Translational Control in Endothelial Cells

Abstract

Cellular phenotype and function is ultimately determined by the synthesis of proteins derived from a genetic blueprint. Control of gene expression occurs at multiple checkpoints including the transcription of DNA into RNA and the translation of RNA into protein. Translational control mechanisms are important regulators of cellular phenotype, controlling up to 10% of overall cellular gene expression, yet they remain relatively understudied when compared with transcriptional control mechanisms. Specific regulation of protein synthesis from mRNA transcripts allows cells to temporally unlink translation from transcription and provides a mechanism for a more rapid response to environmental signals than if transcription were required. We will discuss some of the fundamental concepts of translational control, tools for studying it and its relevance to vascular cells, in particular the endothelium.

Introduction

The central dogma of molecular biology states that genetic information flows from DNA to mRNA (transcription), and mRNA to protein (translation).(1) This model illustrates that genomic DNA does not direct protein synthesis itself, but instead uses mRNA as an intermediary molecule. It is the synthesis of protein from these intermediary mRNA transcripts that ultimately determines cellular phenotype and function. Despite this fact, there is relatively little research into the specific regulation of protein synthesis (as opposed to transcription) in vascular cells.(2)

The classical view of gene expression describes a series of events triggered by some type of signal that stimulates a cell to transcribe its genetic “blueprint,” or DNA, into RNA. This transcript is then processed and transported out of the nucleus and into the cytoplasm. Next, the RNA is translated into protein by ribosomes, yielding a protein which ultimately changes the cell’s structure or function to adapt to the initial signal (Figure 1). This simplistic, “assembly line” view of gene expression suggests that genes are either “on” or “off” as opposed to modulated. Another shortcoming of this model is that it espouses an obligatory temporal relationship in which transcription begets translation as opposed to a scenario in which many regulatory factors may independently (or perhaps simultaneously) work at multiple checkpoints during gene expression.

Figure 1.

Control of gene expression – Classical view of gene regulation in which an extracellular signal triggers a sequential series of temporally related events leading from DNA to mRNA to protein.

Clearly, transcription is an essential process: translation cannot occur in the absence of mRNA. Transcription is also the most common site of regulated gene expression, being targeted up to 90% of the time when a cell responds to a stimulus.(1) However, transcription is time consuming and requires a significant amount of energy. Translational control is an important means of regulating gene expression as it offers an additional level of control in determining which genes are ultimately expressed in protein form, when, how much, and because it can occur temporally independent of transcription. There are important instances in which it is to the cell’s advantage to dissociate transcription from translation; this review focuses on some of those instances.

Translational control defined and its advantages

Translational control is defined as a change in the efficiency or rate of protein translation of one or more mRNAs resulting in a change in the number of synthesized proteins over time. The benefits of controlling gene expression at the level of translation may be summed up in terms of immediacy, precision and redundancy.(3)

Immediacy is self explanatory: if a cell needs to rapidly change its function/phenotype in response to some stimulus, it is much faster to change the translation rate of pre-existing mRNA than it is to synthesize new mRNA before being able to translate proteins from that transcript. Many cells synthesize mRNA and then store it for future use without immediately translating it.(4) These mRNA transcripts are then available for rapid mobilization to the translational apparatus and a rapid change in protein expression given the appropriate stimulus. Such instances are the most obvious example of temporal dissociation of transcription and translation.

Translational control offers increased precision of gene expression by regulating small changes in overall protein levels towards the end of a long, complex pathway rather than at the beginning. Take an automotive assembly line for example. The end product is a car (protein) resulting from a manufacturing process using various raw materials such as steel, rubber, etc (mRNA, amino acids). If the output of the assembly line needs to change by 10%, it makes more sense to exert that control at some point during the assembly process rather than by changing the overall availability of raw materials at the front end.

