Exceptional diversity, non-random distribution, and rapid evolution of retroelements in the B73 maize genome

Abstract

Recent comprehensive sequence analysis of the maize genome now permits detailed discovery and description of all transposable elements (TEs) in this complex nuclear environment. Reiteratively optimized structural and homology criteria were used in the computer-assisted search for retroelements, TEs that transpose by reverse transcription of an RNA intermediate, with the final results verified by manual inspection. Retroelements were found to occupy the majority (>75%) of the nuclear genome in maize inbred B73. Unprecedented genetic diversity was discovered in the long terminal repeat (LTR) retrotransposon class of retroelements, with >400 families (>350 newly discovered) contributing >31,000 intact elements. The two other classes of retroelements, SINEs (four families) and LINEs (at least 30 families), were observed to contribute 1,991 and approximately 35,000 copies, respectively, or a combined approximately 1% of the B73 nuclear genome. With regard to fully intact elements, median copy numbers for all retroelement families in maize was 2 because >250 LTR retrotransposon families contained only one or two intact members that could be detected in the B73 draft sequence. The majority, perhaps all, of the investigated retroelement families exhibited non-random dispersal across the maize genome, with LINEs, SINEs, and many low-copy-number LTR retrotransposons exhibiting a bias for accumulation in gene-rich regions. In contrast, most (but not all) medium- and high-copy-number LTR retrotransposons were found to preferentially accumulate in gene-poor regions like pericentromeric heterochromatin, while a few high-copy-number families exhibited the opposite bias. Regions of the genome with the highest LTR retrotransposon density contained the lowest LTR retrotransposon diversity. These results indicate that the maize genome provides a great number of different niches for the survival and procreation of a great variety of retroelements that have evolved to differentially occupy and exploit this genomic diversity.

Figure 1. Description of the four maize SINE families.

(A) Schematic representation of the four consensus maize SINEs. The size of consensus SINE sequences is indicated for each family and subfamily. The position of A and B motifs that constitute the internal (polymerase III) promoter is shown. The 3′-end similarity of ZmSINE2 and ZmSINE3 is also shown. (B) A sequence comparison of the 3′-ends of ZmSINE2.1, ZmSINE2.2, ZmSINE2.3, ZmSINE3 and the putative LINE partner, LINE1-1, is shown. No significant sequence identity (>50%) was detected between other SINE families and other maize LINE consensus sequences.

Figure 2. Copy number distribution of LTR retrotransposon families in the B73 maize genome.

(A) The result of a homology search using the program RepeatMasker (vers. 3.19) with a library of maize LTR retrotransposon exemplars and (B) the result of a combined structure and homology screen that first uncovered the full-length LTR retrotransposons in the genome and then placed them into families, with 80% identity to an element in the exemplar database required for membership in a family.

Figure 3. The chromosomal distribution of the LTR retrotransposon composition of the B73 maize genome.

The RepeatMasker-identified LTR retrotransposons are summarized as percent composition in 1Mb bins along each of the ten chromosomes. The heatmap was derived by classifying the percent composition values into equal interval quantiles. The distribution of these classified values are illustrated as color tiles superimposed under the empirical cumulative distribution of the observed percent composition values. Asterisks indicate approximate centromere positions.

Figure 4. The insertion-site preferences of maize LTR retrotransposons.

The full-length LTR retrotransposons were placed into bins according to their relative copy number and the results of blast analysis to separate databases of maize genes, cut-and-paste DNA TEs, Helitrons, and LTR retrotransposons were summarized according to their copy number classes.

Figure 5. Abundance and family richness of LTR retrotransposons found on chromosome 1.

(A) The relationship between the % LTR retrotransposon abundance and family richness per 10 Mb bins, and (B) the specific pattern of abundance and richness plotted along the chromosome.

Figure 6. The chromosomal distribution of full-length LTR retrotransposon insertion histories.

The insertion date of each full-length LTR retrotransposon was determined and these values were averaged for all full-length LTR retrotransposons occurring in each 1 Mb bin. The heat map was derived by classifying the average insertion age into equal-interval quantiles. The distribution of these classified ages are illustrated as color tiles superimposed under the empirical cumulative distribution of the average insertion dates for each bin. Asterisks indicate approximate centromere positions.

Figure 7. The average date of LTR retrotransposon insertion for each of the copy-number classes.