Comprehensive analysis of bZIP transcription factors in passion fruit

Summary

The bZIP transcription factors are well-known transcriptional regulators that are essential for regulating resistance to biotic and abiotic stresses in plants. In this study, a total of 56 putative bZIP members were identified in passion fruit (Passiflora edulis). An integrative analysis was performed using bioinformatics. Transcriptome analysis revealed that most PebZIPs respond to drought, salt, cold and heat stress. By combining the transcriptome results of two different resistant genotypes, four representative members were finally selected for differential expression validation in different tissues and cultivars. Furthermore, transcriptome and metabolome association analysis revealed consistent expression trends of PeZIP20 and PeZIP21, with only one difference at 63aa, with different metabolites including flavonoids, lipids and amino acids. This work will contribute to further studies of the functions of bZIPs and their resistance properties, as well as to the development of novel germplasm.

Graphical abstract

Highlights

• •Fifty-six putative bZIP transcription factors were identified in Passiflora edulis
• •PebZIP20 and PebZIP21 have only one amino acid difference at 63aa
• •PebZIP20 and PebZIP21 co-upregulated with different accumulate metabolites

Protein family determination; Plant biology; Omics

Introduction

Passion fruit (P. edulis Sims) is a tropical fruit originating from the Amazon River Basin in South America. The juice of passion fruit is rich in nutrients and functional components such as a variety of amino acids, vitamins, and flavonoids, with a unique and rich fragrance making it a popular source of beverage fruit.¹ In recent years, the planting of passion fruit has developed rapidly in China. The planting area of passion fruit has increased rapidly in southern of China, mainly grown in the provinces of Guangxi, Yunnan, Fujian, Guizhou, Guangdong, and Hainan provinces. Some varieties of Passiflora are not only edible, but also a very good medicinal resource, with sedatives, anxiolytics, anti-diabetic, and anti-inflammatory activities.² With the development of the passion fruit all over the world, the incidence of virus, pathogen diseases, post-harvest diseases, and environmental stress conditions is increasing, which hindering the healthy development of passion fruit cultivation.³ Effective plant breeding techniques can improve tolerance to biotic or abiotic stress, and increase the productivity and quality of fruit.^4,5 Identification of novel biotic or abiotic resistance genes would provide tools for the future of molecular plant breeding.

MYB, bZIP, NAC, WRKY, and AP2/ERF-like transcription factor families are all related to plant stress resistance.⁶ These major plant transcription factor families form a regulatory network in response to stress, among which basic leucine zipper (bZIP) plays a crucial role in plant responses to salinity, drought, chilling, mechanical damage, and osmotic stress.⁷ The bZIP TFs are widely distributed, and relatively conserved in eukaryotes.⁸ About 17 bZIPs in yeast, 27 in fruit flies, 31 in nematodes, and 56 in humans have been discovered,⁹ whereas in plants, more bZIPs were found, 127 in Arabidopsis thaliana, 216 in Zea mays, 210 in Nicotiana tabacum, 214 in Populus trichocarpa, 47 in Vitis vinifera, 70 in Solanum lycopersicum, 88 in Juglans regia, 135 in Malus domestica, and 352 in Glycine max, etc., also have been reported.^10,11,12

Plant bZIP TFs constitute the largest family of ABA-inducible DNA-binding proteins by forming homo- or hetero-dimers.^13,14 The conserved domain of bZIP TFs consists of 60–80 amino acid residues, including a basic amino acid region that binds to a specific DNA sequence and a leucine zipper dimerization region. The less conserved leucine zipper region is located at the C-terminal separate from the N-terminal basic region to form a Y-shaped structure.¹⁵ The bZIP TFs bind to the DNA sequence ACGT as the core DNA motif, including G-box (CACGTG), C-box (GACGTC), and A-box (TACGTA) to regulate the expression of downstream related genes in multiple biological processes such as light signaling, flower development, seed maturation, biotic and abiotic stress responses.¹⁶ The bZIPs TFs are also involved in the transcriptional regulation of secondary metabolites such as terpenoids, flavonoids, and alkaloids.¹⁷

At present, the genome sequence of passion fruit has been published,^18,19 which provides important genome resources for the mining of related genes. In this study, the genome information and transcriptome data of diploid passion fruit were used to identify the whole sequence of the bZIP TFs family of P. edulis by bioinformatics methods, and to study the function of bZIP family genes in response to stress. Our study is the first to characterize the bZIP family genes in P. edulis and the transcriptome in response to abiotic stress, and expression in different genotypes and fruit maturing stage, also the transcriptome and metabolome co-express analysis will serve as a useful resource for studies on the relation of functions of candidate key bZIP TFs with the secondary metabolites that are involved in stress resistance. The mined candidate bZIPs will contribute to the resistance breeding of passion fruit.

Results

Identification of bZIP genes in P. edulis

A total of 56 bZIP family members were identified from the P. edulis genome database based on HMMER, BLASTP, and CDD searches (Table 1). The PebZIPs were named PebZIP1-52 by their chromosome position, and 4 genes (PebZIP53-56) were not located on any chromosome. The length of the amino acids in PebZIP protein sequences ranged from 78 (PebZIP15) to 1406 aa (PebZIP12). The predicted isoelectric point (pI) values of the PebZIPs ranged from 10.38 to 5.3, and the proteins with a pI greater than 7 accounted for 53% of the total number. PebZIPs were found to contain all three phosphorylation sites (Ser, Thr, and Tyr), suggesting that they act through post-transcriptional phosphorylation modification. Subcellular location prediction analysis revealed that most PebZIPs are located in the nucleus, except for PebZIP3, PebZIP15 and PebZIP46, which are located in peroxisomes, mitochondrial matrix, and Golgi apparatus, respectively.

Table 1.

