Polycistronic transcription of fused cassettes and identification of translation initiation signals in an unusual gene cassette array from Pseudomonas aeruginosa

The gene cassettes found in class 1 integrons are generally promoterless units composed by an open reading frame (ORF), a short 5’ untranslated region (UTR) and a 3’ recombination site ( attC). Fused gene cassettes are generated by partial or total loss of the attC from the first cassette in an array, creating, in some cases, a fusion with the ORF from the next cassette. These structures are rare and little is known about their mechanisms of mobilization and expression. The aim of this study was to evaluate the dynamic of mobilization and transcription of the gcu14-bla GES-1 /aacA4 gene cassette array, which harbours a fused gene cassette represented by bla GES-1 /aacA4. The cassette array was analyzed by Northern blot and real-time reverse transcription-polymerase chain reaction (RT-PCR) in order to assess the transcription mechanism of bla GES-1 /aacA4 fused cassette. Also, inverse polymerase chain reactions (PCR) were performed to detect the free circular forms of gcu14, bla GES-1 and aacA4. The Northern blot and real time RT-PCR revealed a polycistronic transcription, in which the fused cassette bla GES-1 /aacA4 is transcribed as a unique gene, while gcu14 (with a canonical attC recombination site) has a monocistronic transcription. The gcu14 cassette, closer to the weak configuration of cassette promoter (PcW), had a higher transcription level than bla GES-1/ aacA4, indicating that the cassette position affects the transcript amounts. The presence of ORF-11 at attI1, immediately preceding gcu14, and of a Shine-Dalgarno sequence upstream bla GES-1/ aacA4 composes a scenario for the occurrence of array translation. Inverse PCR generated amplicons corresponding to gcu14, gcu14-aacA4 and gcu14-bla GES-1/ aacA4 free circular forms, but not to bla GES-1 and aacA4 alone, indicating that the GES-1 truncated attC is not substrate of integrase activity and that these genes are mobilized together as a unique cassette. This study was original in showing the transcription of fused cassettes and in correlating cassette position with transcription.

are rare and little is known about their mechanisms of mobilization and expression. The aim of this study was to evaluate the dynamic of mobilization and transcription of the gene cassette array, which gcu14-bla /aacA4 harbours a fused gene cassette represented by . The cassette bla /aacA4 array was analyzed by Northern blot and real-time reverse transcription-polymerase chain reaction (RT-PCR) in order to assess the transcription mechanism of fused cassette. Also, inverse bla /aacA4 polymerase chain reactions (PCR) were performed to detect the free circular forms of . The Northern blot and real time RT-PCR gcu14, bla and aacA4 revealed a polycistronic transcription, in which the fused cassette bla is transcribed as a unique gene, while (with a canonical /aacA4 gcu14 attC recombination site) has a monocistronic transcription. The cassette, gcu14 closer to the weak configuration of cassette promoter (PcW), had a higher transcription level than / , indicating that the cassette position bla aacA4 affects the transcript amounts. The presence of ORF-11 at , immediately attI1 preceding , and of a Shine-Dalgarno sequence upstream / gcu14 bla aacA4 composes a scenario for the occurrence of array translation. Inverse PCR generated amplicons corresponding to gcu14, gcu14-aacA4 and gcu14-bla / free circular forms, but not to and alone, aacA4 bla aacA4 indicating that the GES-1 truncated is not substrate of integrase activity attC and that these genes are mobilized together as a unique cassette. This study was original in showing the transcription of fused cassettes and in correlating cassette position with transcription. v3 Introduction Class 1 integrons are capable of inserting, excising and rearranging gene cassettes by a site-specific recombination mechanism. These assembly platforms can also act as expression systems due to the presence of a promoter region (Pc), which drives the expression of genes captured by integron 1 . Moreover, naturally occurring integrons may have a second promoter (P2), which is activated by the insertion of three G residues between -35 and -10 hexamers 1 . Gene cassettes are generally promoterless units associated with a recombination site (attC or 59-be), which confers the ability of each structure to be mobilized independently 2 , and the Left Hand (LH -1L and 2L core sites) and Right Hand (RH -1R and 2R core sites) domains from attC sites are crucial for this mobilization 3 . Studies focusing on gene cassette translation in the context of integrons are rare, however, it was recently showed that attC sites regulate the translation of downstream cassettes due to their peculiar sequences composed by imperfect inverted repeats that form stem-loop structures. These secondary structures prevent ribosome progression throughout mRNA, reflecting in a decreased expression of more distal genes regarding Pc 4 . Conversely, TIR (Translation Initiation Region)-deficient gene cassettes could have their expression promoted by the presence of ORF-11 5 . This ORF, when present, is found at the attI site preceding the gene cassette array. It codes for 11 amino acids and harbours its own Shine-Dalgarno (SD) sequence. Therefore, the ORF-11 recruits the ribosomes and, through an event of coupled translation, the subsequent TIR-deficient gene cassette could be expressed 4,5 . Gene cassettes can be found inserted in integrons or in other secondary sites, or free in the cytoplasm as a closed circle, in which the 5' end (5' UTR) and the attC recombination site are covalently linked 6 .

