Structure characteristics of mutation sites in two waxy alleles from Yunnan waxy maize (Zea mays L. var. certaina Kulesh) landraces

A large number of waxy maize landraces are distributed in Yunnan and surrounding areas, and abundant waxy alleles of different types are distributed in these landraces. The identification of waxy alleles is helpful to the protection and utilization of these waxy landraces. This study introduced structure characteristics of waxy genes from two specific landraces of Yunnan, Zinuoyumi and Myanmar Four-Row Wax. Zinuoyumi has two waxy alleles wx-Cin4 and wx-Cin4-2; Myanmar Four-Row Wax has three waxy alleles wx-D10, wx-Reina and wx-D11. The wx-Cin4-2 and wx-D11 are two types of waxy alleles first reported in this study. The wx-Cin4-2 has two mutation sites, deletion of 30 bp in exon 10, insertion of a 1,267 bp non-long terminal repeat (non-LTR) retrotransposon Cin4 in intron 10, and 13 bp extra sequence were found at 5’ end of the Cin4; the mutation site of wx-D11 is a 1,082 bp deletion from exons 11 to 14 of the waxy gene and is replaced with a 72 bp filler sequence. This study enriched the type of waxy allele from Yunnan waxy maize landraces and further discussed the molecular basis for the formation of mutation sites of wx-Cin4-2 and wx-D11.


Introduction
Maize (Zea mays L.) originated in South America and was introduced to China ~400-500 years ago [1].In Southwest China, residents have the habit of waxy food.Maize is under selection pressure since the introduction, and then waxy maize (Zea mays L. var.certaina Kulesh) forms a branch of maize [2].Amylopectin is almost all starch in the endosperm of waxy maize, which makes the grain have a waxy taste [3,4].The waxy taste is controlled by a recessive waxy gene.The maize waxy gene is located on chromosome 9, with a total length of 4.5-Kb and contains 14 exons [5].In the starch synthesis pathway, the waxy gene expresses Granule-Bound Starch Synthase I (GBSS I) protein for amylose synthesis; and the waxy gene is a key gene affecting grain quality.Yunnan is the province with the most ethnic groups in China, which is located on the Yunnan-Guizhou Plateau, bordering Myanmar, Laos and Vietnam.There is a view that Yunnan and its surrounding areas are the origin center and genetic diversity center of Chinese waxy maize [6][7][8][9].There are abundant waxy maize landraces [10], such as an ancient landrace termed Four-row Wax, which has only four rows of seed set in the cob and many characters similar to those of wild species.It was collected from Menghai County of Yunnan in 1970 and has been planted by the local Dai minority since 1890 [7,11].
It is necessary to prevent the loss of rare alleles during the preservation and reproduction of germplasm resources.Identification of alleles is one of the effective measures to prevent the loss of rare alleles.So the identification of waxy alleles is helpful to the protection and utilization of these waxy landraces.So far, a large number of Yunnan waxy maize landraces still contain unknown waxy alleles that need to be identified.In this study, we identified and introduced structure characteristics of two rare waxy alleles from two specific landraces of Yunnan, Zinuoyumi and Myanmar Four-Row Wax.

Plant material
In this study, 405 waxy maize landraces from different regions of Yunnan were used as research material mentioned in our previous work [15].The maize inbred line B73 was used as the non-waxy maize control.Seeds of research material were sampled from the Yunnan Provincial Crops Genebank of Yunnan Academy of Agricultural Sciences, P.R.China, and planted at Songming experimental field.I 2 /KI staining was used to identify the waxy grain of maize [19].

Identification of mutation sites of waxy gene
Genomic DNA was extracted from young leaves and used for waxy gene detection [20].Specific markers are used to identify known waxy alleles (Table 1); Molecular markers that can cover the entire waxy gene are used to identify unknown waxy alleles (Table 2) same as our previous work [13].PCR products were used for 1% agarose gel electrophoresis and direct sequencing.