Characteristics of genes in the bZIP gene family of Passiflora edulis

Gene	Gene ID	Group	start	end	strand	CDS (bp)	Protein (.)	MW (Da)	pI	No phosphate sites Ser Thr Tyr			Subcellular location
PebZIP1	P_edulia010000603.g	D	14727713	14734606	–	1053	350	39.09	6.27	33	18	5	Nuclear
PebZIP2	P_edulia010000619.g	D	17447558	17454574	+	1374	457	50.51	6.53	52	22	6	Nuclear
PebZIP3	P_edulia010000677.g	F	22672061	22672825	–	765	254	28.08	5.55	16	24	2	Peroxisomal
PebZIP4	P_edulia010000785.g	D	40691144	40695449	+	1209	402	44.92	8.38	41	22	6	Nuclear
PebZIP5	P_edulia010001390.g	J	148330447	148334335	+	1335	444	48.5	8.18	52	23	3	Nuclear
PebZIP6	P_edulia010001407.g	A	149439770	149443040	+	1428	475	51.34	6.04	37	25	9	Nuclear
PebZIP7	P_edulia010002485.g	C	172268929	172272303	+	714	237	25.99	9.5	30	13	3	Nuclear
PebZIP8	P_edulia010003315.g	K	180402480	180405104	+	831	276	30.45	4.19	23	8	5	Nuclear
PebZIP9	P_edulia010003789.g	G	183612377	183617679	–	1260	419	44.88	7.27	43	28	12	Nuclear
PebZIP10	P_edulia010004082.g	K	185825540	185827258	–	624	207	23.27	4.23	18	5	3	Nuclear
PebZIP11	P_edulia010004361.g	D	189543626	189549094	–	1440	479	53.47	8.49	49	25	9	Nuclear
PebZIP12	P_edulia010004559.g	A	192146698	192156726	–	4221	1406	155.06	5.36	131	77	34	Nuclear
PebZIP13	P_edulia010004653.g	S	193250718	193251113	+	396	131	14.87	6.54	19	3	1	Nuclear
PebZIP14	P_edulia010004698.g	S	193625132	193625524	+	393	130	14.82	6.54	16	3	1	Nuclear
PebZIP15	P_edulia010005501.g	H	203103366	203103881		237	78	9.21	10.21	6	4	1	Mitochondrial matrix
PebZIP16	P_edulia010005577.g	A	203744634	203746542	+	813	270	29.62	6.55	29	14	6	Nuclear
PebZIP17	P_edulia020006037.g	I	29642327	29645212	–	1026	341	38.18	5.64	37	16	5	Nuclear
PebZIP18	P_edulia020006842.g	D	177369663	177375322	–	1383	460	50.78	8.77	50	20	8	Nuclear
PebZIP19	P_edulia020007325.g	A	183068897	183069832	–	630	209	23.39	10.5	19	8	4	Nuclear
PebZIP20	P_edulia020007400.g	S	183518175	183518642	+	468	155	17.75	6.97	22	6	6	Nuclear
PebZIP21	P_edulia020007442.g	S	183822277	183822744	+	468	155	17.73	6.97	22	6	6	Nuclear
PebZIP22	P_edulia020007475.g	I	183997460	184000854	+	750	249	28.08	9.01	24	11	1	Nuclear
PebZIP23	P_edulia030007689.g	C	3264708	3268524	–	1113	370	40.67	7.69	44	23	6	Nuclear
PebZIP24	P_edulia030008162.g	D	41101515	41106277	–	1122	373	42.2	7.33	28	15	9	Nuclear
PebZIP25	P_edulia030008584.g	G	144470941	144474754	+	1053	350	36.86	5.87	43	15	8	Nuclear
PebZIP26	P_edulia030008649.g	S	146196621	146197091	+	471	156	17.69	6.81	20	6	4	Nuclear
PebZIP27	P_edulia030008865.g	S	150851816	150852394	–	579	192	22.15	6.42	31	6	5	Nuclear
PebZIP28	P_edulia040009596.g	G	636543	641465	–	1224	407	43.01	6.79	46	32	10	Nuclear
PebZIP29	P_edulia040009793.g	S	5722552	5722983	–	432	143	16.55	6.79	11	2	3	Nuclear
PebZIP30	P_edulia040010220.g	G	25135007	25139712	+	1254	417	45.53	6.52	43	22	11	Nuclear
PebZIP31	P_edulia050011439.g	E	9482804	9487478	–	933	310	34.16	8.59	30	6	7	Nuclear
PebZIP32	P_edulia050012622.g	D	130867721	130875844	+	999	332	37.07	9.37	37	13	5	Nuclear
PebZIP33	P_edulia060013045.g	S	644857	645363	–	507	168	19.63	5.3	25	4	4	Nuclear
PebZIP34	P_edulia060014986.g	C	25811738	25815014	+	1158	385	41.93	6.64	42	25	2	Nuclear
PebZIP35	P_edulia060015091.g	I	27932736	27935740	–	1161	386	42.18	7.56	43	20	1	Nuclear
PebZIP36	P_edulia060015334.g	A	34056977	34060980	–	1365	454	48.75	10.09	50	24	10	Nuclear
PebZIP37	P_edulia060015571.g	S	42123188	42123667	+	480	159	18.06	7.73	20	8	3	Nuclear
PebZIP38	P_edulia060015578.g	S	42240050	42240529	+	480	159	18.03	7.73	20	7	3	Nuclear
PebZIP39	P_edulia060015671.g	S	47248718	47249314	–	597	198	22.7	6.42	23	8	2	Nuclear
PebZIP40	P_edulia060015726.g	D	52144105	52147529	+	1293	430	47.41	6.11	40	17	11	Nuclear
PebZIP41	P_edulia060015876.g	S	69730668	69731078	+	411	136	16	10.38	14	7	3	Nuclear
PebZIP42	P_edulia060016514.g	H	129860768	129864616	–	567	188	21.29	10.18	21	16	2	Nuclear
PebZIP43	P_edulia070017340.g	E	80874548	80876546	+	486	161	18.44	10.31	14	5	2	Nuclear
PebZIP44	P_edulia070018253.g	H	119826540	119828858	+	588	195	21.05	10.16	30	13	1	Nuclear
PebZIP45	P_edulia070018285.g	H	120088854	120091184	+	804	267	29.14	9.81	42	15	3	Nuclear
PebZIP46	P_edulia070018432.g	C	121668711	121670827	–	909	302	33.28	5.38	31	9	3	Golgi apparatus
PebZIP47	P_edulia070018710.g	A	123375866	123378559	–	861	286	31.37	7.54	26	18	7	Nuclear
PebZIP48	P_edulia070018729.g	E	123574459	123578635	+	1155	384	43.22	8.14	57	15	12	Nuclear
PebZIP49	P_edulia080019872.g	I	113335955	113339550	+	2103	700	77.09	7.63	100	34	11	Nuclear
PebZIP50	P_edulia080019936.g	D	113646861	113649341	+	1065	354	39.03	6.27	34	17	7	Nuclear
PebZIP51	P_edulia090020496.g	H	45415	50694	+	513	170	18.58	6.53	26	10	1	Nuclear
PebZIP52	P_edulia090022224.g	B	110166951	110170477	–	2082	693	75.01	5.55	80	39	12	Nuclear
PebZIP53	P_eduliaContig270022504.g	A	151299	153840	+	1050	349	38.71	8.38	24	17	4	Nuclear
PebZIP54	P_eduliaContig270022515.g	D	158427	162275	–	942	313	34.59	8.18	34	9	5	Nuclear
PebZIP55	P_eduliaContig310022447.g	H	190675	192656	+	564	187	21.29	6.04	21	16	2	Nuclear
PebZIP56	P_eduliaContig50022970.g	D	403901	410919	–	1374	457	50.52	9.5	52	22	5	Nuclear

Phylogenetic analysis of PebZIPs

Both, P. edulis and A. thaliana are dicotyledonous plants. The hypothetical homologous members of PebZIPs and the known AtbZIPs can be inferred. The phylogenetic tree was generated with the amino acid sequences of bZIP TFs among A. thaliana (74) and P. edulis (56) by neighbor-joining phylogenetic method. The phylogenetic tree classifies all the bZIPs into 12 groups (A-K, S) based on the previous classification criteria²⁰ (Figure 1). In which group A for ABF/AREB/ABI5 motif, C for CPRF2-like motif, G for GBF motif, H for HY5 motif, B and S for big and small protein sizes.²¹ The majority of the PebZIPs were assorted to group S (12 genes), D (9 genes), and A (6 genes), and fewer members were assorted to J and B groups, containing 1 gene each. In group K, PebZIP8 and PebZIP10 gathered closely with AtbZIP60 which play important roles in the endoplasmic reticulum stress with both a bZIP domain and a transmembrane domain (TMD).²² For the S group, PebZIP20, -21, −37, −38, and −26, were the most homogeneous genes of AtbZIP44 and AtbZIP11, PebZIP29 was the best orthologous match of AtbZIP53.

Figure 1.

Phylogenetic tree for the bZIP proteins in Passiflora edulis, and Arabidopsis thaliana

Orange circles and green triangles represent the bZIP family members of A. thaliana and P. edulis, respectively. A–K, S represent different groups.

Analysis of PebZIP gene family structure

The motif structures of the predicted proteins of the PebZIP genes are closely related to their functions. A total of 20 conserved motifs in PebZIP genes were predicted by MEME software (Figure 2). In addition to the motif1 (bZIP domain), some additional conserved motifs were also discovered in PebZIPs. Many motifs were identified in specific groups, which might be related to specific biological functions.²³ Group D members contain more motif types than others, DOG1 and HPB1b like domain are unique in group D. The bZIP-plant GBF1 domain is present in S group members. The plant-bZIP46 motif is particularly present in group A, in which PebZIP12 has the longest amino acid sequence and contains an AdoMet_WTases superfamily domain. The bZIP-plant RF2 domain is unique in E and I groups. PHAD3247 superfamily domain was only present in PebZIP49 (group I). HY5-like motif is particularly present in the H and K groups. All members in G Group contain MFMR and BRLZ domain, and PebZIP50 (group F) contains a BRLZ domain too. PebZIP52 is the only member in Group B that has a Myosin tail 1 superfamily domain.

