Maize (Zea mays) inbred line B73 and Nicotiana benthamiana plants were grown growth in chambers maintained at 24/22°C (day/night) and a 16/8 h (light/dark) photoperiod. SCMV Beijing strain (SCMV-BJ) was from a previously published source (Fan et al., 2003). N. benthamiana leaves were infiltrated with Agrobacterium tumefaciens (at an OD₆₀₀ of 1.5) carrying SCMV infectious clones. Crude extracts from the SCMV-BJ infected maize leaves or SCMV infectious clone-infected N. benthamiana leaves were used to rub-inoculate maize seedling young-expanded leaves as described previously (Chen et al., 2017a).

Total RNA was extracted from individual leaf samples using TRIzol reagent followed by RNase-free DNase I treatment. The cDNA synthesis was performed using 2 μg total RNA per sample, an oligo(dT₁₈) primer, and M-MLV reverse transcriptase in a 25 μL reaction. Quantitative reverse transcription-polymerase chain reaction (RT-qPCR) was performed using 1 μL of 10-fold diluted cDNA per reaction, gene-specific primers ( Supplementary Table 2 ), and a Fast SYBR mixture on an ABI 7500 Real Time PCR system (Applied Biosystems Inc.). The expression level of the maize ubiquitin gene (ZmUbi) was used as the internal control (Chen et al., 2017a). The relative expression levels of the assayed genes were calculated using the 2^-ΔΔCT method (Livak and Schmittgen, 2001).

Total protein was isolated from individual leaf samples and separated in gels through electrophoresis as previously described (Cao et al., 2012). Detections of specific proteins on the immunoblots were performed using antibodies specific for SCMV coat protein (CP) (Xia et al., 2016), Flag (Sigma-Aldrich) and plant β-actin (EASYBIO). The relative expression levels of individual proteins on the immunoblots were quantified using the ImageJ image analysis tool (http://imagej.net/) as previously described (Wyrsch et al., 2015).

At 4- and 5-day post inoculation (dpi), mock-inoculated and SCMV-infected leaf samples from six independent biological replicates were collected. To minimize the effect of variation in metabolite content throughout the plant, maize leaves from SCMV-infected plants and the corresponding mock-inoculated plants were harvested at the same leaf position. Collected tissue was immediately frozen in liquid nitrogen and stored at -80°C until further analysis. Maize leaf powder was weighed to 50 mg and extracted with 450 μL extraction liquid (75% methanol). The mixture was homogenized in a ball mill for 4 min at 45 Hz and then ultrasonically treated for 5 min (incubated in ice water); this process was repeated 3 times. The supernatant was collected after centrifugation at 12,000 rpm for 15 min. The sample extract was filtered through a 0.22 µm filter and added to sample vials. All samples were analysed by a gas chromatograph system coupled with a Pegasus HT time-of-flight mass spectrometer (GC-TOF-MS). The GC-TOF-MS analysis was performed at Shanghai Applied Protein Technology Co. Ltd. and was conducted with an Agilent 1290 Infinity chromatography system and AB SCIEX QTRAP 5500 mass spectrometer. MultiQuant software was used to extract the chromatographic peak area and retention time. The AA standard correct retention time was used to identify the metabolites. To identify the differentially expressed metabolites, statistical analyses between the two sample groups were performed by calculating the variable importance in the projection (VIP) and p values of the metabolites. Student’s t test was used to obtain p values. Metabolites with VIP > 1 and p values < 0.05 were marked as differentially expressed metabolites between sample groups. Pathway enrichment analysis was performed using MetaboAnalyst 5.0 (www.metaboanalyst.ca/MetaboAnalyst/).

Fine powder from leaf samples (~50 mg) was mixed with 600 μL of acetonitrile: chloroform (7: 3, v/v) solution and vortexed for 30 s. The quantitation control was prepared by adding 20 ng L-lactate (Sigma−Aldrich, L1750) into the acetonitrile: chloroform solution. The mixture was sonicated for 1 h on ice and then centrifuged for 5 min at 7,000 g at 4°C. Three hundred microlitres of H₂O was added to each supernatant prior to the two-step liquid−liquid partitioning. The upper aqueous fractions from the same sample were pooled and dried under a nitrogen stream. The dried extracts were then dissolved in 750 μL of H₂O and filtered through a 0.22 μm membrane prior to measurement. LC–MS/MS analysis was performed on a UPLC system (Waters, Milford, Ohio, USA) combined with a 5500 Qtrap MS equipped with an ESI source (AB SCIEX). Each sample (5 μL) was injected onto an HSS T3 C18 column for further analysis.

