Báo cáo khoa học: Analysis of the RNA degradosome complex in Vibrio angustum S14 potx

Previously, we have identified RhlB ATP-dependent DEAD-box RNA helicase-binding, PNPase polynucleotide phosphory-lase-binding and enolase-binding microdomains in the C-terminal half of Vi

Vibrio angustum S14

Melissa A Erce, Jason K K Low and Marc R Wilkins

Systems Biology Laboratory, School of Biotechnology and Biomolecular Sciences, University of New South Wales, Sydney, Australia

Introduction

Post-transcriptional control of gene expression is an

important regulatory mechanism, as the length of time

that a transcript is available for translation limits the

expression of its protein product Messenger RNA

(mRNA) half-lives can differ by as much as two orders

of magnitude within a single cell In Escherichia coli,

where the average message has a half-life of about

5 min, individual mRNA half-lives can be as short as several seconds or as long as 1 h [1] Through the reg-ulation of mRNA stability, patterns of protein synthe-sis in the cell can be modulated in response to changes

in growth conditions [2] In addition, selective mRNA decay results in the differential expression of gene products in some polycistronic mRNAs [3] The

multi-Keywords

degradosome; microdomains; RhlB;

RNase E; two-dimensional Blue

Native-SDS⁄ PAGE

Correspondence

M Wilkins, School of Biotechnology and

Biomolecular Sciences, University of New

South Wales, Sydney, NSW 2052, Australia

Fax: +61 2 9385 1483

Tel: +61 2 9385 3633

E-mail: m.wilkins@unsw.edu.au

(Received 28 June 2010, revised 28

September 2010, accepted 22 October

2010)

doi:10.1111/j.1742-4658.2010.07934.x

The RNA degradosome is built on the C-terminal half of ribonuclease E (RNase E) which shows high sequence variation, even amongst closely related species This is intriguing given its central role in RNA processing and mRNA decay Previously, we have identified RhlB (ATP-dependent DEAD-box RNA helicase)-binding, PNPase (polynucleotide phosphory-lase)-binding and enolase-binding microdomains in the C-terminal half of Vibrio angustumS14 RNase E, and have shown through two-hybrid analy-sis that the PNPase and enolase-binding microdomains have protein-bind-ing function We suggest that the RhlB-bindprotein-bind-ing, enolase-bindprotein-bind-ing and PNPase-binding microdomains may be interchangeable between

Escherichi-a coli and V angustum S14 RNase E In this study, we used two-hybrid techniques to show that the putative RhlB-binding microdomain can bind RhlB We then used Blue Native-PAGE, a technique commonly employed

in the separation of membrane protein complexes, in a study of the first of its kind to purify and analyse the RNA degradosome We showed that the

V angustumS14 RNA degradosome comprises at least RNase E, RhlB, enolase and PNPase Based on the results obtained from sequence analyses, two-hybrid assays, immunoprecipitation experiments and Blue Native-PAGE separation, we present a model for the V angustum S14 RNA degradosome We discuss the benefits of using Blue Native-PAGE as a tool

to analyse the RNA degradosome, and the implications of microdomain-mediated RNase E interaction specificity

Structured digital abstract

l A list of the large number of protein–protein interactions described in this article is available via the MINT article ID MINT-8049250

Abbreviations

BACTH, bacterial adenylate cyclase two-hybrid; CTH, C-terminal half; DMP, dimethyl pimelimidate.2HCl; NTH, N-terminal half;

PNPase, polynucleotide phosphorylase; PVDF, poly(vinylidene difluoride); RhlB, ATP-dependent DEAD-box RNA helicase; RNase E,

ribonuclease E.

enzyme RNA degradosome is involved in the

steady-state regulation of transcripts in E coli Apart from its

role in mRNA degradation, the degradosome is also

responsible for the processing of 5S ribosomal RNA

and tRNAs, as well as the degradation of tmRNAs [4–8]

The principal proteins comprising the degradosome

include the 3¢ to 5¢ exoribonuclease polynucleotide

phosphorylase (PNPase), the ATP-dependent

DEAD-box RNA helicase (RhlB) and the glycolytic enzyme

enolase [9–12] The scaffolding for the degradosome is

provided by RNase E One of the largest proteins

found in E coli, RNase E, is 1061 amino acids in

length [10,13,14] It is defined by two functionally

distinct domains of approximately equal size – the

N-terminal half (NTH, residues 1–498) and the

C-terminal half (CTH, residues 499–1061) [15,16]

