Physiological Adaptation of the Rhodococcus jostii RHA1 Membrane

Jan 29, 2013 - CamM and CamABCD, two systems that take up cholate metabolites, were found exclusively in cholate-grown cells. Similarly, 9 of the 10 M...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/jpr

Physiological Adaptation of the Rhodococcus jostii RHA1 Membrane Proteome to Steroids as Growth Substrates Ute Haußmann,† Dirk A. Wolters,‡ Benjamin Fran̈ zel,‡ Lindsay D. Eltis,§ and Ansgar Poetsch*,† †

Lehrstuhl fuer Biochemie der Pflanzen, Ruhr Universitaet Bochum, Bochum, Germany Analytical Chemistry, Ruhr Universitaet Bochum, Bochum, Germany § Department of Microbiology and Immunology, University of British Columbia, Vancouver, Canada ‡

S Supporting Information *

ABSTRACT: Rhodococcus jostii RHA1 is a catabolically versatile soil actinomycete that can utilize a wide range of organic compounds as growth substrates including steroids. To globally assess the adaptation of the protein composition in the membrane fraction to steroids, the membrane proteomes of RHA1 grown on each of cholesterol and cholate were compared to pyruvate-grown cells using gelfree SIMPLE-MudPIT technology. Label-free quantification by spectral counting revealed 59 significantly regulated proteins, many of them present only during growth on steroids. Cholesterol and cholate induced distinct sets of steroiddegrading enzymes encoded by paralogous gene clusters, consistent with transcriptomic studies. CamM and CamABCD, two systems that take up cholate metabolites, were found exclusively in cholate-grown cells. Similarly, 9 of the 10 Mce4 proteins of the cholesterol uptake system were found uniquely in cholesterol-grown cells. Bioinformatic tools were used to construct a model of Mce4 transporter within the RHA1 cell envelope. Finally, comparison of the membrane and cytoplasm proteomes indicated that several steroid-degrading enzymes are membrane-associated. The implications for the degradation of steroids by actinomycetes, including cholesterol by the pathogen Mycobacterium tuberculosis, are discussed. KEYWORDS: Membrane, cholesterol, cholate, proteomics, catabolism, bacteria

■

erythropolis,7 R. rhodochrous8) and in Comamonas testosteroni.9,10 However, it has also been reported in Pseudomonas11 as well as various Gordonia12 and Mycobacterium species.13 For nonpathogenic soil bacteria, plant and fungal sterols constitute an important carbon source,14 while for M. tuberculosis, assimilation of host cholesterol is important for the establishment of a persistent infection.15 Although steroid degradation pathways have yet to be completely elucidated, they have been proposed to be organized according to the parts of the steroid molecule: the aliphatic side chain at C-17 (if present), rings A/B and rings C/ D, respectively.16 Briefly, side chain degradation resembles the β-oxidation of fatty acids,17 proceeding via CoA thioester intermediates (Figure 1). In the case of cholesterol, side chain degradation is initiated by one of two cytochromes P450 (P450s).18,19 Degradation of steroid rings A and B involves initial oxidation of ring A20 (Figure 1), followed by the sequential actions of a dehydrogenase, KstD,21 three oxygenases (KshAB, HsaAB, HsaC),22,23 and a C−C bond hydrolase24 to yield an indanone that includes the intact rings C and D. Emerging evidence indicates that for cholesterol and cholate catabolism in actinomycetes, side chain and ring A/B

INTRODUCTION Rhodococcus is a genus of catabolically versatile soil bacteria that belong to the suborder Corynebacterineae. Members of this suborder, which includes Mycobacteria, Corynebacteria, Gordonia and Nocardia, are aerobic, G/C-rich, and contain mycolic acids in their cell envelope. Indeed, the mycolic acid-containing outer membrane of rhodococci may contribute to their exceptional ability to catabolize a broad range of organic compounds including pollutants.1 The catabolic versatility of rhodococci, together with their exceptional stress tolerance2 and rapid growth rates have led to their use in numerous applications in bioremediation and biocatalysis, such as the production of acrylamide.3 Rhodococci have also been studied for the production of bioactive steroids, as steroid catabolic pathways appear to be ubiquitous in this genus.4 Rhodococcus jostii RHA1 has become a model bacterium for the study of steroid catabolism. This strain was initially isolated from a lindane-contaminated soil for its ability to degrade polychlorinated biphenyls (PCBs).5 Sequencing studies revealed a genome of 9.7 Mb6 comprising over 9000 proteincoding regions, distributed across one chromosome and three linear plasmids. RHA1 harbors four different steroid catabolic pathways encoded by distinct clusters of genes that include a high number of oxygenases. Steroid catabolism has been characterized best in Rhodococcus species (R. jostii,6 R. © 2013 American Chemical Society

