The goal of this study was to use an unbiased proteomics approach to identify cell surface proteins regulated by transmembrane ubiquitin ligases of the MARCH protein family. Using the viral MARCH homologue KSHV-K5, we had previously demonstrated that it is possible to identify novel targets for both the protein used in the proteomics study, as well as other members of this family 
. For instance, ALCAM and BST2/Tetherin were identified as novel substrates for KSHV-K5. Follow up experiments, however, demonstrated that ALCAM was also down-regulated by MARCH-VIII and Myxomavirus M153R whereas BST2/Tetherin was also down-regulated by MARCH-VIII and HIV-1 Vpu. Similarly, we now show that CD81 was also down-regulated by MARCH-IV. Although a modest downregulation of CD44 was observed with MARCH-IV by flow cytometry, this was not observed by IFA. Interestingly, expression of CD44 or CD81 was not affected by the viral MARCH proteins tested, suggesting that these endogenous substrates are specifically targeted cellular MARCH proteins. Also surprising was the observation that MARCH-I and MARCH-IX, which are closely related to MARCH-VIII and MARCH-IV, respectively, did not affect either CD81 or CD44. This is in contrast to our previous observations which showed that sequence homology strongly correlated with substrate specificity. Why a similar correlation is not observed for CD81 and CD44 is presently not known.
While both CD44 and CD81 are widely expressed on different cell types, their regulation has been shown to play a particularly important role in the developing immune system. Like most targets of MARCH proteins, CD44 is a type I transmembrane glycoprotein. CD44 occurs in many splice variants and connects a variety of extracellular matrix proteins, most notably hyaluronic acid (HA), to the cell surface 
. CD44 plays a crucial role in inflammation 
and in cancer progression 
. It is also involved in leukocyte recruitment 
and regulates T cell interactions during T cell development 
. Since CD44 exists in many isoforms and is extensively spliced 
, MARCH-VIII dependent down-regulation might be variant-specific. However, this possibility has yet to be investigated. Although elimination of MARCH-VIII from fibroblasts did not change the steady state levels of CD44, it is likely that CD44 is affected in more dynamic situations where MARCH proteins are regulated, such as during dendritic cell maturation 
. MARCH- VIII could also play a role for the HA-induced endocytosis of CD44 and its subsequent targeting to the multi-vesicular bodies (MVB) 
since most MARCH-VIII targets reach the MVB 
In contrast to CD44, CD81 surface levels were increased upon depletion of MARCH-IV, suggesting that MARCH-IV is involved in the natural turnover of CD81 in fibroblasts. While depletion of MARCH-VIII had no effect on CD81 levels in fibroblasts, it is possible that higher MARCH-VIII levels in other cell types or tissues might affect CD81. Unlike most MARCH-targets, which are single-span transmembrane proteins, CD81 belongs to the tetraspanin family, four-transmembrane proteins that form the tetraspanin-web as a membrane microdomain which laterally incorporates multiple transmembrane proteins 
. Since several of the CD81-interacting proteins have been shown to be down-regulated by MARCH-IV or MARCH-VIII it is possible that CD81 surface expression is indirectly affected via these bona-fide
targets. Alternatively, it is conceivable that some of these proteins could be down-regulated as a consequence of CD81 ubiquitination and degradation. In fact, targeting a protein with widespread protein:protein interactions, such as CD81, might explain how MARCH proteins down-regulate such a variety of substrates, with no sequence similarity. Since our observations that MARCH-IV lacking a functional RING-CH domain and mutants of MARCH-VIII which are unable to down-regulate B7.2 both fail to down-regulate CD81 are consistent with either model, further work is required to distinguish between these possibilities. The validation of both CD81 and CD44 by independent methods confirms our previous conclusion that three biological replica experiments give a high level of confidence for the specificity of the results 
. Since ALCAM, a known substrate of MARCH-VIII, was identified in only two of the three experiments, however, we selected two additional potential candidate proteins, CD9 and Bap31, for validation based on their sub-cellular localization as well as their known interaction with MARCH-VIII substrates. Only Bap31, however, confirmed using independent methods, indicating that the false-positive rate is high in these types of experiments when biological replicas are limited. This is primarily due to the inaccuracy of quantitation using the mass spectroscope.
Unlike CD44 and CD81, steady state levels of Bap31 were not diminished following expression of MARCH-VIII. Instead, we found that MARCH-VIII bound to Bap31 and prevented Bap31 from reaching the cell surface. We further observed that Bap31 interacts with most MARCH proteins, suggesting that Bap31 chaperones the folding, assembly or intracellular transport of this protein family. Interestingly, depletion of Bap31 resulted in lower amounts of MARCH-VIII expressed on the cell surface. Since Bap31 is known to function as a forward cargo molecule for several other proteins 
we conclude that that Bap31 is involved in sub-cellular sorting of MARCH proteins. Interestingly, the MARCH-family and their viral homologues have been shown to interact with several other intracellular trafficking molecules. Depletion of PACS2 inhibited K5 mediated degradation of newly synthesized CD31, suggesting that it played a key role in K5 function 
. Several members of the MARCH-family have been shown to interact directly with members of the syntaxin family of SNARE proteins 
. However, the functional consequences of these interactions are not currently understood.
Truncated versions of MARCH-VIII that do not interact with syntaxins are still able to interact with Bap31 
suggesting that the interaction with Bap31 occurs in the transmembrane region. This is consistent with previous findings that Bap31's interaction with MHC I, IgD, cellubrevin, and p450 localizes to the transmembrane domains 
. Interestingly, expression of Bap31 is required for surface expression of CD81 
. Thus, inhibition of Bap31 might play some role in MARCH-IV and –VIII down-regulation of CD81.
While depletion of Bap31 affected intracellular transport of MARCH-VIII, it did not prevent MARCH-VIII-mediated down-regulation of transfected CD86 (B7.2) (data not shown). This result, however, is most likely explained by insufficient knockdown of Bap31, resulting in sufficient amounts of MARCH-VIII available at the cell surface to ubiquitinate its substrates.
In summary, we present a systematic identification of endogenous substrates for the MARCH gene family. By using this non-biased approach to substrate discovery, we further demonstrate that the tetraspanin CD81 is down-regulated by both MARCH-IV and –VIII thus expanding the structural range of possible MARCH substrates. We were also able to show that depletion of endogenous MARCH-IV increased the surface expression of CD81, the first time that a substrate of MARCH-IV has been affected by depletion of endogenous protein. Thus, these approaches will be useful in unraveling the function of this class of proteins.