Figure 2.

The grouping, motif, and gene structures analyses of PebZIPs

(A) Phylogenetic tree and grouping (A-K, S) of PebZIPs; (B) Motif analyzed by MEME software, the detailed information of the motif see Figure S1; (C) Motif analyzed by NCBI-CDD; (D) Intron-exon organization.

We also investigated the intron-exon organization of PebZIPs. Our results demonstrated the occurrence of a diverse number of exons (1–23) among all PebZIPs (Figure 2C). Most family members contain 2-4 exons, whereas the S group members have only one exon. Most members had the structure of 5′UTR (untranslated region) and 3′UTR, except for members of the S group and a few bZIPs in other groups that have only CDS structures. Members in groups D and G contain high numbers of introns, and most members in the A, E, I, K, and B groups contain long CDS. The members of the same group share similar intron-exon patterns. Our results support the significant divergence of PebZIPs because of their variation in the intron-exon organization.

The conserved bZIP domain and promoter analysis of PebZIPs

The conserved bZIP domain of PebZIPs was identified by average distance using BLOSUM62 alignment (Figure 3A). Most members of PebZIPs contain an R-X3-NR-X6-RK domain at the N-terminal, and L-X6-L-X6-L at C-terminal. The NR in conserved R-X3-NR-X6-RK domain is KR in PebZIP31, and PebZIP48 is the only one lacking NR in the conserved domain. There are usually X9 amino acids from amino acids R at the end of the N-terminal to the first L at the C-terminal, but PebZIP6 has 41 additional amino acids. In the conserved C-terminal, the first L is changed to I in PebZIP23, the second L is changed to V in PebZIP21, and M in PebZIP7 and PebZIP52. All these differences might be linked to the functional diversity of PebZIPs.

Figure 3.

Conserved domains and predicted cis-elements in promoters of PebZIPs

(A) Phylogenetic analysis and the conserved domains of PebZIPs.

(B) Distribution of related cis-acting elements in PebZIP promoters (2000bp upstream of transcription start site of gene).

Because bZIP regulates the expression of the corresponding gene in the form of homo- or hetero-dimer, the prediction of cis-regulatory elements in PebZIP gene promoters could improve the transcriptional regulatory knowledge (Figure 3B). Among them, TATA-box, CAAT-box, MYB, and MYC were found in all PebZIPs, with a count ratio of 23%, 16%, 13%, and 12%, respectively. Some hormone-related elements were identified, such as abscisic acid responsive element (ABRE, AAGAA), MeJA (Methyl jasmonate)-responsive motif (TGACG, CGTCA), salicylic acid-responsive element (TCA, CCATCTTTTT), gibberellin-responsive motif (TATC, GARE, P box) and the auxin-responsive element (TGA, AuxRR). The drought-responsive elements (MBS, CAACTG), low-temperature responsive element (LTR), WRE3, W box, and WUN motif response to wound-related biotic or abiotic stress were also found in some genes, which suggests that PebZIP genes might be regulated by stress conditions, growth and developmental conditions.

Chromosomal location and synteny analysis of PebZIPs

Using passion fruit genome annotation information, we visualized the chromosomal distribution of the PebZIPs. The 52 PebZIPs were unevenly distributed on 9 chromosomes, with the majority of them were located at both ends of the chromosomes (Figure 4A). The majority of the bZIPs were located on Chr1 (16 genes) and Chr6 (10 genes), followed by Chr2 and Chr7, each of which contained 6 bZIPs, whereas Chr5, 8, and 9 contained only 2 bZIPs each.

Figure 4.

Chromosomal distribution and synteny analysis of PebZIPs

(A) Gene duplication events in P. edulis genome.

(B) Gene duplication events between P. edulis with A. thaliana, O. sativa, P. trichocarpa, and V. vinifera.

PebZIPs gene duplication events were further analyzed. The 26 collinear pairs were found, in which half of them with 1 paralogous gene, and the another half with two homologous genes (Figure 4A). Three genes (PebZIP44, -45, −51) were identified as homologous genes to each other, and seven genes (PebZIP13, -14, −20, −21, −26, −37, −38) came into a homologous network, in which PebZIP20 had five homologous genes, PebZIP14 had four homologous genes. The highest number of duplication was found on Chr 1, whereas Chr9 had the lowest number of duplicated gene pairs, and Chr5 had no duplicated gene pairs. The Ka/Ks values of all PebZIP pairs were found to be less than 1 (see Table S1), which indicates that PebZIPs might have undergone strong purifying selection during evolution.²⁴

Furthermore, we also performed a collinearity relationship between bZIP members in P. edulis, A. thaliana, Oryza sativa, P. trichocarpa, and V. vinifera (Figure 4B). A total of 31, 14, 47, and 48 PebZIP genes have collinearity relationships with 33 (AtbZIPs), 13 (OsbZIPs), 55 (PtrbZIPs), 38(VvbZIPs) genes. When grouping all those collinearity pairs, we found 52 PebZIPs in close relationship with the other bZIPs, in which most number pairs were with PtrbZIPs, then followed by VvbZIPs and AtbZIPs, the least number pairs were observed in OsbZIPs, which is consistent with the evolutionary distance among these species.²⁵ In addition, we found that a large number of the PebZIPs have collinearity relationships with three to four PtrbZIPs, suggesting that these genes may play an important role in the evolution of the gene family.

GO enrichment and KEGG orthology of PebZIPs

A gene ontology (GO) enrichment analysis was performed to further characterize the predicted functions of the bZIP proteins (Figure 5). The PebZIP gene pairs of P. edulis were found to be involved in cellular components, molecular functions, and biological processes. All PebZIPs are associated with intracellular membrane-bounded organelle, intracellular organelle, and intracellular part in the cellular component. The molecular functions of PebZIPs focus mainly on sequence-specific DNA binding transcription factor, heterocyclic compound binding, and organic cyclic compound binding. The essential biological process of PebZIPs includes regulation of the biological process, cellular response to stimulus, response to chemical stimulus, and biotic or abiotic stimulus.

Figure 5.

Gene ontology classification of PebZIPs

The KEGG orthology analysis was also conducted according to the annotation of RNA-seq results. All PebZIPs are classed into 10 KEGG ortholog: K09060, K14431, K14432, K16241, K20557, K04374, K23541, K03248, K14292, and K03452 (Figure 6). There are 16 PebZIPs related in K09060 (G-box-binding factors), the members of which are mainly from S and G group. K14431 (TGA) members are all from the D group; K14432 (ABA) mainly from the A group; K20557 (VIP1) from I group; K04374 (ATF4, CREB2) from E group; K16241 (HY5) from H and K group. Putting the synthetic relationships of bZIP with four other plant (Arabidopsis, rice, popular, and grape) in the network, six clusters were grouped, the groups G, S and C, I and E, H and K are clustered together. The only member of group J clustered together with group G and S, and the only member of group B clustered together with group H and K indicating the function relations of those members. From this network, we can see more clearly the relationship of PebZIP members with other plant bZIPs. This also helps to predict KEGG functions of PebZIPs by related KEGG orthologs (Figure 6).

Figure 6.