The combination of transcriptional and translational control (redundancy) helps to avoid dysregulated expression of potentially harmful molecules. This is analogous to multiple back-up systems in spacecraft and commercial airlines in which catastrophic malfunctions supposedly cannot occur as a result of a single system failure. Translational control is often imposed on critical gene products such as oncogenes, growth factors and signaling molecules.(5)

It is useful to distinguish between global and selective translational control. Global controls govern general processes necessary to translate mRNA and therefore affect translational rates of all classes of mRNA transcripts. Global controls are sensitive to the availability of “raw materials” such as amino acids or energy substrates and are responsible for the overall decrease in protein synthesis that occurs during starvation. Selective controls target features unique to a given mRNA molecule or class of mRNAs possessing that feature. Therefore, a unique subset of mRNA transcripts can be translationally repressed despite an abundance of activated translational components (such as ribosomes) or specialized mRNA transcripts may be increasingly translated despite a condition in which overall protein synthesis is reduced (such as heat shock).(6)

Translational control machinery at the molecular level

Synthesizing protein from mRNA transcripts involves three basic steps: 1) initiation, the recruitment and assembly of intact ribosomes at a start codon; 2) elongation, the sequential addition of amino acid residues and: 3) termination, the dissociation of intact ribosomes from the mRNA transcript. In eukaryotic cells, the initiation of translation is a highly regulated, complex process that is the rate-limiting step where regulation of translation most commonly occurs. Translational control is seldom exerted at the elongation or termination steps and is not further described.

Molecular “machinery” required for initiation includes an appropriately processed mRNA molecule, ribosomes, tRNA molecules with their associated amino acids, and a group of additional proteins known as eukaryotic initiation factors or eIFs.(3) After RNA is transcribed from DNA in the nucleus, it is processed by “capping” the 5’-end with a methylated guanosine, splicing out non-coding intronic sequences and polyadenylation at the 3’ tail (Figure 2). This mature mRNA transcript is then exported into the cytoplasm where translation occurs.

Figure 2.

Initiation of translation in eukaryotic cells - The traditional scanning model illustrates “cap-dependent” initiation of translation in which eukaryotic initiation factors (eIFs) facilitate ribosome binding to the capped 5’ end of an mRNA transcript. “Cap-independent” translation is also depicted in which an internal ribosome entry site (IRES) with a complex secondary structure facilitates ribosome binding in the 5’ untranslated region (UTR).

The initiation of translation can occur in several different ways, but the scanning model is thought to be the most common.(3) In this model, the 40S subunit of a ribosome binds to the capped 5’ terminus of a mature mRNA with the assistance of multiple eIFs (Figure 2). This mode of initiation is thus termed “cap-dependent” translation.

Immediately downstream of the 5’ cap is a section of the mRNA transcript known as the 5’ untranslated region, or UTR, which must be scanned by the 40S ribosome subunit prior to reaching the start codon. The length and secondary structure of the 5’ UTR can profoundly influence translational efficiency by altering access of eIFs to the 5’ cap or by preventing smooth scanning to the start codon.(3) Messenger RNA transcripts that possess 5’ UTR sequences with an extensive secondary structure frequently code for oncoproteins, growth factors, transcription factors and proteins that must be tightly regulated for normal cellular function, an illustration of redundancy.(5) There are multiple opportunities for specific control events to occur during the initiation process. Various signal inputs may be required to unmask the 5’-cap so that the 40S ribosome subunit and eIFs can attach and begin scanning. Additional signal inputs may be necessary to assist scanning through highly structured and complicated 5’-UTRs and may involve regulated association or dissociation of various RNA-binding proteins with particular motifs in this region. Translational control can also be exerted through specific interactions between the RNA-binding proteins with the 3’-UTR but are not further discussed here. The types of control that involve regulation of specific events at the mRNA UTR’s tend to be selective rather than global.

