| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Faculty of Biology and
Faculty of Chemical Engineering, Technion—Israel Institute of Technology, Haifa, Israel;
UVic Proteome Center, University of Victoria, British Columbia, Canada; and
Research Unit for Tropical Diseases, Christian de Duve Institute of Cellular Pathology and Catholic University of Louvain, Brussels, Belgium
1Correspondence: Faculty of Biology, Technion—Israel Institute of Technology, Haifa 32000, Israel. E-mail: danz{at}tx.technion.ac.il07-9254com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: proteomics • metabolomics • intracellular differentiation • gene expression
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Our understanding of these adaptation mechanisms is limited. Recent sequencing of several parasite genomes and advances in high throughput analytical techniques are enabling new insights into these processes. In the present communication we have employed a novel proteomics approach to identify gene products that enable parasite adaptation to a new host during differentiation of the intracellular parasitic protist Leishmania donovani.
L. donovani are the causative agents of kala azar, a fatal disease in humans. These organisms cycle between phagolysosomes of mammalian macrophages and the alimentary tract of sand flies. In the insect vector, they grow as flagellated extracellular promastigotes, whereas in the mammalian host, they proliferate as aflagellated intracellular amastigotes. Promastigotes are introduced into the host during a blood meal taken by the fly and are subsequently phagocytosed by macrophages, where they differentiate into amastigotes (3)
. This process is mimicked in vitro by shifting cultured promastigotes (grown at 26°C, pH 7) to a lysosome-like environment (37°C and pH 5.5; differentiation signal) (4
5
6)
. Although axenic amastigotes are not identical to animal-derived amastigotes (7)
, they closely resemble them and are the best available model for the study of parasite intracellular development.
To date, most studies on Leishmania differentiation have focused on comparing promastigote and amastigote gene and protein expression. These studies have shown that promastigotes and amastigotes are highly adapted to their respective environments. For example, amastigote respiration, catabolism of energy substrates, and synthesis of macromolecules are carried out optimally at acidic pH whereas promastigotes perform these activities optimally at neutral pH. In addition, metabolic pathways such as glycolysis are more active in promastigotes whereas fatty acid oxidation is more active in amastigotes (8)
. Also, gluconeogenesis is essential for the virulence of the amastigotes and their proliferation inside macrophages, but promastigotes can survive without this pathway (9)
. Indeed, numerous studies have identified stage-specific genes, highlighting the importance of gene expression changes during differentiation (10
11
12
13
14
15
16
17
18
19)
. Despite a wealth of knowledge concerning the different gene and protein expression profiles of promastigotes vs. amastigotes, it is not yet possible to envisage the series of molecular events that underlie the differentiation process connecting these parasite forms. Recently, our laboratory initiated time-course analysis of L. donovani differentiation and showed that it is a regulated process involving changes in morphology that parallel discrete stages of protein and gene expression (1
, 4
, 5
; Lahav et al., unpublished results). These studies showed that differentiation is divided into four stages: I) 0–4 h when promastigotes receive and process the differentiation signal; II) 5–9 h, when they cease movement and aggregate; III) 10–24 h, when promastigotes change morphologically into amastigote-shaped cells; and IV) 25–120 h, when maturation into amastigotes is completed. The transition between the first and second stages occurs synchronously, while cells are arrested at G1 (5)
.
Transcriptome analysis revealed that the majority of promastigote-specific gene transcripts were down-regulated early in differentiation, while amastigote-specific transcripts were up-regulated only at the fourth stage of differentiation (i.e., >24 h). However, gene expression in Leishmania is regulated mostly posttranscriptionally (20)
, which means that changes in mRNA abundance are not necessarily reflected faithfully in altered protein profiles. Moreover, it has been shown recently, both in vitro and in vivo, that protein remodeling via autophagy is an essential process for promastigote to amastigote differentiation (21)
. Therefore, in order to form a better picture of the molecular events that occur during the four differentiation stages, we elected to study these stages at the proteomic level.
To follow proteomic changes during Leishmania differentiation with high protein coverage, we employed a novel isobaric tagging methodology, isobaric tags for relative and absolute quantification (iTRAQ; 22
). This method utilizes amine-reactive isobaric tags to label all peptides in a particular protein digest. An advantage of this method is that 4 protein digest samples can each be tagged differently, allowing direct comparison of up to 4 samples. Samples are combined at an equal ratio and subjected to LC-MS/MS, where, on fragmentation, every fragmented peptide tag produces distinct signature ions differing by an m/z value of 114–117. The relative intensities of these signals represent the relative abundance of the analyzed peptide in each sample. Relative abundance values of all peptides attributed to each specific protein are averaged to represent the relative abundance of the entire protein.