Received: August 30, 2012 Published: January 29, 2013 1188

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

Journal of Proteome Research

Article

Figure 1. Overview of the proposed cholesterol degradation pathway and differentially regulated steroid degradation enzymes. (A) Proposed cholesterol catabolic pathway, adapted from Yam et al.,4 to give a survey of the degradation mechanism. Side chain degradation and degradation of rings A and B are likely concurrent. The identity of the R-group is unknown. Enzymes assigned to the catabolic reactions are indicated in gray. The exact mechanisms of side chain degradation following activation of C26 by CYP125 or CYP142, as well as the reactions catalyzed by HsaEFG and the degradation of rings C and D, are unknown (dashed arrows). (B) All steroid degradation enzymes and isoenzymes are listed that were identified or significantly regulated during growth on cholesterol (blue) and cholate (red).

degradation occur concurrently.25,26 The catabolism of rings C/ D remains largely unknown. However, this catabolism appears to be encoded by genes of the KstR2 regulon.27,16 The RHA1 genome harbors four clusters of paralogous steroid catabolic genes: three on the chromosome and a fourth on a plasmid.6 Transcriptomic and gene deletion studies have established that clusters 1 and 3 are responsible for the catabolism of CHO22 and CA,26 respectively. Cluster 1 includes the mce (for mammalian cell entry) genes, which were first described in M. tuberculosis as virulence factors28,29 and were subsequently shown to encode a cholesterol uptake system.30 The mce operon includes two sup genes (yrbE in M. tuberculosis), similar to the ABC type transporter permease, followed by six mce genes that are similar to substrate-binding subunits and, in some cases, two to four mce-associated genes. The ATPase component is usually encoded by a gene, mceG, that is located elsewhere in the genome.31 Mce loci are present in several other actinomycetes including Rhodococcus and Gordonia species, but not in corynebacteria. For its part, cluster

3 includes genes responsible for the degradation of the side chain (cas) and rings A/B of CA.31 This pathway appears to be very similar to that of Pseudomonas sp. Strain Chol1.11 Interestingly, the catabolism of CA rings C/D appears to be encoded by the KstR2 regulon, which is part of Cluster 1.27,16 In contrast to the studies on steroid catabolic pathways and enzymes, very few have addressed the global physiological adaptation of microorganisms to steroids as carbon sources. For example, transcriptomic studies revealed the up-regulation of 572 genes in RHA1 grown on cholesterol versus pyruvate,22 including genes involved in fatty acid metabolism and groEL, encoding a chaperone. Except for steroid catabolism, no regulation of whole gene clusters was observed. Interestingly, for cluster 1 the transcription profile of RHA1 cells grown on CA versus pyruvate displayed almost complementary overlap with the cholesterol transcriptome.27 Finally, no global analysis of the adaptation processes to steroids in the membrane fraction has been described yet. To address this question, the 1189

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

Journal of Proteome Research

Article

system was equipped with a nanoAcquity trap sym C18 5 μm, 180 μm × 20 mm trapping column, a BEH 130 C18 75 μm × 150 mm analytical column. The system was operated at a flow rate of 400 nL/min, and a 210 min gradient was used. After 5 min at 1% acetonitrile, the concentration was increased to 5% within 5 min, followed by a linear gradient to 40% acetonitrile in 165 min. To elute all peptides from the column, the acetonitrile concentration was raised to 99% in 15 min and kept constant for 10 min. The linear ion trap and the orbitrap were operated in parallel; i.e., during a full MS scan on the orbitrap at a resolution of 60 000, MS/MS spectra of the four most intense precursors were detected in the ion trap. Singly charged and more than triply charged ions were rejected from MS/MS. Two biological replicates from cholesterol- and CA-grown cells were analyzed. Two technical replicates for each condition were pooled after normalization for data analysis by spectral counting.

membrane proteome profiles of CHO, CA and pyruvate-grown RHA1 were quantitatively compared.