Once the start codon (AUG) is recognized, the bound eIFs are released from the 40S subunit in order to allow binding of the 60S ribosomal subunit.(7, 8) A separate eIF is necessary to catalyze the formation of a complete 80S ribosome from the two subunits. At this point, the ribosome is fully assembled on the mRNA transcript at the start codon and translation begins as tRNA molecules supply the appropriate amino acids for protein synthesis. As the ribosome progresses along the transcript, the polypeptide product elongates until the ribosome complex reaches the stop codon towards the end of the 3’ terminus, and translation ceases.(9)

Another model for eukaryotic protein synthesis is based upon the concept of “cap-independent” translation. This model does not rely on the 5’ mRNA cap with its associated eIFs to recruit ribosomes as does the scanning model. Instead, internal ribosome entry sites (or IRES elements) exist in which a ribosome can bypass binding to the capped 5’ end of an mRNA and attach directly at a site downstream, within the 5’ UTR (Figure 2).(3) Viral genomes provided the first evidence for IRES elements because their RNA is not processed and capped as it is in eukaryotes.(10) Furthermore, many viruses actually inhibit the host cell’s normal process of cap-dependent translation by disabling key eIFs necessary for cap recognition.(11) Viral protein synthesis typically occurs because the host cell ribosomes are diverted from cap-dependent translation of host cell mRNA to cap-independent translation of viral mRNA. It is now appreciated that there are multiple mechanisms for ribosomal recruitment to IRES elements, and there is a growing body of evidence that cells may utilize cap independent translation for host protein synthesis at specific times: cellular differentiation(12), apoptosis(13) and certain pathological conditions such as Charcot-Marie-Tooth disease and multiple myeloma.(14–16)

Techniques for studying translational control

Regardless of the manner in which translation is initiated, the ribosome is a fundamental component required for protein synthesis. A single, functional ribosome physically occupies only a short stretch of mRNA allowing multiple ribosomes to attach to the same transcript in order to more efficiently produce the protein being synthesized. Transcripts with multiple ribosomes attached are termed polyribosomes, or polysomes, while transcripts associated with a solitary ribosome are termed monosomes.

Recall that the formal definition of translational control involves a change in the efficiency of mRNA translation or a change in the number of completed proteins per unit time. Direct measurement of this parameter is very difficult, so a more convenient surrogate measure is typically used to indirectly assess translational efficiency. This surrogate measurement is the number of ribosomes attached to a given mRNA transcript. Since initiation is usually the rate-limiting step in translation, the number of ribosomes attached to a given mRNA molecule also reflects the efficiency of initiation which is also the most common site of control. Thus, measurement of the number of ribosomes attached to various mRNA molecules under various conditions provides important clues to the regulatory events governing translation in those particular situations. Specifically, mRNA transcripts associated with polysomes are presumed to be efficiently translated while mRNA transcripts associated with monosomes (or not present in the ribosomal preparation at all) are inefficiently translated. The technique of ribosome profiling (Figure 3) is used to assess how many ribosomes are attached to mRNA molecules.

Figure 3.

Ribosome profiling – A) Efficiently translated mRNA transcripts (polysomes) are separated from inefficiently translated mRNA transcripts (monosomes) using sucrose density gradient centrifugation. Because individual ribosomes are so heavy, discrete bands are produced that correspond to an integer number of ribosomes in a polysome. Following centrifugation, the gradients are passed through a spectrophotometer which generates tracings based on the presence of genetic material to facilitate separation of the monosome fraction from the polysome fraction. B) In general, conditions characterized by high translational activity (such as cell growth or proliferation) will have mRNA predominantly associated with polysomes (dashed line) while conditions characterized by low translational activity (such as starvation) will have mRNA predominantly associated with monosomes (solid line).

The most important concept in the experimental study of translational regulation is the idea that signal- or condition-dependent redistribution of mRNA between the polysome or monosome fractions is prima facie evidence of translational control. A widely used parameter to reflect translational efficiency for a given mRNA is the “translation state” (TS) which is simply the ratio of the amount of mRNA in the polysome fraction divided by the amount of mRNA in the monosome fraction. Messenger RNA species with TS > 1 are efficiently translated and those with TS < 1 are not. A significant change in measured TS between different experimental conditions is also evidence of translational control.

