nectar is a complex solution of metabolites and molecules in which carbohydrates, vitamins, lipids, amino acids, proteins, inorganic ions, and secondary compounds like alkaloids and phenolics are diverse and, in some cases, abundant (Adler, 2000; González-Teuber & Heil, 2009; Palmer-Young et al., 2019). Alkaloids, proteins, and hydrogen peroxide inhibit microbial growth in nectar or nectar analogs (Aizenberg-Gershtein et al., 2015; Schmitt et al., 2018, 2021; Koch et al., 2019; Mueller et al., 2023). These findings support the antimicrobial hypothesis, suggesting that specific nectar metabolites may be involved in adaptive defense and filtration of introduced microbes, significantly mediating microbial community composition and assembly in nectar (Adler, 2000; Herrera et al., 2009; Harper et al., 2010; Pozo et al., 2014; Stevenson et al., 2017; Mueller et al., 2023) yet how prevalent such mechanisms may be among plant species is poorly understood.

Hydrogen peroxide (H₂O₂) is a reactive oxygen species that generates free hydroxyl radicals (^•HO) in the presence of metal ions. Hydrogen peroxide generation has been proposed to be a major mechanism of antimicrobial defense (Carter & Thornburg, 2004; Carter et al., 2007) in nectar, and more broadly across host–microbe systems (Allaoui et al., 2009; Smirnoff & Arnaud, 2019; Miller et al., 2020). Indeed, among potential antimicrobial compounds examined in laboratory studies, H₂O₂ had the broadest inhibitory effect across many microbial species (Mueller et al., 2023). In Nicotiana, nectarin proteins including superoxide dismutases and glucose oxidases (GOXs) drive a ‘nectar redox cycle’ that produces hydrogen peroxide and also provides a mechanism for detoxifying the free radicals generated, allowing plants to avoid self-toxicity (Carter & Thornburg, 2000, 2004; González-Teuber & Heil, 2009; Harper et al., 2010). Hydrogen peroxide has been detected in the nectar of many species of Nicotiana, ranging between 23.4 and 4000 μM but also at lower levels in Cucurbita and the legumes Mucuna and Robinia (Table 1) (Carter & Thornburg, 2000; Bezzi et al., 2010; Liu et al., 2013; Nocentini & Guarnieri, 2014; Silva et al., 2018). However, whether antimicrobial levels of peroxide are also found broadly in the nectar of phylogenetically diverse plant species remains unknown. Hydrogen peroxide is also a common antimicrobial defense used by bees, found at high levels in bee-related habitats including the honey food stores of social Hymenoptera, produced via bee-secreted GOX proteins (Burgett, 1974).

Table 1. Peroxide concentrations measured in nectar from 58 different floral species spanning 25 families, both in this study and reported in the literature.