■

METHODS

Bacterial Growth

RHA1 was grown as described in ref 27 at 30 °C in minimal media5 supplemented with either 2 mM CHO, 2 mM CA, or 20 mM pyruvate as sole organic growth substrate. Cells were harvested in midexponential phase (OD600 of 1.0 for pyruvate and 2.0 for steroids) and frozen at −80 °C until further use. Triplicate cultures were grown for each growth substrate. Membrane Protein Preparation and Digestion

RHA1 cells were disrupted by French Press treatment, and membrane fractions were prepared as described for Corynebacterium glutamicum32 for both membrane proteome analysis and the localization experiments for cholesterol degradation proteins. To remove cytoplasmic contaminants for membrane proteome analysis, membranes were washed twice with 0.1 M (NH4)2CO3, pH 11, as described in ref 33. The resulting membrane pellet was resuspended in 60 μL of methanol per 100 μg of protein by sonication (10 min). 40 μL of 25 mM NH4HCO3, pH 8.6, were added. To start the digest, trypsin (Promega) was added in an enzyme/protein ratio of 1:100. The digest was incubated overnight at 37 °C. Membrane remnants were removed by ultracentrifugation (100000g, 4 °C, 30 min) prior to MudPIT analysis. For the localization experiment, after cell disruption equal amounts of protein from cell extract, cytoplasm, and membrane fraction were digested with trypsin in urea as follows. Urea was added to a final concentration of 6 M in 0.1 M Tris-HCl, pH 8. TCEP was then added to a final concentration of 5 mM. After 20 min incubation at room temperature, iodoacetamide was added to 10 mM. After 20 min incubation in the dark, the urea concentration was diluted to 2 M, and trypsin was added to a concentration of 0.01 μg/μL. After digestion overnight at 37 °C, the reaction was stopped by 1% formic acid. The peptide samples were desalted using Spec PT C18 AR solid phase extraction pipet tips (Varian, Lake Forest, CA, USA) and analyzed by 1D-nLC−MS/MS.

Data Analysis

All database searches were performed using SEQUEST algorithm, embedded in BioworksTM (Rev. 3.3.1, Thermo Fisher Scientific Inc., Waltham, MA), with a RHA1 database containing 9145 sequences.6 Only tryptic peptides with up to two missed cleavages were accepted. No fixed modifications were considered. Oxidation of methionine was permitted as variable modification. The mass tolerance for precursor ions was set to 10 ppm; the mass tolerance for fragment ions was set to 1 amu. For protein identification a threshold for deltaCn (0.08) and for XCorr values was defined, depending on the peptide charge (>2.5 (+2); > 3.5 (+3)). A protein was considered identified if at least two different peptides met these criteria. To assess the false discovery rate (FDR) of protein identification, a database with reversed protein sequences was searched retaining the search parameters and filter criteria. The FDR was calculated by dividing the absolute number of hits from the reversed database through the sum of hits from both database searches (reversed database and original database). Using the stringent criteria described above, no reversed database hit was found; therefore, the FDR was 0% in all measurements. This low error rate can be attributed to the additional demand for two different peptide matches per protein, which eliminated all otherwise observed protein hits against the decoy database. Spectral counting was applied for relative protein quantification. MS2 spectra per protein were counted from Bioworks results tables using an in-house Perl script. Spectral counts from different samples were normalized by dividing the spectral count of each protein by the total spectral count in this sample and multiplying the result with the total spectral count of the sample with the fewest spectra. The normalized spectral counts of the technical replicates were summed up, and their log2 was calculated. Spectral counts of proteins that were identified in all 3 biological replicates of at least one condition (CHO, CA, or pyruvate) were submitted to a single factor analysis of variance (ANOVA). If a difference between the mean spectral counts from the 3 conditions was detected in the ANOVA (p-value 1, the protein was considered to be significantly regulated. The last two criteria

Multidimensional Protein Identification Technology (MudPIT)