Approximately 21% of the entire theoretical proteome was detected at >95% confidence. We observed major changes in the abundance of metabolic enzymes as well as altered levels of translation machinery and DNA condensation proteins. These observations shed light on how promastigotes retool themselves for life as amastigotes during differentiation. Remarkably, only a few proteins displayed significant changes in their abundance during the initial 10 h of differentiation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Promastigotes and amastigotes were grown as described (5)
. Differentiation of promastigotes to amastigotes in axenic culture was carried out as in Barak et al. (5)
. Briefly, late-log promastigotes were transferred from promastigote medium at 26°C to amastigote medium at 37°C and 5% CO2. Three days after initiation of differentiation, cells were split 1:2 in prewarmed amastigote medium.
Protein extraction
Cells were washed once in ice-cold 0.9% NaCl, then twice in ice-cold PBS and finally suspended in PBS. Subsequently, cell suspension was sonicated (3 cycles of 1 min sonication and 1 min incubation on ice). Resulting lysate was centrifuged (12,000 g, 4°C, 10 min), and supernatant was precipitated in cold acetone.
Peptide labeling
Labeling of peptides with the iTRAQ reagent was carried out essentially as suggested by the manufacturer (Applied Biosystems, Inc., Foster City, CA, USA). Briefly, from the commercial kit, denaturant was added to each of up to 4 sample tubes containing 100 µg of protein, followed by reducing reagent and incubation at 60°C for 1 h. Cysteine residues were then S-alkylated for 10 min followed by tryptic digestion overnight at 37°C. This was followed by N-alkylation of peptides with the appropriate iTRAQ reagents for 10 min at 37°C and combining of differently labeled samples for analysis. Two labeled peptide mixes were created; one mix included equal amounts of protein from promastigote, 2.5, 5, and 10 h of differentiation (early differentiation), and the other included promastigotes, 15 and 24 h of differentiation and mature amastigotes (late differentiation).
Multidimensional chromatography and mass-spectrometry
Labeled and multiplexed sample peptides were fractionated using 2-D LC employing strong-cation exchange (SCX) chromatography for the first dimension separation. The latter was carried out on a Vision Workstation (Applied Biosystems) equipped with a Polysulfoethyl A (Poly LC, Columbia, MD, USA) 100- x 4.6-mm, 5-µm, 300-Å SCX column. Buffer A was 10 mM KPO4 (pH 2.7) in 25% ACN, while buffer B was 10 mM KH2PO4 with 25% ACN in 0.5 M KCl. The column was developed with a linear gradient of 0–35% B in 30 min at a flow rate of 0.5 ml/min. SCX fractions were collected, and each fraction subsequently was analyzed using an integrated online RP LC/MS/MS system. The latter consisted of an autosampler, switching pump, and micropump (LC Packings, Amsterdam) interfaced to a hybrid Quadrupole-TOF LC/MS/MS Mass Spectrometer (QStar Pulsar I, Applied Biosystems). The latter was equipped with a nanoelectrospray ionization source (Proxeon, Odense, Denmark) and fitted with a 10 µm fused silica emitter tip (New Objective, Woburn, MA). RP LC was performed on a 75-µm x 15-cm C18 PepMap Nano LC column (LC Packings, Amsterdam, The Netherlands) with a 300-µm x 5-mm C18 PepMap guard column (LC Packings) in place. Samples were loaded onto a trap or guard column in a volume of 25–50 µl and were equilibrated for 10 min in 95% solvent A/5% solvent B at a flow rate of 100 µl/min. On switching inline with the MS, a linear gradient at 200 nL/min from 95–40% solvent A was developed for 40 min, and in the following 5 min the composition of mobile phase was increased to 20% A before decreasing to 95% A for a 15-min equilibration before the next sample injection. Fractions were eluted with a linear 60-min (low peptide concentration) or 120-min (high peptide concentration) gradient from 5–60% solvent B. MS data were acquired automatically using Analyst QS 1.0 software Service Pack 8 (Applied Biosystems MDS SCIEX, Concord, Canada). A 1-s TOFMS survey scan was conducted over 400–1200 amu, followed by two 2.5-s product ion scans over mass range 100–1500 amu. The two most intense peaks showing 20 counts/s with charge state of 2–5 were selected for fragmentation. A precursor ion within a 6-amu window, once selected for fragmentation, was excluded from detection for 180 s. Curtain gas was set at 23 (gas flow 1.22 L/min), nitrogen was used as the collision gas, and the ionization tip voltage used was 2700 V.