The catalytic activity of E coli RNase E resides within

its globular NTH The sequences within this essential

region are highly conserved amongst eubacteria [17]

In contrast, the CTH lacks sequence conservation, has

little structural character, has no known catalytic

func-tion [17,18] but provides the scaffolding for the

recruit-ment of the degradosome through short recognition

motifs [19] The RNase E CTH contains two regions

which can bind RNA and interact with substrates such

as the 9S precursor for 5S ribosomal RNA [18]

Apart from the canonical degradosome components,

polyphosphate kinase, poly(A) polymerase, ribosomal

proteins and chaperone proteins GroEL and DnaK

have been found to be present in the degradosome,

but in substoichiometric amounts [11,20,21]

The molecular interactions of RNase E with other

proteins of the RNA degradosome are proposed to be

mediated by short microdomains of 15–40 amino acids

on RNase E [22] These microdomains exhibit higher

amino acid sequence conservation than the rest of the

RNase E CTH It is through these regions that specific

molecular interactions are believed to occur to direct

RNase E function, and hence confer adaptive

reorgani-zation of protein complex formation [17,22–25]

In order to study RNase E and its interaction

part-ners, it would be useful if it could be isolated The

purification and characterization of RNase E have

pro-ven challenging [15], and obtaining an amount

suffi-cient for analysis is even more difficult Methods for

its purification as part of the RNA degradosome, as

well as the reconstitution of an active degradosome

through the purification of its individual components,

have been developed [11,26–29] Large-scale

prepara-tions have involved purification under denaturing

con-ditions; others have involved the overexpression of

individual components and subsequent reconstitution

of the degradosome Recently, a method for the

preparation of a recombinant degradosome has been described [30] Co-immunoprecipitation methods have also been used to study the RNA degradosome This technique is especially useful when studying bacteria in which the introduction of genetic tags is difficult [31] These methods, however, require a large amount of starting material Furthermore, as they often require the overexpression and⁄ or tagging of the RNA degradosome components, the complexes formed may not truly represent what is occurring in the cell In view of these limitations, we decided to use Blue-Native PAGE (BN-PAGE), a ‘charge shift’ technique employed in the separation of mitochondrial mem-brane proteins and complexes [32,33] This technique has been successfully employed recently to study the

E coli complexome [34] It requires far less starting material and, when used in conjunction with immuno-blotting and⁄ or mass spectrometric analysis, may provide a better view of the dynamic nature of the RNA degradosome

In this study, we characterized the RNA degradosome from an environmental species: the marine heterotro-phic, Gram-negative bacterium, Vibrio angustum S14 [35] V angustum S14 is a model organism for the study of starvation, as it exhibits remarkable physio-logical changes and increased mRNA stability during carbon starvation [36,37] We have demonstrated pre-viously, using two-hybrid screening, that RNase E from V angustum S14 contains sites for interaction with enolase and PNPase [38] Here, we use a combi-nation of proteomic techniques, such as BN-PAGE, co-immunoprecipitation and tandem MS to identify the components of the RNA degradosome in this organism We also show, using two-hybrid analysis, that the CTH of RNase E from V angustum S14 inter-acts with RhlB

Results RhlB binds to V angustum S14 RNase E and PNPase

Previously, we predicted that the RNase E CTH from

V angustum S14 possessed interaction sites for PNPase and enolase, and demonstrated these interactions through two-hybrid analysis [38] We also predicted that RhlB should bind to RNase E at residues 719–

753 Here, we undertook two-hybrid analysis to test this; the results of this analysis and our previous analy-sis [38] are shown in Table 1 This two-hybrid system introduces two proteins of interest fused to the T18 and T25 domains of Bordetella pertussis adenylate cyclase into E coli cya) on plasmids When physically

RhlBsite 684–784

RhlBsite 714–758

CTH 526–1094

PNPsite 844–1061

PNPsite 1015–1094

Enosite 833–851

Enosite 885–909

c Two-hybrid

separated, the T18 and T25 domains are inactive, but

the interaction of the hybrid proteins results in the

functional complementation of adenylate cyclase in

E coli cya) and the subsequent expression of the lac

operon [39]

RhlB demonstrated positive interactions with the

CTH of V angustum S14 RNase E (grids 3K and

11C) In order to better define the region of

interac-tion of RNase E with RhlB, we tested the putative

RhlB-binding domain of V angustum S14 RNase E

(residues 719–753, plus five amino acid residues

flank-ing on either side); it was found to interact with RhlB

(grid 2K and its reciprocal cross in 11B)

