U.S. Department of Health and Human Services
Andy Golden

 Contact Info

Tel: +1 301 594 4367
Email: andyg@mail.nih.gov

 Training and Experience

Senior Investigator, NIDDK, NIH 2006-present

Investigator, NIDDK, NIH, 2000-2006

Program Fellow, NCI-Frederick, 1994-1999

Post-doctoral Fellow, California Institute of Technology, 1990-1994

Ph.D., State University of New York at Stony Brook, 1990

B.S., Muhlenberg College, 1982

 Related Links

Specialties: Developmental Biology, Genetics/Genomics

Research Summary

Research Goal

The goal is to identify regulators and interactors of conserved genes that, when mutated, can cause human disease.

Current Research 

The mat mutants

Over the past 15 years, we have studied the maternal components necessary for the completion of the meiotic divisions of the oocyte. C. elegans oocytes remain in prophase of meiosis I as they pass through the spermatheca, where they are fertilized by sperm. After fertilization, the fertilized oocyte completes both meiotic divisions, extruding two polar bodies. The haploid oocyte pronucleus then migrates to join the haploid sperm pronucleus. The 1-cell embryo then undergoes its first mitosis. RNAi with a number of conserved cell cycle genes causes embryonic lethality. Some of these dead embryos arrest at the 1-cell stage, during metaphase of meiosis I: the oocyte chromosomes organize on a metaphase plate, set up a meiotic spindle, and the sperm chromosomes remain condensed at the posterior of the embryo.

In collaboration with the Seydoux Lab (Johns Hopkins University School of Medicine), the Bowerman Lab (University of Oregon), and the Shakes Lab (College of William and Mary), we isolated temperature-sensitive (ts) mutants that arrested as 1-cell embryos when mothers were shifted to the nonpermissive temperature of 25°C (Golden et al., 2000). Almost 40 such mutants were recovered in this screen; all but five were arrested in metaphase of meiosis I. The remaining five arrested as 1-cell embryos due to cytokinesis defects.

The meiotic-arrested mutants defined five complementation groups. Members from two groups were found to be allelic with​ emb-27 and emb-30 (emb = embryonic lethal), two genes that were previously identified in screens from 30 years ago. The remaining alleles defined three new complementation groups, which we named mat for metaphase to anaphase transition-defective (Golden et al., 2000). The mat-1, mat-2, and mat-3 genes are each represented by multiple ts alleles. Each of the mat and emb genes was mapped and molecularly cloned. All five genes encode orthologs of Anaphase Promoting Complex/Cyclosome (APC/C) subunits (Furuta et al., 2000; Golden et al., 2000; Davis et al., 2002; Shakes et al., 2003). In addition to emb-27, emb-30, and the three mat genes, Diane Shakes had shown that a sixth gene, emb-1, belongs in this class of mutants. Similar to emb-27 and emb-30, this gene was previously identified by others in a screen for embryonic lethality. Though our screen did not yield any new emb-1 alleles, we continued to characterize the two existing ts alleles because of the similarity in phenotype to the mat mutants. In 2011, we reported that the emb-1 gene encodes a small subunit of the APC/C (Shakes et al., 2011). Recent studies from another group suggested that EMB-1 is orthologous to the newly discovered APC16 subunit (Kops et al., 2010). Biochemical studies in C. elegans further demonstrated that the EMB-1 protein associates with the APC/C (Green et al., 2011).

Of the 15 APC/C subunits that we, or others, have identified by sequence homology, only the above six are represented by ts alleles. Many of the others are represented by deletion alleles; homozygous embryos hatch due to maternal rescue and develop into sterile adults. This sterile phenotype seems to be common to all APC/C subunit deletions and many other cell cycle regulators in C. elegans. However, further testing (by RNAi) of the APC/C subunits for which we did not isolate ts alleles revealed that most APC/C subunits are required for the metaphase to anaphase transition during meiosis I in C. elegans. One of our recent studies demonstrated that there are two APC5 orthologs and that they act redundantly in meiosis I; when both are depleted, a metaphase I arrest is observed (Stein et al., 2010).

One-cell embryos from our mat mutants all share the same arrest phenotype when adults are shifted to 25°C: (a) the oocyte chromosomes are arranged on a metaphase plate or "pentagonal array;" (b) the oocyte chromosomes organize a morphologically normal meiotic I spindle; and (c) the sperm chromosomes remain highly condensed at the other end of the embryos. The embryo does not progress further. No anaphase figures are observed and no polar bodies are extruded. The chromosomes do not decondense or undergo DNA synthesis. Maternal and paternal pronuclei do not form. The sperm centrosomes remain quiescent and do not nucleate microtubule asters. Staining with a number of M-phase marker antibodies further suggests that these meiotic arrest mutants persist in an M-phase-like state. The meiotic metaphase arrest observed in all these mutants suggests that key cell cycle regulator genes are functionally defective at the nonpermissive temperature.