Results-dependent data acquisition and peptide analysis
Using additional aliquots of the SCX fractions, exclusion analyses were carried out. For this procedure, data files were processed using the Protein Pilot software (Ver. 1.0) (Applied Biosystems). A list of peptides (i.e., m/z values) that were detected with confidence greater than 95% was generated by Protein Pilot, saved as an Excel file, and imported into the Analyst method as a text file. The same data acquisition method (.dam file) was used to acquire the first and second LC/MS/MS data. Duplicate aliquots of SCX fractions were analyzed by the first and second methods. Peptides were excluded throughout the entire run of a specific SCX fraction based on their m/z value. This procedure allowed identification of significantly more peptides than a single LC/MS/MS analytical cycle.
Data processing and analysis
Data files were processed using the Protein Pilot software. The LC/MS/MS data were used to search GeneDB L. infantum V3.0 database. Proteins were considered to be confidently identified if they had an unused Protscore > 1.3 (Conf.>95%). The Protein Pilot software then calculated the ratios for the reporter ion fragment masses for each peptide.
Data from early and late differentiation iTRAQ analysis were united, using the promastigote time points as reference values, and log2-transformed. For clustering analysis, only proteins quantified at all time points were used. Proteins were grouped according to self-organizing maps algorithm, with parameters set at 2 x 2 grid and 100,000 iterations using Expander software (23)
.
Protein coexpression significance of each functional group was calculated using the Pearson correlation coefficient. Each pairwise correlation average was then compared with a distribution of 1000 random groups with the same number of proteins. Western blot analysis was done as in Barak et al. (5)
. Protein loading was based on tubulin load but confirmed using Ponceau-red stain densitometry and NanoDrop® spectrophotometry.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, a gradual increase in the abundance of mRNA coding for ascorbate-dependent peroxidase (1)
, and a transient increase in spliced leader RNA abundance (Lahav et al., unpublished results). In the present study, promastigotes were differentiated into amastigotes on two separate occasions, representing independent biological repeats of complete axenic differentiation. All three time-course markers were exhibited similarly in both repeats, and thus, according to these differentiation criteria, were complete and consistent in both repeats. L. donovani promastigotes were subjected to the differentiation signal (pH 5.5 and 37°C; time zero) and soluble proteins collected at 2.5, 5, 10, 15, and 24 h, as well as at 6 days postsignal when the parasites are mature, fully differentiated amastigotes. Protein samples were divided into two groups of four for labeling. The first labeling group was unstimulated promastigotes (no differentiation signal) and promastigotes 2.5, 5, and 10 h postsignal; the second labeling group was unstimulated promastigotes, promastigotes 15 and 24 h postsignal, and mature amastigotes. Each sample was digested and labeled with one of the four iTRAQ tags. Equal amounts of peptides from each sample were mixed together (according to the aforementioned labeling groups) and subjected to multidimensional LC-MS analysis (see Materials and Methods). In addition, a control experiment was performed where two independent soluble protein samples derived from unstimulated promastigotes were mixed and subjected to multidimensional LC-MS analysis. The mean difference in protein abundance for the control experiment was ±0.1-fold, indicating the iTRAQ method accuracy. Peptide masses obtained from mass-spectrometry analysis were characterized using the theoretical L. infantum version 3.0 proteome (http://www.genedb.org/genedb/linfantum/). This genome was used because L. infantum and L. donovani are closely related Old World species.
Summing all the signals from both repeat experiments, a total of 12,021 distinct peptides were detected. These were grouped to identify 1713 proteins at 
95% confidence out of 8184 theoretical proteins (Supplemental Table S1). The detected proteins represent 20.9% of the entire L. infantum proteome. Assuming that at any given time point about half of the proteome is expressed, this analysis detected almost half of the expressed proteome. On average, 7 distinct peptides were detected for each identified protein. Of the identified proteins, 934 have a known or predicted function, whereas 779 are proteins with no known function. The analysis revealed 969 proteins at all 7 time points (Supplemental Table S1a), while the rest were detected only in the first (i.e., promastigote, 2.5, 5, 10 h, Table S1b) or second labeling group (promastigote, 15, 24 h, amastigote, Table S1b). Of note, out of the 1713 proteins identified, only 931 were detected in both biological repeats. The reason for the limited overlap is that iTRAQ is a shotgun-proteomic method, and, as such, the assortment of peptides identified in each independent assay is random. Focusing on the 931 proteins detected in both repeats, a correlation plot was constructed (Fig. 1
A), demonstrating high correlation between the experimental repeats, ranging from 0.38 in promastigotes to 0.85 in mature amastigotes. This plot indicated relatively small variations in a given protein’s abundance between repeats and corroborated the reproducibility of data generated using iTRAQ methodology. Since we were interested in characterizing the entire time course of differentiation, we decided to include in our data analysis proteins identified in only one of the repeats.
|
The protein expression profiles began to diverge slightly during the first two differentiation stages (up to 9 h), but diverged more prominently during the third and fourth stages (Fig. 1B
). Analysis of only proteins detected at all 7 time points, using a self-organizing map algorithm, yielded four expression clusters (Fig. 1C
). The amastigote-specific cluster I contains 289 proteins (16.9% of the proteome), while the promastigote-specific cluster II contains 310 proteins (18.1% of the proteome). The other proteins (370) do not change expression significantly during differentiation (clusters III and IV). Both amastigote- and promastigote-specific expression gradually increased and decreased, respectively, with the largest changes occurring during the third and fourth differentiation stages.