Interest-ingly, this region was also capable of interacting with

E coli RhlB (grid 2J and its reciprocal cross in 10B)

It should be noted, however, that, when the

T18 domain was fused to the C-terminus of the

V angustumS14 RNase E RhlB-binding site (residues

714–758), the interaction with RhlB was negative,

probably because of a difference in conformation

(grid 2K) or because the T18 domain occluded the

RhlB-binding site Previous reports have suggested

that RhlB and PNPase can interact independently of

RNase E in E coli [16] Our two-hybrid analysis

fur-ther confirmed this (grid 8J and its reciprocal cross in

10H) We also observed an interaction between

V angustumS14 PNPase and RhlB; however, this was

weaker than that seen in E coli (grid 11I, C-terminal

fusion of the T18 domain) We did not expect PNPase

to interact with the RhlB-binding site, and we

observed negative interactions for this (grids 1H, 1I,

2H, 2I, 8A, 8B, 9A and 9B) Negative interactions

were also seen between RhlB and the PNPase-binding

site of RNase E (grids 4J, 4K, 5J, 5K, 10D, 10E, 11D

and 11E), which was expected These, together with

the negative interaction of PNPase with the

RhlB-binding site, served as negative controls for the

exper-iment (grids 1H, 1I, 2H, 2I, 8A, 8B, 9A and 9B)

A series of cross-species’ interactions was also tested

here We observed strong interactions for E coli RhlB

and the CTH of V angustum S14 RNase E (grids 3J

and 10C) Weaker interactions were seen for other

cross-species’ crosses involving RhlB and the

RhlB-binding site (grids 1K, 2J, 10B and 11A) Further, we

observed the self-interaction of E coli RhlB (grid 10J),

suggesting that this protein can self-interact [40]

PNPase and RhlB copurify with V angustum S14

RNase E

Having shown that PNPase and RhlB can interact

with RNase E microdomains as well as the RNase E

CTH and with each other in a two-hybrid assay, we

investigated whether they interacted in vivo First, we determined whether V angustum S14 proteins can be detected by antisera against E coli RNase E, PNPase and RhlB (Fig 1A) Following that, we then carried out several immunoprecipitation experiments to deter-mine whether V angustum S14 RNase E forms a complex with V angustum S14 PNPase and RhlB, and to determine whether other possible interaction partners were present V angustum S14 PNPase and RhlB were found to coprecipitate with RNase E when antiserum against RNase E was used for immuno-precipitation (Fig 1B) Similarly, when antiserum against PNPase was used for immunoprecipitation,

V angustum S14 RhlB and RNase E were found to associate with V angustum S14 PNPase (Fig 1C) As enolase comigrated in the same region in the gel as

188

98

62

49

MW (kDa)

RhlB

250

148

64

50

MW (kDa)

RNase E

PNPase

IgG RhlB

250

148

64

50

Mw (kDa)

C

Fig 1 Immunoprecipitation of RNase E and PNPase in V angu-stum S14 by antisera against E coli RNase E and PNPase (A)

V angustum S14 lysate probed with antisera against RNase E, PNPase and RhlB (B) Immunoprecipitation using RNase E antise-rum The RNase E antiserum was incubated with V angustum S14 lysate and then isolated by protein G-conjugated Sepharose The eluted fraction was separated by SDS ⁄ PAGE and analysed by immunoblotting using a series of antisera PNPase and RhlB were found to copurify with V angustum S14 RNase E The antibody heavy chain is indicated Composite image obtained when the PVDF membrane was probed serially with antisera against

RNa-se E, PNPaRNa-se and RhlB (C) Immunoprecipitation using PNPaRNa-se antiserum Antiserum against PNPase was incubated with V angu-stum S14 lysate and then isolated by protein G-conjugated Sepha-rose Following SDS ⁄ PAGE separation, the immunoprecipitates were analysed by western blot and probed in a serial fashion with antisera against RNase E, PNPase and RhlB RNase E and RhlB were found to copurify with PNPase.