Initially, methods for direct analysis of protein expression (or proteome analysis) were cumbersome, insensitive and limited in their ability to assess large numbers of genes for translational activity.(17) Classically, translational control was recognized in experiments when a given condition could induce changes in protein levels without corresponding changes in mRNA levels. A high-throughput method for simultaneously monitoring the translational state of large numbers of individual mRNA species was first described in 1999 by Zong et al.(18) This technique, known as translation state array analysis (TSAA; Figure 4), combines microarray technology with ribosomal profiling to determine the TS of thousands of mRNA species simultaneously. Poorly translated mRNA transcripts associated with monosomes are separated from efficiently translated mRNA transcripts in polysomes by ribosomal profiling (Figure 3). In this instance, all fractions containing two or more ribosomes are pooled to form the polysome fraction. After isolation of RNA from monosome or polysome fractions, fluorescent labeled cDNA copies of the mRNA transcripts are synthesized and used to interrogate DNA arrays on which thousands of known gene sequences are bound. Since monosome cDNA is labeled with a different fluorophore than polysome cDNA, competitive hybridization yields a measure of TS for each gene on the array (Figure 4).

Figure 4.

Translation state array analysis (TSAA) – This high-throughput method for assessing the translational state (TS) of a large number of mRNAs combines ribosomal profiling and microarray technology. Cellular mRNAs are separated into inefficiently translated fractions (monosomes) or efficiently translated fractions (polysomes). The mRNAs are then converted into cDNA labeled with a fluorescent marker (Cy3 is green and Cy5 is red) and hybridized to an array chip containing thousands of genes. If there is a larger amount of mRNA for a certain gene sequence in the monosome fraction, that spot on the chip (representing that gene) will fluoresce green and will have a corresponding TS <1 (where TS represents the mRNA in the polysome fraction divided by the mRNA in the monosome fraction). If there is a larger amount of mRNA for a certain gene sequence in the polysome fraction, that spot on the chip will fluoresce red and will have a corresponding TS >1. If there are equivalent amounts of mRNA for both the monosome and polysome fractions, then that spot will fluoresce yellow and have a corresponding TS≈1. If there is no hybridization of the probes to the gene in question, the spot will not fluoresce.

Once the translational states are determined for each gene on the control array, they can be compared with the translational states for each gene on the experimental array. A ratio of the TS for each gene under experimental conditions divided by the TS for control conditions will yield a translational index (TLI). This value represents the redistribution of mRNA between monosome and polysome fractions for the given experimental condition.

If separate arrays are used for control and test conditions, simultaneous measurement of experimentally-induced changes in TS for thousands of genes is possible. The change in TS for a given experiment can be expressed by the translation index (TLI) which is simply the ratio of the measured TS under the experimental conditions divided by the measured control TS. A TLI > 1 implies translational upregulation since the conditions of the experiment have redistributed mRNA to the polysome relative to the control situation. A TLI < 1 implies translational repression since the experiment has resulted in a redistribution of mRNA out of the polysome and into the monosome. A key attribute of TSAA is that it can recognize translational control even when there is concomitant transcriptional control since the TS and the TLI only reflect the proportions (not total amounts) of mRNA in the two fractions.

If a third array chip is added to a given experiment, transcriptional control can also be directly assessed by traditional microarray methods. Total RNA is isolated from cells in both the treatment and control conditions and fluorescently labeled cDNA probes from both conditions are used to competitively hybridize with the third chip. Thus, with three arrays, it is possible to simultaneously assess both transcriptional and translational changes. Based on the results of TSAA experiments, we have categorized nine different patterns in which gene expression is potentially regulated in response to a given stimulus in terms of transcriptional and translational indices (Table 1).

Table 1.