MudPIT analysis was performed as described in ref 34, except that the number of steps was reduced from 13 to 9: A quaternary Accela U-HPLC pump was connected to a Thermo LTQ XL Orbitrap from Thermo Fisher Scientific Inc. (Waltham, MA) ion trap mass spectrometer. The digested samples were loaded onto a triphasic microcapillary (Ø=100 μm), which was packed with 12 cm of Luna C18 (Phenomenex, Germany) RP material, 4 cm of strong cation exchange SCX material (Phenomenex, Germany) and finally with 3 cm of Luna C18 RP material. In the instrument method files the LTQ Orbitrap was set to detect a full MS scan between 400 and 2000 m/z for the precursor ion (R = 60 000) followed by full MS/ MS scans of the top four ions from the preceding MS scan to detect fragment ions by the ion trap. Three biological replicates from CHO-, CA-, and pyruvate-grown cells were analyzed with two technical replicates each. 1D-nLC−MS/MS

1D-nLC−MS/MS was performed as described in ref 35. A nanoAcquity UPLC system (Waters) coupled to a LTQOrbitrap (Thermo Fisher) was used. The nanoAcquity UPLC 1190

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

C C C

C C C C C C C C

C C C C MS SL MS C

MS SL

C C C

S M C C C C C

Ro02104 Ro03680 Ro03747

Ro04055 Ro04078 Ro04482 Ro04532 Ro04539 Ro04586 Ro04593 Ro04604

Ro04649 Ro04679 Ro04688 Ro04689 Ro04698 Ro04702 Ro04703 Ro04885

Ro04886 Ro04888

Ro05200 Ro05648 Ro05790

Ro05791 Ro05792 Ro05794 Ro05797 Ro05798 Ro05802 Ro05810

loc.

M C MS

Ro00645 Ro01121 Ro01137

1191

HsaD3 KstD3 HsaA3

Hsd4A3 CamM

PlsB

CamC CamA

EchA19 FadD19 Mce4A Mce4E Mce4F CamD

FadE31

KstR KstD HsaA

GyrA

name

description

hypothetical protein probable long-chain-fatty-acid-CoA ligas probable cytochrome c oxidase subunit II hypothetical protein DNA topoisomerase subunit A probable long-chain-fatty-acid-CoA ligase possible phosphatase probable FAD-linked oxidoreductase transcriptional regulator, TetR family 3-ketosteroid-delta-1-dehydrogenase 3-HSA hydroxylase, oxygenase possible acetoacetate decarboxylase acyl-CoA dehydrogenase Ni-dependent hydrogenase large subunit 2-nitropropane dioxygenase cytochrome P450 CYP125 enoyl-CoA hydratase fatty-aci-CoA ligase MCE family protein MCE family protein MCE family protein HHIDP transporter, ATP-binding component HHIDP transporter, ATP-binding HHIDP transporter, substrate-binding component probable thiolase glycerol-3-phosphate O-acyltransferase probable 3-oxoacyl-[acyl-carrierprotein] reductase dehydrogenase THSBNC transporter, MFS superfamily hypothetical protein alpha/beta-fold C−C bond hydrolase possible dehydrogenase hydroxylase probable dehydrogenase 4.65 2.15 2.20 1.87 4.73 3.09 4.15