Referee Status:
As demonstrated previously, several stress conditions could evoke the activation of the SOS response resulting in integron-integrase expression 7 . Therefore, under stress, the integrase activity increases, favoring the occurrence of integration/excision/rearrangements events.
Although rare, fused cassettes may be generated by partial or total loss of the first attC, retaining both complete coding regions and, therefore, creating permanent gene arrays comparable to bacterial operons 8 . The functionality of such structures has been indirectly inferred by the resistance profile of transformants carrying the fusion 9 ; however, the transcription itself has never been verified.
This study showed the dynamics of fused cassette mobilization, the co-transcription of the gcu14-bla GES-1 /aacA4 cassette array and the effect of cassette position on transcription levels in Pseudomonas aeruginosa wild lineages carrying class 1 integrons. Moreover, the presence of translation signals in this gene cassette array was determined.

Material and methods
An unknown Open Reading Frame (ORF), gcu14 (gene cassette of unknown function), followed by the fused cassette bla GES-1 /aacA4, created by partial loss of GES-1 attC were present in integrons from clinical P. aeruginosa isolates (PS1 and PS26) 10 . Total RNA was extracted and purified according to the manufacturer's instructions with the SV 96 Total RNA Isolation System (Promega). Northern blot using 7 μg of total RNA from PS1 and PS26 was performed in order to detect the transcript originated from gcu14-bla GES-1 /aacA4 cassette array. After electrophoresis in a denaturing-formaldehyde 1.5% agarose gel, the total RNA was transferred to the Hybond-N + nylon membrane (GE Healthcare) by upward capillary transfer. An amplicon of 519bp corresponding to part of the bla GES-1 gene was used as a probe (Table 1) in hybridization assay. The GES probe was labeled with the AlkPhos Direct Labelling kit (GE Healthcare) and hybridized with the target RNA immobilized on the Hybond-N + membrane as recommended. The chemiluminescence was detected with the CDP-Star detection reagent (GE Healthcare) according to manufactures. Immediately after applying the detection reagents, the blot was drained, incubated five minutes at room temperature and exposed to the Hyperfilm ECL (GE Healthcare) for 60 minutes at room temperature.
In order to verify whether the relative position of gene cassettes on the variable region plays a role in transcription level, real-time RT-PCR reactions using the TaqMan System (Applied Biosystems) were performed with primers and probes detailed in Table 1. The rpsL gene of the P. aeruginosa chromosome was amplified by PCR (Table 1) and used as a reference gene for normalization. The relative quantification (RQ) results were presented as ratios of gene transcription between the target gene (cassettes) and the reference gene (rpsL), which were obtained according to the following equation: RQ=2 -ΔCT , where CT is the value corresponding to the crossing point of the amplification curve with the threshold and ΔCT=CT target gene minus CT reference gene. The effect of cassette position on gene transcription was considered significant when the ratios obtained between RQ values (RQ value of cassette 1/RQ value of cassette 2) were ≥2.0, taking into account the standard deviation intervals.
In order to induce cassette excision from integrons, PS1 and PS26 strains 10 were submitted to thermal stress during the log growth phase to induce integrase activity. Cells were grown on Luria-Bertani (LB) broth medium (OXOID) at 37°C for two hours. Subsequently, the bacterial cultures were submitted to a heat shock at 4°C for 30 minutes and immediately incubated at 42°C for another 30 minutes. Briefly, the total DNA from PS1 and PS26 cultured under thermal stress were obtained with the Wizard Genomic DNA purification kit (Promega) following manufacturer recommendations and used as templates in inverse PCR reactions. The inverse PCR was performed with primers facing outwards towards the ends of gcu14, bla GES-1 and aacA4 so that only circular gene cassette configurations would be amplified. The reactions targeting the circular forms of gcu14, bla GES-1 , aacA4 and bla GES-1 /aacA4 fusion was performed with primers and combinations described in Table 1. The inverse PCR was performed using Platinum Taq DNA Polymerase reagents (Invitrogen), and the following components were added to a sterile 0.2-mL tube: 5 μL of 10X PCR buffer (1X final concentration); 1 μL of 10mM dNTP mixture (0.2 mM each); 1.5 μL of 50mM