Types of transcriptional and translational control as predicted by TSAA

Category	Description	Translation	Transcription
Positive Redistribution	Shift of mRNA from monosome to polysome; no change in total mRNA abundance	Increased	Static
Negative Redistribution	Shift of mRNA from polysome to monosome; no change in total mRNA abundance	Decreased	Static
Co-ordinate Activation	Shift of mRNA from monosome to polysome out of proportion to increased total mRNA abundance	Increased	Increased
Co-ordinate Repression	Shift of mRNA from polysome to monosome out of proportion to decreased total mRNA abundance	Decreased	Decreased
Paradoxical Activation	Shift of mRNA from monosome to polysome despite a reduction in overall mRNA abundance	Increased	Decreased
Paradoxical Repression	Shift of mRNA from polysome to monosome despite a reduction in overall mRNA abundance	Decreased	Increased
Obligatory Upregulation	Increased mRNA in both monosome and polysome paralleling overall increase in mRNA abundance	Static	Increased
Obligatory Downregulation	Decreased mRNA in both monosme and polysome paralleling overall reduction in mRNA abundance	Static	Decreased
No Regulation	No change in mRNA in monosomes or polysomes and no overall change in mRNA abundance	Static	Static

Data derived from TSAA experiments should be validated using molecular biology techniques such as quantitative PCR (polymerase chain reaction). In addition, TSAA-derived data does not give any information about the function of the gene or genes in question. Traditional cell biological studies are still necessary to determine the importance of the observed changes in translation (and transcription).

Translational control in Vascular Cells

Vascular endothelial cells prominently utilize translational control mechanisms. Despite early beliefs that theses cells were merely “bystanders” lining the inside of the blood vessels, it is now clear that these cells play a dynamic role in determining the ultimate biologic behavior of the vessel wall.(19) Endothelial cells respond to signals from the environment with rapid functional and phenotypic changes. These alterations in endothelial phenotype are inducible by a variety of agonists which act at various receptors. Unregulated endothelial activation is found in numerous pathological conditions including sepsis, inflammation and ischemia-reperfusion injury.(20) With the explosion of endovascular interventions, there is also an increasing amount of direct mechanical trauma to endothelial cells injured by balloon catheters, stents and endografts. The rapidity with which endothelial cells can alter their phenotype in response to the environmental stressors listed above supports the notion that translational control mechanisms play a significant role in regulating their function.

The response of endothelial cells to external stimuli may occur within a broad timeframe ranging from seconds to days. Second to second responses generally involve phosphorylation or dephosphorylation modifications of proteins. Transcriptional control mechanisms may require many hours or even days. Translational control responses tend to occur in a matter of minutes to hours, providing the cell with the opportunity to mount a phenotypic response to an environmental challenge in an intermediate time frame (immediacy). Fluid shear stress is a particularly relevant environmental stimulus to endothelial cells with well known effects on endothelial phosphorylation events and transcription.(21) Our group has shown that fluid shear stress also influences translational activity in vascular endothelium.(22, 23)

The mTOR (mammalian target of rapamycin) pathway is a ubiquitous signaling system that regulates translation in many cell types.(24) mTOR is a protein kinase that directs phosphorylation of related protein kinases including S6K1 (S6 kinase 1, previously known as P70/P85 S6 kinase).(25) Activation of the mTOR pathway is crucial to the initiation of protein synthesis in many circumstances. Rapamycin (or sirolimus) is a peptide isolated from the bacteria Streptomyces hygroscopicus. It is an important adjunct in the study of translational control because it directly inhibits mTOR activity. Rapamycin also has clinical applications as an immunosuppressant and an inhibitor of cell growth when eluted from specialized vascular stents. Fluid flow activates the mTOR pathway in endothelial cells, resulting in activation of S6K1 which facilitates 40S ribosomal recruitment to specific mRNAs (initiation). (23, 26) In addition, fluid flow induces a rapid increase in the synthesis of Bcl-3, a transcription factor that is a member of the NF-κB (nuclear factor kappa B) family of transcription regulators. A key finding in these studies was that rapamycin effectively blocked both the activation of S6K1 and the synthesis of Bcl-3 while transcriptional inhibition with actinomycin did not.(23)