the IgG heavy chain, any signal from probing with

enolase antiserum was masked by IgG at that

posi-tion in the gel

Identification of the RNA degradosome in

V angustum S14 using BN-PAGE

Previously, we predicted that V angustum S14

RNa-se E contains interaction sites for RhlB, PNPaRNa-se and

enolase [38] Using two-hybrid analysis here and in a

previous study, we demonstrated the interactions of

V angustumS14 RNase E microdomains and the

CTH of RNase E with RhlB, enolase and PNPase

From the results of our immunoprecipitation

experi-ments, we found that RhlB and PNPase associate with

RNase E in V angustum S14 Together, these results

strongly suggest, but do not prove, that all of these

proteins form a degradosome complex To ascertain

that these proteins all co-associate into a

degrado-some-like complex in V angustum S14, we separated

the cell lysate using BN-PAGE (Fig 2A) This

revealed a range of protein complexes of various sizes

up to 1.2 MDa To identify the RNA degradosome

amongst all these complexes, the proteins were

transferred onto a poly(vinylidene difluoride) (PVDF)

membrane and probed with antisera against RNase E, PNPase and RhlB (Fig 2B) On probing with RNase E-specific antibodies, RNase E was detected in two high-molecular-mass bands above 1 MDa and in bands corresponding to approximately 480 and

650 kDa The latter is consistent with RNase E’s ho-motetrameric state (Fig 2B) It should be noted that the true molecular mass of RNase E is 116 kDa, but its apparent molecular mass is approximately 180 kDa [41] PNPase was observed in the same two high-molecular-mass bands (above 1 MDa) as detected by antiserum against E coli RNase E PNPase was also observed at approximately 750 kDa (a potential trimer

of trimers), as well as in a band running at approxi-mately 240 kDa, which is consistent with its trimeric form (Fig 2B) The presence of RhlB was detected in the same two high-molecular-mass complexes above

1 MDa, which also contained RNase E and PNPase (Fig 2B), but was not seen below 1 MDa The detec-tion of lower mass bands containing RhlB may have been reduced as the proteins were in their native rather than denatured state, making the epitopes more difficult to recognize by the antiserum It is interesting

to note that RNase E, PNPase and RhlB are part of the high-molecular-mass complexes, but did not co-associate in other lower mass heteromultimers Probing with antiserum against enolase proved to be uninformative as enolase is present in high abundance

in the cell and comigrates with other protein complexes Antibodies were not available for the identification

of other possible interactor proteins; we therefore used

MS to analyse Bands A and B (Fig 2A) This verified the presence of enolase and PNPase, but not RNase E (Table 2) This was not unexpected because of the low abundance of RNase E in the cell The amino acid sequence of RNase E is arginine-rich, giving rise to peptides that are either too short or too long for frag-mentation when subjected to digestion by trypsin This may have occluded the identification of RNase E in the mass spectrometric analysis of Bands A and B However, RNase E was detected by MS of immuno-precipitates Ribosomal protein L4 was detected in Bands A and B, together with GroEL, a large number

of ribosomal proteins and a high abundance of meta-bolic proteins (data not shown) It is unclear whether these ribosomal proteins are truly complexed with the RNA degradosome in V angustum S14 or whether parts of the ribosome or, indeed, other protein com-plexes co-electrophorese with the degradosome; this is likely when one considers the abundance of these pro-teins within the cell and the large complexes which comprise the ribosome The proteins identified, which have been described to associate with the RNA

MW

(kDa)

1236

1048

1236

MW (kDa) RNase E PNPase RhlB

Band A

720

1048

146

66

242

Fig 2 Separation of protein complexes through BN-PAGE (A)

Analytical analysis of protein complexes in V angustum S14.

V angustum S14 lysate (17 lg) was separated by BN-PAGE and

silver stained (B) Western blot of V angustum S14 RNA

degrado-some components V angustum S14 lysate (200 lg) was

sepa-rated through BN-PAGE and electroblotted onto a PVDF

membrane Bands were detected by probing serially with antisera

against RNase E, PNPase and RhlB Molecular mass markers are

shown in kDa.

degradosome in E coli, are shown in Table 2 Only

one peptide was identified for PNPase and ribosomal

protein L4, but these have been found previously to be

associated with RNase E [42] Fragmentation spectra

for both of these showed a strong series of y-ions and

highly significant identification scores (Fig 3)