Amendments from Version 2
Taking into account the Dr. Partridge recommendations, a new version of Figure 2 was provided in order to make clearer the differences between the canonical and truncated GES-1 attC sites. In this new version of the figure, all points raised by Dr. Partridge were addressed.

Results and discussion
The integrons analysed in this study harbored the weak Pc configuration (PcW) 11 , and they carry the gcu14 as the first gene cassette (see reference 12 for nomenclature), which has not been reported so far. Considering that transcription initiates from the Pc promoter placed upstream the cassette array, both monocistronic and full length polycistronic transcripts could be identified. In fact, Northern blot and hybridization assays revealed a unique signal of approximately 2,300 bases, which corresponds to the co-transcription of the entire array (gcu14-bla GES-1 /aacA4) ( Figure 1). This result is in agreement with previous work in which the occurrence of transcripts containing more than one gene cassette was observed by Northern blot analysis 1 . Moreover, this finding gives support to the lack of attC function in terminating transcription of downstream gene cassettes as demonstrated previously 4 .
This fusion retained both entire coding regions and, due to a possible erroneous recombination event, the bla GES-1 attC site was replaced by part of the attI1 site (DQ236170) 10 , reducing it to the 6bp from the 1L core site ( Figure 2). Taking into account that the region responsible for stem-loop formation was missing in GES-1 attC and the participation of this site in terminating translation 4 , our findings indirectly suggested that bla GES-1 and aacA4 translation is occurring in a unique step.   The strains were submitted to thermal stress conditions in order to verify the dynamics of mobilization of gcu14-bla GES-1 /aacA4 gene cassettes. Since the excision event depends on the recognition of the LH and RH domains of the attC site, and that the 2L core site and the entire RH domain are missing in GES-1 attC (Figure 2), it is expected that the bla GES-1 /aacA4 excision occurs only at the aacA4 attC site, and that this structure is excised together as a unique cassette.
Positive results were obtained for the gcu14, gcu14-aacA4 and gcu14-bla GES-1 /aacA4 circular forms, but not for bla GES-1 and aacA4 alone. This finding indicates that the GES-1 attC is not functional and that the fused gene cassette is excised as a unique cassette. Moreover, the presence of gcu14-aacA4 circular form suggests that this strain carries a second integron containing this gene cassette arrangement. Sequencing assessed the recombination point where excision occurred, confirming the occurrence of free circular forms (Figure 3). This is in agreement with the presence of the PcW configuration, in which the corresponding intI1 gene codes for a high efficient integrase 11 . The lack of activity of a truncated attC had also been observed before when associated with aadA10 13 . However, Ramirez and colleagues 14 showed that the integrase was able to recognize and mediate excision of a truncated site associated to aadA1, indicating that the genetic context of such truncated sites could influence their role in IntI1 recognition and mobilization.
The relative quantification performed by real time RT-PCR revealed that PS1 and PS26 presented very similar RQ values for gcu14-bla GES-1 /aacA4 transcription (Figure 4). This result was expected since integrons from these two strains have the same backbone, including the Pc promoter, and are at the same genetic environment 10 . gcu14, the first cassette in integrons with the weak PcW configuration, presented approximately two-fold higher transcription when compared to bla GES-1 and aacA4 separately or when the fused cassette bla GES-1 /aacA4 was considered (Figure 4). The same RQ value obtained for bla GES-1 , aacA4 and the fusion reveals that these two ORFs are transcribed as a unique gene. The lower transcript amount of bla GES-1 /aacA4 compared to gcu14 lies on the distance between these gene cassettes and Pc, which is one of the determinants influencing cassette transcription 1,7 , and it shows the effect of cassette position on expression levels.
A putative promoter for gcu14 (-35 TTGATG [17 bp] -10 TGTTAC) was found 45 bp upstream from its start codon, which has the potential to influence transcription. Moreover, the ORF-11, which enhances the translation efficiency of downstream TIR-deficient cassettes inserted in integrons 5 , was found at the attI1 region preceding the TIR-deficient gcu14 gene cassette. This ORF contained its own Shine-Dalgarno (SD) sequence placed 8 bp upstream of the ATG codon. The ribosome at the ORF-11 stop codon could, therefore, be carried along the mRNA by lateral diffusion, reinitiating translation at the gcu14 start codon. A potential SD sequence was identified 10 bp upstream of the fused cassette bla GES-1 /aacA4. In addition, the loss of the GES-1 attC region, which is involved in stem-loop formation, may enhance the chances of aacA4 translation, since this attC, reduced to the 6bp of the 1L core site, no longer constitutes a physical barrier to ribosome progression 4 . Together, these findings create a scenario for the occurrence of gcu14-bla GES-1 / aacA4 expression in PS1 and PS26, which then provides a possible explanation for their resistance profile to β-lactams and aminoglycosides that has been observed elsewhere 10 .