Network of KEGG orthologs and collinearity of PebZIPs with AtbZIPs, OsbZIPs, PtrbZIPs, and VvbZIPs

Protein-protein network interaction of PebZIP gene encodes proteins

The co-expression network was created using an ortholog-based method to better understand the interactive relationship between PebZIPs and target genes (Figure 7). The 30 PebZIPs and 27 target proteins constitute the left and right circles of the protein interaction network, respectively. The top 8 genes with the highest degree (number of connected proteins) are PebZIP9, -51, −42, −6, −8, −20, −31, and −29. The top six target proteins with the highest degree include COP1, PIF3, PHYA, PHYB, CRY1, and SPA1, which are associated with the circadian rhythm of plants. Among these target proteins PIF3 (HLH), PP2CA, BIN2, SNRK2.3, KEG, and NPR1 are known participants in plant hormone signal transduction, particularly in response to ABA signal. PP2CA, MPK3, and SNRK2.3 participate in the MAPK signaling pathways. PIF3, PHYB, PHYA, CRY1, SPA1 (WD40), UVR8, RUG1, RUG2, RUG3, DET1, and BBX21 participate in the light responses or photoreceptor activity. DREB2A, SNRK2.3, HD1 (ERF), and LAF1 (MYB) participate in abiotic stimulus responses, such as cold and drought stress. COP1 (WD40), MPK3, NPR1, HD1, and GFX480 participate in the immune system to defense response to diseases and other organisms.²⁶

Figure 7.

The Predicted Protein Interaction Network of PebZIP proteins with target proteins

Left circle is PebZIP proteins, right circle is target proteins, the area size indicates the degree of each protein’s correlation with other proteins.

The four PebZIPs with the highest degree (PebZIP9, -51, −42, −6) have high homology with GBF3G, HY5, HYH, and ABI5, respectively, that may associate with light or ABA-mediated stress response. Three PebZIPs caught our attention, namely PebZIP8, PebZIP20, and PebZIP29, which correspond to AtbZIP60, AtbZIP44, and AtbZIP53. AtbZIP60 consists of a bZIP DNA binding domain followed by a putative transmembrane domain that plays a role in plant immunity and abiotic stress responses.²² AtbZIP44 is involved in the positive regulation of seed germination by binding to the DNA G-box motif of the MAN7 promoter.²⁷ AtbZIP53 binds to DNA with the C-box-like, G-box-like motif, ABRE elements, DOF, I-box, BS1, and MY3 in target gene promoters and heterodimerizes with other bZIP proteins.²⁸ These genes may play crucial roles in biotic or abiotic stress resistance.

Expression analyses of PebZIPs under abiotic stresses and pulp development process

The expression profiles of PebZIPs under various abiotic stresses were investigated using RNA-seq data. The result showed that the 12 group genes of PebZIPs had different levels of response to different abiotic stresses (Figure 8A). The expression pattern of PebZIP29, PebZIP21, and PebZIP8 were upregulated under L1 (0°C for 20 h), L2 (0°C for 48 h), G5 (45°C for 24 h), D2 (10% soil moisture), and N1 (300 mM NaCl solution 3 days) stress. PebZIP38 and PebZIP19 were upregulated under D1 (50% soil moisture), N2 (300 mM NaCl solution 10 days), G1 (45°C for 2 h), and G2 (45 °C for 4 h) stress. PebZIP8, PebZIP10, PebZIP13, and PebZIP14 showed high expression levels in three pulp maturing times of P. edulis (Figure 8B). Almost half of the PebZIPs expression was decreased in Time 3 (1 week after harvest, the peel has shrunk). These differential expression genes related to the expression of proteins and metabolic compounds, indicated that it may associate with biotic or abiotic stress resistance.

Figure 8.

Heatmap of transcription expression levels of PebZIPs

(A) Gene expression data under abiotic stress, CK: control; D1:50% soil moisture; D2: 10% soil moisture; N1: 300 mM NaCl solution 3 days; N2: 300 mM NaCl solution 10 days; L1: 0°C for 20 h; L2: 0°C for 48 h; G1: 45°C for 2 h; G2: 45 °C for 4 h; G5: 45°C for 24 h; (B) Gene expression data in three pulp maturation time.

Time 1: 2 weeks before harvest, the peel is green; Time 2: the harvest time, the peel turns red without shrinkage; Time 3: 1 week after harvest, the peel has shrunk.

qPCR analysis of 4 key PebZIP genes

By combining the results of the PebZIP family gene analysis, four genes (PebZIP8, PebZIP14, PebZIP20, and PebZIP29) that may play key roles in the stress resistance were selected for qRT-PCR analysis. The expression of these genes was analyzed and verified in different varieties of Passiflora (purple fruit P. edulis, yellow fruit P. edulis, Passiflora alata, Passiflora foetida) (Figures 9A and 9B), and different tissues of P. edulis (Figures 9C and 9D) and P. foetida leaves infected with telosma mosaic virus (TeMV) (Figures 9E and 9F).

Figure 9.

Relative expression of key genes in different tissues and varieties of Passiflora

(A) Relative expression of 4 key genes in four varieties of Passiflora.

(B) Pictures of leaves of four varieties (purple fruit P. edulis, yellow fruit P. edulis, P. alata, and P. foetida).

(C) Relative expression of 4 key genes in different tissues of purple fruit P. edulis.

(D) Pictures of leave, root, flower, stem, peel, and pulp of purple fruit P.edulis.

(E) Relative expression of 4 key genes in healthy and Telosma mosaic virus (TEAV) infected P. foetida leaves.

(F) Pictures of healthy and TEMV virus infected P. foetida leaves. Bar = 2 cm.

In different tissues of P. edulis, PebZIP21 was more higher expressed in flowers, followed by pulps, stems, and peels. The expression of PebZIP8 was higher in flowers, peels, and stems than in leaves, whereas PebZIP14 was mainly expressed in leaves and peels, and the expression of PebZIP29 was higher in pulps and leaves.

In two species (P. alata, P. foetida) and two genotype of P. edulis (purple fruit, yellow fruit), PebZIP8 and PebZIP21 were highly expressed in P. foetida and P. alata, respectively. The expressions of PebZIP8 and PebZIP21 in purple fruit P. edulis was higher than that in yellow fruit P. edulis, which was consistent with the transcriptome results. The expression of PebZIP14 was expressed highest in yellow fruit P. edulis. The expression of PebZIP29 in different leaves did not show much difference, but PebZIP29 responded to cold exposure and was highly expressed in maturing fruit, which may be related to anthocyanin synthesis in purple fruit.

We also found that, compared with the healthy P. foetida, that the expression of PebZIP8 and PebZIP14 was barely detectable in the leaves of virus-infected P. foetida. Therefore, we speculate that PebZIP8 and PebZIP14 may negatively regulate and PebZIP20/21 positively regulate the virus resistance of the passion fruit.

Metabolomic and transcriptomic co-expression analysis to identify the structural properties of PebZIP20 and PebZIP21

We also analyzed the expression of PebZIPs in the leaves of two genotype of P. edulis, the relatively susceptible purple fruit P. edulis, and the relatively disease-resistant yellow fruit P. edulis.^29,30 Expression levels of PebZIP14 and PebZIP20 were significantly higher in the purple fruit P. edulis leaves than in the yellow fruit P. edulis leaves, whereas PebZIP21 was highly expressed in the yellow fruit P. edulis leaves (Figure 10A).

Figure 10.

Heatmap of transcript expression levels of PebZIPs in two genotype and a network of key PebZIPs co-expressed with DAMs (differential accumulation metabolites)

(A) Heatmap of transcript expression levels of PebZIPs between the relative susceptible purple fruit P. edulis and the relatively resistance yellow fruit P. edulis.

(B) Co-expression network of metabolome and transcriptome. Detailed information is shown in Table S9.

The metabolomic data detected 294 compounds with significant difference (⎟LogFC⎟≥1) in purple fruit P. edulis and yellow fruit P. edulis leaves (see Table S8). Of these, 136 compounds were downregulated and 158 were upregulated in purple fruit P. edulis leaves compared to yellow fruit P. edulis. Flavonoids, others, and amino acids and derivatives accounted for 29%, 13%, and 11% of all DAMs (differential accumulation metabolites), respectively.

