Genome-wide detection of predicted non-coding RNAs in Rhizobium etli expressed during free-living and host-associated growth using a high-resolution tiling array

Abstract

Background: Non-coding RNAs (ncRNAs) play a crucial role in the intricate regulation of bacterial gene expression, allowing bacteria to quickly adapt to changing environments. In the past few years, a growing number of regulatory RNA elements have been predicted by computational methods, mostly in well-studied gamma-proteobacteria but lately in several alpha-proteobacteria as well. Here, we have compared an extensive compilation of these non-coding RNA predictions to intergenic expression data of a whole-genome high-resolution tiling array in the soil-dwelling alpha-proteobacterium Rhizobium etli.

Results: Expression of 89 candidate ncRNAs was detected, both on the chromosome and on the six megaplasmids encompassing the R. etli genome. Of these, 11 correspond to functionally well characterized ncRNAs, 12 were previously identified in other alpha-proteobacteria but are as yet uncharacterized and 66 were computationally predicted earlier but had not been experimentally identified and were therefore classified as novel ncRNAs. The latter comprise 17 putative sRNAs and 49 putative cis-regulatory ncRNAs. A selection of these candidate ncRNAs was validated by RT-qPCR, Northern blotting and 5' RACE, confirming the existence of 4 ncRNAs. Interestingly, individual transcript levels of numerous ncRNAs varied during free-living growth and during interaction with the eukaryotic host plant, pointing to possible ncRNA-dependent regulation of these specialized processes.

Conclusions: Our data support the practical value of previous ncRNA prediction algorithms and significantly expand the list of candidate ncRNAs encoded in the intergenic regions of R. etli and, by extension, of alpha-proteobacteria. Moreover, we show high-resolution tiling arrays to be suitable tools for studying intergenic ncRNA transcription profiles across the genome. The differential expression levels of some of these ncRNAs may indicate a role in adaptation to changing environmental conditions.

Figure 1

Sampling conditions and rRNA fragmentation. (A) Three samples were taken during free-living growth based on the OD of the culture; early exponential phase (OD₆₀₀ = 0.3), late exponential phase (OD₆₀₀ = 0.7) and stationary phase (6 hours after reaching the maximum OD₆₀₀). (B) An example of high quality total RNA, illustrating the fragmentation 23S rRNA: (a) small RNA peak including 5S rRNA and the ncRNAs (b) 23S fragment of ~135 bp (c) two 23S fragments of ~1300 bp (d) 16S rRNA (e) intact 23S rRNA.

Figure 2

Expression profiles of selected ncRNAs. The individual probe intensities of six selected ncRNAs are shown during free-living growth (early exponential, late exponential and stationary phase) and symbiosis (two and three weeks after plant inoculation). The summary expression values of the probe intensities (left) and RT-qPCR confirmations of the ncRNA expression levels (right) are given below each profile. The mean values and standard deviation of three RT-qPCRs are shown. Genes are represented by bars and arrows indicate the direction of transcription. (A) TPPb RS is a thiamin pyrophosphate-sensing riboswitch located on plasmid p42B; (B) ReC11 is highly similar to a previously experimentally verified sRNA in S. meliloti 1021 [41,45]; (C, D, E, F) ReC43, ReC49, ReC27, ReC62 are novel candidate ncRNAs.

Figure 3

Heat map of the candidate-ncRNAs. The heat map visualizes the individual ncRNA transcription profiles of the detected ncRNAs that were differentially expressed under our experimental conditions. The expression values in each row were standardized by subtraction of the mean and division by the standard deviation and hierarchically clustered using R. ncRNAs showing similar expression patterns are grouped as follows: group 1, stationary phase; group 2, symbiosis; group 3, exponential phase. The letters b, d and e indicate gene location on the respective plasmids.

Figure 4

Novel ncRNAs detected by Northern analysis. Four novel ncRNAs were analyzed to validate the tiling array results. All four ncRNAs were detected during exponential and stationary phase grown in defined medium; expression during one growth phase is shown for each ncRNA. Primary transcripts are indicated by an asterisk. Transcription initiation sites identified by 5' RACE are shown in bold and the difference with the estimated transcription initiation sites using the array data are shown between brackets. Genes are represented by bars and arrows indicate the direction of transcription. (A) ReC12 and (B) ReC14 detected during early exponential phase; (C) ReC56 and (D) ReC64 detected during stationary phase.

Figure 5

Genomic map of ncRNAs. White, grey and black arrows show the location of functionally characterized, previously reported but uncharacterized and novel ncRNAs, respectively. The majority of the identified ncRNAs are chromosomally encoded. An apparent 'hot spot' with increased ncRNA gene density is located around 1.7 to 2.0 Mb on the chromosome.

Figure 6

Conservation of R. etli ncRNAs in α-proteobacteria. Similarity analysis was performed using BLASTN (NCBI). Black squares indicate E-values ≤ 10^-5, grey squares represent E-values ≤ 10^-3. The percentage of conservation for each species is shown at the bottom of the figure. Only ncRNAs conserved beyond A. tumefaciens str. C58, R. etli CIAT652, R. leguminosarum bv. viciae 3841, S. meliloti 1021 and Mesorhizobium loti MAFF303099 are shown. Conservation analysis results of all identified ncRNAs are included as supplementary data, see Additional file 5, Table S3.