Maize plants were separately sprayed with 100 nM UK5099 (Sigma−Aldrich; dissolved in DMSO), 100 μM DCA (Sigma−Aldrich; dissolved in double-distilled water), or 10 μM lactate (Sigma−Aldrich; dissolved in double-distilled water) containing 0.2% Tween-20, or with a 0.2% Tween-20 solution with no chemical agents as a blank control.

Putative ZmLDH1 and ZmLDH2 genes were obtained from the updated maize B73 genome website (AGPv4, http://ensembl.gramene.org/Zea_mays/Info/Index). Multiple sequence alignments were performed by DNAMAN 7.0 (Lynnon Biosoft, San Ramon, CA, USA).

Extractions were carried out on ice. Maize leaves were ground in extraction buffer (0.1 M Tris-HCl, pH 8.5, 10 mM Na borate, 10 mM DTT and 5 mg/ml BSA). The sample was centrifuged, and aliquots of the supernatant were taken for the enzyme assay. LDH enzymatic activity was assayed by monitoring the pyruvate-dependent NADH oxidation based on the absorbance at 340 nm spectrophotometrically. The assay mix (final volume 1.3 ml) contained 1 ml of 0.13 M Tris-HCl (pH8.0), 150 μg NADH, 3 μmol 4-methylpyrazole, 3 μmol NaCN, 15 μmol Na pyruvate, and 0.2 ml of enzyme extract.

The pGD-6K2-VPg-Pro-mCherry vector used in this study was described previously (Xie et al., 2021). For transient expression assays in maize protoplasts, coding region sequence (CDS) of ZmLDH2 was cloned into the pGD-eGFP vector. For overexpression via SCMV, the LDH2_3Flag and GFP_3Flag fragments replaced the GFP fragment of pSCMV-GFP to produce pSCMV-LDH2_3Flag and pSCMV-GFP_3Flag, respectively. For the luciferase complementation imaging (LCI) assays, CDS of ZmLDH2 was inserted into pCAMBIA-Cluc vectors at the Kpn I and Sal I sites to generate pUC-CE-ZmLDH2; and 6K2 was inserted into pCAMBIA-Nluc vectors at the Sac I and Sal I sites to generate pUC-NE-6K2, respectively. For the biomolecular fluorescence complementation (BiFC) assays, ZmLDH2 and 6K2 were inserted into pUC-CE or pUC-NE vectors at the BamH I and Sal I sites to generate pUC-CE-ZmLDH2 and pUC-NE-6K2, respectively. For cucumber mosaic virus (CMV)-virus-induced gene silencing (VIGS) assessment, a 200 bp DNA fragment representing a conserved partial sequence of ZmLDH1 and ZmLDH2 was amplified via RT-PCR using specific primers ( Supplementary Table 2 ). The resulting fragment was cloned into the pCMV201-2b_N81 vector resulted in pCMV201-2b_N81:LDH. All constructs were checked by sequencing prior to use.

Agrobacterium cultures carrying pCMV101, pCMV301 or one of the two constructs (pCMV201-2b_N81:LDH, and pCMV201-2b_N81:GFP) were grown, mixed, and infiltrated into the leaves of N. benthamiana plants as described (Wang et al., 2016). At 4 days post infiltration, the infiltrated N. benthamiana leaves were processed into crude leaf extracts, which were then sap-inoculated individually into maize seedlings using the vascular puncture inoculation (VPI) method (Wang et al., 2016).

Maize seeds were inoculated with crude leaf extracts from SCMV-BJ-infected or noninfected (control) maize plants using the VPI method. The germinated seedlings were kept in the dark at 24°C to produce etiolated plants. Maize protoplasts isolation and transfection were performed as described (Zhu et al., 2014).

At 14 hours post transfection, maize protoplasts were examined with a Zeiss LSM 800 confocal microscope for subcellular localization assays. For EGFP, the excitation wavelength was set at 488 nm and the emission wavelength at 510–550 nm. For mCherry protein, the excitation wavelength was set at 552 nm and the emission wavelength at 562–632 nm.

The yeast two-hybrid (Y2H) assay was performed in accordance with the procedures provided by the manufacturer (Clontech). Yeast expression plasmids were introduced into the yeast strain Gold (Clontech), and all transformants were cultured at 30°C for 72 h on synthetic dextrose (SD) medium without Leu and Trp. Afterwards, they were switched to a medium lacking Leu, Trp, His, and Ade.