According to the transcriptome and metabolome association analysis of purple fruit P. edulis and yellow fruit P. edulis leaves, we found that 6 PebZIPs were co-expressed with upregulated compounds and 7 PebZIPs were co-expressed with downregulated compounds in 3 (co-upregulated) and 7 (co-downregulated) quadrants of nine quadrants results (⎟log2 Fold Change⎟, ⎟logFC⎟≥2, Pearson’s correlation coefficient, PCC>0.95) (Figure 10B). The expression of most flavonoids (70% of all flavonoids) was significantly upregulated in purple fruit P. edulis, and most flavonoids were in the class of anthocyanins, which were related to the dark purple color of the fruit (see Table S9). All lignin and coumarins, and tannins were upregulated in DAMs compared with yellow fruit P. edulis, whereas the expression of most amino acids and derivatives, and lipids was all downregulated in purple fruit P. edulis (Figure 10B). These co-downregulated genes and metabolites in purple fruit P. edulis means that they were co-upregulated in yellow fruit P. edulis, In them, PebZIP21 co-expressed with most of the upregulated DAMs and PebZIP20 co-expressed with most of the downregulated DAMs. The top five metabolites highly accelerated metabolites in yellow fruit P. edulis were: choline alfoscerate, quercetin −3, 4′-O- di-glucoside, L-ornithine, D-threonic acid, apigenin -7-O- rutinoside - 4'-O- rhamnoside (Table 2). Although the top five highly accelerated metabolites in purple fruit P. edulis leaves were: cyanidin-3-O-galactoside, cyanidin-3-O-glucoside (Kuromanin), luteolin-7-O-glucoside (Cynaroside), 1-O- (3, 4 - Dihydroxy -5-methoxy-benzoyl)-glucoside, and cyanidin-3-O-rutinoside (Keracyanin).

Table 2.

Top 20 DAMs and co-expression PebZIPs in two types of P. edulis

No.	PebZIPs	Purple fruit P.edulis	Class	LogFC	PebZIPs	Yellow fruit P.edulis	Class	LogFC
1	PebZIP21	Cyanidin-3-O-galactoside	Flavonoids	22.20	PebZIP20	Choline Alfoscerate	Lipids	−17.00
2	PebZIP21	Cyanidin-3-O-glucoside (Kuromanin)	Flavonoids	22.11	PebZIP37	Choline Alfoscerate	Lipids	−17.00
3	PebZIP21	Luteolin-7-O-glucoside (Cynaroside)	Flavonoids	19.43	PebZIP56	Choline Alfoscerate	Lipids	−17.00
4	PebZIP21	1-O-(3,4-Dihydroxy-5-methoxy-benzoyl)-glucoside	Phenolic acids	18.54	PebZIP20	Quercetin-3,4′-O-di-glucoside	Flavonoids	−15.84
5	PebZIP21	Cyanidin-3-O-rutinoside (Keracyanin)	Flavonoids	18.17	PebZIP37	Quercetin-3,4′-O-di-glucoside	Flavonoids	−15.84
6	PebZIP21	Dihydrokaempferol-7-O-glucoside	Flavonoids	16.77	PebZIP20	L-Ornithine	Amino acids	−15.04
7	PebZIP21	Apigenin-7-O-glucoside(Cosmosiin)	Flavonoids	16.30	PebZIP37	L-Ornithine	Amino acids	−15.04
8	PebZIP21	Geniposide	Terpenoids	16.20	PebZIP56	L-Ornithine	Amino acids	−15.04
9	PebZIP21	Epigallocatechin	Flavonoids	15.99	PebZIP20	D-Threonic Acid	Others	−15.01
10	PebZIP21	Quercetin-3-O-rutinoside-7-O-glucoside	Flavonoids	14.79	PebZIP37	D-Threonic Acid	Others	−15.01
11	PebZIP21	N1,N8-Bis(sinapoyl)spermidine	Alkaloids	14.61	PebZIP56	D-Threonic Acid	Others	−15.01
12	PebZIP21	p-Coumaroylputrescine	Alkaloids	14.24	PebZIP20	Apigenin-7-O-rutinoside-4′-O-rhamnoside	Flavonoids	−14.77
13	PebZIP21	Eriodictyol-7-O-glucoside	Flavonoids	14.17	PebZIP37	Apigenin-7-O-rutinoside-4′-O-rhamnoside	Flavonoids	−14.77
14	PebZIP21	Casuariin	Tannins	14.11	PebZIP56	Apigenin-7-O-rutinoside-4′-O-rhamnoside	Flavonoids	−14.77
15	PebZIP21	Dihydrokaempferol-3-O-glucoside	Flavonoids	14.09	PebZIP20	Diosmetin-7-O-rutinoside (Diosmin)	Flavonoids	−14.37
16	PebZIP21	6-O-Caffeoylarbutin	Phenolic acids	14.04	PebZIP37	Diosmetin-7-O-rutinoside (Diosmin)	Flavonoids	−14.37
17	PebZIP21	Quercetin-3-O-arabinoside (Guaijaverin)	Flavonoids	13.34	PebZIP20	4-Hydroxy-2-oxoglutaric acid	Organic acids	−13.64
18	PebZIP21	1,3-O-Dicaffeoylglycerol	Lipids	13.25	PebZIP37	4-Hydroxy-2-oxoglutaric acid	Organic acids	−13.64
19	PebZIP21	Tercatain	Tannins	12.64	PebZIP20	Chrysin-6-C-arabinoside-8-C-glucoside	Flavonoids	−13.59
20	PebZIP21	Aromadendrin-7-O-glucoside	Flavonoids	12.64	PebZIP37	Chrysin-6-C-arabinoside-8-C-glucoside	Flavonoids	−13.59

The KEGG ortholog involved in PebZIPs20 and PebZIPs21 is K09060 (GBF; plant G-box-binding factor). PebZIP8, PebZIP11, and PebZIP44 are involved in K16241(HY5) ortholog, co-expressing mainly with lignans and coumarins and phenolic acids. The above indicates that the PebZIPs may regulate metabolite content through these KEGG ortholog involved pathways.

PebZIP21 co-expressed with most upregulated compounds, and PebZIP20 co-expressed with downregulated flavonoids, organic acids, and nucleotide and its derivatives. To explore the difference between these two genes, the structure of the predicted proteins was compared. PebZIP20 and PebZIP21 are tandem repeat genes and have similar amino acids, both have 155 aa, but only have 3 alleles and one amino acid difference, in which one allele changing in from C to G at positions 187, resulting in a change from leucine to valine at position 63 amino acid (Figures 11A and 11B). The 3D model of these two proteins was generated by SWISS-MODEL protein structure homology-modeling (7x5e.2). As shown in Figure 11C, the heterodimeric bZIP binds to DNA through the N terminus of the protein. The 63aa of PebZIP20 and PebZIP21 is located at the center of both bZIPs, so we hypothesized that the mutations at 63aa would affect the formation of homo- or heterodimers of PebZIP20 and PebZIP21, which would then affect the transcription and translation. This leads to differential expression of downstream proteins and catalytic synthesis of metabolites. The changes in one allele related to the functions difference of PebZIP20 and PebZIP21 indicate that these two genes are the key bZIPs in regulating the downstream proteins and secondary metabolites, which is worthy of further studies.

Figure 11.

The cds, amino acids, and protein models of PebZIP20 and PebZIP21

(A) Full sequences of cds and amino acids of PebZIP20 and PebZIP21, the mutant bases pairs and amino acids are marked in red color.

(B) Predicted protein model of PebZIP20 and PebZIP21. The 63 amino acid are L(Leucine) in PebZIP20, and V(Valine) in PebZIP21.

(C) SWISS-MODLE predicted hetero-dimer model of PebZIP20 and PebZIP21 structural domains bound to DNA(CsMBE1) template (7x5e.2). Bar = 1 nm.