Expression pattern of metabolic enzymes
iTRAQ analysis detected many enzymes. Most major metabolic pathways were represented in the data, including glycolysis, gluconeogenesis, β-oxidation, the tricarboxylic acid cycle, oxidative phosphorylation, the pentose phosphate pathway, mitochondrial respiration, and amino acid catabolism (Table S2; Fig. 2
).
|
Glycolysis is down-regulated during late differentiation
The glycolytic pathway between glucose and 3-phosphoglycerate occurs in the glycosomes, whereas the final steps that lead to formation of pyruvate are cytosolic (8)
. In trypanosomatids, the regulation of glycolysis takes place in the cytosol, with pyruvate kinase being the regulatory enzyme (24)
. We observed that the expression of glycosomal glycolytic enzymes was up-regulated gradually during differentiation, such that it was 
1.4-fold higher in mature amastigotes (Fig. 2A
). In contrast, the expression of cytosolic glycolytic enzymes was down-regulated significantly and coordinately from 10 h postsignal, with phosphoglycerate mutase, enolase and pyruvate kinase expression reduced 2-fold, on average, in amastigotes (Fig. 2B
). This concerted down-regulation of three consecutive glycolytic enzymes, one of which is the rate-limiting step, suggests active, comprehensive down-regulation of glycolysis during the late stages of differentiation.
β-oxidation is up-regulated during late differentiation
iTRAQ analysis detected 13 proteins directly involved in β-oxidation (Fig. 2C
; Table S2). Except for one acyl-CoA dehydrogenase isoform (LinJ28_V3.2700), which exhibited increased expression as early as 2.5 h postsignal, the expression of these proteins did not change significantly during the initial 15 h of differentiation. However, during the late third and fourth differentiation stages, 11 of the 13 proteins displayed significantly increased expression. iTRAQ also identified 7 proteins that participate in β-oxidation indirectly, including lipases, fatty acyl-CoA synthetases/ligases and acetyl-CoA carboxylase. The amount of long chain fatty acyl CoA synthetase, the rate-limiting enzyme of β-oxidation responsible for associating long-chain free fatty acids with coenzyme A, increased 3-fold after 15 h of differentiation and 13-fold in mature amastigotes, while the other ligases were down-regulated. The large increase in abundance of this enzyme was accompanied by a significant increase in the expression of three acyl-CoA dehydrogenase isoenzymes, predicted by homology to be specific for very-long-, long-, and short/branched-chain fatty acids. In addition, the level of 2,4-dienoyl-CoA reductase, an enzyme responsible for processing unsaturated fatty acids prior to oxidation, was increased 1.8-fold at 15 h and 7.5-fold in the mature amastigote. These data substantiate the hypothesis that fatty acid oxidation is more active in amastigotes (25)
and is adapted primarily to very-long- and long-chained and/or unsaturated fatty acids.
More support for this premise is provided by two other observations. First, acetyl-CoA carboxylase, which synthesizes a key inhibitor of β-oxidation that is a precursor of fatty acid synthesis (malonyl-CoA) (26)
, exhibited 1.6-fold down-regulated expression after 15 h of differentiation, retaining this reduced level in amastigotes. Second, 2-fold increases were detected in the abundance of lipolytic enzymes, namely phospholipase A2 (PLA2) and monoglyceride lipase (Supplemental Table S2). PLA2 releases free fatty acids from phospholipids and diacylglycerols, while monoglyceride lipase releases fatty acids from monoacylglycerols. Monoglyceride lipase increases gradually throughout differentiation, while expression values of PLA2 only are available for late differentiation. Incidentally, the 1.8-fold up-regulated level of glycerol kinase in mature amastigotes suggests strongly that the glycerol produced by the lipases is utilized for gluconeogenesis (8)
.
Tricarboxylic acid cycle enzymes are up-regulated during differentiation
All enzymes of the tricarboxylic acid (TCA) cycle were identified at all differentiation time points (Fig. 2D
; Supplemental Table S2). Although there were no major changes in abundance detected for most TCA enzymes during the first 15 h of differentiation, the levels of these enzymes increased on average 1.7-fold during the fourth differentiation stage. While two fumarate hydratase isoenzymes displaying opposite expression trends were detected, sequence alignment and in silico subcellular localization by TargetP indicated that LinJ24_V3.0310, the up-regulated isoenzyme, is probably mitochondrial, whereas LinJ29_V3.2080 is probably glycosomal. Thus, the generally increased abundance of TCA cycle enzymes by the fourth differentiation stage suggests that TCA activity is higher in amastigotes.