LCI assays were performed as previously described (Chen et al., 2008). All of the tested combinations were agroinfiltrated into N. benthamiana leaves. At 3 days after infiltration, the leaves were sampled, sprayed with 1 mM luciferin (Invitrogen), and photographed using a low-light cooled CCD imaging apparatus (iXon, Andor Technology, Belfast, UK). The pictures were taken 15 min after exposure.

Maize protoplasts were transfected with different combinations of expression vectors. At 14 h post-transfection, YFP was excited at 514 nm with emission detected at 565–585 nm using a Zeiss LSM 800 confocal microscope.

Measurement of H₂O₂ content in assayed maize leaves was performed using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen, Carlsbad, USA) as instructed. Fluorescence was excited at a wavelength of 530 nm and detected at a wavelength of 590 nm.

Maize leaf tissues were cut and incubated for 10 h in a water solution supplemented with 10 nM CM-H₂DCFDA [5-(and-6)-chloromethyl 2’,7’-dichlorodihydrofluorescein diacetate, Invitrogen] to measure H₂O₂ levels and then washed twice before imaging.

The statistical significance of the data was determined by GraphPad Prism 7.0 (GraphPad Software Inc., USA: http://www.graphpad.com/). Comparisons between two groups of data were calculated by Student’s t-test or ANOVA (*P<0.05; **P<0.01; ***P<0.001).

Previous studies found that maize plants mainly showed systemically mosaic symptoms at 5 dpi of SCMV (Chen et al., 2017a; Chen et al., 2017b; Du et al., 2020). In this study, we also discovered that only 1.3% of maize plants (3 of 221-inoculated plants in three replicates) started to show mosaic symptoms at 4 dpi, whereas almost all SCMV-infected maize plants developed mosaic symptoms on the first systemically infected leaf (referred to as 1 SL) at 5 dpi ( Figure 1 and Supplementary Figure 1 ). RT-qPCR and western blotting showed the quite low accumulation levels of SCMV genomic RNA and CP at 4 dpi while their dramatical increase at 5 dpi ( Figure 1C ). Therefore, we determined that the time from 4 to 5 dpi was a key period for robust SCMV multiplication and infection.

To elucidate the metabolites and metabolic pathways that might be crucial for effective SCMV infection, we sampled maize leaves at two time points for untargeted metabolomics analysis. One was before the manifestation of mosaic symptoms (on 4 dpi), and another was immediately after the appearance of symptoms (on 5 dpi). Results of metabolomics analyses revealed that thirteen metabolites showed significant changes (VIP > 1, p < 0.05) at 4 dpi ( Figure 1 and Supplementary Table 1 ). Topological analysis revealed that the functions of the 13 differentially expressed metabolites at 4 dpi were mainly enriched in sulfur metabolism and riboflavin metabolism ( Figure 1 ). Thus, SCMV infection could affect the biosynthesis of sulfur-containing defense compounds and antioxidant compounds before symptoms appearance, both of which involved in plant resistance (Gao et al., 2012; Deng et al., 2014; Zhao et al., 2014). At 5 dpi, with the manifestation of mosaic symptoms and robust multiplication of SCMV in maize leaves, 26 metabolites were significantly changed ( Figure 1 and Supplementary Table 1 ). Topological analysis revealed that the differentially expressed metabolites at 5 dpi were primarily linked to pyruvate metabolism, glycolysis, and the TCA cycle ( Figure 1 ). Thus, SCMV infection at this stage primarily affected plant energy metabolism. Notably, both lactate and palmitoleic acid changed significantly among all the differentially expressed metabolites in both periods ( Figure 1 ). We further, by using LC−MS/MS, confirmed that lactate accumulated much higher levels in SCMV-infected plants, i.e. 1.7- and 2.4-fold higher at 4 and 5 dpi, respectively, than those in mock-inoculated plants ( Figure 1 ). These results indicate that lactate accumulation may play an essential role in supporting robust SCMV infection.

Since lactate is a product of anaerobic glycolysis, we firstly determined whether anaerobic glycolysis plays a role in SCMV infection. Oxidative phosphorylation (OxPhos) and anaerobic glycolysis are the two major catabolic glucose pathways in plants ( Figure 2 ). To dissect the key step of glucose metabolism involved in SCMV infection, we treated SCMV-infected maize plants once at 3 dpi with UK5099, which is known to promote anaerobic glycolysis, or dichloroacetate (DCA), which controls the anaerobic glycolysis shift to OxPhos (Zhou et al., 2021). By 5 dpi, UK5099 treatment of SCMV-infected maize plants caused more severe mosaic symptoms on maize leaves compared with that water-treated ( Figure 2 ). Both RT-qPCR and western blotting results showed significant increment of SCMV RNA and CP by UK5099 treatment ( Figure 2D ). In contrast, treatment with DCA alleviated the severity of mosaic symptoms on maize leaves, and decreased SCMV RNA and CP accumulation levels compared with that water-treated ( Figure 2D ). Taken together, induction of anaerobic glycolysis promotes SCMV infection.