Two-dimensional BN-SDS⁄ PAGE analysis

To better understand the subunit composition of the

degradosome protein complex, we used

two-dimen-sional BN-SDS⁄ PAGE In this technique, protein

com-plexes can be separated into subunits in an

SDS⁄ PAGE second dimension, following first

dimen-sion BN-PAGE We excised six bands from the first

dimension BN-PAGE separation (Fig 4A), and ran

these bands separately in a second dimension reducing

SDS⁄ PAGE to separate their constituent proteins

according to size (Fig 4B) As a control, cleared lysate

was loaded (Fig 4B, lane CL) It can be seen that the

composition of each of the bands excised from the

BN-PAGE separation varied (Fig 4B, lanes 1–6) and

was different from the control It was evident that each

band contained proteins and complexes apart from the

RNA degradosome This is to be expected as

BN-PAGE was separating a whole-cell lysate After second

dimension SDS⁄ PAGE separation, the proteins were

electroblotted onto a PVDF membrane and serially

probed with antisera against RNase E, PNPase, RhlB

and enolase (Fig 4C) We found that all BN-PAGE

bands excised between 900 kDa and 1.2 MDa

contained RNase E (Fig 4C, lanes 3–5) However,

RNase E was not present in Band 6 In all instances in

which RNase E was present, PNPase, RhlB and eno-lase were also present, but in varying quantities Although the antisera used have different affinities for their target protein, Band 4 showed more RNase E and PNPase than RhlB and enolase By contrast, Bands 1, 2, 3 and 5 showed roughly equivalent amounts of each of these proteins in the four samples Interestingly, Band 6, which corresponded to a band excised in the vicinity of the 380-kDa region of the first dimension BN-PAGE separation, only contained PNPase, enolase and RhlB There was no full-length RNase E However, there was evidence of a band migrating at approximately 80 kDa, which migrated above PNPase This corresponds to a fragment of RNase E which is seen in one-dimensional SDS⁄ PAGE analyses and immunoblots of whole-cell lysate of

V angustum S14 RNase E (Figs 1B and 4D), and which has been observed previously to migrate in this region [18] This suggests that RhlB, enolase and PNPase may be found together in association with a fragment of RNase E containing its CTH This would

be consistent with the known positions of the microdo-mains in RNase E Although more unlikely, we cannot rule out that, if this fragment of RNase E is its NTH, and not its CTH, RhlB, enolase and PNPase may be able to form an association independent of RNase E

Discussion Through sequence analysis, we have shown previously that the CTH of V angustum S14 RNase E contains binding sites for RhlB (residues 714–758), enolase (resi-dues 885–909) and PNPase (resi(resi-dues 1015–1094) We

Table 2 Identification of V angustum S14 RNase E-associated proteins through mass spectrometric analysis.

V angustum S14

protein

Mass

Individual peptide scorea Band A

50S Ribosomal protein

L4 (Q1ZJA7) b

Band B

Immunoprecipitation

a

For this analysis, peptides with scores > 10 are considered to be statistically significant.bSwiss-Prot accession number.

have also shown, using two-hybrid analysis, that

the enolase- and PNPase-binding sites are capable of

interacting with enolase and PNPase [38] Here, we

report that the putative V angustum S14 RNase E

RhlB-binding site (residues 714–758) interacts with

RhlB Further, results from our BN-PAGE analysis

and immunoprecipitation experiments confirmed that

RNase E complexes with RhlB, enolase and PNPase in

V angustumS14 V angustum S14 PNPase exists as a

trimer and, through two-hybrid experiments, we

showed that enolase and PNPase in V angustum S14

can self-associate Based on our observations, we present a model for the RNA degradosome complex in

V angustumS14 which shows the positions of the microdomains in the CTH of RNase E, the proteins that bind to them, as well as their dimerization state (Fig 5) We have predicted previously that there are five regions of increased structural order in the

V angustumS14 RNase E CTH: the first corresponds

to Segment A, which is involved in membrane binding (residues 565–584); the second is a putative RNA-bind-ing domain (residues 604–661); the third has a function

1 114.0913 L 13

2 227.1754 I 1315.7256 12

3 326.2438 V 1202.6416 11

4 425.3122 V 110.3.5732 10

5 540.3392 D 1004.5047 9

6 654.3821 N 889.4778 8

7 801.4505 F 775.4349 7

8 872.4876 A 628.3664 6

9 985.5717 L 557.3293 5

10 1114.6143 E 444.2453 4

11 1185.6514 A 315.2027 3

12 1282.7042 P 244.1656 2

13 K 147.1128 1

1 114.0913 I 8

2 185.1285 A 731.3934 7

3 314.1710 E 660.3563 6

4 427.2551 L 531.3137 5

5 498.2922 A 418.2296 4

6 627.3348 E 347.1925 3

7 698.3719 A 218.1499 2

8 K 147.1128 1

A

B

y(6) y(7)

y(10) y(11)