Family	Species	Range peroxide (μM)	Mean peroxide (μM)	Var		N	Assay method	Study
Acanthaceae	Ruellia simplex	300–3000	1000	–		4	Peroxide test strip	Current
Apocynaceae	Amsonia tabernaemontana	< 15–60	30	–		3	Peroxide test strip	Current
Asparagaceae	Hesperaloe parviflora	< 15	ND (< 15)	–		4	Peroxide test strip	Current
Asphodelaceae	Aloe sp. 1	11–19	15	±4	SD	3	Amplex Red kit*	Current
	Aloe sp. 2	7–19	12	±5		8	Amplex Red kit*	Current
	Kniphofia	15	15	–		3	Peroxide test strip	Current
Bignoniaceae	Chilopsis linearis	< 15	ND (< 15)			3	Peroxide test strip	Current
	Cerinethe major	24	24	–		1	Amplex Red kit*	Current
	Cynoglossum grande	30–150	70	–		4	Peroxide test strip	Current
	Echium candicans	3–7	5	±7	SD	2	Amplex Red kit*	Current
	Phacelia heterophylla	200–300	300	–		3	Peroxide test strip	Current
	Phacelia hydrophylloides	300	300	–		2	Peroxide test strip	Current
	Phacelia tanacetifolia	16–57	29	±16	SD	5	Amplex Red kit*	Current
Brassicaceae	Brassica oleracea	32–131	69	±34	SD	3	Amplex Red kit*	Current
	Erysimum capitatum	< 15–15	5	–		3	Peroxide test strip	Current
Caprifoliaceae	Lonicera japonica	300	300	–		2	Peroxide test strip	Current
Cleomaceae	Cleomella arborea	7–43	23	±12	SD	12	Amplex Red kit*	Current
Crassulaceae	Cotyledon orbiculata var. engleri	7–134	59	±36	SD	20	Amplex Red kit*	Current
Cucurbitaceae	Cucurbita pepo	3–296	61	±75.5	SD	16	FOX method*	Nocentini (2015)
Ericaceae	Arbutus unedo	1500	1500	–	SD	3	Peroxide test strip	Current
	Arctostaphylos sp.	< 15	ND (< 15)	–		3	Peroxide test strip	Current
Fabaceae	Calliandra eriophylla	< 15	ND (< 15)	–		3	Peroxide test strip	Current
	Erythrina crista-galli	15	15	–		3	Peroxide test strip	Current
	Mucuna sempervirens	–	68.4	±10.7	SD	30	Beyotime kit*	Liu (2013)
	Robinia pseudoacacia	–	48.8	±0.9		3	Carter, 2000 method*	Silva (2018)
Francoaceae	Melianthus comosus	22	22	–	SD	1	Amplex Red kit*	Current
Grossulariaceae	Ribes cereum	< 15–15	5	–		3	Peroxide test strip	Current
Lamiaceae	Mondardella odoratissima	15–500	200	–		3	Peroxide test strip	Current
	Salvia mellifera	13.1	13.1	–		1	Amplex Red kit*	Current
	Phlomis fruticosa	< 1	ND (< 1)	–		36	Amplex Red kit*	Current
Onagraceae	Clarkia unguiculata	3–13	7	±4	SD	3	Amplex Red kit*	Current
	Clarkia unguiculata	< 15–15	5	–		4	Peroxide test strip	Current
	Epilobium canum ‘silver’	81–162	103	±25	SD	10	Amplex Red kit*	Current
	Epilobium canum ‘canum’	31–99	72	±26		9	Amplex Red kit*	Current
	Epilobium canum ‘calistoga’	22–97	60	±26		9	Amplex Red kit*	Current
	Epilobium canum ‘everrit’	25–54	39	±9		10	Amplex Red kit*	Current
	Oenothera elata	< 15	ND (< 15)	–		2	Peroxide test strip	Current
Orobanchaceae	Castilleja applegatei	< 15	ND (< 15)	–		1	Peroxide test strip	Current
	Castilleja miniata	15–60	30	–		3	Peroxide test strip	Current
Phrymaceae	Diplacus aurantiacus	< 15–15	10	–		2	Peroxide test strip	Current
Plantaginaceae	Gambelia speciosa	15	15	–	SD	3	Peroxide test strip	Current
	Penstemon centranthifolius	9–21	15	±4		10	Amplex Red kit*	Current
	Penstemon gracilentus	60–150	100	–		2	Peroxide test strip	Current
	Penstemon heterophyllus	4–24	18	±7	SD	8	Amplex Red kit*	Current
	Penstemon heterophyllus	30	30	–		1	Peroxide test strip	Current
	Penstemon speciosus	60–300	190	–		7	Peroxide test strip	Current
Polemoniaceae	Ipomopsis aggregata	< 15	ND (< 15)	–		2	Peroxide test strip	Current
	Phlox diffusa	< 15	ND (< 15)	–		1	Peroxide test strip	Current
Ranunculaceae	Aquilegia formosa	< 15	ND (< 15)	–		3	Peroxide test strip	Current
	Delphinium gracilentum	< 15	ND (< 15)	–		2	Peroxide test strip	Current
Sapindaceae	Aesculus californica	7–14	11	±3	SD	4	Amplex Red kit*	Current
	Aesculus californica	< 15	ND (< 15)	–		1	Peroxide test strip	Current
Scrophulariaceae	Russelia equisetiformis	< 15	ND (< 15)	–		4	Peroxide test strip	Current
Solanaceae	Nicotiana alata	–	673	±41.55	SD	3	FOX method*	Silva (2018)
	Nicotiana attenuata	–	116.2	±17.3	SE	7	Luminal chemiluminescence method*	Bezzi (2010)
	Nicotiana attenuata	–	23.4	±1.86		7	Amplex Red kit*	Bezzi (2010)
	Nicotiana benthamiana	–	288	±50.5	SD	3	FOX method*	Silva (2018)
	Nicotiana bonariensis	–	1840	±8.90		3	FOX method*	Silva (2018)
	Nicotiana clevelandii	–	869	±136		3	FOX method*	Silva (2018)
	Nicotiana glauca	–	407	±184		3	FOX method*	Silva (2018)
	Nicotiana langsdorffii	–	833	±127		3	FOX method*	Silva (2018)
	Nicotiana rustica	–	2140	±582		3	FOX method*	Silva (2018)
	Nicotiana sylvestris	–	543	±50.5		3	FOX method*	Silva (2018)
	Nicotiana sp.	< 20–> 4000	771	–		10	Carter, 2000 method*	Carter (2000)
Verbenaceae	Lantana montevidensis	30–1500	600	–		4	Peroxide test strip	Current