Sequence analysis
Gene structure was drawn using Gene Structure Display Server (GSDS) software http://gsds1.cbi.pku.edu.cn/[21,22].BlastN in NCBI is used to analyze the type of insertion sequences and find reference sequences, and CDD is used to analyze the domain of transposons [23].ORFfinder in NCBI is used to analyze the open reading frame (ORF).Soft-ware Clustal X (1.8) was used for sequence alignment [24].The genome sequence of the maize inbred line, B73 (http:// ftp.maizesequence.org/current/assembly/)[25], was used to identify copy number of transposon with blast tool (version 2.2.23+) [26].The secondary structure of the waxy gene was predicted using RNA structure 6.2 software [27]; the neighbor-joining (NG) tree was constructed using Mega 5 software [28].The primer design software was Primer3 [29].

Results
This study identified the waxy allele composition of the Yunnan-specific waxy maize landraces Zinuoyumi and Four-Row Wax.

Zinuoyumi
Zinuoyumi from Lancang County, Yunnan Province, except purple seed and cob, the purple pigment was deposited in the whole plant during the adult stage.It is an ideal purple gene source in black waxy maize breeding (Fig 2).It is the typical waxy maize with purple seed and cob among all landraces of purple waxy maize in Yunnan, so it is selected as one of the specific landraces used in this study.
Zinuoyumi has two waxy alleles, wx-Cin4 and wx-Cin4-2.Unlike the known waxy alleles, wx-Cin4-2 has two different mutation sites, 30 bp deletion in exon 10 same as wx-D10 and 1,267 bp insertion of Cin4 transposon in intron 10 (Fig 1).However, the allele wx-D10 was not found in this material.It is the first time that waxy alleles with two mutation sites have been identified in waxy maize landraces of the Yunnan area.
Cin4 is a class of non-long terminal repeat (non-LTR) retrotransposon in the maize genome [30].The full-length Cin4 transposon is about 7 Kb in length and has two open reading frames (ORF) to encode the necessary enzymes for jumping in the maize genome.Non-full-length Cin4 is usually truncated at the 5'-end of the full-length Cin4.Both full-length and non-fulllength Cin4 have target site duplications (TSDs) sequence and poly (A) tail at 3'-end.The length of Cin4 in wx-Cin4-2 is 1,267 bp, which is a truncated transposon of Cin4 (Fig 3 ); The 5'-end contains part of the sequence of the second ORF of the full-length Cin4, and its 3'-end is a poly (A) tail of 9 bp; Its TSDs sequence is 5'-GCAACGCGATGGATAA-3'.
In total 53 Cin4 insertion sites were identified from the B73 genome, of which 23 had complete TSDs structure (Table 3).Subsequently, the phylogenetic tree was established with 23 Cin4 sequences, and The Cin4 of wx-Cin4, wx-Cin4-2 and Cin4-2 were clustered into a single clade (Fig 4).The Cin4-2 is a full-length Cin4 with 7,199 bp, which has the complete structure of Cin4 and is used as the reference sequence in this study.Furthermore, the sequences of Cin4      of waxy gene (Fig 6).In addition, we noted that the mutation site of wx-D11 was consistent with wx-Reina, but the TSDs sequence was not the same as wx-Reina.

Myanmar Four-Row Wax
Specific markers for wx-Cin4-2 and wx-D11 were further developed in this study, and agarose gel showed amplicons from Zinuoyumi and Myanmar Four-Row Wax (Fig 7).

Yunnan and its surrounding areas are the secondary origin center of waxy maize
Vavilov's theory on the origin center of crops [32]: The primary origin center is the primary origin center.When the crops spread to a certain range, they will form a new recessive gene- So far, it can be seen that there is no wild ancestor of maize in the Yunnan region; however, a large number of waxy maize resources were accumulated in this area with recessive waxy genes.Some of them are unique to this region, and these materials have special morphological characteristics; abundant recessive gene variation was observed from the waxy site.In addition, we suggest that the phenomenon of multiple alleles is a major molecular characteristic of ancient landraces, and the accumulation of waxy alleles is caused by selection pressure in Yunnan waxy landraces.