Mitochondrial respiration and oxidative phosphorylation are up-regulated throughout differentiation
Fourteen proteins involved in mitochondrial respiration were detected (Fig. 2E
; Supplemental Table S2). A slow but steady increase in expression of these proteins reached a final 1.7-fold change, on average, in amastigotes. The coordinated increased expression of respiratory chain proteins implies that mitochondrial respiratory activity increases during differentiation.
Two proteins of the F0/F1 ATPase complex, an F1 
subunit and a β subunit, were detected. The abundance of these proteins increased slightly over the time course, the rise beginning at 15 h of differentiation and reaching 1.5-fold in mature amastigotes (Fig. 2F
). Though modest, this increase was consistent across several time points, and we therefore consider it significant. The ATP/ADP translocase, a carrier that couples ATP export from mitochondrion with ADP import, and a mitochondrial phosphate transporter responsible for phosphate transport from the cytosol to the mitochondrion exhibited the same expression pattern. The concurrently increased expression of F1 ATPase subunits, the ATP/ADP translocase, and a mitochondrial phosphate transporter suggests that oxidative phosphorylation increases during differentiation. Moreover, the similarity between the differentiation expression patterns of oxidative phosphorylation and mitochondrial respiratory proteins suggests that expression of these proteins is linked functionally.
Amino acid catabolic enzymes are up-regulated during late differentiation
Five enzymes that catabolyze amino acids were detected (Fig. 2G
): two glutamate dehydrogenase isoforms, an alanine aminotransferase, a branched chain aminotransferase and a glycine cleavage system H protein. All of these enzymes displayed an up-regulated differentiation expression pattern, reaching 3.3-fold expression for glycine cleavage system H protein, 1.9- and 3.8-fold for the two glutamate dehydrogenase isoforms, 3.2-fold for branched-chain aminotransferase and 2-fold for alanine aminotransferase in mature amastigotes.
Pentose phosphate pathway
iTRAQ detected five enzymes involved in the pentose phosphate pathway (Fig. 2H
). The abundance of these proteins decreased slightly, but significantly (P<0.01), during the fourth differentiation stage.
Expression pattern of gluconeogenic enzymes
Sixteen proteins of the gluconeogenesis pathway were detected (Fig. 2I, J
; Supplemental Table S2). Two key enzymes of this pathway were up-regulated more than 2-fold in amastigotes. The first is phosphoenolpyruvate carboxykinase (PEPCK), which converts oxaloacetate to phosphoenolpyruvate (PEP), and the second is fructose 1,6-bis-phosphatase (F16BP), which dephosphorylates fructose 1,6-bis-phosphate to fructose-6-phosphate. The levels of PEPCK and F16BP increased steadily throughout differentiation, reaching 2.1- and 2.9-fold in amastigotes, respectively (Fig. 2I
). Other gluconeogenic enzymes exhibited moderate up-regulation, with the exception of the cytosolic enzymes, enolase, and two phosphoglycerate mutase isoenzymes, which also are part of the glycolytic pathway and were down-regulated (Fig. 2J
).
The observed decrease in glycolysis results in insufficient pyruvate for gluconeogenesis in the maturing amastigote. A possible alternative source of pyruvate could be the transamination of alanine; indeed, alanine aminotransferase expression was up-regulated 2-fold in the mature amastigotes. Pyruvate is likely to be further phosphorylated by pyruvate phosphate dikinase (PPDK), the abundance of which began increasing early in the time course, finally reaching 4.4-fold in amastigotes (Supplemental Table S1).
Proteins of the translation machinery
iTRAQ identified 90 ribosomal protein subunits, of which 78 were detected at all differentiation time points (Fig. 3
A, Supplemental Table S3). These proteins remained unchanged during the initial 10 h of differentiation and subsequently gradually decreased in abundance, reaching 2.4-fold decrease in amastigotes. Interestingly, we observed an almost uniform expression pattern of these proteins, suggesting strict regulation of the translation machinery. This is further emphasized by the similar expression pattern of translation elongation and initiation factors detected (Fig. 3B
, Supplemental Table S3) as well as tRNA synthetases (Fig. 3C
, Supplemental Table S3). Twenty-seven translation factors and 22 tRNA synthetases were detected and quantified. On average, both groups remained unchanged during initial differentiation and were coordinately down-regulated from 15 h postsignal, to half the promastigote expression level in mature amastigotes.
|
Other expression trends
Twelve histones were detected. The expression of these proteins did not change during the first 10 h of differentiation, and then increased, reaching 2-fold in mature amastigotes (Fig. 3D
). One isoform of histone H2B displayed higher expression throughout differentiation; its abundance increased 5.2-fold in amastigotes. The elevated expression of histones in the amastigote stage may indicate transcriptional changes.