In the lactate fermentation pathway of anaerobic glycolysis, lactate is produced from pyruvate by cytosolic NAD-dependent LDH (Maurino and Engqvist, 2015). Here, we determined whether LDH participates in SCMV infection. We found that the enzymatic activity of LDH was significantly up-regulated immediately before and after robust SCMV infection, i.e. at 4 and 5 dpi ( Figure 3 ).

Next, we analysed the sequences of ZmLDH genes and then investigated their transcriptional patterns following SCMV infection. Two homologs of maize LDH were obtained from the updated Z. mays B73 genome (AGPv4, http://ensembl.gramene.org/Zea_mays/Info/Index): ZmLDH1 (Zm00001d014944) and ZmLDH2 (Zm00001d014945). Sequence alignments indicated that the two LDH genes shared approximately 90% and 89% identities on nucleotide and amino acid sequences, respectively ( Supplementary Figure 2A ). We used RT-qPCR to determine the relative expression levels of different ZmLDH transcripts in mock-inoculated and SCMV-infected maize plants. At 4 dpi, the expression level of ZmLDH1 was not significantly changed compared with that of mock-inoculated plants, while the expression level of ZmLDH2 was strongly up-regulated 40-fold by SCMV infection ( Figure 3 ). At 5 dpi, the expression levels of ZmLDH1 and ZmLDH2 in systemically infected leaves were approximately 7.5-fold and 400-fold higher, respectively, than that in the equivalent leaves of mock-inoculated plants ( Figure 3 ). These results suggest that LDH might play important roles for SCMV infection. Considering the high identities between ZmLDH1 and ZmLDH2 and the much higher expressional change of ZmLDH2 caused by SCMV infection than that of ZmLDH1, subsequently, we chose ZmLDH2 for the following studies.

To explore the role of LDH in viral infection, we overexpressed ZmLDH2 using an LDH2-expressing SCMV infectious clone designated pSCMV-LDH2_3Flag, with pSCMV-GFP_3Flag as a control ( Figure 4 ). We used the sap of N. benthamiana leaves agroinfiltrated with each of these two virus infectious clones to mechanically inoculate maize plants. SCMV-LDH2_3Flag-infected plants showed mosaic symptoms (4 dpi) earlier than SCMV-GFP_3Flag-infected plants (5 dpi) ( Supplementary Figure 3 ). At 5 dpi, maize plants infected by SCMV-LDH2_3Flag exhibited more severe mosaic symptoms on 1 SLs than that of SCMV-GFP_3Flag-infected plants ( Figure 4 ). The accumulation of SCMV genomic RNA in the SCMV-LDH2_3Flag-infected plants increased by 3.3-fold at 5 dpi compared with that in SCMV-GFP_3Flag-infected plants ( Figure 4 ). Western blotting results were consistent with the RT-qPCR results and showed 2.0-fold increases of SCMV CP accumulation at 5 dpi ( Figure 4 ). Detection of the expression of Flag-tagged proteins in systemically infected leaves by western blotting demonstrates that both SCMV-LDH2_3Flag and SCMV-GFP_3Flag can successfully infect maize plants ( Figure 4 ). These results show that overexpression of ZmLDH2 enhances SCMV infection.

Meanwhile, we silenced both ZmLDH1 and ZmLDH2 using the ZMBJ-CMV-based gene silencing vector (Wang et al., 2016). A DNA fragment (200 bp) that is conserved between ZmLDH1 and ZmLDH2 was selected after analysis for minimal off-target silencing, and then was cloned into pCMV201-2b_N81 ( Figure 4 ). Maize B73 seeds were inoculated with ZMBJ-CMV harboring the ZmLDH₂₀₀ or GFP₂₅₄ fragment using VPI (Wang et al., 2016). Maize plants inoculated with CMV-GFP₂₅₄ were used as controls. RT-qPCR results showed that the relative expression of ZmLDH1 and ZmLDH2 reduced c. 56 and 66%, respectively, in silenced plants at 7 days post-SCMV infection compared with the CMV : GFP control plants ( Figure 4 ). Silencing of ZmLDH expression did not affect maize growth, and milder, later (7 dpi) mosaic symptoms were observed in the ZmLDH-silenced plants than in the non-silenced control plants (5 dpi) ( Figure 4 and Supplementary Figure 3 ). Meanwhile, SCMV genomic RNA levels in the ZmLDH-silenced plants decreased by c. 78% compared with those in the control plants ( Figure 4 ). SCMV CP accumulation also reduced by 75% in the ZmLDH-silenced plants compared with that in the control plants ( Figure 4 ). Taken together, expression of ZmLDH is essential for robust SCMV infection.