Fig 3 Fragmentation spectra for peptide sequences from V angustum S14 proteins identified in Bands A and B (A) Fragmentation spectra for peptide sequence LIVVDNFALEAPK from ribosomal protein L4 identified in Band A For this peptide, the theoretical ion mass was 1427.802 Da and the experimental ion mass was 1427.767 Da (B) Fragmentation spectra for peptide sequence IAELAEAK from PNPase identified in Band B For this peptide, the theoretical ion mass was 843.470 Da and the experimental ion mass was 843.454 Da The boxes illustrate the fragment ions for these peptide sequences as predicted by Mascot Fragments that match between our data and the predicted peptide fragments are shown in red The fragment ions for both peptide sequences include a near-complete y-ion series and two from the b-ion series, indicating a good match of the peptide fragments to the predicted sequence.

which has yet to be described (residues 792–797); the

fourth corresponds to the enolase-binding

microdo-main (residues 885–909); and the fifth corresponds to

the region which binds PNPase (residues 1015–1094)

The RhlB-binding site, which has been shown here to

be capable of binding RhlB, did not correspond to a

region of increased structural order (residues 714–758)

From the results obtained in all of our analyses, it

would be reasonable to conclude that the core of

the V angustum S14 RNA degradosome comprises

RNase E, PNPase, RhlB and enolase, all of which

are also components of the so-called canonical degradosome in E coli This is not surprising as

V angustum S14 belongs to the same subset of c-proteobacteria as E coli Organisms belonging to this group have previously been classified to have Type A RNase E homologues [22]

The interactions observed in the immunoprecipita-tion experiments and two-hybrid analyses are pairwise interactions; therefore, we sought to prove that these proteins actually form a single protein complex using BN-PAGE Previous techniques to isolate the RNA

D

Fig 4 Analysis of the V angustum S14 RNA degradosome complex through two-dimensional BN-SDS ⁄ PAGE (A) Coomassie-stained gel of the preparative separation of V angustum S14 protein complexes in the first dimension by BN-PAGE Bands 1–6 were excised and sepa-rated in a second dimension SDS ⁄ PAGE Bands were chosen for excision based on the results obtained from our BN-PAGE blot (Fig 2B).

We chose to excise Band 1 as we suspected that it consisted mainly of protein aggregation Band 2 was chosen as we wanted to deter-mine whether the excision of a band slightly higher than Bands 3 and 4 (where RNA degradosome components were detected previously) would also contain degradosome proteins Bands 3 and 4 were excised as they corresponded to the high-molecular-mass bands, Bands A and B in Fig 2A, where the degradosome components were detected previously (Fig 2B) Band 5 was excised to determine whether the degradosome components seen in Fig 2B were still present across the mass range Band 6 was chosen on the basis of the apparent migra-tion of the PNPase trimer on the BN-PAGE blot (B) Coomassie-stained 4–12% Bis-Tris gel of the second dimension SDS ⁄ PAGE separation

of the excised bands (Bands 1–6) Cleared lysate (CL) is shown for comparative purposes There is a clear difference in proteins present in each band, indicating that the bands excised from the BN-PAGE gel are composed of different proteins (C) Western blot of the second dimension SDS ⁄ PAGE separation of the excised bands The blot was probed simultaneously with a combination of RNase E, PNPase, RhlB and enolase antisera Bands 1 and 3 are more enriched in the components of the degradosome than the others Band 6 does not contain full-length RNase E Cleared lysate (CL) is again shown for comparison The lower half of the blot is not shown as there was no signal detected below the 40-kDa region A band migrating at approximately 80 kDa (*) cross-reacts with the RNase E antiserum and may be the RNase E CTH (D) Western blot of V angustum S14 lysate probed with RNase E antiserum Molecular mass markers are shown in kDa.

CTH NTH

RNase E

RhlB (714–758) Putative

RNA-binding domain

? (792–797)

“Segment A”

(565–581)

Enolase (885–909)

PNPase (1015–1094) (604–661)

Fig 5 V angustum S14 RNA degradosome model The CTH of V angustum S14 RNase E contains sites for interaction with RhlB, enolase and PNPase In addition, we have identified regions with sequence homology to ‘Segment A’ and the RNA-binding domain in E coli A remaining region (residues 792–797) which may be involved in interactions has yet to be characterized Figure not drawn to scale.