Discussion

The bZIP TFs are involved in pathogen defense, stress and light signaling, flower development, and seed maturation of plants.³¹ The diversified function of bZIP TFs are associated with their structures of gene or protein and are achieved by interacting with other TFs and functional proteins in cells. The members of group A of bZIP are mainly involved in abiotic stress responses such as ABA and drought; the C group includes mostly stress-related genes; the G group is mainly involved in light signal transduction and seed maturation; and the H group is mostly involved in photosynthesis; the I group includes genes mostly involved in the synthesis and metabolism of gibberellin; and the S group may be involved in the stress response and carbohydrate signal metabolism.⁸ Members of the same group have certain similarities in amino acid sequence length, recognized DNA sequence, and functions.

As a larger group of bZIP TFs, the S group plays an important role in abiotic stress response, defense against pests and diseases, and carbohydrate signal transduction. In this study, we found several key genes in the S group that may be related to disease or abiotic stress resistance. PebZIP20 and PebZIP21 are highly co-expressed with up or downregulated compounds in P. edulis. Those two genes together with PebZIP13, -14, -37, and -38 are phylogenetically closest to AtbZIP44 and AtbZIP11 (Figure 1). Same in the S group, PebZIP29 is phylogenetically closest to AtbZIP53. Although the promoter of AtbZIP44, AtbZIP11, and AtbZIP53 contained the conserved HSE (nGAAnnTTCn) binding to heat-stress element (HSE) of LlHsfA3A and LlHsfA3B in lily (Lilium longiflorum). Overexpression of LlHsfA3A or LlHsfA3B in A. thaliana affected proline oxidation via regulation of AtbZIP11, AtbZIP44, and AtbZIP53, finally conferring increased thermotolerance and salt sensitivity.³² One HSE (aGGAgaTTCc) motif was also found in the promoter (−2000bp) of PebZIP20 and PebZIP21, and the expression of these two bZIPs was upregulated under stress induction (Figure 9). Therefore, we speculate that PebZIP20 and PebZIP21 may have similar functions to regulate high temperature or salt resistance. The promoter of PebZIP20 and PebZIP21 also contained ABA-related motifs (AAGAA-motif, ARE, ABRE), 4 MYB, 4 MYC, 3 WRE3, 3 GARE-motif (gibberellin-responsive element), 2 TCA-element (CCATCTTTTT, involved in salicylic acid responsiveness), 2 MeJA-responsiveness (TGACG-motif, CGTCA-motif) motifs, 1 DRE1(drought response element) and 1 RY-element (CATGCATG, involved in seed-specific regulation), which supports its multifunctionality in gene binding and regulation.

PebZIP8 and PebZIP10 are members of the K group, phylogenetically closest to AtbZIP60. AtbZIP60 acts as during endoplasmic reticulum (ER) stress regulator involved in drought tolerance with a putative transmembrane domain near its C terminus. AtbZIP60 also regulates the expression of Ca2+-dependent protein kinase genes,³³ and is involved in regulating virus resistance.³¹ AtbZIP60, heterologously expressed in tobacco, rice, and swamp pine (Pinus elliottii) have enhanced resistance to salt stress, enhanced superoxide dismutase activity, and slowed lipid peroxidation. AtbZIP60 interacts with the transcription factor MYB7 to negatively regulate ABA-induced seed dormancy and promote seed germination.³⁴ Silencing or knockout of NbbZIP60 inhibits accumulation of virus, suggesting that NbbZIP60 promotes viral infection.¹ NbbZIP60 is a homology of NtbZIP60. PebZIP8 is phylogenetically closest to AtbZIP60, whereas NtbZIP60 is phylogenetically close to AtbZIP60.³⁵ The expression of PebZIP8 was downregulated in TeMV-infected P. foetida than in healthy ones. Therefore, we speculate that PebZIP8 may play a regulatory role in virus infection.

Secondary metabolites have a variety of complex biological functions and play an important role in coping with environmental stress, competition, and co-evolution between species, and in attracting insects for pollination. There are many kinds of secondary metabolites, such as flavonoids, alkaloids, terpenes, phenols, etc. bZIP TFs can regulate the expression of genes related to secondary metabolic pathways, activate their biosynthetic pathways, and promote the efficient accumulation and directional synthesis of active ingredients in plants. The bZIP TFs in rice can positively or negatively regulate the expression of terpenoid phytochemical synthesis genes.³⁶ Overexpression of the OsbZIP79 gene can inhibit the expression of diterpene phytoalexin biosynthesis genes, and the content of diterpene phytoalexin decreased sharply, which reduced the disease resistance of rice.³⁷ The key chemicals of Chinese herbal medicines such as Artemisia annua, ginseng, and Salvia can be effectively improved by regulating the expression of bZIP TFs.³⁸ The CaLMF TFs in Camptothecus negatively regulates the expression of camptothecin.³⁹ The CrGBF1 and CrGBF2 TFs can specifically bind to the G-box in the promoter of the terpenoid indole alkaloid synthesis key gene and inhibit the synthesis of terpenoid indole alkaloids.^40,41 Metabolites co-expressed by PebZIPs include flavonoids, lipids, amino acids, terpenoids, and alkaloids. Some of those metabolites, especially flavonoids, showed significant differences between yellow fruit and purple fruit P. edulis. These metabolites may be indirectly regulated by PebZIPs and can enhance plant resistance.

Flavonoids are derived from the phenylpropane metabolic pathway, and bZIP TFs regulate the synthesis of different kinds of flavonoids, which have important functions such as defense against pathogens, adaptation to environmental stress, and improvement of crop quality.⁴² Anthocyanins belong to flavonoids and are widely found in the plant. The HY5 or HYH TFs promote the accumulation of anthocyanins and increase the cold resistance of Arabidopsis by upregulating the key gene DFR downstream of the anthocyanin biosynthesis pathway.⁴³ The DkbZIP5 TF binds to the ABRE element in the DkMyb4 gene promoter region to directly regulate the expression of the DkMyb4 gene and affects the synthesis of proanthocyanidins.⁴⁴

In our research, we found even in the same structure group, the functions of different bZIPs vary greatly. Some PebZIPs were separated into two groups which co-expressed with up or down-expressed DAMs of relative resistance and susceptible genotypes. Although the relationship and mechanism between the function and structure of PebZIPs are still unclear.

The function of TFs is closely related to their structure. Group A bZIP has a phosphorylation site in the conserved sequence, group C bZIP has a hydrophobic or acidic transcriptional activation domain at the N-terminus, and group D bZIP can bind to the TGACGT core sequence in DNA.⁴⁵ The function of PebZIPs was predicted by comparing their phylogenetic evolution with AtbZIPs (Table 3). The function of several key genes is highly valued, in which PebZIP14, -20, -21, -29 (S group) could regulate the function of multiple stress response; PebZIP31, -43 (E group) could be involved in pathogen defense and abiotic stress response; and PebZIP8, -10 (K group) could participate in salt resistance. The functional prediction of PebZIPs will give us a better reference for the future mining and utilization of these genes.

Table 3.