0.75 0.21 2.14

2.11 2.69

2.79 −0.87 −1.00 −0.14 −1.00 −1.00 −1.00 2.11

1.83 1.94 −0.25 −1.00 −1.00 −1.00 2.09 −1.00

0.51 0.07 −1.00 0.71 −0.46 −1.00 −1.00 −1.00 −0.46 −1.00

1.87 3.33 −1.00 −0.11 −1.00 −1.00 −0.73 0.01 −0.43 −0.55

−0.82 −0.54 0.00 0.27 1.01 0.04 0.45

1.36 3.26 0.00

−1.12 −3.13 3.14 4.75 3.15 3.20 2.59 4.72 3.51 4.71

0.24 1.48

−1.00 −1.00

−0.76 0.48

3.38 2.79 3.01 4.84 2.80 2.98 3.35 0.50

1.26 1.48 3.14 4.86 2.98 3.33 1.86 2.38

−0.57 −1.07 3.12

2.80 2.74 −0.94

CHO:PYR

2.87 2.21

0.41 −2.66 −3.01 −3.98 −2.80 −2.98 −3.35 2.61

−1.00 −1.00 −1.00 −1.00 −1.00 −1.00 −1.00 −1.00

2.38 1.79 2.01 3.84 1.80 1.98 2.35 −0.50

1.00 2.38 2.58 3.86 1.98 2.33 0.86 1.78

0.83 −0.44 −2.82 −4.86 −2.98 −3.33 1.23 −2.78

−2.80 −1.40 −0.23

CA:CHO

−0.27 0.90 −0.56 −1.00 −1.00 −1.00 −1.00 −0.61

2.31 1.07 3.49

−0.23 1.17 0.19

PYR −1.00 −0.37 4.54 −2.54 0.10 −3.30

1.80 2.37 3.60

3.93 2.61 3.20 2.87 5.73 3.55 5.15

0.24 0.13 3.14

3.11 3.69

3.79 0.13 0.00 0.86 0.00 0.00 0.00 3.11

2.10 1.04 0.31 0.00 0.00 0.00 3.09 −0.39

−3.11 −0.98 −0.18

0.00 1.34 −1.18

CA:PYR

regulation factors as log2 ratios

2.88 2.15 0.37

CHO

CA −1.00 0.97 3.37

means of normalized log2 values

Table 1. Significantly Regulated Proteins in Steroid-Grown Cells of RHA1a

8.98 1.36 8.88 4.35 1.09 1.58 1.17 1.85

10−06 10−05 10−05 10−04 10−08 10−04 10−08 10−04

2.70 8.19 1.18 1.77 5.06 2.14 5.00

× × × × × × ×

10−02 10−03 10−03 10−02 10−02 10−02 10−03 × × × × × × ×

10−02 10−03 10−08 10−05 10−03 10−03 10−05 2.33 1.54 4.36 4.38 2.07 2.27 1.70

7.32 × 10−02 7.44 × 10−03 7.46 × 10−04

7.20 × 10−03 1.21 × 10−04 1.12 × 10−08

× × × × × × × ×

1.97 × 10−02 4.81 × 10−02

× × × × × × × ×

6.82 × 10−05 8.42 × 10−04

3.21 4.25 1.82 1.99 3.49 1.00 4.31 2.51

10−02 10−02 10−03 10−02 10−03 10−02 10−03 10−03

10−01 10−03 10−03 10−04 10−02 10−03 10−01 10−03 × × × × × × × × 2.59 2.32 1.69 7.60 4.32 2.57 3.04 9.93