degradosome have required large amounts of starting

material, or have involved the denaturing purification

of individual components followed by their

reconstitu-tion into the RNA degradosome As emerging

evi-dence points to the RNA degradosome as a dynamic

complex with many possible components, these

tech-niques provide limited avenues for truly assessing

what is occurring in vivo We report the use of

BN-PAGE, a technique commonly employed to separate

protein complexes, two-dimensional BN-SDS⁄ PAGE

and immunoprecipitation to analyse the RNA

de-gradosome from V angustum S14 As little as 17 lg

of protein from whole-cell lysate is sufficient to

visu-alize the separation of protein complexes through

BN-PAGE Our results from the BN-PAGE

separa-tion and subsequent immunodetecsepara-tion indicate that

PNPase and RhlB associate with RNase E in a

com-plex above 1.2 MDa, but that these proteins also

co-associate in smaller complexes of approximately

900 kDa This is in agreement with current findings

that the RNA degradosome may be an assembly

which can range from 500 kDa to possibly in excess

of 4 MDa [30] Mass spectrometric analysis of the

bands at 1.1 and 1.2 MDa identified GroEL and

ribosome proteins as possible interaction partners

This was not surprising as ribosomal proteins have

been found in association with RNase E and the

de-gradosome, and play a role in its regulation,

espe-cially in times of stress [42] However, it could also

be that components of the ribosome are comigrating

with the RNA degradosome in the native gels

Together with ribosomal proteins, we identified

eno-lase and PNPase in these high-molecular-mass

com-plexes, confirming the results obtained in our

previous E coli two-hybrid analysis and

immunopre-cipitation experiments [38] It remains to be

eluci-dated whether the multiple locations of degradosome

components observed in BN-PAGE was a result of a

heterogeneous population of RNA degradosomes

in the cell, or whether the degradosome subunits

dissociated in preparation steps preceding BN-PAGE

separation

We have identified sequence homologues for the

degradosome proteins in V angustum S14 and showed

that they are capable of interacting with the CTH of

V angustumS14 RNase E as well as their respective

binding sites ([38] and this work) Previously, we have

shown that the RNase E enolase-binding and

PNPase-binding sites from V angustum S14 and E coli may

be interchangeable Enolase and PNPase from

V angustumS14 and E coli can bind to RNase E

microdomains (used in two-hybrid analysis) in V

an-gustumS14 RNase E and E coli RNase E Here, we

have expanded on these results and shown that RhlB can interact with the CTH of V angustum S14

RNa-se E and the predicted RhlB-binding microdomain Interestingly, we found that, despite low sequence identity between the CTH of E coli and V angu-stumS14 RNase E (28% identity) and the high varia-tion in sequences flanking the microdomains, E coli RhlB, enolase and PNPase are able to bind to the

V angustumS14 RNase E CTH These results suggest that specific protein interactions with RNase E can occur despite the disordered nature and variability of the sequences flanking the microdomains This high-lights the importance of microdomain sequence conser-vation; the cross-species’ setting of the two-hybrid experiments lends further strength to this The fact that the CTH of RNase E appears to adopt little struc-ture probably plays a role in providing the flexibility that is required by many molecular interactions [18,38] Further, there is decreased evolutionary con-straint, allowing the sequences to adapt to the specific requirements of the organism Future experiments may

be aimed at the further investigation of how microdo-mains direct the specificity of RNase E and how they influence the assembly of different types of degrado-somes in the cell

Experimental procedures Bacterial strains and plasmids

for BN-PAGE analysis and co-immunoprecipitation of RNase E E coli strain DHM1 was used in E coli two-hybrid assays to study protein–protein interactions as described below

Media and growth conditions

Vibrio angustumS14 was grown in high-salt Luria–Bertani broth (LB20) (10 gÆL)1 tryptone, 5 gÆL)1 yeast extract,

20 gÆL)1NaCl) E coli strains were grown in Luria–Bertani broth (LB10) (10 gÆL)1 tryptone, 5 gÆL)1 yeast extract,

10 gÆL)1NaCl) Solid medium was made by the addition of

15 gÆL)1 of agar Where appropriate, isopropyl thio-b-d-galactoside (IPTG), the lac promoter inducer, was added to

a final concentration of 0.5 mm, and

b-galactosidase, was added to a final concentration of

40 mgÆL)1 Liquid bacterial cultures were inoculated with day-old single colonies and grown in the appropriate medium on a rotary shaker at 180 r.p.m V angustum S14 was grown at

otherwise specified Where appropriate, the medium was

supplemented with antibiotics (ampicillin, kanamycin and

nalidixic acid at 100, 50 and 30 mgÆL)1, respectively)