Conclusions
Fused cassettes have been found in class 1 integrons 9,13-18 ; however, the transcription of such structures has rarely been addressed. This work showed the transcription pattern of a fused cassette as a polycistronic mRNA and that these unusual structures are excised as a unique cassette. The mobilization in block of the entire gcu14bla GES-1 /aacA4 array together with its active transcription, and the presence of translational signatures demonstrate the potential for dissemination and expression of multidrug resistance, in a one-step fashion, to other bacteria. Therefore, such events could represent a threat to public health and to the establishment of efficient antibiotic regiments.

Nucleotide sequence accession number
The sequence of the cassette array composed by the fusion has been deposited in the GenBank database under accession number DQ236170. The sequence obtained from the inverse PCR amplicon, showing the circular form gene organization, was submitted to GenBank under accession number KT336477. Author contributions ELF and ACPV conceived the study and designed the experiments. ELF carried out the research and prepared the first draft of the manuscript. All authors were involved in the revision of the draft manuscript and have agreed to the final content.

Competing interests
No relevant competing interests were disclosed.

Grant information
This work was supported by the CNPq and FAPERJ fellowship and Oswaldo Cruz Institute Grant. 1.

2.
3. The grey shading to indicate the attC site should not start until the start of 1L and the portion derived from attI1 is still not indicated.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. ttagat of the IR site shown for the complete blaGES-1 attC is really part of the adjacent 3'-CS in AF355189 (most of the 1R core site belonging to the blaGES cassette (TTAGAC) is found at the start of the linear cassette). I would suggest ending this sequence after the first g of this site.
The partial 1L of the truncated blaGES cassette (924-9), the nucleotides derived from the attI1 site (930-6) and 1R of the aacA4 cassette (937-42) should be indicated. I would suggest ending the sequence of the truncated cassette after 1R of aacA4, as the alignment of the complete attC with the start of the aacA4 cassette beyond this point is not really useful.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed.