Function prediction of PebZIPs

Group	PebZIPs	No.	AtbZIPs	Prediction Function	Motif	Reference
A	PebZIP6, 12, 16, 19, 36, 47, 53	7	AtbZIP27, 36, 40	Involved in phosphorylation, ABA stress responsive	ABF2/AREB1, GBF4	Choi et al.46; Rabara et al.47; Schütze et al.48
B	PebZIP 52	1	AtbZIP17, 28	Salt stress		Liu et al.49
C	PebZIP7, 23, 34, 46	4	AtbZIP10, 63, 9, 25	Basal defense, positive regulator of pathogen response. Repression of oxidative cell death, seed germination	BZO2H1	Kaminaka et al.50
D	PebZIP1, 2, 4, 11, 18, 24, 32, 40, 50, 54, 56	11	AtbZIP20, 21, 22, 26, 45, 50	Defense against pathogens and development, NPR related resistance	DOG1, TGA, HBP1b_like	Jakoby et al.45
E	PebZIP31, 42, 48	3	AtbZIP61, 76, 78	Pathogen defense, abiotic stresses, auxin- mediated response	RF2	Xiang et al.51; Miao et al.52
F	PebZIP3	1	AtbZIP19, 23, 24	Zinc deficiency response		Castro et al.53
G	PebZIP9, 25, 28, 30	4	AtbZIP41, 54, 55	Seed storage, protein production, phosphorylation, stress response	GBF	Oñate et al.54; Choi et al.46
H	PebZIP15, 42, 44, 45, 51, 55	6	AtbZIP56, 64	Promotes photomorphogenesis	HY5	Ang et al.55
I	PebZIP17, 22, 35, 49	4	AtbZIP29, 18, 30, 51	Controls the activity of gibberellin; height of the plant.	VIP1	Van et al.56
J	PebZIP5	1	AtbZIP62	Drought stress response	GBF1	Rolly et al.57
K	PebZIP8, 10	2	AtbZIP60	Salt, cold,virus resistance	HY5	Tang et al.58
S	PebZIP20, 21, 29, 13, 14, 26, 27, 33, 37, 38, 39, 41	12	AtbZIP53, 11, 44, 23, 58, 75	Stress response (drought, cold, wound), sucrose signaling, carbohydrates, repress gibberellin production.	ATB2, GBF5	Alonso et al.28; Fukazawa et al.59

Plant bZIPs are crucial for regulating abiotic and biotic stresses. bZIP also handles the salicylic acid, jasmonic acid, and ABA signaling pathways, and are indispensable for maintaining plant tolerance to pathogen and pest attacks. The structural and taxonomic properties of PebZIPs in response to abiotic stress are investigated in this work. It is important to understand the complex regulatory network of bZIP in response to different stresses. The functions of bZIP factor can be validated by genetic engineering techniques to transgenically over-express bZIP or to regulate the expression level of specific bZIPs, which may lead to resistance or higher quality crops. Recent reports on CRISPR/Cas9-mediated targeted mutagenesis of bZIP TFs in Salvia miltiorrhiza and V. vinifera indicate the importance of bZIP TFs in crop breeding.^60,61 Given the diversity and importance of bZIP biological functions, research on PebZIPs is encouraged to test the hypothesis of putative functions.

Conclusions and future outlook

In this study, we provide a comprehensive analysis on the bZIP TFs of P. edulis through genome-wide analysis. A total of 56 PebZIPs were identified from passion fruit genome and their structural features were investigated. Furthermore, phylogeny relationships with A. thaliana, rice, grape, and popular, chromosomal distribution, duplication events, conserved motifs, and expression levels under various abiotic stresses were all investigated. This was done in two major genotypes of P. edulis. In addition, key gene expression levels of bZIPs may help to provide a preliminary knowledge of their function under biotic or abiotic stress situations. Moreover, we can predict key PebZIPs and related metabolites that may play a role in resistance through transcriptome and metabolome co-expression analysis. These key PebZIPs may be applied in the future for molecular resistance breeding in passion fruit.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples
Purple fruit Passiflora edulis	This paper	N/A
Yellow fruit Passiflora edulis	This paper	N/A
Passiflora alata	This paper	N/A
Passiflora foetida	This paper	N/A
Critical commercial assays
RevertAid First Strand cDNA Synthesis Kit	Thermo Scientific	CAT# K1622
SYBR ® Premix Ex Taq™(Tli RNaseH Plus)	TaKaRa	CAT# DRR420A
Deposited data
Genome data	https://ngdc.cncb.ac.cn/search/?dbId=gwh&q=GWHAZTM00000000	GWHAZTM00000000
Raw data for RNA-seq	This paper	SRP410034
Metabolite data	This paper	https://doi.org/10.17632/b5pbybr6db.1
The co-expression data of PebZIPs with metabolites	This paper	https://doi.org/10.17632/b5pbybr6db.1
Oligonucleotides
qPCR primers	This paper	Table S7
Software and algorithms
MEGAX	Kumar et al.62	https://www.megasoftware.net/
TBtools	Chen et al.63	https://github.com/CJ-Chen/TBtools/releases
Orthovenn2	Xu et al.64	https://orthovenn2.bioinfotoolkits.net/home
Cytoscape software	Shannon et al.65	https://cytoscape.org/
String	Szklarczyk et al.25	https://cn.string-db.org/
Swiss-Model	Marco et al.66	https://swissmodel.expasy.org/
Other
UPLC-ESI-MS/MS	Metware Biotechnology	N/A

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Ma Funing (mafuning@catas.cn).

Materials availability

This study did not generate new unique reagents.

Experimental model and subject details

Plant materials and abiotic stress treatment

The healthy passion fruit seedlings of the purple fruit P. edulis, the commercial name “Tainong”, were grown in the chamber (30°C; 200 μmol·m-2·s−1 light intensity; 12-h light/12-h dark cycle; 70% relative humidity) until a height of about 0.5 m. Each abiotic stress treatment had 8-10 viable, robust seedlings (approximately 2–3 months old, seedlings were grown from tissue culture) with functional leaves. For drought stress analysis, the soil moisture is 50% and 10%. For salt stress tolerance analysis, the seedlings were treated with a 300 mM NaCl solution after 3 and 10 days. For the low-temperature treatment, the plants were treated at 0°C for 20 h and 48 h respectively. For the high-temperature treatment, the plants were treated at 45°C for 2h, 4h, and 24h.

Method details

Identification of bZIP gene family members in passion fruit

The genome data of passion fruit (P. edulis SIMs) were downloaded from the Phytozome V13 (https://ngdc.cncb.ac.cn/search/?dbId=gwh&q=GWHAZTM00000000). PebZIPs were identified by the HMMER (hidden Markov model, http://www.hmmer.org/) using bZIP protein domain (PF001700, PF07716, PF03131) from the PFAM database (http://pfam.xfam.org/) as template and local Blastp (NCBI- Blast-2.11.0+) searching against the genome database of P. edulis using A. thaliana bZIP protein sequence (TAIR, http://www.arabidopsis.org/). The retrieved protein sequences were further confirmed using the online Conserved Domains search tool in NCBI database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). The PebZIPs were finally considered by integrating the above methods and parsed by manual editing to remove the redundancy. Furthermore, the protein iso-electric point (pI) values and molecular weight (MW) for all candidate PebZIPs were computed using the ExPASy proteomic website (http://www.expasy.org/tools/protparam.html). The NetPhos 3.1 Server (http://www.cbs.dtu.dk/services/NetPhos/) was used to predict the protein phosphate site. The subcellular localization of all identified PebZIPs was also predicted through the WoLF PSORT software (https://www.genscript.com/wolf-psort.html).

Phylogenetic, gene structure and conserved motif analysis

The full-length protein sequences of the PebZIPs (56) and Arabidopsis (74) were aligned using MEGAX software (Kumar et al., 2018) to build the neighbor-joining phylogenetic tree. The criteria were adopted with a pairwise deletion option and Poisson correction model with 1000 bootstrap replicates.

The conserved motifs among all the PebZIPs were analyzed by the online tool MEME (Multiple EM for Motif Elicitation, http://meme-suite.org/tools/meme), and NCBI CDD searching (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). The intron-exon organization of all the identified PebZIPs was determined by TBtools software,⁶³ as were the combined visualization of the motifs and the gene structure map. The conserved bZIP domain was analyzed by MEGAX and Jalview software.⁶⁴ A 2000bp of genomic DNA sequence upstream of the transcriptional start site in each PebZIPs was obtained from the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to identify the cis-acting elements in the promoter region of PebZIPs.