Calpains are cysteine peptidases that have been demonstrated to play a role in calcium-regulated functions such as signal transduction and cell differentiation in other organisms (27)
. iTRAQ analysis detected 10 calpains (Fig. 3E
) out of 33 putative calpains in the L. infantum genome. Five of these displayed 2- to 4-fold decreased expression in amastigotes, while three showed 1.4-fold to 1.92-fold increased expression in amastigotes. These observations indicate a development-dependent switch in the repertoire of calpains expressed.
Nineteen heat shock proteins and chaperones were detected by iTRAQ. These displayed a very small increase during the first and second differentiation stages, then returned to promastigote or lower levels (Fig. 3F
, Supplemental Table S4). One exception was HSP100, which was previously reported to be amastigote-specific (28)
. Since HSP83 is encoded by a multicopy gene family (29)
, it is not clear whether the iTRAQ results represented the products of all the HSP83 genes. Using polyclonal antibodies specific to HSP83 (5)
, Western blot analysis (Fig. 4
A) demonstrated that the overall HSP83 expression pattern was similar but not identical to the iTRAQ data. On the other hand, Western blot analysis of HSP100, which is single gene (28)
, displayed an expression pattern identical to that reported by iTRAQ (Fig. 4B
).
|
Proteins up-regulated at the beginning of differentiation
Although most of the changes in protein abundance occurred during the third and fourth stages of differentiation, the expression of 20 proteins was up-regulated more than 2-fold within 2.5 h of differentiation (Table 1
). Of these, 13 are proteins with unknown function, and the other 7 are as follows: D1 myo inositol transporter (up-regulated 7.2-fold at 2.5 h), universal minicircle sequence binding protein (UMSBP) (up-regulated 6.7- and 10.4-fold at 2.5 and 5 h, respectively), a flavoprotein subunit, a methionine synthase, a histone H1, a kinetoplast-associated protein, and ascorbate-dependent peroxidase. In the case of ascorbate-dependent peroxidase, the increase in protein abundance is mirrored by the increase in its encoding mRNA (1)
. Notably, UMSBP was detected only in one repeat and only at 2.5 and 5 h postsignal, in both cases at marginal confidence. However, as its expression increased dramatically, we used specific antibodies raised against Crithidia fasciculata UMSBP to verify the differentiation expression pattern by an alternative protocol (30)
. In agreement with the iTRAQ data, Western blot analysis shows that a low level of UMSBP expression is detected in unstimulated promastigotes, but expression subsequently increases dramatically, remaining high throughout the first 24 h of differentiation, only decreasing in mature amastigotes back to unstimulated promastigote levels (Fig. 4C
). These Western blot data not only corroborate the developmental regulation of UMSBP expression but also validate the reliability of iTRAQ data.
|
Correlation analysis of functional group expression
To analyze and quantify the relationship between pathways, a dendogram was created using Expander software whereby the average expression values for each group of proteins with related functions were calculated, correlated, and depicted (Fig. 5
) (23)
. This dendogram analysis highlights several features of differentiation. Catabolic pathways that lead to energy production were up-regulated concurrently during the third and fourth differentiation stages. Specifically, the catabolic pathways TCA cycle, mitochondrial respiration, and oxidative phosphorylation exhibited highly correlated expression. The expression pattern of one biosynthetic pathway, glycosomal gluconeogenesis, was found to correlate with the expression patterns of β-oxidation and amino acid catabolism enzymes. Ribosomal protein subunits and translation factors displayed a coordinated pattern of down-regulation. Thus, in general, anabolic functions were down-regulated, whereas catabolic functions were up-regulated.
|
The dendogram analysis confirms that, as expected, when considering all the time points, amastigote protein expression (6 days postsignal) is the most diverged from promastigote expression (unstimulated). The protein expression patterns detected at 15 h and 24 h postsignal are significantly different from those observed at 2.5, 5, and 10 h, which reflects the increasing magnitude of expression changes during the third differentiation stage.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Using high-coverage, comparative proteomic analysis of soluble parasite proteins, we collected expression information on 21% of the L. donovani proteome. Our data reveal changes in the expression level of most proteins involved in core metabolic pathways and of many molecules involved in DNA condensation and translation. This study, for the first time, enables a comprehensive view of the biogenesis processes underlying differentiation from promastigote to amastigote. This is summarized in the dynamic metabolome of L. donovani differentiation (Fig. 6
). Our major observations include: a) during differentiation, the parasites shift from glucose to fatty acid oxidation as the main source of metabolic energy; b) in line with the increase in β-oxidation capacity, TCA cycle enzymes, mitochondrial respiratory chain, and oxidative phosphorylation proteins are up-regulated; c) concomitantly, differentiating parasites up-regulate gluconeogenesis, producing sugars from glycerol and amino acids; d) translation slows down in maturing amastigotes, and DNA condensation increases; and e) most changes in proteins of physiological significance occurred during the third and fourth differentiation stages, when promastigotes undergo a morphological transition to amastigotes.