Suppressor screen for regulators and substrates of the APC/C

In order to identify other proteins that functionally interact with or regulate the APC/C, we undertook another genetic approach to identify such proteins in a suppressor screen.We simply mutagenized mat-3 mutants and looked for animals that could then survive at the nonpermissive temperature. We identified 29 such suppressors, the majority of which are extragenic and semi-dominant. Only three alleles were recessive. We identified a number of these suppressors as alleles of the spindle assembly checkpoint (SAC).  The SAC in most organisms consists of at least six components. We have identified alleles of mdf- 1mdf-2, and  mdf-3 , the orthologs of MAD1, MAD2, and MAD3 of yeast.  We also have shown that gain-of-function alleles of the CDC20 gene, called fzy-1 in C. elegans, can also suppress the 1-cell arrest phenotype of mat-3 (Stein et al., 2007).  Many of these alleles do suppress mutations in the other mat genes (mat-1, mat-2emb-1emb-27, and emb-30), suggesting that the SAC serves as a negative regulator of the APC/C during meiotic divisions.  Perhaps the SAC serves to regulate the proper timing of the meiotic divisions upon fertilization in C. elegans.  In addition to SAC components, we also identified gain-of-function alleles of an APC5 ortholog (Stein et al., 2010). 

SPE-11, a paternally provided protein essential for embryonic development

More recently, the lab has become interested in the paternal factors that impact early embryonic development. Only one strict paternal-effect lethal mutant exists in C. elegans. The gene is spe-11. The spe-11 gene was originally identified in a screen for sperm-defective mutants (Spe) (L’Hernault et al., 1988). Unlike most Spe mutants, which are incapable of fertilizing an oocyte, spe-11 mutants are competent for fertilization. However, the resulting embryos fail to extrude polar bodies, fail to establish a chitinous eggshell, and fail cytokinesis (Hill et al., 1989Browning and Strome, 1996). All such embryos fail to hatch. This mutant is a strict paternal-effect lethal; mutant oocytes develop normally if fertilized by wild-type sperm, and wild-type oocytes fertilized by mutant sperm fail to develop. We are interested in understanding the role that SPE-11 plays in embryonic development. Since SPE-11 is a novel protein with no recognizable motifs, we have little information about its interactions and mechanism of regulation. We are taking a multidisciplinary approach to investigating SPE-11. First, we are further characterizing its phenotypes to understand how much of early development is affected. For such studies, we are examining the expression of many maternal proteins involved in the oocyte-to-embryo transition and in eggshell formation.

We are also looking to identify interactors through biochemical studies using tagged SPE-11 proteins. In addition, we are doing structure-function analyses using variants of spe-11 to determine their ability to rescue spe-11 mutants and support normal development. We are also trying to generate additional alleles of spe-11 that would be suitable for a suppression screen. Currently, the existing alleles all bear nonsense mutations except for one temperature-sensitive allele that is too leaky for a screen. Alternatively, an enhancer screen might also be a feasible approach to further identify genes that act in the spe-11 pathway.

Mutations in the pph-5 gene suppress the embryonic lethality of separase mutants

Separase is the protease that cleaves cohesin to allow sister chromatid separation at anaphase. We became interested in separase since one of its regulators, securin, is a substrate of the APC/C. In order for the metaphase-to-anaphase transition to occur, the APC/C targets securin for degradation, thus freeing separase to cleave cohesin. To further understand this pathway and identify factors that interact with or regulate separase, we took a genetic approach. We mutagenized a strain carrying a temperature-sensitive mutant allele of sep-1, the C. elegans separase gene, to identify suppressor mutants that would allow this strain to develop at the nonpermissive temperature. We recovered three suppressors, one of which was intragenic. Of the two extragenic mutations, we have characterized one as a mutation in the phosphatase gene, pph-5 (Richie et al., 2011). This missense mutation, as well as two in-frame deletion mutations, suppresses the embryonic lethality of two of our three sep-1 alleles. RNAi of the pph-5 gene also suppresses the embryonic lethality of sep-1 mutants. Surprisingly, depletion of PPH-5 via mutation or RNAi in an otherwise wild-type genetic background has no obvious phenotypes; animals are viable and fertile. The pph-5 gene thus appears to be a nonessential gene, and mutations in this gene are only apparent in specific sep-1 mutant backgrounds.

In collaboration with Josh Bembenek (now at the University of Tennessee), we have shown that our  sep-1 alleles are defective for cortical granule exocytosis (CGE), an event that occurs at the metaphase-to-anaphase transition of meiosis I (Bembenek et al., 2007). Therefore, it appears that separase not only controls chromosome segregation during this stage of meiosis, but it also coordinates chromosome segregation with CGE. The products released from the cortical granules are required for eggshell formation and fittingly, sep-1 mutants have very defective eggshells. In the two sep-1mutants whose lethality is suppressed by pph-5, the pph-5 mutants are able to suppress the CGE defects. In addition, the SEP-1 protein localizes to these cortical granules in wild-type embryos, but not in the sep-1 mutants, and this localization is restored in the pph-5 suppressed embryos.