10−02 10−04 10−04 10−08 10−03 10−07 10−03 10−04 × × × × × × × ×

4.08 × 10−02 7.95 × 10−01 3.32 × 10−02

1.73 × 10−02 1.91 × 10−02 1.93 × 10−03 2.16 1.79 1.59 1.18 1.80 4.51 3.50 8.15

CA:CHO 2.56 × 10−02 1.20 × 10−01 2.79 × 10−01

ANOVA 1.09 × 10−03 2.53 × 10−03 1.39 × 10−03

t test

× × × × × × × ×

× × × × × × × ×

10−03 10−03 10−03 10−03 10−03 10−02 10−03 10−01

10−01 10−02 10−02 10−04 10−02 10−03 10−02 10−03

4.23 4.23 9.72 4.23

× × × ×

10−01 10−01 10−01 10−01

7.21 × 10−01 4.23 × 10−01

6.22 × 10−03 4.80 × 10−03

4.23 × 10−01 1.62 × 10−02

3.83 9.05 8.88 3.47 1.09 1.58 1.17 1.84

1.92 1.31 1.25 7.60 4.32 2.57 8.07 4.70

1.98 × 10−01 3.74 × 10−02 1.12 × 10−02

2.56 × 10−02 5.00 × 10−02 5.85 × 10−03

CHO:PYR

p-values

8.65 5.35 1.18 2.01 1.04 3.17 5.77

× × × × × × ×

10−02 10−02 10−03 10−03 10−02 10−02 10−04

5.90 × 10−01 7.84 × 10−01 7.46 × 10−04

7.21 × 10−03 2.29 × 10−02

1.28 × 10−02

2.74 × 10−01

1.44 × 10−03 4.23 × 10−01

1.21 × 10−02 4.23 × 10−01

2.00 × 10−02 2.72 × 10−02 4.84 × 10−01

5.62 × 10−03 1.49 × 10−01 8.48 × 10−01

5.07 × 10−02 5.29 × 10−03

CA:PYR

3 3 3 3 3 3 3

3

1 1 1 1 1 1 1

1 1 1 1 1

steroidassoc. cluster

Journal of Proteome Research Article

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

1192

C

C C M C

C

S MS

C C S S C

M M

Ro06073

Ro06190 Ro06201 Ro06482 Ro06637

Ro06823

Ro06856 Ro08826

Ro08996 Ro09023 Ro09040 Ro10044 Ro10045

Ro10047 Ro11210

name

KstD4 KstD4b

SdhC

SigA

GroEL ChoD

KstD3b CasB CasC CasD CasG CasI CasL

description

probable 3-oxosteroid 1-dehydrogenas probable thiolase possible acyl-CoA dehydrogenase possible MaoC family dehydratase cholyl-CoA synthetase Steroid-22-oyl-CoA synthetase acyl-CoA dehydrogenase acyl-CoA dehydrogenase reductase ABC sugar transporter, permease component probable FMN-dependent (S)-2hydroxy-acid oxidase chaperone protein cholesterol oxidase hypothetical protein 3-oxoacyl-[acyl-carrier-protein] reductase sigma factor, sigma 70 type, probable group 1 (principle sigma factor) possible esterase/lipase succinate dehydrogenase Cytochrome b subunit probable flavin reductase probable fumarate reductase 3-ketosteroid delta(1)-dehydrogenase hypothetical protein possible type II/IV secretion system protein hypothetical protein probable copper-exporting ATPase 1.12 3.39