Competing Interests:
Author Response: Reviewer is right. This was corrected in the figure. Actually, the figure was modified in order to make clearer the differences between the canonical GES-1 attC and the truncated GES-1 attC found in this study.
Reviewer: ttagat of the IR site shown for the complete blaGES-1 attC is really part of the adjacent 3'-CS in AF355189 (most of the 1R core site belonging to the blaGES cassette (TTAGAC) is found at the start of the linear cassette). I would suggest ending this sequence after the first g of this site. Response: We agree and this was corrected in the new figure.
Reviewer: The partial 1L of the truncated blaGES cassette (924-9), the nucleotides derived from the attI1 site (930-6) and 1R of the aacA4 cassette (937-42) should be indicated. I would suggest ending the sequence of the truncated cassette after 1R of aacA4, as the alignment of the complete attC with the start of the aacA4 cassette beyond this point is not really useful.
Response: We agree and this was modified in the new figure.
There is no competing interests Competing Interests: 28  This work is conducted to address the transcription of the fused gene cassette / This bla aacA4. arrangement is rare but not exceptional in a class 1 integron. In this work, the experimentation confirmed that the three gene cassettes harbored in this class 1 integron were expressed together from a unique polycistronic transcript.
It seems that the authors have observed a circular form containing the gene cassette associated gcu14 with the gene cassette. Could they explain this arrangement? Is it possible that the aacA4 Pseudomonas isolates harbored another class 1 integron containing the gene followed by the gene? gcu14 aacA4 The authors have to provide in the paper the classical length of the site, which is 110bp and bla attC contained the 1L, 2L, 2R and 1R sequences-shown in bold.
In the SP26 and PS1 isolates, the site is reduced to the 6 bp of 1 L box. Pseudomonas Bodilis et al., 2012 -In the conclusion, page 6, change "never" to "rarely".

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed.

Competing Interests:
Author Response 05 Aug 2015 It seems that the authors have observed a circular form containing the gcu14 gene cassette associated with the aacA4 gene cassette. Could they explain this arrangement? Is it possible that the Pseudomonas isolates harbored another class 1 integron containing the gcu14 gene followed by the aacA4 gene?
Yes, the presence of the gcu14-aacA4 circular form indicates that this strain harbours Response: a second integron composed by the gcu14-aacA4 arrangement. However, this result does not invalidate our main conclusion, which is that the truncated GES-1 attC site is not functional and this gene is only mobilized when recombination occurs in aacA4 attC. A brief explanation was included in the text (Results section).
The authors have to provide in the paper the classical length of the site, which is bla attC 110bp and contained the 1L, 2L, 2R and 1R sequences-shown in bold.
A figure showing the truncated and the complete canonical from GES1 was Response: attC included in the new version.
So, the distance from of the stop of blaGES-1 to the start codon of the aacA4 gene is 46bp. The translational start codon of aacA4 gene is erroneous in the Genbank data base (DC236170). Indeed, it has been determined by N-terminal amino acid sequencing to be a GTG (in italic in the sequence above) and the beginning of the cassette was 24 bp before [Hanau-Bercot B. et al., 2002].
-Could the Genbank annotation be modified as explained above?
We performed this modification and the updated sequence regarding the beginning of Response: aacA4 gene is already accessible on GenBank database under the same accession number (DQ236170).
-The Appendex is not necessary and should be deleted.
The appendix is only for helping referees in their evaluation. It will not be published. this was modified in the new version of the manuscript.

Response:
In the conclusion, page 6, change "never" to "rarely". this was modified in the new version of the manuscript.