To explore the mechanism by which ZmLDH promotes SCMV infection, we used Y2H arrays to investigate potential interactions between SCMV-encoded proteins and LDH2. The Y2H analysis showed that 6K2 interacted with ZmLDH2 ( Figure 5 ). To further confirm the interaction between 6K2 and ZmLDH2, we performed LCI and BiFC assays. For the LCI assay, the restoration of luciferase activity by the 6K2-ZmLDH2 interaction led to the detection of a positive luciferase signal in the leaf area co-expressing 6K2 and ZmLDH2. Negative controls showed no luciferase signal ( Figure 5 ). For the BiFC assay, N-terminal YFP-tagged 6K2 (6K2-nYFP) and C-terminal YFP-tagged LDH2 (cYFP-LDH2) were co-expressed in maize protoplasts. Aggregated YFP fluorescence (positive interaction signals) were observed only in maize protoplasts co-expressing 6K2 and ZmLDH2 ( Figure 5 ). Together, these results suggest that SCMV 6K2 interacts with ZmLDH2 in yeast and in planta.

Since 6K2 is a key protein that induces the formation of the ER-derived vesicles for potyviruses replication (Wei et al., 2010), we determined whether ZmLDH2 is recruited into the VRCs of SCMV. Transiently single expression of ZmLDH2-GFP localized to the cytoplasm in maize protoplasts ( Figure 6 ). In contrast, in SCMV-infected cells, ZmLDH2 aggregated in the cytoplasm ( Figure 6 ). To identify the precise compartment for the interaction of 6K2 and LDH2, cYFP-ZmLDH2 and 6K2-nYFP were co-expressed with 6K2-VPg-Pro-mCherry, which was previously used to indicate SCMV VRCs (Xie et al., 2021). Confocal microscopy observations showed that the site of the 6K2-ZmLDH2 interaction colocalized with aggregated 6K2-VPg-Pro ( Figure 6 ). In addition, the aggregation of ZmLDH2 under SCMV infection also colocalized with aggregated 6K2-VPg-Pro ( Figure 6 ). Altogether, ZmLDH2 can be relocated into SCMV VRCs.

Lactate has been reported to function in animal immune escape of hepatitis B virus (HBV) infection (Zhou et al., 2021). We next determined whether lactate, the product of LDH catalysis, also functions in SCMV infection by interfering with plant immunity. To evaluate the role of lactate in SCMV infection, we sprayed SCMV-infected maize plants with lactate at 3 dpi. Plants inoculated with SCMV and then sprayed with water were used as controls. By 5 dpi, the lactate-treated and SCMV-infected plants exhibited more severe mosaic symptoms than water-treated control plants ( Figure 7 ). The results of RT-qPCR and western blotting analyses showed significantly enhanced accumulation of SCMV genomic RNA and CP in maize plants treated with lactate compared to control plants ( Figure 7C ).

To investigate whether immunity response correlated with enhanced SCMV infection by lactate treatment, we measured the expression levels of the pathogenesis-related (PR) genes and the accumulation of H₂O₂ in SCMV-infected maize leaves at 24 h post lactate or water treatment. As previously reported (Chen et al., 2017a; Yuan et al., 2019), SCMV infection stimulated the expression levels of PR genes (ZmPR1, ZmPR3, ZmPR4 and ZmPR5) and the accumulation levels of H₂O₂ ( Figure 7F ). Intriguingly, lactate treatment significantly decreased the expression levels of PR genes in SCMV-infected plants ( Figure 7 ). Meanwhile, the abundance of H₂O₂ and fluorescence intensity of CM-H₂DCFDA (DCF) were dramatically lowed in lactate-treated maize leaves than that in water-treated maize leaves during SCMV infection ( Figure 7F ). Altogether, these results suggest that lactate facilitates viral infection via suppressing plant defense responses.