The mechanism of suppression remains to be determined. Given the observations from other systems that all players in the separase pathway are phosphorylated at one time or another, there are numerous candidate substrates for PPH-5. We plan to take genetic and biochemical approaches to determine the substrates of PPH-5 and the mechanism by which hypomorphic mutants and RNAi depletion of PPH-5 restore viability to  sep-1​ embryos.​

Applying our Research

Understanding the genes that interact with a disease-causing gene will shed light on the genetic pathway in which that disease gene acts. By modulating the activity of interacting genes, it may be possible to alleviate symptoms of the disease. Thus, genetic screens to identify interacting genes may lead to novel therapeutic targets.

Strategies for studying rare diseases in C. elegans

For all diseases in which the responsible gene is known, we need to understand the function of that gene in the normal and disease condition. And for those diseases in which the responsible gene(s) is not known, we need more research designed at finding and defining the responsible gene(s).

There are approximately 7,000 rare (or orphan) diseases known in humans. By definition, a rare disease affects fewer than 200,000 people in the United States (~1:1500). However, given the number of rare diseases that exist, about 10 percent of the population has a rare disease. Treatments for less than a 1,000 of these diseases currently exist.​

We are using C. elegans as a tool to determine the genetic interactions of a given rare human disease gene ortholog. The factors identified in such screens could then be examined in other model systems and human cells to determine whether such factors might be useful therapeutic targets.

For human autosomal recessive monogenic diseases in which the responsible gene is known, we would like to use C. elegans to study the function of that gene and to genetically identify other factors that act in the same pathway. There are a number of criteria that would have to be met in order for this strategy to work. First, there would have to be a convincing and clear C. elegans ortholog. Second, there would have to be a mutation or deletion in this gene that already exists (alternatively, we could use current technologies to generate mutant alleles). Third, there would have to be a scorable phenotype. The more penetrant the phenotype, the better. If these criteria are met, genetic suppressor and enhancer screens could be performed to identify interacting factors that function with that given gene and the biological process in which it functions.

For diseases caused by dominant mutations, our strategy would be different. We could attempt to express a dominant variant of the C. elegans ortholog. If there is a penetrant phenotype, we can characterize it and use genetic suppressor and enhancer screens to identify other factors that act in the same pathway. Likewise, we could even express the human gene in C. elegans, characterize its phenotypes, and subject such transgenic animals to genetic screens.

For genes in which mutations do not yet exist, RNAi works extremely well to elicit phenotypes. We can screen for suppressors and enhancers of a given RNAi phenotype to identify other factors interacting with the given gene (our unpublished observations).

Ideally, suppressors are identified and confirmed. Their phenotypes will then be determined in an otherwise wild-type background, one devoid of the original disease mutation. These suppressor mutations may reveal phenotypes on their own, or may appear wild-type. Such “silent” suppressors have been identified in genetic screens in many model organisms. Silent suppressors are mutations that, in an otherwise wild-type background, have no overt phenotype and are thus indistinguishable from wild type. For example, some of the genes that suppress vulval phenotypes in C. elegans have no phenotype on their own (Sundaram and Han, 1995; Kornfeld et al., 1995). In another example, my lab has recently shown that alleles of pph-5 have no obvious phenotype, but suppress the embryonic lethality of separase mutants (Richie et al., 2011). We interpret these silent suppressors as mutations in genes that fine-tune a molecular pathway without drastic consequences. Perhaps PPH-5 is not an essential protein and functions redundantly in a number of processes, but is not redundant in the separase pathway. If we accept such interpretations, the potential exists for silent suppressors to be useful therapeutic targets. Partially depleting their expression, altering their biochemical function, or affecting their interactions with other proteins might be the explanation for their suppression. Thus, drugs that mimic the effects of such suppressor mutations might be worthy new targets for therapies of diseases in which recessive mutations are known to be responsible for suppression. In the absence of effective gene therapy or stem cell therapy, drugs that target specific proteins may be beneficial to patients with such diseases.

Suppressors that are not silent but have phenotypes of their own may even prove beneficial. A gene that, when mutated, has early developmental defects may not be excluded as a therapeutic target if the treatment is targeting animals that have already developed. Thus, a therapy that targets an essential gene may be useful if used on late-stage or adult animals. For many human diseases, the onset of disease symptoms is quite variable, and therapies targeted at essential genes may not disrupt development if treating older children or adults.

Of the more than 1,000 diseases for which the responsible mutated gene has been identified, the molecular mechanism for a majority of those diseases has not been clearly defined. Often it requires the collaboration between basic researchers and clinicians to establish the molecular pathway and mechanism of the disease. Once we have identified new players in the rare disease pathway (through the methods described above), it is our intention to utilize the NIH Clinical Center and contacts with external clinicians to attempt to put the basic research knowledge we have gained to practice in a clinical setting. Such knowledge and collaboration will hopefully lead to a rational (and perhaps novel) design for therapy that will benefit individuals suffering from the rare disease being studied.​