Response:
There is no competing interests Competing Interests: 02  This paper looks at expression of genes in a gene cassette array that includes a fused cassette and at the excision of cassette from this array. Polycistronic transcripts from cassette arrays and the excision of other fused cassettes have previously been noted by others, as cited in this paper. The sequences in Appendix 1 need to be annotated properly, rather than just the results of searches given, and the point illustrated by each sequence needs to be explained to make it possible to assess whether they support GES-1 illustrated by each sequence needs to be explained to make it possible to assess whether they support the conclusions drawn in the paper.
Some reorganisation (moving some information in the Results and Discussion to the Introduction) would help to make the paper easier to follow and the Conclusions section is very short and could be expanded. The English could also be improved and the manuscript checked for typos etc. A few specific points also need correcting or clarifying: Abstract The sequences in Appendix 1 need to be annotated properly, rather than just the results of searches given, and the point illustrated by each sequence needs to be explained to make it possible to assess whether they support the conclusions drawn in the paper.
The sequence resulted from inverse PCR product, showing the circular form of the Response: gcu14-blaGES-1-aacA4 array, was submitted to GenBank under accession number KT336477. This information was included in the text. Moreover, we provided a new figure showing a schematic representation of primer targeting sites used in inverse PCR in order to make easier for the reviewers comprehend our strategy and results.
Some reorganisation (moving some information in the Results and Discussion to the Introduction) would help to make the paper easier to follow and the Conclusions section is very short and could be expanded. The English could also be improved and the manuscript checked for typos etc.
We agree that some points from results and discussion section would be more Response: adequate in the introduction section, and they were moved in the new version of the manuscript (Page 3, lines 59-60 and 65-70; page 4, lines 71-76 in the new version). The English was reviewed by a native spoken-English, and the conclusion was expanded.
A few specific points also need correcting or clarifying: Abstract Line 5-fused cassettes do not always have fused orfs (e.g. aacA1/orfG) We agree and it was modified: We affirmed that only in some cases fused orfs can be Response: created.
Abstract 6th Line from end and Results and Discussion p. 5 1st Line of 2nd paragraph -what is in the gcu14-aacA4 circular form? Should this be blaGES-1/aacA4?
The presence of the gcu14-aacA4 circular form indicates that this strain harbours a Response: second integron composed by the gcu14-aacA4 arrangement. However, this result does not invalidate our main conclusion, which is that the truncated GES-1 attC site is not functional and this gene is only mobilized when recombination occurs in aacA4 attC. A brief explanation was included in the text (Results section).
Materials and Methods Line 1 -this needs rewording to explain that gcu is a gene cassette of unknown function.
It was included in the text.

Response:
Results and Discussion p. 4 -the fusion here is the type where part of the blaGES-1 attC site is replaced by part of the attI1 site (see ref. 14). The LH and RH domains are part of the attC site, rather than flanking it. The 5′ UTR contains 6 bp of the attC site.
We change the text emphasizing that the attC was replaced by part of attI1.

Response:
The referee is right; LH and RH domains are part of the attC site, rather than flanking it. This general idea concerning attC site and its domains was modified throughout the text.
Results and Discussion p. 5 -most of the 1L core site and the 1R core site of the blaGES attC site are present.
In fact, only the 1L core site is present in the truncated attC. We included a figure Response: (figure 2) in this new version showing this.
Results and Discussion p. 6 -the deletion of part of the attC site doesn't bring the aacA4 gene much closer to Pc and ref. 14 makes slightly different point (that expression of the downstream cassette may be enhanced).
The referee is right. This statement was removed.

Response:
Formatting "14" of gcu14 and "A4" of aacA4 should be in italics. Transposon, gene and species names etc are not correctly formatted in the references.
All formatting errors were properly corrected.

Response:
There is no competing interests Competing Interests:

, Instituto Oswaldo Cruz, Brazil Erica Fonseca
Dear Authors, Besides the comments already made by the reviewers, I would like to specify some points: The C to G mutations that converts the usual weak PcW variant into PcWTGN-10 has been shown to increase (instead of 'decrease', first part of the Results section).
This sentence was removed.