|
Our investigations revealed that during late differentiation, the abundance of glycosomal glycolytic enzymes increases slightly, whereas the expression of three consecutive cytosolic enzymes is down-regulated significantly. Therefore, we propose that the decrease in abundance of cytosolic glycolytic enzymes must result in the formation of a bottleneck, which slows the glycolytic flux. Our conclusion concurs with and validates early in vivo studies by Hart and Coombs (25)
, who reported low activities for several amastigote glycolytic enzymes, particularly pyruvate kinase.
In contrast to the down-regulated expression observed for glycolytic enzymes, the expression of two key regulatory gluconeogenic enzymes, PEPCK and F16BP, is gradually but significantly up-regulated throughout differentiation. Since gluconeogenic flux is one order of magnitude smaller than glycolytic flux (8)
, the up-regulation of key enzymes likely indicates a significant increase in gluconeogenic flux. These data fit with recent in vivo studies demonstrating that gluconeogenesis is essential for amastigote proliferation inside murine macrophages (9)
. Naderer et al. (9)
also indicated that the parasitophorous vacuole is poor in glucose, requiring parasites to synthesize their own sugars from other sources, such as amino acids and glycerol. In line with this, we observed that the expression of glycerol kinase and NAD-dependent glycerol-3-phosphate dehydrogenase is increased significantly in amastigotes, accelerating the formation of glycosomal triosephosphate, two molecules of which are used as substrate for gluconeogenesis. The reduced NAD required for this process is provided by the observed increased amount of PEPCK and malate dehydrogenase, which also was reported by others (31
, 32)
. Notably, consumption of glycerol occurs in the glycosomes, which ensures that it is not affected by the cytosolic bottleneck likely created by the observed low expression of enolase and phosphoglycerate mutase. Under glycerol limitation, PEP alone or in combination with glycerol may be used as substrate for glucose formation. In the latter case the glycosomal PPDK reaction (together with adenylate kinase) is responsible for the production of the additional ATP required. Indeed, PPDK and adenylate kinase also are up-regulated in the amastigote stage. Based on our observation that the abundance of cytosolic enzymes common to both glycolysis and gluconeogenesis is decreased significantly during amastigote maturation, we hypothesize that residual enzymatic activity must be dedicated mainly to gluconeogenesis.
In this case, as differentiation progresses, an alternative to glucose must be utilized as the main energy source. A logical candidate for such a source during the late stages of differentiation is fatty acids. Early in vivo studies showed that enzymatic activities associated with β-oxidation of fatty acids were significantly higher in L. mexicana amastigotes than in promastigotes (25
, 33)
. In agreement with this, we found in the present study that the expression of L. donovani β-oxidation pathway enzymes is up-regulated gradually from 10 h postsignal, suggesting an increase in β-oxidation activity. Up-regulated expression of long-chain fatty acid acyl-CoA synthetase and lipolytic enzymes on the one hand, and down-regulated expression of acetyl-CoA carboxylase on the other, further support a strong up-regulation of β-oxidation flux in amastigotes. Moreover, the increased abundance of three different chain-length specific acyl-CoA dehydrogenases and a 2,4-dienoyl-CoA reductase indicates that maturing amastigotes increase their capacity for utilizing a wide range of fatty acid chain lengths, including very-long chain and unsaturated fatty acids. The multiple acetyl-CoA units produced by β-oxidation are handled by the TCA cycle in the mitochondrion, and the reducing equivalents so generated are oxidized by the mitochondrial respiratory chain,
Dendogram analysis revealed highly correlated expression patterns for β-oxidation, TCA cycle, mitochondrial respiration, and oxidative phosphorylation throughout differentiation. As this correlation reflects data from multiple time points and since β-oxidation has been demonstrated to be up-regulated in vivo, we conclude that likely all these pathways are up-regulated during differentiation. Moreover, we propose that the increase in β-oxidation is linked functionally to increased TCA cycle activity, mitochondrial respiration, and oxidative phosphorylation.
In addition to β-oxidation of fatty acids, amastigotes also are geared toward the catabolism of certain amino acids. A branched chain aminotransferase and alanine aminotransferase, two glutamate dehydrogenase isoenzymes, and at least one component of the glycine-cleavage system are all up-regulated, suggesting that isoleucine, valine, alanine, glutamate, and glycine all can be used as substrates by amastigotes.