1.96 1.25

1.00 −0.55 −1.00 1.93 1.43

3.72 0.44

−0.68 −0.73 4.89 3.64 1.92 2.40 1.75

2.98

1.11 2.64 2.52 −1.00

2.11

−0.67 −1.32

−0.84 2.14 2.63 2.57

2.00 −0.42 0.00 −1.95 −1.12 3.89 4.19 2.92 0.47 0.32

−1.00 −0.13 −1.00 3.88 2.55

4.72 −1.86

−4.40 −1.17

0.92

−1.23

−1.51 0.82

5.89 3.77 2.92 −1.49 −0.80

0.32 −3.03

−0.31

0.23 1.01 −1.90 2.93

−1.18 1.69 −0.81 0.00

1.42 −0.68 −1.09 2.93

2.84 3.33 3.03 3.47 4.32 3.53 5.03 3.95 3.98 −1.23

CA:PYR

2.27

0.00 0.00 −0.97 0.00 0.54 0.00 0.27 0.00 0.00 1.27

CHO:PYR

0.72

1.55

2.84 3.33 4.00 3.47 3.79 3.53 4.76 3.95 3.98 −2.51

CA:CHO

regulation factors as log2 ratios

−1.00 2.30

2.06

2.29 0.94 3.33 −1.00

1.39

PYR −1.00 −1.00 0.16 −1.00 −1.00 −1.00 −1.00 −1.00 −1.00 1.18

CHO −1.00 −1.00 −0.80 −1.00 −0.46 −1.00 −0.73 −1.00 −1.00 2.45

1.75

2.52 1.96 1.43 1.93

3.66

1.84 2.33 3.20 2.47 3.32 2.53 4.03 2.95 2.98 −0.06

CA

means of normalized log2 values

2.04 × 10−03 5.91 × 10−03

× × × × ×

1.33 × 10−01 6.92 × 10−03

1.89 2.34 2.48 7.99 3.22

10−03 10−03 10−02 10−02 10−01 × × × × ×

10−07 10−03 10−04 10−03 10−03 2.60 5.05 3.67 1.03 6.04

3.03 × 10−03 2.55 × 10−02

× × × ×

5.55 × 10−06 4.17 × 10−03

2.23 2.10 4.67 1.34

2.71 × 10−02

× × × ×

1.34 × 10−03

4.88 9.77 8.19 6.44

10−02 10−01 10−02 10−03

10−02 10−03 10−02 10−03 10−03 10−02 10−03 10−02 10−03 10−02

10−02 10−03 10−03 10−08

× × × × × × × × × ×

CA:CHO 1.57 6.50 1.25 7.89 3.15 1.86 5.29 1.30 7.10 3.40 4.01 × 10−02

10−05 10−06 10−02 10−05 10−04 10−04 10−06 10−05 10−06 10−03

4.08 × 10−03

× × × × × × × × × ×

ANOVA 9.77 7.19 1.95 1.28 2.91 1.61 4.57 5.56 9.36 7.69

t test

1.01 × 10−01 7.80 × 10−02

4.37 × 10−02 1.24 × 10−02

9.39 × 10−03 7.50 × 10−01

4.39 × 10−03 8.24 × 10−02

5.16 × 10−02

6.21 × 10−03 1.06 × 10−01 2.01 × 10−01

3.04 × 10−01

1.63 × 10−02

4.23 × 10−01

4.23 × 10−01

5.28 × 10−01

CHO:PYR

p-values

10−02 10−03 10−02 10−03 10−03 10−02 10−03 10−02 10−03 10−01

× × × ×

10−01 10−02 10−02 10−03

× × × × ×

10−04 10−02 10−02 10−02 10−01 6.69 × 10−03 2.53 × 10−01

5.66 6.02 2.48 3.53 1.60

1.93 × 10−01 4.46 × 10−02

3.19 × 10−02

4.95 4.63 9.28 1.34

5.87 × 10−02

× × × × × × × × × ×

CA:PYR 1.57 6.50 7.65 7.89 5.69 1.86 2.37 1.30 7.10 1.58

4 4

3 3 3 3 3 3 3 3 3

steroidassoc. cluster

a Besides protein IDs, localization (predicted by TMHMM38 and SignalP39), and protein names, the mean of normalized spectral counts (as log2 ratio) is given for each growth condition. Regulation factors are given as log2 ratio, together with the p-values from the ANOVA and the subsequent t tests. For proteins that are part of one of the steroid-associated clusters, the number of the cluster is given as well. Loc. = predicted localization (C = cytoplasm, M = membrane, S = secreted, L = lipoprotein, according to TMHMM,38 SignalP39 and LipoP55).

C C C C C C C C C M

Ro05813 Ro05815 Ro05816 Ro05817 Ro05820 Ro05822 Ro05825 Ro05827 Ro05832 Ro05989

loc.

Table 1. continued

Journal of Proteome Research Article

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

Journal of Proteome Research

Article

Different Sets of Enzymes for Cholesterol and Cholate Degradation

were chosen to reduce the number of false positives and to focus on considerably regulated proteins.

■

Although proteins belonging to the same functional categories are up-regulated during growth on CHO and CA, only 4 proteins were significantly regulated under both conditions: Ro04078, Ro04649, Ro04888 and Ro08996. Ro04649, a predicted flavin-containing protein, is the only protein involved in steroid degradation that is significantly up-regulated under both conditions. This is consistent with ro04649 belonging to the KstR2 regulon that has been proposed to encode rings C/ D-degrading enzymes that act on all steroids used as a growth substrate by RHA1. As mentioned above, RHA1 possesses four clusters of steroid catabolic genes. While in CHO-grown cells 10 proteins coded in cluster 1 (several clusters of genes coded in the region between ro04482 and ro04707 22) were significantly up-regulated, 17 proteins coded in cluster 3 (ro05787−ro05833), 2 proteins from cluster 4 (ro09002− ro09040), and Ro04649 from cluster 1 were significantly upregulated during growth on CA (see Figure 2). The numbers of

RESULTS

Cholesterol and Cholate Membrane Proteome of RHA1

A total of 1703 proteins were identified in this study, of which 734 were detected under all three experimental conditions (CHO, CA, and pyruvate), 233 in cholesterol-grown cells only, 255 in CA-grown cells only, and 125 in pyruvate-grown cells only. With spectral counting quantification, 223 proteins yielded a p-value below 0.05 in the ANOVA of which 168 had significant p-values in a subsequent t test. Finally, 59 proteins also met the cutoff values for the number of peptides and the regulation factor. Eleven of the proteins displaying significant abundance changes are annotated as membraneintegral proteins; six are annotated as secreted proteins. Table 1 lists all significantly regulated proteins, and further information about the spectral counting quantification and statistical analysis, as well as the MS results can be found in the Supporting Information “SpectralCounts-Data MudPIT”, “PeptideIdentifications_MudPIT”, and “ProteinIdentifications_MudPIT”. Growth on steroids as sole carbon and energy source led to the induction of numerous proteins: 31 of the 59 significantly regulated proteins were not detected under the control condition (PYR). Although the two steroids are structurally similar, the proteome responses to CHO and CA were quite distinct: of 24 proteins that were significantly up-regulated during growth on CHO, 16 were detected under this growth condition only or up-regulated compared to both CA and PYR. Similarly, of 32 proteins that were up-regulated during growth on CA, 21 were up-regulated compared to both other growth conditions. The majority of proteins up-regulated during growth on steroids belonged to the functional categories of lipid metabolism, energy production and conversion, and secondary metabolite metabolism according to COG annotation.6