• In the current study (45 species; 23 families), nectar samples from individual flowers were pooled to provide sufficient volume for either Amplex Red hydrogen peroxide or colorimetric test strip peroxide detection assays. SD and SE are abbreviations for standard deviation and standard error, respectively. For the Amplex Red assay, the limit of detection (LOD) is 50 nM, while our limit of quantification (LOQ) is 1 μM. Meanwhile, for the test strip assay, the LOD and LOQ are both 0.5 ppm (15 μM). Plant species that we sampled in the present study are colored blue, while those from the literature are indicated in gray. Additionally, species for which hydrogen peroxide (as opposed to peroxide, generally) was specifically measured are marked with an * in the ‘Assay Method’ column.

Concentrations of hydrogen peroxide at and above 2000 μM have been shown to reduce the growth of some common plant pathogens and yeasts, and bacteria isolated from nectar, pollinators, and the environment (Carter & Thornburg, 2004; Carter et al., 2007; Mueller et al., 2023). However, some microbes may tolerate high hydrogen peroxide concentrations (Herrera et al., 2009; Álvarez-Pérez et al., 2012; Vannette et al., 2013), including the nectar-specialist yeast Metschnikowia reukaufii. Catalases defend against oxidative stress by catalyzing the decomposition of hydrogen peroxide, detoxifying the environment and potentially making it more favorable for microbial growth (Vannette et al., 2013; de Vega et al., 2022), for both the catalase producer and possibly co-occurring microbes (Mueller et al., 2023). This has yet to be evaluated in the complex and dynamic nectar environment of living flowers, or at concentrations of hydrogen peroxide lower than 2000 μM. Investigating whether lower concentrations of hydrogen peroxide, which may be more commonly found in nectar, are also antimicrobial is necessary to assess the range of efficacy. Additionally, evaluating microbial taxa that are specialized to nectar and pollinator-associated environments and grow within a community, rather than as individual isolates, will be necessary to assess the range of conditions under which hydrogen peroxide may be an effective defense.

Here, we assess the generality of hydrogen peroxide as a potential antimicrobial compound in floral nectar, including its inducibility and variation in microbial tolerance and compositional response to peroxide. In Aim 1, we surveyed 45 plant species and compiled previous values for 13 additional species from the literature, spanning 25 families to measure the range of nectar peroxide concentration and examine if plant phylogeny predicts hydrogen peroxide concentration. In Aim 2, we assessed if plant defense hormones methyl salicylate (Me-SA) and methyl jasmonate (Me-JA) induce hydrogen peroxide upregulation in nectar. Salicylic acid plays a central role in priming plant defenses against potential microbial pathogens, and we predicted that this could lead to increased nectar defenses and a possible mode of hydrogen peroxide regulation against microbes introduced into nectar. Methyl jasmonate, while typically associated with defense responses to herbivory and tissue damage, has also been found to mediate floral trait expression as well as plant traits that affect microbial growth (Pak et al., 2009; Zuñiga et al., 2020). Next, we assessed microbial responses to field-relevant concentrations of hydrogen peroxide. We approached this question from multiple perspectives. In Aim 3, we tested how yeast identity or isolation source, including bee, floral tissue, and nectar-associated habitats, impacts tolerance to hydrogen peroxide using in vitro assays. Next, in Aim 4, we examined how enzyme-increased hydrogen peroxide in the field shapes microbial community assembly in nectar. Finally, in Aim 5, we examined how individual vs community context determines microbial growth and effects on hydrogen peroxide concentration using lab assays. We predicted that differences we might observe could be mediated by microbial competition or the benefits of co-growth, where the detoxification of hydrogen peroxide by more tolerant microbes could facilitate improved growth of less resilient microbes.