We observed a differentiation-associated, significant increase in the abundance of two enzymes that synthesize PEP, PPDK, and PEPCK, which utilize pyruvate and oxaloacetate, respectively. In line with this, our analysis indicates that levels of these substrates should rise as differentiation progresses. Elevated oxaloacetate levels are expected from the deduced up-regulation of TCA cycle activity, and increased pyruvate levels can be assumed in light of the observed up-regulation of amino acid catabolism. In addition, increased expression of both alanine aminotransferase and glutamate dehydrogenase during differentiation enhances production not only of pyruvate but also NADH.
Interestingly, we observed that during the third stage of differentiation, at the time when promastigotes transform morphologically into amastigotes (15 h), a highly synchronized decrease occurred in the abundance of translation machinery proteins, including ribosomal protein subunits, translation factors, and tRNA synthetases. This synchronous decrease was associated with an estimated 5-fold decrease in cell volume and a 2.5-fold decrease in growth rate when comparing amastigotes to promastigotes. The fact that these changes occur during the third differentiation stage and not before suggests strongly that they are part of the parasite differentiation program and underlie adapting to a generally lower protein synthesis rate. Indeed, previous studies indicate that protein synthesis is lower in cultured amastigotes than in promastigotes (4)
and is low also in animal-derived amastigotes (34)
. Of note, the observed decreased expression of translation machinery proteins correlates inversely with the observed increase in histone abundance. The increased expression of histones suggests transcriptional changes during late differentiation.
In general, the present study reveals that most metabolic and translational changes take place synchronously during the third and fourth differentiation stages. Therefore, it stands out that the expression of major heat shock proteins is up-regulated transiently during the first two differentiation stages. This transient expression may allow the parasite time for signal perception and processing before morphogenesis. Taken together, our data support strongly the notion that amastigotes are highly adapted to life within the phagolysosomes, which are likely to be poor in carbohydrates but rich in amino acids, triglycerides, phospholipids, and their lipolytic byproducts such as fatty acids and glycerol.
The results also indicate the high level of promastigote adaptation to the sand fly midgut. The sand fly diet is based on fruit nectar, which is sucrose rich. They continuously release sucrose from their crop into the midgut, where it is converted to glucose (35)
. Occasional bloodmeals supplement the nectar diet of female sand flies. Thus, midgut promastigotes are continuously exposed to high sugar levels. Evidently, late log and stationary promastigotes that were used in this work to initiate differentiation were shown to express enhanced glycolytic activity relative to amastigotes (36)
.
We did not expect to observe so few (only 20) significant changes in protein abundance during the early stages of differentiation. These proteins may initiate or reflect the initiation of differentiation. Unlike the expression changes observed in the third and fourth differentiation stages, the changes in protein abundance observed during the first two stages do not appear coordinated. Based on these observations, we surmise that the mechanisms underlying the observed metabolic retooling of amastigotes must be mostly posttranslational. Given that metabolic retooling occurs late in differentiation, the following question is prompted: how do incompletely differentiated parasites manage to reside in the same host environment as mature amastigotes? We speculate that the parasites carry into the new environment endogenous substrate pools, which are utilized during the period of transition and enable survival during this time. Future research can address this speculation directly.
To summarize, the present study suggests that a prominent feature of the promastigote-to-amastigote differentiation process is a switch to producing more metabolic energy. We propose that the energy is important for driving primary proton pumps and secondary carriers, required to battle the large proton gradient between the acidic phagolysosome and the amastigote cytosol. Thus the energy enables pH homeostasis to be maintained. The significantly increased expression of a number of proton-translocating ATPases and carriers that we observed provides further support for this model.
The present study focused on changes in protein abundance occurring during differentiation from promastigote into amastigote, the latter being the highly infectious parasitic form observed in diseased patients. As such, this study is the first proteome-wide time-course analysis of the metabolic and physiological adjustments that take place during the developmental cycle of parasitic protozoa.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 18, 2007. Accepted for publication August 23, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Paape, C. Lippuner, M. Schmid, R. Ackermann, M. E. Barrios-Llerena, U. Zimny-Arndt, V. Brinkmann, B. Arndt, K. P. Pleissner, P. R. Jungblut, et al. Transgenic, Fluorescent Leishmania mexicana Allow Direct Analysis of the Proteome of Intracellular Amastigotes Mol. Cell. Proteomics, September 1, 2008; 7(9): 1688 - 1701. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) 5 h, (
) 10 h, (
) 15 h, (
) 24 h (
) mature amastigotes. C) Expander self-organizing maps algorithm clustering of proteins quantified at all 7 time points, using a 2 x 2 grid at 100,000 iterations.