Additional studies have demonstrated that the translation of E-selectin is modulated by shear stress.(27) E-selectin is a cell surface molecule inducibly expressed by endothelial cells, and it plays a critical role in leukocyte adhesion and overall endothelial cell activation. Expression of E-selectin protein by endothelial cells was induced using the traditional inflammatory agonist TNF-α (Tumor necrosis factor alpha). Exposure to fluid flow attenuated the expression of E-selectin in the presence of TNF- αwhen compared to cells not exposed to fluid flow. Fluid flow did not reduce overall E-selectin mRNA levels, but did reduce the amount of E-selectin mRNA associated with polysomes implying the existence of a regulatory step that specifically regulated access of the mRNA to the protein synthesis machinery, a form of translational control. Interestingly, neither rapamycin nor nitric oxide synthase inhibitors eliminated the modulatory effect of flow on E-selectin expression.(23, 27)

This series of discoveries illustrates the concept that endothelial cells can rapidly alter protein synthesis by discrete translational control mechanisms independently of transcription. These observations also emphasize the importance of translational control in endothelial cells because of the existence of multiple regulatory pathways, some of which are independent of the mTOR system and the classic flow-dependent nitric oxide signaling pathway.

Recently, translational control mechanisms have been identified in other important cellular components in the vascular system. Platelets, leukocytes and vascular smooth muscle cells all exhibit some degree of translational control.(28–30) Platelets are intriguing cells for the study of translational control because they are anucleate and lack DNA. Despite this fact, thrombin stimulates platelet synthesis of a number of proteins from pre-existing mRNA stores derived from the parent megakaryocyte.(31) In particular, platelet expression of the transcription factor Bcl-3 is induced by thrombin activation. This was a confusing finding initially since platelets would appear to have no use for a transcription factor. Subsequent studies revealed Bcl-3 to have activities apart from transcriptional regulation, participating in platelet-mediated clot retraction.(32) Expression of Bcl-3 protein is diminished by the translational inhibitor rapamycin, demonstrating that platelets regulate protein synthesis through signal-dependent activation of translation despite a lack of transcriptional activity.

Conclusions

Regulation of gene expression is a complex process, and although a great deal of work has been done on transcriptional regulation in vascular biology, there are many other factors that determine whether or not a gene ultimately produces a functional protein which can alter cellular phenotype and function. Translational control mechanisms represent one manner in which cells can regulate gene expression, and although research efforts have increased in this field, it remains vastly understudied. Translation is a complex process involving interactions between outside signals, cellular machinery, enzymes and genetic material. Manipulation of these interactions can drastically alter cellular function, even in the absence of transcriptional changes. With the realization that translational control is separately targeted by extracellular signals, the classical notion of sequential control of gene expression show in Figure 1 has been refined to reflect the complex and simultaneous nature of parallel signal input for both transcriptional and translational regulation shown in Figure 5.

Figure 5.

Parallel signal inputs can regulate transcription and translation independently - Contemporary view of factors controlling gene expression in which there may be multiple signaling inputs occurring simultaneously to modulate changes in a cell’s gene expression profile.

Endothelial cells are an intriguing model for studying translational control mechanisms because they must respond rapidly to changes in blood flow, chemical signals and other forms of stimulation such as mechanical trauma. Translational control is a particularly useful type of regulation that allows endothelial cells to respond to signals relatively rapidly when compared to the time required for a transcriptional response. Continued investigation into translational control mechanisms in endothelial cells may provide important insights into their pathogenic responses to stimuli and novel strategies for modulating their function. Utilization of techniques like TSAA offers an efficient means for screening large numbers of candidate genes for evidence of translational control with the caveat that verification and functional assays will still be necessary to place the results in proper biologic perspective.

Abbreviations

Bcl-3

B cell lymphoma 3 (a transcription factor)

eIF

eukaryotic initiation factor

IRES

internal ribosome entry site

mTOR

mammalian target of rapamycin

NF-κB

nuclear factor kappa B (a transcription factor)

S6K1

S6 kinase 1

TLI

translation index

TNF-α

tumor necrosis factor alpha

TS

translation state

TSAA

Translation state array analysis

UTR

untranslated region