E coli two-hybrid analysis of interactions of

RNase E with RhlB

The Bacterial Adenylate Cyclase Two-Hybrid (BACTH)

system [39] was used to test for an interaction between

RhlB and RNase E All genes and gene fragments were

cloned into the BACTH parental plasmids, whereby the

T25 domain was fused to the N-terminus (pKT25) of the

fragment and the T18 domain was fused to both the

N-(pUT18C) and C-termini (pUT18) (see Table S1 for a

com-plete list of strains and constructs, and Table S2 for the

primers used) The genes used included V angustum S14

RhlB (Swiss-Prot Accession No Q1ZNA5) and V

the V angustum S14 RNase E putative RhlB-binding

micr-odomain (residues 714–758) Experimental details and

posi-tive and negaposi-tive controls were the same as in a previous

study [38] The negative interaction of RhlB with the

RNase E PNPase-binding site also served as a negative

control Briefly, reciprocal crosses were performed where

appropriate by cotransforming chemically competent E coli

DHM1 cells and plating on LB–X-gal plates supplemented

with appropriate antibiotics The transformants were grown

new plates The patched transformants were then grown

results were scored by a ‘++’ for strong interactions, ‘+’

for moderate interactions and ‘)’ for very weak or no

inter-actions, based on colour intensity

Analysis of protein complexes

Native PAGE buffer (50 mm Bis-Tris, 6 m HCl, 50 mm

EDTA-free protease inhibitor cocktail (one tablet per

Germany) and lysed by sonication (Branson sonifier,

Bran-son UltraBran-sonics Corporation, Danbury, CT, USA) The

resulting lysate was clarified by centrifugation (22 000 g,

Native-PAGE Novex Bis-Tris gels (Invitrogen Life Techonologies,

Carlsbad, CA, USA) Gels were run at a constant voltage

of 100 V using 1· NativePAGE running buffer (50 mm

Bis-Tris, 50 mm Tricine, pH 6.8) at the anode and a light blue

cathode buffer (50 mm Bis-Tris, 50 mm Tricine, 0.02%

Coomassie G-250, pH 6.8) until the dye front migrated to

the end of the gel The apparent molecular mass of protein

complexes was estimated by comparison with very

high-molecular-mass markers of range 20–1236 kDa (Invitrogen Life Techonologies) Protein bands were visualized either

by silver staining for analytical gels or Coomassie blue staining for preparative gels Bands of interest were excised and analysed through mass spectrometric analysis

For subsequent analysis in a second dimension via SDS⁄ PAGE, following first dimension separation, the BN-PAGE gel was equilibrated in 1· Mes buffer (Invitrogen Life Technologies) for 10 min Bands of interest were excised and placed into the wells of a 4–12% Novex Bis-Tris gel (Invitrogen Life Technologies) Electrophoresis was performed at a constant voltage of 150 V until the dye front migrated to the end of the gel Protein bands were visualized with Bio-Safe Coomassie (Bio-Rad Laboratories, Hercules, CA, USA)

Western blotting and immunodetection

For western blot analysis, proteins were separated on a 3–12% NativePAGE Novex Bis-Tris gel, 4–12% Novex Bis-Tris gel, or both (Invitrogen Life Technologies), and transferred onto PVDF membranes (Millipore, Bedford,

MA, USA) using the Invitrogen XCell II blot apparatus in 1· NativePAGE Transfer buffer (25 mm Bicine, 25 mm Bis-Tris, 1 mm EDTA, pH 7.2) for native gels and 1· Native-PAGE transfer buffer in 20% methanol (v⁄ v) for Bis-Tris gels Transfer was carried out at a constant 600 mA for

90 min After blocking with 5% (w⁄ v) dried skimmed milk

mem-branes were probed by incubation with a 1 : 5000 dilution

of primary antisera (against E coli RNase E, RhlB, enolase

or PNPase, which were generous gifts from Dr A J

NaCl⁄ Pi with 0.1% Tween 20 (v/v), followed by incubation with a 1 : 5000 dilution of anti-rabbit IgG conjugated to horseradish peroxidase (GE Healthcare, Little Chalfont,

(v/v) Protein bands recognized by each specific antibody were then detected using chemiluminescence (GE Health-care) and visualized using a Fujifilm LAS 3000 imager (Fuijifilm, Tokyo, Japan)

Immunoprecipitation

Escherichia coli RNase E antiserum (a generous gift from

Dr A J Carpousis) was crosslinked to Dynabeads Protein

A (Invitrogen Dynal, Oslo, Norway) using dimethyl pim-elimidate.2HCl (DMP, Thermo Scientific, Rockford, IL, USA) according to the manufacturer’s instructions Briefly, the Dynabeads were incubated with antibodies for 4 h on a rotating wheel at 4C Excess antibodies were washed off with NaCl⁄ Pi (0.16 mm Na2HPO4.H2O, 5.51 mm NaH