The influence of Chinese waxy food culture on crop selection
Maize originated in the Americas continent and was introduced to China about 400-500 years ago [1,33].It is generally believed in the academic community that waxy maize is formed by selection after gene mutations in common maize [34].Waxy maize has been planted in China for over 200 years [6].Maize undergoes genetic mutations within one to two hundred years after being introduced into China.These mutations were noticed by Zhuang, Dong, Buyi, Dai and other ethnic minorities in southwest China who lived in the "waxy rice cultivation area" or "waxy rice cultural circle" or "waxy food cultural circle", and these mutations were selected and retained.These ethnic minorities were developed from their ancestors, the Baiyue people, during the Ming and Qing dynasties.The Baiyue people were the earliest rice cultivators in China, as well as the earliest waxy rice cultivators and eaters.These ethnic minorities have learned and borrowed from their ancestors' experience and knowledge in the selection of waxy rice for the selection of waxy maize [35,36].The waxy mutant of maize, waxy maize, can be preserved and accumulated, forming hundreds of landraces or germplasm resources.Therefore, the waxy food culture and knowledge of ethnic minorities in southwest China, especially in Yunnan and surrounding areas, is the main driving force for the formation of waxy maize in China.

Transposon activity is the main driving force for the formation of waxy allele diversity
Transposon insertion is a major form of waxy gene mutation in maize.Transposons can be divided into DNA transposons and RNA transposons based on the difference in the transposon jumping medium within the genome [37].RNA transposons can further divide into long terminal repeat (LTR) retrotransposon and non-LTR retrotransposon.Compared with non-LTR retrotransposon, LTR retrotransposon has long forward repeat LTR at 5'-end and 3'-end.Non-LTR retrotransposons usually have a 3'-poly (A).Whereas LTR retrotransposon can be further divided into Ty1-Copia and Ty3-Gypsy according to the order of their domains [38].Miniature inverted-repeat transposable element (MITE) is a kind of DNA transposon which is mainly non-self-transposable [39,40].The insertion of different types of transposons represents the structural characteristics of different types of waxy alleles.Our previous research found that, the waxy gene mutation type of waxy maize in Yunnan province involves the above-mentioned transposons (Table 4).The mutations of wx-Cin4-2 were also caused by transposon activity.This further indicates that the activity of transposon is the main cause of the formation of the waxy allele.

Formation of an extra sequence of Cin4 and its biological role
Insertion of the non-LTR retrotransposon into chromosomal DNA is reported to be initiated by a mechanism called target-primed reverse transcription (TPRT), in which the target DNA was cleaved to generate a free hydroxyl (OH).This hydroxyl acts as a primer for reverse transcription using retrotransposon RNA as a template [41,42].However, more details of the second strand synthesis are not very clear [43].In some reports, the following step of transposition was the cleavage of the second strand.The newly synthesized cDNA jumps from the retrotransposon RNA to the second strand at the target site.The second-strand cleavage creates a primer for second-strand synthesis.The primer was a microhomology (MH) sequence between the 5'-end of non-LTR retrotransposon and its target site [42,44].The formation of an extra sequence can change the terminal sequence of transposons, which is conducive to the formation of MH sequences and promotes template switching (Fig 8).

Formation of wx-D11 mutation site
Previous studies have found that wx-B, wx-BI, wx-B6, and wx-C4 are mutations due to waxy gene sequence deletions from exons 1 to 7, and the deletion sequences are replaced by filling sequences [45].These structures are the result of more than one molecular mechanism in the genome.One possibility is that the sequence of the deletion site can form a stable secondary structure.The secondary structure cannot be effectively opened during DNA replication.During the DNA replication process, the absence of replication results in the deletion of the DNA sequence at this site.The deletion of sequences triggers the repair mechanism of the genome, and the repair process introduces filling sequences [45].We analyzed the secondary structure of the deletion sequence of wx-D11 and found that this site could form a typical secondary structure (Fig 9).The other is that, DNA transposition can carry some of its flanking sequences jump to other parts of the genome; sometimes the transposition could induce double-strand DNA breakage, and then the homologous recombination repair mechanism of chromosome would be triggered to repair the DNA in this position, the repair may not be so perfect, caused the formation of filler sequence at the break sites [46].The gene sequence in this article has been submitted to the genebank database and the sequence number wx-Cin4-2 (OR161363) wx-D11 (OR161364) has been obtained.