Figure 2. Identified and significantly up-regulated proteins coded in steroid-associated clusters 1−4. Clusters described in refs 22 and 37. Up-regulation refers to pyruvate and/or the other steroid carbon source.

Induction of Steroid Degradation Enzymes

identified proteins encoded in the different steroid-associated clusters underline that CHO almost exclusively induces expression of proteins from cluster 1, while CA mainly induces expression of proteins encoded in cluster 3 and to a lesser extent of proteins coded in cluster 4. Since up to 6 isoenzymes for several steroid degradation enzymes are encoded in the different gene clusters, the amino acid sequences of significantly regulated steroid-degrading enzymes and their isoenzymes were submitted to an in silico digest (using ProteinProspector MS-digest: http://prospector. ucsf.edu) to identify ambiguous peptides. However, none were detected in the mass spectrometric analysis. Thus, the different isoenzymes were identified unambiguously.

Many significantly regulated proteins are directly involved in steroid degradation. Thus, 11 out of 24 proteins significantly up-regulated during growth on CHO are involved in cholesterol degradation: 22 cytochrome P450 CYP125 (Ro04679), EchA19 (Ro04688) and FadD19 (Ro04689) are involved in degradation of the C17 aliphatic side chain, KstD (Ro04532) and HsaA (Ro04539) catalyze steps of rings A/B degradation (see Figure 1). Ro04649 is part of the KstR2 regulon36 and is predicted to be involved in degradation of rings C/D of the parent steroid.16,27 Additionally, KstR (Ro04482), the putative transcriptional regulator of the kstD gene, was up-regulated in CHO-grown cells. The 3 Mce4 proteins (Ro04698, Ro04702, and Ro04703), which were induced during growth on CHO, are subunits of the multicomponent CHO transport system.30 Similarly, 22 of 32 proteins significantly up-regulated in CAgrown cells are encoded by steroid catabolic genes.22,37,27 Of these, 17 are encoded by CA catabolic genes occurring in Cluster 3,27 including KstD3, KstD3b, KstD4, and KstD4b, all of which are 3-ketosteroid-Δ1-dehydrogenases, and HsaD3, a 4,9-DHSA hydrolase (see Figure 1). Two other proteins, Hsd4A3 and Hsd4B3 are homologous to enzymes degrading the C17 aliphatic side chain.37

Localization of Steroid Degradation Enzymes

Numerous proteins annotated as steroid degradation enzymes were identified and quantified in the membrane fraction, although most proteins were predicted to be cytoplasmic proteins according to TMHMM38 and SignalP.39 For significantly regulated proteins involved in steroid catabolism, CELLO40 and PSORTdb41 results (available on www.BioCyc. org) were used to complement and verify the predictions for subcellular localization. For 10 out of 27 proteins the predicted localization varied depending on the prediction tool. The 1193

dx.doi.org/10.1021/pr300816n | J. Proteome Res. 2013, 12, 1188−1198

Journal of Proteome Research

Article

results obtained with TMHMM,38 SignalP39 and with CELLO40 only differed in 2 cases (Mce4E and KstD4B). In contrast, with PSORTdb,41 for 4 KstD isoenzymes (KstD, KstD3, KstD3b, and KstD4) membrane localization was predicted instead of cytoplasmic localization. To assess the localization of steroid-degrading enzymes experimentally, membrane and cytoplasmic fractions were compared to the total cell extract, as obtained after removal of intact cells and cell debris following French Press disruption. If a protein was membrane-associated, it should be enriched in the membrane fraction and depleted in the cytoplasmic fraction (compared to the cell extract) after sedimentation of membranes by ultracentrifugation. Normalized spectral counts from membrane fraction and cytoplasm were compared to spectral counts from cell extract. A protein was considered to be membrane-associated if the mean membrane to cell extract ratio was >1, while the mean cytoplasm to cell extract ratio was