Response:
I have not checked which IntI1 had been used for the excision assay, usually these are made with the most efficient "PcW" IntI1 (IntI1R32_H39); note that in our PLoS Genetics paper (Jove et al. ) we evidenced the IntI1 "PcWTGN-10" (IntIP32_H39) were much less efficient for excision of 2010 gene cassettes, which could had been taken into account in your discussion.
As properly noticed by you, our Pc configuration is the weak one, and this was changed in the Response: text. However, we agree that including this discussion we will improve our manuscript (lines 182 and 184).
The hypothesis of a synergistic effect between the Pc and P(gcu14) tandem promoters does not fit the observed phenotypes: since the resulting level of transcription observed for the downstream genes is lower I suggest that some transcriptional interferences occur rather than synergistic interactions. By the way in absence of any experimental evidence for the functionality of the gcu14 promoter, such hypothesis is overstated. It would have been interesting to check the level of transcription of the blaGES-1,aacA4 gene cassette when gcu14 is deleted.
we totally agree with this observation, in fact our conclusion is overstated since no Response: experimental assay was performed to test the promoter synergy. Therefore, we remove this part from the text. We also agree that would be interesting to verify the expression of the fused cassette in the absence of gcu14, however it was not the focus of this paper.
-First, the class 1 integron sequence of the PS1 strain in Genbank DQ236170 does not display the rare C to G mutation that converts a weak "PcW" promoter into a stronger "PcWTGN-10" as stated in the paper. Consequently you should write "the weak PcW configuration" instead of "weak PcWTGN-10". Also it means that its own IntI1 integrase is likely to be optimally efficient.
Dr. Jové is absolutely right, the DQ236170 accession number presented the PcW Response: configuration and not the PcWTGN-10 as stated in the manuscript, we apologize about this mistake. This was corrected in the text.
-Then, the sequence of the class 1 integron as deposited in the Genbank does not display the complete array of gene cassette (the end of the attCaacA4 is not available) which means that it is ambigous whether there is a downstream third gene cassette or not. Consistently this array of gene cassette, although a novel one, has not been numbered in the INTEGRALL database ( ) http://integrall.bio.ua.pt/?acc=DQ236170 We know that aacA4 was the last cassette in the array because this variable region was Response: obtained using primers annealing in the attI1 and in the beginning of qacE∆1 from 3'CS. It is true that the aacA4 attC is not completed in the GenBank accession number DQ236170. Therefore, spite of the incomplete attC sequence we are sure that the aacA4 is the last cassette from this array.
incomplete attC sequence we are sure that the aacA4 is the last cassette from this array.
-Lastly, as an element of reply to the reviewer Dr Bercot, I would like to notify that the so-called gcu14 GC has not been reported in other reports except in an environmental strain of Citrobacter (and the sequence is not 100% identical, Genbank FM998050).
thanks for this observation. It was included in the text.

Response:
There is no competing of interests Competing Interests: -First, the class 1 integron sequence of the PS1 strain in Genbank DQ236170 does not display the rare C to G mutation that converts a weak "PcW" promoter into a stronger "PcWTGN-10" as stated in the paper. Consequently you should write "the weak PcW configuration" instead of "weak PcWTGN-10". Also it means that its own IntI1 integrase is likely to be optimally efficient.
-Then, the sequence of the class 1 integron as deposited in the Genbank does not display the complete array of gene cassette (the end of the attCaacA4 is not available) which means that it is ambigous whether there is a downstream third gene cassette or not. Consistently this array of gene cassette, although a novel one, has not been numbered in the INTEGRALL database ( ) http://integrall.bio.ua.pt/?acc=DQ236170 -Lastly, as an element of reply to the reviewer Dr Bercot, I would like to notify that the so-called gcu14 GC has not been reported in other reports except in an environmental strain of Citrobacter (and the sequence is not 100% identical, Genbank FM998050).

Competing Interests:
Reader Comment 17 May 2013 , LGPB, Belgium Thomas Jové Dear Authors, Besides the comments already made by the reviewers, I would like to specify some points: The C to G mutations that converts the usual weak PcW variant into PcWTGN-10 has been shown to increase (instead of 'decrease', first part of the Results section). I have not checked which IntI1 had been used for the excision assay, usually these are made with the most efficient "PcW" IntI1 (IntI1R32_H39); note that in our PLoS Genetics paper (Jove et al. ) we evidenced the IntI1 "PcWTGN-10" (IntIP32_H39) were much less efficient for excision of 2010 gene cassettes, which could had been taken into account in your discussion.
The hypothesis of a synergistic effect between the Pc and P(gcu14) tandem promoters does not fit