Fig 1 .
Fig 1. Known waxy alleles in waxy maize landraces.In maize, the wild-type waxy gene is about 4.5 Kb and has 14 exons numbered e1 to e14.The arrows above the schematic of the gene point to insertion mutations; the lines below indicate deletion mutations.Allele wx-Cin4-2 is a 1,267 bp insertional mutation in intron 10.Allele wx-D11 is a 1,082 bp deletion mutation from exon 11 to exon 14. https://doi.org/10.1371/journal.pone.0291116.g001

Fig 3 .
Fig 3. Sequence and structure of the Cin4 in wx-Cin4-2.The full-length Cin4 transposon is about 7 Kb in length and has two open reading frames (ORF) to encode the necessary enzymes for jumping in maize genome.The length of Cin4 in wx-Cin4-2 is 1,267 bp, which is a truncated transposon of Cin4.Both fulllength and non-full-length Cin4 have target site duplications (TSDs) sequence and poly (A) tail at 3' end.https://doi.org/10.1371/journal.pone.0291116.g003

Four-
Row Wax has long been regarded as one of the specific waxy maize landraces in the Yunnan area because its rows number of ears is four (Fig 2).Four-Row Wax was first found in the Dai nationality residential area of Menghai, Yunnan Province, and it is also distributed in other areas.Yunnan provincial repository for corps germplasm now stores Four-Row Wax collected from the Menghai, Menglian, Yingjiang and Namhkan areas of the China-Myanmar border, and the Four-Row Wax from Namhkan was named Myanmar Four-Row Wax by us.It was previously reported that the waxy gene of Four-Row Wax is wx-D10 [16, 17, 31].However, our study found that Myanmar Four-Row Wax has three different alleles, including wx-D10, wx-Reina and wx-D11 which is a new mutation found in this study.Through sequence alignment with the wild-type waxy gene, it was found that the wx-D11 is a 1,082 bp deletion mutation from exon 11 to exon 14 (Fig 1), and replaced by a 72 bp filler sequence.The filler sequence has an 11 bp TSDs structure like to transposon of the genome, and the sequence is 5'-TGGTACGTGTG-3'.The filler sequence between TSDs can be divided into three parts: Part I is an 18 bp forward sequence from exon 10 of waxy gene; Part II is a 17 bp unknown source sequence; Part III is a 37 bp reverse complementary sequence from exon 14

Fig 5 .
Fig 5.The extra sequence of Cin4.(a) Sequence alignment of wx-Cin4 and Cin4-2; (b) Sequence alignment of wx-Cin4-2 and Cin4-2; (c) the structure of extra sequence.There were overlapping sequences between the extra sequences, the target sequence and Cin4 sequence.The red box indicated that the 5' end sequences of wx-Cin4-2 and wx-Cin4 were found not to match Cin4-2, which are extra sequences of Cin4.https://doi.org/10.1371/journal.pone.0291116.g005

Fig 8 .
Fig 8.The possible process of Cin4 jump.(a) The first strand was cleaved at the transposition target site, producing a free hydroxyl (OH) at the nick.(b) The cDNA synthesis used the free hydroxyl (OH) as a primer following template RNA, and the second strand was cleaved at transposition target site.(c) Template RNA leaved from the cDNA during the synthesis of cDNA.After the 3' end of cDNA annealed to a new site of the template RNA, cDNA synthesis restarted.(d) The cleavage of the second strand annealed to cDNA, the second strand synthesis started following newly cDNA.(e) A copy of element was integrated at a new genomic location and was flanked by target site duplications (TSDs).The extra sequence is formed in process (c) and remains in Cin4 after transposition.https://doi.org/10.1371/journal.pone.0291116.g008