The distribution of PebZIPs across all the nine chromosomes of P. edulis genome and the PebZIPs duplications Ka/Ks were mapped and calculated using TBtools software too. Genome data of rice (O. sativa subsp. japonica), popular (P. trichocarpa), and grape (V. vinifera) were downloaded from the NCBI (https://www.ncbi.nlm.nih.gov/) and Phytozome V13, respectively. Evolview software (https://www.evolgenius.info/evolview/) was used to get beautify graphics.

GO enrichment and KEGG orthology

The GO enrichment and the KEGG orthology analysis of PebZIPs were performed by the annotation of the RNA-seq results. WEGO (Web Gene Ontology Annotation Plot, https://biodb.swu.edu.cn/cgi-bin/wego/index.pl) was used to plot GO annotation results. Cytoscape was then used to visualize the KEGG orthology results.⁶⁷

Protein collinearity network analysis of PebZIPs

The orthovenn2 tools (https://orthovenn2.bioinfotoolkits.net/home) were used to compare the PebZIPs with AtbZIPs. The protein interaction network was analyzed by String software (https://string-db.org/). The Multicollinearity Scanning Toolkit (MCScanX) was used to analyze genetic duplication events within the same species and between species.

Transcriptome analysis

The samples of different abiotic stress treatments, three fruit maturity stages (Time 1, two weeks before harvest, the peel is green; Time 2, at harvest time, the peel turns red purple without shrinkage; Time 3, one week after harvest at 30 °C, the peel has shrunk), and two genotypes of P. edulis: the relatively disease-resistant yellow fruit P. edulis, and the relatively susceptible purple fruit P. edulis leaves were used for RNA seq analysis. The FRKP (Fragments Per Kilobase of transcript per Million fragments mapped) data of RNA seq are shown in supplementary information (see Tables S2–S4). The raw data of the transcriptome sequencing of two genotypes was uploaded to NCBI SRA data with SRP 410034. The normalized expression data were used to generate a heatmap using the TBtools software.

qRT-PCR analysis of key PebZIPs

Six different tissues (root, stem, leave, flower, peel, and pulp) of purple fruit P. edulis, two Passiflora species (P. alata, P. foetida), and two genotypes of P. edulis (purple fruit, yellow fruit), and P. foetida infected with Telosma mosaic virus (TeMV) leaves were chosen to detect the relative expression of key PebZIPs. In those plants, the fruit of yellow fruit P. edulis is more sour than that of purple fruit P. edulis, mainly used for beverage juice, and purple fruit P. edulis mainly for fresh eaten. The shoots of yellow fruit P. edulis have always been used as rootstocks for purple fruit P. edulis to improve disease resistance. P. foetida is a wild variety with high resistance, whileP. alata has a big fruit and shaped like a papaya. The tissues and leaves were frozen and the total RNA was extracted using a plant RNA isolation kit with three biological replicates. The expression of PebZIPs in different tissues and cultivars was detected by quantitative real-time PCR (qRT-PCR) analysis using SYBR Premix Ex Taq (TaKaRa, Japan, Tokyo) chemistry on Light96 (Roche). The raw data of qRT-PCR are shown in (see Table S5) (Different tissues), and (see Table S6) (Different cultivars). Relative expression levels were calculated using the 2^−ΔΔCt method. Primer sequences were designed using the Primer 5.0 tool (see Table S7).

Telosma mosaic virus (TeMV) was detected from P. foetida leaves with mottled mosaic symptoms. The RNA was extracted using the plant RNA isolation kit and transferred to cDNA by RevertAid First Strand cDNA Synthesis Kit (Thermo). The forward primer is: 5′GCACTTCACCAGATGTCAATG-3′, and the reverse primer is: 5′AATCGAATGCATATCGCGCCA-3′ were used to amplify a fragment of approximately 816bp. PCR conditions were as follows: initial denaturation at 94°C for 4 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 58 °C for 1 min and an extension step at 72 °C for 1 min, extension at 72 °C for 10min. The fragments are separated by agarose gel electrophoresis (see Figure S2).

Metabolomic analysis

The metabolome was performed by Metware Biotech (Wuhan, China). The 100 mg freeze-dried leaf powder was weighted and extracted overnight at 4°C with 0.6 mL 70% aqueous methanol. Following centrifugation at 10,000g for 10 min, the extracts were absorbed (CNWBOND Carbon-GCB SPE Cartridge, 250 mg, 3 mL; ANPEL, Shanghai, China, www.anpel.com.cn/cnw) and filtrated (SCAA-104, 0.22 μm pore size; ANPEL, Shanghai, China, http://www.anpel.com.cn/) before UPLC-MS/MS analysis. The sample extracts were analyzed using an UPLC-ESI-MS/MS system (UPLC, Shim-pack UFLC SHIMADZU CBM30A system, www.shimadzu.com.cn/; MS, Applied Biosystems 4500 Q TRAP, www.appliedbiosystems.com.cn/). The analytical conditions were as follows, UPLC: column, Agilent SB-C18 (1.8 μm, 2.1 mm∗100 mm); The mobile phase consisted of solvent A, pure water with 0.1% formic acid, and solvent B, acetonitrile. Sample measurements were performed with a gradient program that employed the starting conditions of 95% A, 5% B. Within 9 min, a linear gradient to 5% A, 95% B was programmed, and a composition of 5% A, 95% B was kept for 1min. Subsequently, a composition of 95% A, 5.0% B was adjusted within 1.10 min and kept for 2.9 min. The column oven was set to 40°C; The injection volume was 4 μL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-M. LIT and triple quadrupole (QQQ) scans were acquired on a triple quadrupole-linear ion trap mass spectrometer (Q TRAP), API 4500 Q TRAP UPLC/MS/MS System, equipped with an ESI Turbo Ion-Spray interface, operating in positive and negative ion mode and controlled by Analyst 1.6.3 software (AB Sciex). The ESI source operation parameters were as follows: ion source, turbo spray; source temperature 550°C; ion spray voltage (IS) 5500 V (positive ion mode)/-4500 V (negative ion mode); ion source gas I (GSI), gas II(GSII), curtain gas (CUR) were set at 50, 60, and 30.0 psi, respectively; the collision gas(CAD) was high. Instrument tuning and mass calibration were performed with 10 and 100 μmol/L polypropylene glycol solutions in QQQ and LIT modes, respectively. QQQ scans were acquired as MRM experiments with collision gas (nitrogen) set to 5 psi. DP and CE for individual MRM transitions was done with further DP and CE optimization. A specific set of MRM transitions were monitored for each period according to the metabolites eluted within this period.

Correlation analysis between transcriptomics and metabolomics

Transcriptomic and metabolomic data were uniformly normalized by log2 transformation. The Pearson correlation analysis between differential expressed genes (DEGs) and differential accumulation metabolites (DAMs) was evaluated based on the core function of R language (https://www.r-project.org/) with normalized data. The 3 (co up-regulated) and 7 (co down-regulated) quadrants of ninequadrant results were picked for co-expression analysis, with ⎟log FC⎟≥2, PCC>0.95. (see Table S9 in Medeley Data https://doi.org/10.17632/b5pbybr6db.1).

The Swiss-Model interactive tool⁶⁸ was used to predict the 3D structure of the PebZIP protein and display the 3D structure through Pymol software (https://pymol.org/2/#opensource).

Quantification and statistical analysis

All data are the average of three replicates. Statistical analysis was carried out using software SPSS v 23.0. ANOVA followed by a Duncan’s multiple range test (DMRT) was used to evaluate the differences among different treatments at a significance level of p < 0.05.

Published: April 1, 2023

Data and code availability

• •RNA-seq data have been deposited at GEO and are publicly available (SRP410034). Metabolomic data have been deposited at Mendeley Data (https://doi.org/10.17632/b5pbybr6db.1). Accession numbers are listed in the key resources table.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.